Cross-Reference

The present application claims priority to U.S. Provisional Patent Application No.

62/471 ,637, .filed; March 15, 20.17, incorporated by reference herein in its entirety. Background

Influenza virus is a member of Qrthomyxovirklse family. There are three subtypes of influenza viruses designated A, B, and C. The influenza virion contains a segmented negative-sense NA genome, encoding, among other proteins, hemagglutinin (HA) and neuraminidase (HA). Influenza vires infection is initiated by the -.attachment of the virion surface HA protein to a sialic- acid-containing cellular receptor (glycoproteins and giycolipids). The NA protein mediates processing of the sialic acid receptor, and vims- penetration into the cell depends on HA-dependent receptor-mediated endoeytosis. So fax, chemical analogs of the receptor have not been successful as viral-entry blockers. Current treatment options include therapeutic- antibodies, srnall-moieeules drugs and vaccination. These therapies allow protection against circulating subtypes, but may not protect against newly emerging strains. Hence, general or quickly adaptable solutions for cheap treatment options represent a constant need. Additionally, in orde to rapidly diagnose early whether a patient, indeed suffers from influenza, sensitive diagnostics are desirable, a treatment at the onset of the infection have been shown to be more efficient influenza presents a serious public-health challenge and new therapies are: needed to combat viruses that are resistant to existing antivirals r escape neutralization by the immune system.

Small (4-12 kDa) binding proteins have the potential to bridge the gap between monoclonal antibodies and small molecule drugs, with advantages of stability and chemical synthesis over monoclonal antibodies, and in selectivity and designability over small molecules. Directed evolution has been used starting from naturally occurring, small protein scaffolds to generate new binding proteins. While, powerful, such approaches have

limitations: they cannot modify the overall shape of the starting scaffold protein(s), they can only sample a very small fraction of sequence space-, -and naturally occurring disulfide- mint- proteins can be difficult to express. Computational protein design has the potential to overcome these limitations by efficiently sampling both shape -and -Sequence space on a much larger scale, an ^' d generating readily producible proteins, as recently demonstrated by the design of stapled mini protein scaffolds with a wide range of shapes. Despite this potential, the high cost to synthesize gea.es for each designed, protei has generally limited, testing to -small numbers (tens) of designs for any one application, which is too few to systematically explore the capability of this approach ami to provide feedback to mprove - ^' the c mputaiio l model.

Snnimai

in one aspect are provided isolated polypeptides comprising an amino acid sequence having at least 70% ^" sequence identity over its length to the amino acid sequence of any one of SEQ ID HQS: 1-462. in various embodiments, the polypeptide comprises an amino acid sequence having at least 70%, 75% SQ%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97 %, 98%, 99%, or 100% sequence identity over its length to the amino acid sequence of any one of SEQ ID NOS: .1 -462. in various embodiments, the polypeptide is 7S, 70, 65, 60, 55, or fewer amino acid residues in length, in other embodiments-, the polypeptide comprises a tag.

In other aspects are pro vided i solated nuc leic acids encodin the pol ypeptide of any embodiment of the disclosure, recombinant: expression vectors eompfising.the nucleic acid of the disclosure operatively linked to a suitable control, sequence, .recombinant host cell comprising the recombinant expression vector of the disclosure, antibodies that selectively binds to a polypeptide of the disclosure, and pharmaceutical compositions, comprisin one or more polypeptides of the disclosure and a pharmaceutically -acceptable carrier.

In a.further aspect, methods are provided for limiting and/or treating an influenza snfecti o.il, comprising administering to a subject in need thereof an therapeutically effective amount of one or more polypeptides of the disc losure, salts thereof conjugates thereo f, or pharmaceutical compositions thereof to treat -and/or limit the influenza infection. In various embodiments, the polypeptide may be administered mitcosaily, such as by intranasal or inhaled administration, or orally, in other embodiments, the _. subject is immune-compromised and/or is 65 years of age or older.

In another aspect, methods are provided for diagnosing an- influenza ^' infection, or monitoring progression of an influenza infection, ^' comprising

(a) contacting a biological sample from a subject suspected of having an influenza infection with a diagnosiically effective amount of one or more polypeptides of the disclosure, under conditions suitable for bidding of the polypeptide to a viral hemagglutinin protein present in the sample; and

(h) detecting pelypeptide-viral hemagglutinin, binding complexes,

where die presence of such binding complexes indicates that the subject lias an mfjuen¾a infection, or provides measure progression of an influenza infection.

In an her aspect, methods are provided for identifying candidate influenza vaccines, comprising

contacting test compounds with one or more polypeptides of the disclosure under conditions _* suitable for ^' polypeptide binding;

removing unbound test compounds; and

identifying those test compounds that bind to the polypeptide of the disclosure, wherein such test com ounds are candidate influenza vaeeines.

hi one aspect methods are provided for identifying candidate compounds fo preventing, treating, limiting, and/or diagnosing influenza infection, comprising

contacting an influenza hemagglutinin protein with (i) test compounds and S ^' ii a polypeptide of the disclosure, under conditions suitable for binding of the hemaggluti nin protein to the polypeptide of tire present disclosure; and

identifying those test compounds- that outconrpete the polypeptide for binding to the hemagglutinin protein, wherein such test compounds are candidate compounds for preventing, treating, limiting, and/or diagnosing influenza infection.

Description ^' of the Figures

Figure L BioefEkaey of A13 m a Ferret Mode! of Influenza J eetiou _* Ferrets

( - /group) weighing 6X0-110 grams were challenged by aerosol with pathogen ie H 1 I (CA09) influenza vims. Twenty-four hours post-challenge, Al 3 treated animals received a .total of iOmg dose of A 1.3 protein (resulting in a dose of 1.4-16 mg/kg, average 15 rng/kg) using an A13 formulation at - 19.5 mg/nil prepared a. Control animals received the same volume of buffer only (placebo controls). The. 10 mg A 13· .dose was administered via two routes of delivery consisting of .combination of intranasal droplet (5 ^' mg) and intratracheal droplets (5 mg) to achieve distribution of the Al 3 niinibinder into both upper and lower respiratory tract. Ferrets were monitored twice daily for clinical signs of disease using a standardize .scoring scale of 1-10 for various clinical signs including pasture, activity level, ocular discharge, pulmonary function (sneering,, coughing, labored breathing) and changes i food and water consumption and eHminaiton (Left- Panel}. In addition, weight km. {Right Panel) was measured twice daily. Both control and test animals showed t¾o significant changes in temperature (measured twice daily) throughout the study (not shown). Technicians performing scoring evaluation were blinded to group assignment.

Detailed Description

All references cited are herein iacorporafed by reference in their entirety. Within this application, unless otherwise stated, the techniques utilized may be found in .any of several well-known references soc as: Molecular. Cloning: A litberafory.M tnicti (Sambrook, et al, 3989, Cold Spring Harbor Laboratory Press), Gene Expression Technology (Methods in Enzymoiogy, Vol 185, edited by D. Goeddel, 1991. Academic Press,. San Diego, CA), "Guide to Protein Purification" in Methods in Pnzymology (MP. Deutshcer, ed., (1990) Academic Press, Inc.); PGR Protocols; A Guide to Methods ami Application (hinis, et al 1 90, Academic Press, San Diego, CA), Cul re of Animal Cells: A Mam l of Basic Technique, * Ed. ( .I, Freshney. 1987. Liss, Inc. Mew York, NY), Gene Transfer and Expression Protocols, pp. 109-128, ed. E Murray, The Humana S ^t ress Inc., Clifton, NX), and the Ambson 1 98 Catalog (Ambson, Austin,. TX).

As used herein, the singular forms, "a", "an" and "the" include plural referents unless the context clearly dictates otherwise. "And" as used herein is interchangeably used with "or" unless expressly stated otherwise.

As used herein, the amino aeid.residues are abbreviated as follows: alanine (Ala; A), asparagme (Asn; W), aspariic acid (Asp; D), arginine (Arg; ¾ cysteine (Cys; Q, glutamic, acid. (GIu; E), glutarfline (Gin; Q), glycine (Giy; G), histidine (His; H), Jsoleudw (lie; I), leucine (Leu; L), lysine (Lys; K , methionine (Met; M), phenylalanine (Phe; F) _< proline (Pro; P), serine (Set; S), threonine (Thr; T), tryptophan (Trp; W), tyrosine (Tyr; Y) _s and valine (Vai; V).

Ail embodiments of any aspect- of the disclosure can he used in combination, unless the context clearly dictates otherwise.

Unless the contest clearly requires otherwise, throughout the description and the claims, the words 'comprise' , 'comprising', and the lik are to be construed in afl inclusive sense as opposed to -art exclusive or exhaustive: sense; that is to say, in the sense of

"including, but not limited to" Words using the singular or plural number also mclisde the plural and singular number, respectively. Additionally, the words "herein," "above; ^' ' and "below" and words of similar import, when used in this application, shall refer to this- application -as a whole and not to any particular portions of the application.

The description of embodiments of the disclosure is not intended to be exhaustive or to. limit the disclosur to the precise form disclosed. While the specific embodiments of and examples for, the disclosure are described herein tor illustrative purposes, various equivalent modifications are possible within the scope of the disclosure, as those skilled in- the relevant art will recognize.

In a first aspect are disclosed isolated polypeptides comprising an amino acid sequence having at least 70% sequence identity over its length to the amino acid sequence Of any one of SEQ ID OS: 1 -462,

The polypeptides of all aspects/embodiments of the disclosure bind to heraaggfu& n (HA.); binding of the polypeptides to the HA protein can be determined using binding assays as detailed in the examples that follow. The polypeptide of the disclosure can thus- be- used, for example* to treat or detect/diagnose influenza infection. Exemplar polypeptides- of the disclosure have been extensively tested and demonstrated in the examples that follow to strongly bind to the HA protein and inhibit viral entry into ceils. Polypeptide binding to the stern-region of th HA protein is alone sufficient for highly effective in vi o protection, against influenza, which is preferable to antibody-based therapeutics because it avoids immune activation and antibody-dependent: enhancement of disease.

The polypeptides of the disclosure also provide a cheaper, more selecti ve alternative to currently used hemagglutinin binding antibodies, which are costly to produce. The polypeptides of the disclosure can also be used lor in viv& biosensing applications, whereas the antibodies cannot because of their structurally necessary disulfide bonds and difficulty to express robustly.

As disclosed in the -examples that follow, exemplary HA-binding polypeptides of the disclosure have been identified and subjected to extensive mutational analysis against a variety of viral strains. These studies have identified residues where modifications are tolerated, and where they may lead to additional functionality, in vitro testing via deep mutational scanning shows. that a number of these mutations lead to. increased binding specificity against distinct subtypes of influenza, w nch could be highly useiui in a diagnostic role or for therapeutic use against existing, new, or emerging strains of influenza.. In various embodiments, a given amino acid can. be replaced by a residue having similar physioehemJcal characteristics, e.g., substituting one aliphatic residue for another (such as He, Val, Leu, or Ala lor one another), or substitution of one polar residue for another (such as between Lys and Arg; Giu and Asp; or Gin and Asn). Other Such conservative substitutions* e .g.,

substitutions of entire regions having similar hydrophobicity characteristics, are well known. Polypeptides comprising conservative -.ammo acid substitutions can be tested in any one of the assays described herein to confirm thai a. desired activity, e.g. antigen-binding activity and specificity of a nat ve or reference polypeptide is retained. Amino acids can be grouped according to ^' .similarities in the properties of their aide chains (in A. L. Lehninger, in

Biochemistry, second ed,, pp. 73-75, Worth Publishers, New York (1975)): (!) non-polar; Ala (A), Val (V), Leu (L), He (1), Pro (P), Phe (F), Ir (W), Met (M); (2) uncharged polar: Gly (Gl Ser ($), Thr (T), Cys (€), Tyr (Y), Am (N), Gin (Q); (3) acidic: Asp (¾ Giu (B); (4) basic; Lys CK), Arg (R), His (H). Alternatively, naturally occurring residues can be divided into groups based on common side-chain properties; ( I ) hydrophobic; Norieucme, Met, Ala, Val, Leu, lie; (2) neutral hydrophiftc: Cys, Ser, Thr, Asn, Gin; (3) acidic: Asp, Giu: (4) bask: His, Lys, Arg; (5.) residues that influence chain orientation.: Giy, Pro; (6) aromatic: Trp, Tyr, Phe. Non-conservative- ubstitutions .will entail exchanging a member of one of these classes for another class. Particular -conservative substitutions include, for example; Ala into Gly or into Ser; Arg into Lys; Asn into Gin or into 1-1 is; Asp into: Giu; Cys into Ser; Gin into Asn; Giu into Asp; Gly into Ala ^' or into ro; His into Asn or into Gin; lie into Leo or int Val; Leu into He or into Val; Lys into Arg, into Gin or into Giu; Met into Leu, into Tyr or into lie: Phe into Met, into Leu or into Tyr; Ser into Thr; Thr into Ser; Tip into Tyr; Tyr into Trp; and/or Phe into Val, into lie or into Leu.

in various embodiments, the isolated polypeptides comprises an amino acid sequence having at least 75%, 89%, 8-5%, 90%, 9.1 %, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% sequence identi ty over its length to the amino acid sequence of any one of ^' SEQ ID NOS: 1.-462 (Table 1 ), which have been designed as disclosed in the examples that follow, Table !

I» specific embodiments, the isolated polypeptide comprises an amino acid sequence having an ammo acid sequence having at least 70%, 75%, 8¾ _s 85%, 90%, 91%, 92% _s 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% sequence identity over its length to the amino acid sequence of any one of SEQ ID HQS SEQ ID NOS:i, 351-352, 378-379, .381-382, and 455-462.

aB(Q/f)S(F/V)lT(l P}FACQ(^^

(SEQ ID NO: 1}

TSRVRATSKF AAU AAE1AREFGYTVD VQE VNGQ W EVTFD (SEQ ID NO: 331) TSGVRATSKFAAL!AAEIAREFGYTVDYQEKNGEWR FD (SEQ ID NO; 352} ClEQSF1TLFACQ AAEiWRAFGYTVKiMVDNGNCRLHVC(SEQ ID NO: 379) CIBISVTT PFACQiAAEfWRAFGYE V IDDD GNCRUIVC (SEQ ID NO: 78) CQDYI^TDFFACQlAAEiLRDFCiYDCTVQTN GECRVRCC (SEQ ID NO; 381) CQB YRFTMPF ACQ! ALEIERDFGYACTVQT! GECRVRCC (SEQ SO NO; 382)

TS(R/G)VRA SKFAALIAAE1AREFGY VDVQE(V/K

IDNO:455

(SEQ ID NO; 456)

C^¾? ^

G-NM2P) (SEQ iD N ;457)

CSI5 «9- A13jr31 AWQ C-4«~C

aEQSFTTEFAA TAAEiWRAFGYWKIMQ0^GNWRI,HVC (SEQ ID NO;459)

AI3.._r3I _„ AWQT T-40-T

'TIEQSFITLFAAQTAAEIW AFGYTVKIMQDNG WROlVr S:EQ ID NO;460)

CS16342 AO- pE-CG-NI-O

pEIEQSFl EFACQTAAEiWR^FGYTV l VD GN RLHVG-NlJl (SEQ I ^' D NO: 461) CSI6343 A13-NG.. CP

(51EQSFTTLFACQTAAEiWRAF<JYTVKIMVD GNCRLHVF (SEQ ID NO; 462)

In a specific embodiment, the isolated polypeptide; comprises an amino acid sequence having an amino acid sequence having at least 70%, 75%, 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 09%, or 100% sequence identit over its length to the amino acid sequence ofSEQ ID NO: 379 and 459-462.

in various farther embodiment.;, the polypeptide is 75, 70, 65, 60, 55, or fewer amino acid, residues in length, to ether embodiments, tie polypeptide is at least 30, .35, of 40 amino acids in length.

The polypeptides of the disclosure may include additional residues at the N emhn«s, C-tetniinus, or both. Such residues may he any residues suitable for an intended use, including but not limited _. to tags. As used ^' herein, "tags" include general detectable moieties (i.e.: fluorescent proteins, .antibody epitope tags, etc j, therapeutic agent, purification tags (His lags, etc.), linkers, iigauds suitable for purposes of purification and. peptide domains that add functionality to the polypeptides, etc,

in a further aspect, the present disclosure provides isolated ^' nucleic acids encoding a poly peptide of the present disclosu re. The isolated nucleic acid sequence may comprise RNA or DMA, As used, herein, "isolated nucleic acids" are those thai have been removed from their normal surrounding nucleic acid sequences in the genome or in cDNA. sequences. Such, isolated nucleic acid sequences may comprise additional sequences useful for promoting expression and/or purification of the encoded protein, including but not limited to poly A sequences, modified Kozak. sequences, and: sequences encoding epitope tags, export signals, and secretory signals, nuclear localization signals; and plasma membrane localization signals. It will, be apparent to those of skill in the art, based on the teachings herein, what nucleic acid sequences will encode the polypeptides of the disclosure.

In another aspect, the present disclosure provides recombinant expression vectors comprising the isolated nucleic acid of any aspect of the disclosure operativeiy linked to a suitabl control sequence. "Recombinant expression vector" includes vectors that operativeiy link a nucleic acid coding region or gene to any control, sequences capable of effecting expression of the gene product. "Control sequences" operab!y linked to the nucleic acid sequences of the disclosure are nucleic acid sequence capable of effecting the expression of the nucleic acid molecules. The control sequences need, not be contiguous with the nucleic acid sequences, so long as they function to direct the expression thereof. ^' hns, for example, intervening ^■ untranslated yet transcribed sequences can be present between a promoter sequence and the nucleic acid sequences and the promoter sequence can still be considered "operably linked" to the coding sequence. Other such control, sequences include, but are not limited to, polyadenylation signals, termination, signals, and rihosome binding sites. Such expressio vectors ^' can: be of any type known in the art, -including but not limited piasmid and viral-based expression vectors. The control sequeac usedtd drive expression of toe disclosed nucleic acid sequences in a mammalian system may be constitutive (driven by any of a variety of promoters, including, but. not limited to, CMV, SV40, RSV, actin, BP) or inducible (driven by any of a number of inducible promoters Including, but. not. limited to, tetraeyc line, eedy sone, steroid-responsive) . Th construction, of expression vectors for use in irartsfecthvg host cells is well known in the art, and thus can. be accomplished via standard techniques. (See, for example, Sambrook, Fritsch, and Maniatis, in: Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory Press, 1.989; Gene Transfer and ■Expression Protocols, pp. 1 9-128, ed. EX Murray, The Humana Press Inc., Clifton, NJ,), and the Amnion 1998 Catalog (Amnion, Austin, TX). The expression vector must be repSicable in the host organisms either as an episome or by integration into host chromosomal UNA. in various embodiments, the expression vector may comprise a plasniid, viral-based vector, or any other suitable expression, vector.

In a &rtber aspect, the present disclosure provides host cells that comprise the.

recombinant expression vectors disclosed herein, wherein the host cells can be either rakaryotie or eukar otic. The cells can be transiently or stably engineered to incorporate the expression, vector of the disclosure, using Standard techniques in the art including but not limited to standard bacterial transfor ations;, calcium phosphate co-precipitation,

ekctroporation, or liposome mediated-, DEAE dextraa mediated-, polycat me mediated-, or viral .mediated transfection. (See, for example, Molecular C oning: A Laboratory Manual (Samhrook, et. aL 1989. Cold Spring Harbor Laboratory Press; Culture of Animal Celts: A Manual of Basic Technique., 2 ^l Ed. (RJ. Freshney. 1 87. Liss, Inc. New York, NY). A method of producing a polypeptide accordi ng to the d i sclosure is an. add itional part of the disclosure, in. one embodiment,, the method comprises the steps of (a) cuituring a host according to thi s aspec t of the disclosure under conditions conducive to the ex pression of the polypeptide, and (b) optionally, recovering the expressed polypeptide. The expressed polypeptide can be recovered from the cell .free extract, but preferably they are recovered from the culture medium. Methods to recover polypeptide from, cell free extracts or culture medium are well known to the person skilled in the art. in another embodiment, the method comprises chemicall synthesizing the polypeptides.

in a further aspect, the present disclosure provides antibodies that selectively bind to the polypeptides of the disclosure. The antibodies can be polyclonal, monoclonal antibodies, humanized antibodies, and fragments thereof, and can be made using techniques known to those of skill in the art. As used -herein, "selectively bind*' -means preferential binding of the antibody to. the polypeptide of the di sclosure, as opposed to one or more other biologi cal molecul es, structures, ceils, tissues, etc., as is well understood by those of skill in the art. in another aspect the present disclosure provides pharmaceutical compositions, comprising one or more polypeptides of the disclosure and a pfaarmaeetuically acceptable carrier. The, harmaceutical compositions of the disclosure can be used, for example,- in the methods of the ^' disclosure ^' described below, The ^' pharmaceutical composition may comprise in addition to the polypeptide of the disclosure (a) a lyoprotectant; (b) a surfactant; (e) a bulking agent; (d) a tonicity adjusting agent; (e) a stabilizer; i f) a preservative and/or (g) a buffer.

In some embodiments, the buffer in the pharmaceutical composition is a Tris buffer, a bisttdine buffer, a phosphate buffer, a citrate buffer or an acetate buffer. The pharmaceutical composition may also include a lyoprotectant, e.g. sucrose, sorbitol or trehalose. In certain embodiments, the pharmaceutical composition includes a preservative e.g. benzalkonhnn chloride, benzethomum, ehlorobesidine, phenol, m-cresol, .benzyl -alcohol, inetnylp&raberi, propyl araben, ehlorobutanoi, o-cresol, p-eresol, ehioroeresol, phenylniercuric nitrate, thiffierosaJ, benzoic acid, and various mixtures thereof In other embodiments, the pharmaceutical, compositio includes a bulking agent, like glycine. In yet other embodiments, the pharmaceutical composition includes a surfactant e.g., polysorbaSe-2 , polysorbate-40, polysorbate- 60, po!ysorbate-65, polysorbate-80 poiysorhate-83, polo.xame.r- 188, sorbitaii monolaurate. sorbite monopal itate, sorbitan monostearate, sorbitan monooleate, sorbitan trilaurate, sorbitan tristearate, sorbitan irioleaste, or a combination thereof The

pharmaceutical composition .may also include a tonicity adjusting agent, e.g., a compound that renders the-. formulation substantially isotonic or isoosmotic with human blood.

Exemplary tonicity adjusting agents include sucrose, sorbitol, glycine, methionine, mannitol, dextrose, inositol, sodium chloride, arginine and arginine hydrochloride. 1» other embodiments, the pharmaceutical composition additionally includes a stabilizer, e.g., a molecule which, when combined with a protein of interest substantially prevents or reduces chemical and/or physical instability of the protein of interest in lyophilized or liquid form. Exemplary stabilizers include sucrose, sorbitol,, glycine, inositol, sodium chloride, methionine, arginine, and arginine hydrochloride.

The polypeptides ma be the sole active agent In the pharmaceutical composition, or the composition, may further comprise one or more other active agents suitable for an intended use. in a further aspect, the present disclosure: provides methods for .treating and/or limiting aii .influenza infection, mmprising administering to a subject in need thereof a therapeutically effective amount of one or more polypeptides of the disclosure, salts thereof, conjugates thereof, or pharmaceutical- compositions thereof to treat- and/or limit the influenza infection. When the method comprises treating an influenza infection, the one-or ^' rnore polypeptides are administered to a subject that has already bee infected with the ..influenza virus, and/or who is suffering from symptoms ^' (including but not limited to chills, fever, sore throat muscle pains, coughing, weakness, fatigue, and general discomfort) indicating that -the subject is likely to have been infected with the influenza virus. As used herein, "treat" or "treating" means accomplishing one or more of the foJ lowing: (a.) reducing influenza viral titer in the subject; (b) limiting any increase of influenza viral, titer in the subject; (c) reducing the severity of flu symptoms; (d) limiting or preventing development of flu symptoms after -infection; (e) mhihiiing- worsening of flu symptoms; (f) limiting ^' r preventing ..recurrence of flu symptoms in subjects that were previ usly symptomatic for influenz infection.

When the method compri ses ii miting an inii uenza. infection, the one or more polypeptides are administered prophylaciically to subject that is not known to have been infected, but may be at risk of exposure to the influenza virus. As used herein, "limitin ^ means to- limit influenza infection in subjects at risk of influenza: infection. Given the nature: of seasonal influenza outbreaks, virtually all subjects are at risk of exposure, at least at certain times -of the year. Groups at particularly high risk include children under age 18, adults over the age of 65, and individuals suffering from one or more of asthma, diabetes, heart disease, or any type of immunodeficiency.

Whi le not being hound by any mechanism of action, it is believe that prophy lactic protection by the polypeptides of th -disclosure appears to be mediated by limiting or blocking viral replication at the respiratory site of exposure whereas therapeutic protection ma be achieved by curtailing the spread of the vims Into the lower respiratory tract and limiting .inflammation and disease.

in one embodiment, the subject is immune-compromised (including, but not limited to, subjects taking immunosuppressants, subjects with a- disease that compromises the immune system,: such a acquired immune deficiency syndrome, etc.) and/or is 65 years of age or older. The therapeutic and prophylactic activity of the polypeptides of the disclosure is not dependent on die subjec having a properly functioning immune system, and thus the methods are of particular benefit from any subject that does, not have a properly functioning immune system. The methods of the disclosure can be used to treat any individual infected with influenza vims, including but not limited to influenza virus A group 1.

As used herein, a ^therapeutically effective amount" refers to an amount of the polypeptide that is effective for treating and/of limiting influenza infection. The polypeptides are typically formulated as a pharmaceutical, composition, such as those disclosed above, and can be administered via an suitable roiite, including orally, by inhalation spray, ocularly, in dosage unit formulations containing conventional pliaimaceutica!ly acceptable carriers, adjuvants, and vehicles. In one particular embodiment,, the polypeptides are administered ffiucosally, including but not limited to intraocular, Miaied, or intranasal administration. In another particular embodiment, the polypeptides are administered orally. Such particular embodiments can be administered via droplets, nebulizers, sprays, or other suitable formulations.

Dosage regimens can be adjusted to provide the op imum desired response (e.g., a therapeutic.: or prophylactic response). A suitable dosage range may,. tor instance, be 0.1 ug/kg- 100 mg/kg body weight; alternatively, it may be 0.5. ug kg to 50 mg/kg 1 ug/kg to 25 mg kg, or 5 ug kg to 10 mg/kg bod weight The polypeptides .can be delivered in a single bolus, or m y be administered more than once (e.g., 2, 3, 4, 5, or mor times) as determined by an attending physician.

in certain embodiments, the polypeptides of the disclosure- neutralize ^' influenza virus infeeti ity. In various embodiments, the polypeptides of the disclosure prevent influenza virus from infecting host cells by at least 99%, at least 95%, at least 90%, at least 85%, at least 80%, at feast 75%, at least 70%, at least 60%, at least 50%, at least 45%, at least 40%, at least 45%, at least 35%, at least 30%, at least 25%, at least 20%, or at least .10% relative to infection, of host cells by influenza virus in the absence of the polypeptides. Neutralization can, for instance, be measured as described in "Laboratory techniques in influenza." edited by F. -X. Meslin, M. , Kaplan and H. Koprowski (1996), 4th. edition. Chapters 15-1:7, World Health Organization, Geneva.

The polypeptides according to the disclosure can bind to the HA protein with any .suitable- ffinity constant (¾ value) that provides therapeutic, or prophylactic- benefit, in various embodiments, the ¾ value is: lower than 0.2*3 O ^"4 M, 1.0* 10 ^" ¼ 1 ,θ ^" ¾, 1.0* 10 ^" ¾ i .0*10¾, I . 0¾ 1.0*10 ^' ¾, ! .0* I O" ^¾ M, or 1,0*10- ^¾2 M. Affinity constants can for instance be. measured using surface plasmon resonance, i.e., an optical phenomenon that allows for the analysis of real-time biospeeific interactions b detection of alterations in protein, concentrations within a biosensor matrix, for example* using the B! ACO E system (Pharmacia Biosensor AB, Uppsala, Sweden).

The polypeptides made be administered as the sole prophylactic or therapeutic agent, or ma be administered together with (i.e.: combined, or separately) one or more other prophylactic or therapeutic agents, including but not limited to oseliamivir, zanamivir, and ianinamivir.

In another aspect, the present, disclosure provides methods for diagnosing an- influenza infection, or monitoring progression of an influenza infection, comprising

(a) contacting a biological sample from a subject suspected of having an influenza infection with a diagnost iea!i effective amount of one or more polypeptides of the disclosure under conditions suitable for binding of the polypepti de to a viral HA protein present in the sample; and

(b) detecting oi peptide-vkal HA binding complexes,

where the: resence of such binding complexes indicates that the subject has an influenza infection, or provides a measure progression of an influenza, infection.

The..method* -of this aspect of the disclosure can be used, to more .accurately identif patients that may be suf ering from an influenza infection and to thus provide more informed determination of treatment options by an attending caregiver. Individuals at risk: of an influenza infection are as described, above. The methods can also be used to monitor progressio of aft. influenza infection; in this embodiment, the subject is known to be infected, and the methods can be used, for example, as a data point for an attending caregiver to

'determine whether to initiate, modify, o continue a particular course of therapy..

The- biological, sample may be any suitable biological sample including, but not limited to blood, serum, nasal secretions, tissue or other biological material from a subject at risk of infection.

The sample may .first be manipulated to make it more suitable for the method of detection. "Manipulation" includes, but is not limited to treating the sample in such a wa that any influenza .virus in the sample will disintegrate into antigenic components such as proteins, polypeptides or other antigenic fragments. The poly peptides of the disclosure are contacted with th sample under conditions which, allo the formation, of an complex between the human polypeptides and influenza virus or antigenic components thereof that may be present in the sample. The formation of such complexes, if any, indicating the presence of influenza virus in the sample, is then detected and measured by suitable .means. Such methods include, but are not limited to■homogeneous and heterogeneous binding, immunoassays, such as r oimmuneassays (RIA), ELISA _* immunofluorescence, immuno isioc emistry, FACS, BiACORE and Western blot analyses. Suitable conditions to promote binding of the test compounds to one or more polypeptide of the disclosure can be determined by those of ski l l in the art, based on the teachi ngs herein .

The polypeptides of the disclosure for use in this aspect may comprise a conjugate as disclosed above, to provide a tag useful for any detection technique suitable for a given assay. The tag used will depend on the specific detection analysis diagnosts techniques and or methods used. The methods may be carried in solution, or the poiypep ideis) of the disclosure may be bound or attached to a carrier or substrate, e.g., m.scrotiter plates (ex: for EUSA), membranes and beads, etc. Carriers or substrates may be made of glass, plastic (e.g., polystyrene), polysaccharides, nylon, nitrocellulose, or teflon, etc. The surface of such supports ma be solid or porous and of any convenient shape. In one embodiment, conditions are selected to identity test compounds that bind to the polypeptide of the disclosure with a R value lower than 0.2* 10 ^" , LQ*Hr¾ l .0*10 ^"6 M, 1.0*10 ^"7 M, 1.0*10 ^" ¾ L0*I0 ^* ¼ L0*1O ^"U; M, .1.0*!ø ^' *¾ or .! .0*I0 ^'I2 M.

In a further aspect, the present disclosure provide methods for identifying candidate influenza vaccines, comprising

(a) contacting test compounds with a polypeptide of the present disclosure under conditions -suitable for polypeptide binding; and.

(b) identifying those test compounds that bind to the polypeptide of the disclosure, wherein such test compounds are candidate influen vaccines..

As discussed above, the polypeptides of the present disclosure were designed to target the HA protein. Thus, the polypeptides of the disclosure can be viewed as specific binders to HA epitope, similar to antibody binding to a specific epitope. Vaccines can be produced, for example, by selecting .small molecules (ie: mimotopes) that bind to an antibody specific- to a viral epitope. Thus, the present methods involve substituting one or more polypeptides of the present disclosure for the antibody in. such assay to identify candidate influenza vaccines.

Suitable conditions to promote binding, of the test compounds to one or more polypeptide of the disclosure can be determined b those of skill, in the art, based on the teachings herein. The polypeptides of the disclosure" for use in this aspect may comprise a conjugate as disclosed above, to provide a tag useful for any detection technique suitable for a given assay. The tag used will depend on die specific detec-tion a«alysis diagnosis techniques and or methods used, as discussed above. The methods may be carried in solution, or the polypeptide's) of the disclosure may be bound or attached t a carrier or substrate, as discussed above. Based on the teachings herein, it is within the level of skill in the art to determine specific conditions for the various -types of diagnostic assays disclosed in ibis aspect of the disclosure. In one embodiment, conditions are selected to identify test compounds thai bind to ih polypeptide o the di sclosure with & _< s value lower than 0.2 * 10~* M _s i. ,0*iO ^"5 M ₅ 1.0 ^" ¼ 1.0*10 ^" ¾ 1.0*10¾, 1.0*10%, L0*i0 ^"5 ¼ J ,€t* U ^" %1, o 1.0* iO ^*i2 M.

Any suitable test compound may be used, including but not limited to polypeptides, antibodies, nucleic acids, etc,

in another aspect, th present disclosure provides methods for identifying candidate compounds tor treating, limiting, and/or diagnosing influenza infection, comprising

(a) contacting a influenza HA protein with (i) test compounds and (is) a polypeptide Of the- present disclosure, under conditions suitable for binding of the HA protein to the polypeptide of the present disclosure; and

(b) identifying those test compounds that outcompete the polypeptide for binding to the HA protein, wherein such test compounds are candidate compounds for treating, limiting, and/or diagnosing influenza infection,

in thi aspect, the methods identify test compounds that compete with the

polypeptides of the disclosure for binding to HA, and thus such candidate compounds may be useful in any of the other methods of the disclosure disclosed herein. Any suitable test compound can be used, as disclosed above in the eleventh aspect of the disclosure.

in general, competitive, inhibition is measured by means of aft assay, wherein an HA composition is admixed with the polypeptide's) of the disclosure and the test compounds, to be screened. In one embodiment, the test compounds to be screened are present in excess. Protocols. based upon BLISAs are suitable for tise in such competition studies. In certain. embodiments, one may pre-raix the polypeptide's) of th disclosure with varying: amounts of test compounds to be screened (e.g., 1 :10, 1:20, 1:30, 1 :40, 1 :50, .1 :60, 1:70, 1 :80, 1 :90 or 1 :1 0} for a period of time prior to applying to the H A composition. In other embodiments, the polypeptide's) of the disclosure and varying amounts of test compounds to be. screened are admixed during exposure to the HA composition,. Any suitable detection means can be used, binding, in one embodiment; the polypeptide's} of the disclosure are tagged for detection, as discussed above. In this embodiment, the detectable label will decrease in the presence of competitive test compounds. The reactivit of the (labeled) polypeptide of the disclosure in the absence of test compound could serve as one suitable control Preferably, competitive test compounds will, when present in excess, inhibit specific bindin of the olype tide's} of the disclosure to HA by at least 10%, preferabl by at least 25%, more preferably by at least 50%, and most preferably by at least 75% to 90% or even greater.

Exemplary conditions for HA .binding studies can be carried out as disclosed in the ex m les thai follow.

Example 1

Be now protein design holds promise for creating small stable proteins with shapes customized to hmd therapeutic targets. We describe a. massively parallel approach for designing, .manufacturing and screening mini-protein binders thai integrate large-scale computational design, oligonucleotide, synthesis, yeast display screening and next generation sequencing. We designed and tested thousands of ~40-resid«e proteins targeting influenza hemagglutinin, along with 6,286 control sequences probing contributions to folding and binding, and identified high affinity binders. Biophysical characte iza ^' ti on of a subse showed they are extremely stable and, unlike antibodies, do not lose activity following high temperature exposure. Comparison of the sets of binding and non-binding designs, two orders of magnitude larger than an before, enables systematic improvement of the computational model. The designs elicit little or no immune response, and provide potent proph lactic and therapeutic protection against influenza even, after extensive repeated dosing.

Here, we describe an integrated computational and experimental approach that enables the rapid design and iestin of tens of thousands ofcie η ^' όνο mini-protein binders. Our approach exploits advances in both DNA nianu&ciuring and protein design that have led to a fortunate convergence between the upper Emit of the size of oligonucleotides (230bp) that can be synthesized as pools of 1.0,000 or larger, and the Sower limit of the ske of genetically encodable computationally designed proteins (-40 amino acids). To generate binders for a given target, we use R.osetta ^{i M} to design thousands of protein scaffolds with varying shapes, dock these onto the target, optimize the residues at the interface tor high affinit binding, and identify, front die resulting: pool of hundreds of thousands of designs, ~- 1.0,000 with high predicted stabili ty and affinity . This large pool of computational designs, together with controls probing different aspects of the design procedure, are then experimentally evaluated by encoding each individual sequence in a single oligonucleotide, manufacturing the oligonucleotides in parallel, sorting yeast libraries displaying the designs labeled with ^' fluorescent target, and identifying the designs most enriched: for binding using deep

sequencing .

High-throughput computational tlesign of mini-pr»t€½ binders

We .selected influenza Hi hermgglutinin (HA) as a target, as this virus remains a serious public .health concern. We generated virtual scaffold libraries with over 4000 backbone geometries in five different topologies (liHH, Εί-ΪΕΕ, HEE, EBHJB, and HEEH; where H indicates t helix and E β-strand) with or without a diversity of disulfide

connectivities, and then, designed binding. interfaces into these. The core of the interface was seeded with 4-9 notspo residues, while the remainder of the bindin residues were designed novo: The different design s interact with the targets in a myriad of di fferen t ways, resulting in a wide range of interface buried surface area.

We selected for experimental characterization 3,594 disulfide and 3,682 non-disulfide containing HA-hinding designs with an average sequence. identity of -44% (see Methods). To probe the contribution of different aspects of the design process, we also included versions with the amin acids outside the helical interface motif randomly permuted. Oligo pools encodin the design and control sequences were sy thesized, amplified, and co~ tmisibrmed into yeast along with a linearized yeast display vector. The resulting, yeast libraries displaying the proteins were incubated with different concentrations of uorescently labeled target, in some cases after protease treatment to remove poorly folded designs, and cells displaying designs which bound target were retrieved by Fluorescence-Activated Cell Sorting (FACS). After deep sequencin of the various pools, each binder was categorized based: on. the stringency of the condition under whicn it was maximally enriched.

Deep sequencing of the initial yeast transformed pools showed, near complete representation of full length genes; the HA pool contained f 1 ,002, of the 1 3 ,65? sequences ordered (94,4%). Sorting of the library ¹ under conditions of increasing stringency (decreasin concentration of target) sharply reduced the number -of distinct sequences recovered. For HA, the population following sorting against A/PoertoRico/8/1 34 HA concentrations of 3000, 100 and 30 nM contained 1 15, 41 and 29 distinct sequences _* The population fraction of the computational designs increased over that of the scrambled control sequences with increasing selection stringency; -computational design considerably increased the probability of binding the target, with high affinity. The simplest explanation of this increase is that, a significant fraction of the protein fold as designed. The design, population included 3,594 HA designs with .multiple disulfides, i geometrically-allowed -positions. Disulfide containing designs did not have a higher success rate than the non-di sulfide designs (0,5% vs 0.8%) _r consistent with a late .non-instructive role for -disulfides in protein folding. However, w e the design libraries were 'treated, with tr psin before binding selection, only disiilfide-sfabiiized designs were recovered; while not guiding folding, the disulfides clearly confer ^' stability against proteoly sis.

Assessment of computational model

The measured binding activity of a design reflects both th extent to which the protein is folded and the binding affinity of the folded state to the target. Sequences with binding activity in general had lower computed folding energies and binding energies. This is perhaps the largest-scale confirmation of the ability of a computational model to recapitulate protein-protein, interaction to date. The second-order features most strongly associated with binding were; local sequenee-strnctnre compatibil ity and the numbers of contacts across the interface. Based on. these results, we updated the design protocol (see Methods)., and generated: 1 1 ,420 new HA designs for a second round experimental testing in which the success rate increased, torn 1.4% to 3 J (342 new HA binders). The improvement was ^• particularly dramatic in the subset of HB2 seeded designs, improving the success rate 10-fold from 0.19% to 1.9%.

The large dataset provides an opportunity to determine whether extensive molecular dynamics (MD) simulations in explicit, solvent can reprodueibly distinguish bindin and nonhinding designs. We simulated randomly selected non-binders and binders .for a total simulation time of l.OSps (see Methods). There was a strong correlation between the extent of fluctuations of residues at the binding interface and experimental binding acti ity, suggesting that binding site pre-organization is important for binding and that MD simulations ^■' capture this property reasonabl well.

To interrogate the sequence dependence of fold ing and function, we also tested designs containing every single point mutant variant of 6 HA active designs. Residues at the binding interface and in. the protein core were more conserved than surface residues and cysteines tend to be highly conserved in proteins containing disulfides. The contributions of the: noo-hotspot designed contacts differentiated the highest affinity designs from the remainder. For example, mutation of non-hotspot HB1.6928.2 residues Alal I, ^" ftp 19, and Tyr24 drastically decreased binding affinity. The effects of the SSMs were also

computationally modeled considering both changes in binding energy and monomer stability. A reasonable correlation was -found between the predicted and experimentally determined susceptibility of positions to .mutation for three of the six designs for i-IBl, suggesting that they adop structures close to the designed modeU This low resolution structure validation is powerful considering tha no protein purification or traditional biochemical characterization Is required. Finally* the single-sit saturation -mutagenesis (SSM) dataseis were used to. guide generation of higher affinity HA binders (see Methods).

Individual characterization of designed hinders

Six HA binders-, a mix of af ¾ity~ma ui¾d and original designs, were ^■ chemically- -synthesized or expressed in E. coll, purified, and characterized in solution. All designs had circular dtchroism spectra consistent with the design, models, and melting temperatures greater Mian ?0*C. For designs containing disulfides, there was little unfolding at 95 l, white m the presence of the reducing- agent TCEP their stability and thermal reversibility were seriously compromised (data not shown), suggesting correctly formed disulfide bonds. These disulfide containing designs were resistant to trypsin (data not shown). The HA binders bound to HA proteins from two HI ! influenza strains, A i ^> uetto ico/8/i 34 (PR8) and A/Califomia/04 2Q09 (Cat09) _i and the three affinity-matured binders had affinities against C-al0 below 10 nM, Co-crystal structures were determined fo a binder HB1.6928,2.3 and were found to be in excellent agreement with the computational models (mononrer-Q MSD - 0.94 + for HB i .6928.2.3,.

To compare the ability of the designs to survive high temperature exposure with that -of antibodies, we incubated MB 1.6928.2,3 and Ab FI6v3 at 80°C for various ti mes prior to performing binding assays. The .mini-protein binder showed no detectable loss of binding after lhr at high temperature, while Fl6v3 binding activity was reduced -?4% (data not. shown). These results suggest small protein-based therapeutics could overcome the current retirement ^' of cold chain management for -.monoclonal -antibodies:

in vitro assays were carried out tor HB t .6928,2,3,; an affinity- ma tured, disulfide- contalning design. HB I , 6928.2.3 strongly neutralized PR8 and CA0 influenza viruses after 48 hours in culture with EC50 values more titan 100-fold lower than the broadly neutralizing antibody F!bvS^ and H836.6 on a mass basis (data not shown; the B€s« are comparable on molar basis). The increase in protection likely reflects both the reduction in conformational entropy of the binding motif and additional designed interface contacts. HBi ,6928.2.3 protected; mice from influenza both pre- and post- exposure, Intranasal administration (in.) ofMBl .6928.2.3 24hrs prior to lethal, challenge with. CA09 iafiuenza gave 10 % survival at doses- as low as 0.03- mg kg, 100-foid lower on a. mass basis than the Fl ^' 6v3 dos required tor equivalent protection (data not shown). Therapeutic administration of HBL6928.2.3 24hrs post virus challenge resulted in. .100% survival and little (<10%) weight loss (data not shown), and a single 3rag kg dose administered 72 hoars post-challenge gav complete protection and 100% survival (data not shown). Intravenous administration of HBI .6928.2.3 exhibited little protection (not shown), indicating that similar to the on-n arket drug Zanamivir intranasal administration is likely the optimal delivery route for these mmi- proteins.

tmrn unogenici ty stud ies showed that 3 sequential doses of the maw-proteins administered by i.n. or intravenous (i.v.) delivery every two weeks induced Utile or no antibody response (data not shown). The low level of antibody detected was comparable to levels induced by a moose l:gO (negative control) and substantially less than human IgO (positive control). The influenza mini-protein binders retained 1.00% prophylactic efficacy even after four repeat in or i.v doses over a six week period prior to virus challenge- (data sot .shewn) indicating that any immune response and clearance is minimal and not sufficient to interfere with antiviral potency. The lo inimunogeriiciff is likely due to a combination of the very small size and iiypemability of the mini-proteins, and suggests that the binders could be used for prophylactic protection against influenz ove an extended -period of time. To our knowledge this is the first investigation of the immuiiogenichy of de now designed proteins.

Conclusions

Our high throughput computational desi n-experimeniai testing pipeline enables the characterisation of computationally- designed binding proteins ou a scale orders of magnitude greater than previous studies, and generated hundreds of potential anti Flu drug leads. This approach enables probing of individual contributions to folding and binding on thousands of test cases simultaneously. For example, the finding that substitution of designed loop sequences with generic gly-ser linkers reduces binding fitness more than scrambling the order of the designed core residues or substitutin them all with valine. suggests a perhaps underappreciated instructive role for loops in the folding of protein in thi size range.

Different topologies were found to be optimal for the HA interfaces, supporting the hypothesis that no single protein topology/shape is ideally suited to best it all interfaces. The massi vely parallel enabled success despite uncertainties in the design of 40 residue proteins ^■ with multiple surface hydrophobic residues potentially compUcataig folding. Iteration between, data-driveti model improvement and experimental testing should improve both the computational design methodology and our understanding of the determinants of folding and binding—the limited number of native protein structures from which much Of current knowledge is derived is dwarfed by the nearly unlimited number of de novo .proteins that can be designed and tested.

Be novo protein design has the promise to generate pharmaceutically superior molecules that combine the specificity of antibodies with the high stability and

tnaniifacturabiliiy of small molecules. The d novo designed binders described here- exhibit much greater stability to incubation at elevated temperatures and better neutralizatio than the FI6v3 niAh, have -l/30t of the molecular weight, and are readily chemjeally synthesizafoie, which enables the introduction of a wide variety of chemical functionalizaiions.. Likely due to their small size and very high stability, they elicit little immune response even without explicit negative design ^4' ' ^'1 , and. the best of the E A designs provides prophylactic and merapeittic: protection against tnfiuensa infection in viva with potency rivaling or surpassing antibodies. The antibody Fe region is clearly not required for potent protection against in fluenza, and lack; of the Fe could reduce problems with, an tibody enhanced infeciivi ty. Mare generally, hypers-table designed mini-proteins show promise for both therapeutic- and diagnostic applications,

Methods

Mioi-proietn binders design: Mini-protein design began by defmiag a variet of mixed .<$ and OHahiy ^' scaffold topologies using the osettaR-emodel ^m 'blueprint' format ^"1 with the requirement of at least one 10-14 residue helix. The blueprints were used to generat backbones using the Rbsetia Monte Carlo-based fragment assembly tocol^ ^' *. One to three disulfides were added to a subset of these backbones at geometrically allowed positions. Sequence design was performed using the Fast esig ^s protocol with layer control active, alternating between side-chain rotamer optimization and gradient-descent based energy minimization. For each topology, over 10,000 structures were generated and filtered on overall energy per residue and score terms related to backbone qualify, compactness, and disulfide quality..

To match the. mini-proteltt-scaffbids into the desired target heik-binding motifs, we used the osetta Motiffiraft™ Mover" ^'2"5 . The inputs were composed of: ( !) HB36, HB80 or Syt-ILheiical binding motife CPDB IDs: 3 2X, 4BEF and 2NM1 ₅ respectively):; (2) The context target i-o n (Ftu-HA or BoNT ¾Β), ¾»d (3) the above described li rar Of<&? nam mini-protein scaffolds. Matching parameters were set to perform Ml. backbone alignment of the input motif _f with a maximum, backbone RMSE .OA, endpoinis RMSD-LOA, clash_scorej£ut f£=5: and enabling revert_graft_iO _m nativ^_sequence. in the ease of

BoNT/B's Syt-II binding domain, the hotspois were defined as; Met-L Phe-2, Leu-5, Lys-6, Lys-8, Phe >, Phe-I 0, Glu-12, He- 13 (see Extended Data Fig. 1% while for Fla-HA HB80.4 (flB!) binding domain, the hotspots were defined as: Pbe-1 , fie~5, e-9, Phe 3, and for Flu- HA MB36.6 (HB2) binding domain, the hotspots were defined as: Pbe-I, Met-S, Trp- , Phe- 1 . Moii.ftjraft ^TM was followed by Roberta's sequence repack of interface .neighboring residues (except ^' hotspots), cartesian, minimization and filtering using the scoring function. Talaris.2013 or Talaris2014.

After the first round of HA design and testing die Koimogarov-Sminio 2~sample test was used to determine ^: .p-values fo die null hypothesis that the computational metrics of the binding vs _: non-binding designs were drawn from the same underlying distribution. Based on me metrics that correlated strongly with s«ccess(such as those shown Figur 3b) a second round of H A design was performed which incorporated more stringent .filtering o a broader range of metrics. The metrics used, to select the first found of HA designs were delta <3 of bindingiddg filter), shape eoMplementarityCse), and interface buried surface areafSASA). The additional metrics used to select the second generation HA designs and shown to he highly predictive of round one success in the logistic regression model(Figure.3c inset) were: average degree (degree), side-chain probability given phi-psi (p_ aaj>p), percent core by side chain neighbors, phi-psi probability given si.de~ehai« (ratna) and more stringent shape- complementarity _^

Software analysis: All ammo-acid sequences were reverse translated a id eodon optimized for yeast using DNAw.orks ' ^M 2& , _. ^' Sequence identity calculations were performed with a subset of designs using PRALINE ² * after PSI-BLAST global, alignment Sequencing pairing after Illumina deep sequencing was performed by PEAR ² ', Plots and visualizations were created using Seaborn statistical visualization tools ² *,. Python (Python Software Foundation), and python's seifcit-ieara (INRIA).

Gene- pools; Oiigo library pools ordered from either CustomArray or Agilent with all genes 3 ^* and 5 ^? flanked with corrutwn 20bp adaptor segments to allow for amplification.. We obtained conventional oligonucleotides (PGR primers and sequencing primers) from

Integrated D ^' NA Technologies. The raw oligonucleotide pools were amplified with Kapa PfiFi ^' Hotstart™ Ready Mix ( apa Biosystems) using ^' extension primers to add pETCOM™ yeast homologous recon nation segments(4©bp) to each end. AO amplifications were performed using real-time FC on a MiniOpticon ^m (Bio-Rad) for between.9-20 cycles. qPC amplification was critical as over-amplification of gen pools resul ted in low

inmsforniari n efficiency. Amplified pools were size selected on a 2% agarose gel and cleaned fQiagen QiA.quick™ Get Extraction Kit). A second round of qPCR amplification was performed with the same primers on the size-selected pools to generate 2~4pg of ON A. Yeast EBYIOO ceils were transformed with library DNA and linearized pBTCOH™ plasnuc using an established protocol ^' '". After transformation (minimum 1 E?

transtbtoiants),, cells were grown overnight in SDCAA media in 30 nil cultures at 30 °C. passaged once, and stored in 20 mM HBPES 150 mM NaCi pB 7.5, 20% (w/v) glycerol in le? a!iqnots at -80 °C.

Yeast display and deep sequencing: Cell a!iquots were thawed on Ice, eentrifuged at 13,000 r,p.m. far 30 s, restispemfed in 1 e? cells per mi of C-Trp-Ora media and grown- at 30 •*C ^' tor 16 h. Cells were then, cenirifnged for 13,000 r.p.m. and res spended at Ie7 cells per ml SGCAA medi and induced at 30°C for 16-24 h. Ceils were labeled with either PR8

hemagglutinin, or CA/04/09 hemaggluttmn, washed, secondary labeled with SAFE

(!nvitrogen) and anti-cmyc FITC (M.iltenyi Biotech), and sorted by fluorescent gates under various stringency conditions using a Sony SH800. HA target proteins were produced -as previously described** ^' . Cells were recovered overnight at 2.5e5 collected- cells per ml SDCAA. media, whereupon at least 1 e? ceils were spun down at 13,00© r.p.rn. for 1 .nun and stored as. ceil pellets at ^™ S0 before library prep for deep sequencing. Between le? and 4e? yeasts cells were bareoded and prepared for deep se ue cing for each library as previously described ¹ '.

SSM and affinity maturation; SSM libraries for 8 designs were constructed from Agilent gene pools and yeast display selections performed as described: above. Upon, deep

.sequencing the 5 post beneficial mutations at 9 positions in each of the HA designs predicted to result in higher affinity were combined into high-diversity iibraries(<! E?) using wobble bases as guided by SwiflLifa- \ A DNA librar for each design was constructed from assembly PCR using Uitramer ^fM Qiigomre!eoiides ί Integrated DNA Technology) to encode the variable regio . These libraries went : through three increasing stringency sorts rd 1

I ^' OOnM, rd2 I ^' OnM, and rd3 1 M against CA/04/09. Promising constructs were identified through. Sanger sequencing of a subset of the final rd3 pool . Mini-protein xp essio and peptide synthesis: Genes encoding due designed protein sequences were synthesized and cloned into pET-28b(- ) £. coh plas id expression vectors (GenSeripL N-teraikal (SexHis tag and thrombin cleavage site), Piasmids were then transformed into chemically ^■ competent E. colt Lemo2 i ¹ cells (NEB). Protein expression was then induced with 1 mM of isopropyl p-D-thiogaJactopymnoside (1FTG) at 18 °C. After overnight expression, cells were collected and purified by nickel affinity follo ed by F.PI.C siEe-exefusion chromatography (Sisperdex™ 75 10/300 GL, (IE Heal&eare) and Mass Spectrum (MS) verification of the molecular weight of the species in solutio (Thermo Scientific ). Peptide sequences were synthesized from commercial vendors Biomarik or CS Bio in 5 ra quantities with 70% purity requirements. Sequences containing cysteines underwent -standard natural oxidation performed by vendor.

Circular dJehroism. (CD): Far-ultraviolet CD measurements were carried out with an AVIV spectromete model 420 in. PBS buffer (pH 7.4) in a 1 mm path-length cuvette with protein concentration of -0.25 rog/ml (unless . otherwise . mentioned in the text). Temperature melt where from 25 to 95 °C and monitored absorption signal at 222 nm (steps of 2 °C/mm, 30 s of eq iiibration by step). Wavelength scans (195-260 nm) were collecte at 25°C and 95° and again at 2S°C after fast refolding -(~'5 min). four chemically-synthesized, disulfide- containing, mini-proteins were also characterized at a concentration of ~0.2mg/mi -in the absence or presence of 2.5 mM f the reducing agent TCEP

Biolayer interferometry: Binding dat were collected in a Octet ^■ RED96 ¹ (ForieBio, Menlo Park, CA) and processed using the instrument's integrated software using a i : 1.

binding model. For influenza binding proteras., su¾ptavidin-eoated biosensors were dipped -in wells containing biotiny!ated HA proteins (.100 nM) in binding buffer for inimobilkation for 300 seconds, while the binding proteins to assay were diluted from a conce trated stock into binding buffer ( ix PBS pH 7.4 0.01% BSA, 0.002% ^' ϊ ween 20). After baseline -measurement in binding buffer alone, the bindin kinetics were monitored by dipping the biosensors in wells containing defined concentrations of the designed protein (association) and then dipping the sensors back into baseline wells (dissociation). For heat-time courses, the proteins were incubated for defined times at a concentration of 160 nM in PBS buffer ( 150 nM NaC!, -pH ^~ 7.4) and then diluted to 8 » in .the.- final buffer and assayed -as described above influenza HI co-crystal structure For the HA-MBi .6928.2.3 complex,

A/PuertoRico/8 1934 HA and HB i .6928,23 peptide (in 25 mM Iris pH 8.0, 150 mM NaCl) were mixed at a 1:4 molar ratio at a final concentration of 10 mg/ml HA in 25 mM . Tr.is pH 8.0, 1.50 mM NaCl. Crystals, were. rown with a well solution of 5% PEG .3000, 30% PEC} 200, 100 mM MES pFI 6. -using the sitting drop vapor diffusion method and directly flash cooled in liquid nitregen. Date were collected at ALS beam!ine 5.0.3 and processed with H L200& ^S . Phaser* * ³ *^ was used for molecular replacement within enix™* usmg.a single protomer of HI PR8 HA (PDB ID: IRV ⁴¹ as a search model The OB I .6928,2.3 peptide was manually built into Fo-Ff and 2/¾-j¾ maps using clearly defined aromatic residues and. disulfide bonds ^' to confirm the register. The model was refined through iterati ve rounds of model building in COOT ^'4 and refinement in Phenix, TLS groups were automaticall identified by Phenix. Giycans and waters were .manually added and edited in COOT. The final model was assessed with quality metrics within the

Phenix.refine interface which utilizes MoS Probity™ ³⁷ , ]

Meteetitar dynamics sioitilatlons: A total 1 6 independent mini-protein binders were simulated (see Extended Data Fig. 3) using Oromacs™ 5..04 ⁴² and Amber9¾b-1LDN force field ⁴³ . Each protein was simulated n a trielfeic box with explicit water solvent (ΉΡ3Ρ ⁴ ), with box edges at least iOA from the protein, Counterions (Na^/CJ-) were used to neutralize the system. Integratbn time step was 2,0 ps and UNCS ⁴⁵ was applied to constrain all the bonds. Long-range electrostatics (> 12 A) were treated with the Particle-Mesh Bwald method ⁴⁶ . Van der Waals interactions ^' were smoothly switched off between lOA-llA. After inimizafion (10,000 steps}, the system was position restrained, for 200 ps in an VT ensemble (only heavy atoms, restraint^! 0 kJ ^¾ mor ^s ₅ 1 =33» K), followed by 500ps of MPT (T ^::: 3 I0K, restrai ts fcJxmol * A ^"2 , 1 bar) using Berendsen thennostat and batostat ⁷ . For each protein, we then performed 5 independent PT production simulations (T-3 K, ib r) with 5CK ps of initial temperature annealing

thermostat ⁴⁸ and Patxineilo-Rahnmn Barostat . Each production simulation was in the length of 50ns for Flu binders. Snapshots were recorded every SOps, and all of them, were used for .subsequent data analysis.

Inllue ^a ne traikatton assays; 1.00 T0Dse units of virus and half-log dilutions of binders wer incubated in quadruplicate at 37* for t wo hours in 00 ul of neutralization, assay media ('NAM"- media 199, 0,3% BSA, lOmM MEPBSJroM CaClj, pen/strep). 96 well plates with ^■ confluent monolayers of MadhvDarby Canine Kidney Epithelial Ce.ils.{ATCC) cells- were washed twice with PBS fo!Iowed by addition of 50ul of 5ug/mi TPC -Trypsin in NAM and th virus/ binder neutralization mix. Plates were incubated 48 hours and virus detected by combining 50 ul each of assay supernatant and 0.5% turkey red blood cells (TRBG). Virus positive wells that hemaggiutina e the TRBC were identified and the BCm was calculated using Peed- Muench method. Ift ^' vivo im uiteg&iictty and isftae«¾a challenge: Animal studies approved by the University of Washington Institutional Animal Care nd Use Committee. Female, 6-8 week- old AL /c mice (female, 6-8 weeks old, n~5 0/groop) were randomly separated iuto ^' groups, ^' anesthetized ahd were dosed either raaasaliy (La.) or intteveiiously (i.v. ) with PBS (negative control), the broadly neutralizing antibody FI6 (SFFV-FI6V3 IgG, Molecular desi n & Therapeutics, Fred Hatch, Seattle, WA) or mini-protein binders (KB 1 .6928.2,3 or HB36.6). 24-96 hours afterwards the mice were anesthetized with 2.5% isofiurane- and challenged Lti. with 2 ML¾> (fifty percent mouse lethal dose) of A/Calif omia t}4/09 (H1. 1) (CA09). Following challenge, the mice were monitored twice daily for weight loss and survival until up to 14 days post-infection. Animals that reached 30% of weight lost (respect of their initial body weight) were euthanized by carbon dioxide in accordance with our animal protocols. For the immuriogenicit experiment, female.. 6-8 _. week-old BALB/c mice (n=S gro p) were ^' randomly separated into groups, anesthetized and dosed with (in. or i.v) !¾S, or mini-proteins (HB L5702.3 J, MB! .6928.2.3, flB I .6394.2.3, HB3&6, Bot2H0.4 or Bot3194.4) ₉ or monoclonal antibodies mlgCi (Innovative I -MSBC-GF) or hlgG (innovative I -HU-GF-BD . A total of three 3 r 4 do s were administered two weeks apart for in. and iv. administration.: Blood was collected two weeks after each dose by retro orbital bleed using micro-hematocrit capillary tubes (Fisher), Serum was separated by eeniri&gtng the blood samples in polymer gel chemistry tubes (8D>. For mouse experiments, researchers were not blinded to animal identity.

Enzyrne-lissked iatmaa sorben assay (ILISA) BB36.6, HB1.69212.3, MB 1.6394.2.3, migG, hlgG and BSA-specifie IgG an tibody levels in. mouse serum were assessed by ELISA. Maxisorp ^'m (Thermo Scientifk-Nunc) were coated with .100 ngtoett of HB36.6 _f

I IB 1 ,5 ^" 02 J 3. HB 1.69212,3, HB ! .6394.2.3, mlgG ( innovative 1IGM8SC-GF), hlgG

(Innovative 1R-HO-GF-ED) or BS (L A PIEB Biological laboratories, cat no, 7500804} in PBS overnight at 4*C. Plates were blocked with 5% nonfat milk powder in PBS for Ih at room temperature, and then washed three times with wash buffer (FSS-T; phosphaie-bui ered saline containing 0,05% Tweers 20). Samples were diluted in a buffer containing 1% non&t milk, powder in PBS-T, added t the wells, and incubated tor Ihr at room temperature.

Following three washes with PBS-T, plates were incubated with horseradish-peroxidase conjugated goat anti-mouse IgG ( 1/5,000 dilution) secondary antibodie (I ^' hermoFisher 62- 6520) for ih at room temperature. After five washes with PBS-T, 1MB substrate (KPL 52- 00-03) was added to the wells for 30 -mitt. at mom temperature. Color development was stopped by t e ddition of 50 L MCI (1M) ₅ and the plates were read at.450 nm to measure relative optica! densities (O.D.). The average O. ^' D. of blank wells was subtracted to calculate the reported valu s, Statistical and me ^' r .anal ses: Survival analyses were performed, by using the f aplan- Meier log-tank test, A P value of <0.05 was considered to be significant For tnice, the minimu m roup size was determined using weight loss data with 70% of control mice becoming infected with CA09. Comparisons m antibody responses were performed using unpaired student t-test. Based on a standard deviation of 2% m weight loss, a group size of n ~ 5 yields >80% power to detect minimum of a 10% difference between groups in weight loss using a two-sided t-test with an alpha value of 0,05.

Exam le

Four novel derivatives of anti-influenza rnirnbinder HS1.6¾8.2.3 (aka A 13) have been designed, ni&raifactored / synthesized,, purified, and characterized in vitro for their antiviral activity, fcr addition, protocols to purify and formulate the lead peptide A13 to have been developed to support multiple routes of delivery {intranasal, intratraehea.lv subcutaneous and intravenous) and have been validated in a ferret, challenge model of infection for 1IB L6928.2.3.

A 13 Atiti-Fht Peptide:

The amino acid sequence for fiBL69:28,2.3 (aka AI3) is as follows;

€815134 Al3jrd3I Cys-40.Cys

C¾¾H e i1 -Gla-Ser-P-h^

Pher ly-T^r-Thr-Va^

Where Cys.l is disulfide bonded to Cys40, and ^' Cyst 2 is disulfide bonded to Cys35 (SEQ D O:379) The _: production of this peptide was achieved by so l id phase peptide synthesis- and air oxidation, followed by HPI purification. Extensive purification, formulation and mass spectrometry analysis of Al 3 revealed that the majority of the synthetic A13 peptide was properly configured wit both disulfides formed€ysi~€ys40 and Cys!2-Cys3S. However,, minor quantities: of alternatively disulfide bonded -forms of A.13. could also be detected, and the quantities of such alternate forms of A.13.appeared to vary between independent synthesis and purification of different Sots of A 1.3. These .observations, together with bio-potenc data on. A 13 tor blocking influenza vims infection led to the design of A.13 variants thai could retain anti-flu activity, yet had a reduced number of disulfide bonds, l re following four (4) A.13 derivatives have been synthesized.

Two 12) A.13 Analogues:

The sequences of two (2) A13 derivatives axe listed below. CSIS989- Ai3 _„ r3 l_A WQ C s- 8-Cys

€)w-lie-Glu-Gin-Se

Arg~Aia-Phe iiy-Tyr-1!u~

€ys

where CI is disulfide bonded to C40 (SEQ ID NO:459).

This design eliminates the CysI2-Cys35 disulfide b replacing Cys! 2 with Ala and

Cys35 with Trp,

CS19990 A13 t-31 _TA QT Tfcr O-Ttsr

l¾r lfi-Giu-(In-Ser-Fhe-Tbr-Tbr-Leu-Phe-AIa-Aja-G

Arg-Aia-Phe--G!y~Tyr-Tlu~Va!4

Thr (SEQ ID O:460)

This design eliminates- both disulfides by replacing Cys-ί with Thr, Cys 12 with Ala, Cys35 with Trp, and Cys40 with Thr. The I vitro anti-flu activity of these two (2) A13 derivatives is as follows (See Table 2).

I vitro formulation and anti-flu testing, of the acti vibes of A.13 x31 _AWQ. Cys-40* Cys and ,I3_r31._ AWQT Thr~40-T.hr revealed that they had reasonable potency against flu virus strains Ca0409 and Ft. Mtmmoat with ICSO's ranging from 0,01 to 0.002 ug ml. While these two: peptides were not as potent as A13 (1C50 -2 I ^* \ measured for two independent A 1.3 lots), they have the have the feature of fewer { ! .) or no disulfid bonds; with A ! 3jr31_AWQ Cys~40~Cys having only a single f 1) disulfide bond (confirmed: by mass spectrometry), and A13j3 IJTAWQT Thr- 0-Thr has no.disuirl.de bonds. Table 2, Invitro ftio-efflcac of A] 3 -and Analogues. EC50 values for Mhibition of viral replication in vitro reported at ug/rnl for each Test Article.

Two (2) Additioaal A13 Analogues:

Two additional A 13 derivatives were designed, CS16342 Αί 3- pGi.u-CGly~NH2 and ^' CS16343 Al3-NGiyjC:Pro..

Both of the■ following A 13 derivatives preserve the internai Cys l2-Cys35 disulfide which is predicted to be most important to stabilize the binding helix, of AI3 to influenza hemagglutinin. However, since the - and C-teraima! ends are no longer "protected" with the Cys!~Cys40, we designed natural modifications pGlo on N-tena and Giy-MM2 on€-terrn (CS 16342), which are comoxrn natural modifications to bioactive peptides: .that improve their resistance to degradation by exoproteases and can be used on any of the amino acid sequences described herein. In addition, we designed a companion A13 derivative with designed Gly! and Pro40 (CS 16343) residues expected to stabilize the ends by folding / stacking.

C$16342 Al3-Nptitn-CGiy- 2

pGfo-¾le-Giu-Gln>-Ser-Phe-1^

A]a-.i¾>G1y~Tyr-Thr-Va^

.H2 (SEQ ID NO; 461), wherein CI2 is disulfide bonded to C35,

This design eliminates the Cys 1 -Cys4G disulfide by replacing Cys! with pyrog utamk acid. pGiu, and Cys40 with glyeinamxie Giy~NR2. Both modifications are natural modifications of proteins, conimooiy found in bioaetive neuropeptides though to stabilize peptides from digestion by exoproteases. C$.16343 A13-NG1yjCPi*

ID NO: 462 , wherein C!2 is disulfide bb ed to C35.

This design eliminates the Cys -Cys40 disulfide b replacing Cys 1 with Giy, and

Cys40 with Pro, natural amino acids predicted to stabi1i¾e the N- an C- termini of the peptide.

These two peptides CS 16342 A!3- pGiu-CGIy-Ml-i2 aiid CS 16343 AI 3~NGly _„ CPro were synthesized and formulated Summary of At 3 arid tfce Four (4) Analogue Sequences:

Table 3, Amino Acid Sequences of A 13 and Four (4) Analogues.

Mass Spectrometry Analysis of the Disulfides of ΑΪ3:

Two independent lots of synthesized and purifie A 13 CS 159134 have been characterized by BPLC, and mass spectrometry; both prior to aad after exposure to reducing agent dithiothreitoi (DTT). These studies .confirm ^' that the majority of Al 3 peptide behaves .as if it has two disulfides prior to reducing with DTT and that these disulfide are ^. reduced after exposure to DTT. .Furthermore,.. ttempts to oxidize any free thiols in GS 1.59134 A 13 revealed that only a very minor amount of the peptide contains any free Cys residues, which is consistot with MS/MS analysis of intact and trypsin or chymotrypsin digested. A 1.3.

MS/MS analysis of AO, A13 r31 TA ^' QT l¾r-4 Thr,€81599» A13 r31 TAWQT

Thr-40-Thr 13 (CS15134), A t 3_ r3 IJFAWQT Thr-40-Thr (CS 15990), and A l3_r3l_AWQ

Cys-40-Cys (CS 15989), were all analyzed by trypsin digest MS/MS with and without prior reducing agent treatment. The results are primarily consistent with A 13 having two.

disulfides Cysl~Cys40 and Cysl 2~Cys35. A 13_r3 I JVYWQT Thr-40-Thr has no disulfides as expected. CSi 5989 A 13_r31_AWQ Cys-40-Cys has a single disulfide as Cysl -Cys40. These data are consistent with the predicted structures of these peptides.

SDS-PAGE and HFLC Analysis of A13:

A J 3 f om different lots of f rmulated peptide have been analyzed by SDS-PAGB both prior to and after exposure to reducing agent dithiothreitoi (DTT). These results demonstrate that the majority of A13 contains ^' disulfide bonds which make the peptide more compact and fun. in SDS-PAGE with higher mobility than when it is reduced. The majority of the peptide runs as a single hand with mino contaminants. Similar results were observed, by BPLC analysis of A 13 before and after reduction, with DTT where a dear .mobility shift in bulk peak elution. is observed.

The SEC-HPLC of the lots of AO formulated followed by dilution indicates a single ffionomerfc peak, corresponding to the expected retention time of a 4.5kDA peptide. The SDS-PAGB results demonstrate a shift in molecular weight in reduci ng versus non-reducing conditions, indicating ^, formation .of the disulfide bonds.

Q> Analysis of ΑΪ3;

Two independent lots of synfhesked and purified ^' Al 3 were formulated by diluting from 5 rng/mi in water to 0.2 mg ml in PBS, and then analyzed, by circular dkhroism (CD) includin thermal melt study, with identical CD properties observed for the tw Sots (DS).

Ferret Study A 13:

A13 was investigated: for the abilit to protect ferrets from influenza challenge.

^' Ferrets are the standard preclinical model, for evalu ting new influenza aativirals and success in this .model generally translates to success in human clinical testing. Ferrets were challenged by aerosol exposure to H i Ml (CA09) and. then 24 hours after challenge, a total of 10 mg of A 13. was admhiistered via both intranasal (5 mg) and intratracheal inoculation (5 mg) to mimic routes of deliver achieved by nose/oral nebulizers used in humans. The results in Figure 1 show that Al 3 afforded profound protection ^' from clinical signs of disease in ferrets. Control animals that received only saline exhibited significant clinical signs of disease (lethargy, sneezing coughing, labored breath, reduced appetite. Figure I Left Panel) and weight loss (Fig-are I Eight Pane!), in contrast; animals thai recei ved ΑΓ3 exhibited minimal eMnleai signs of disease (Figu re I Left Panel) and tm significant weight loss

(Figure Ri ht Panel). Evaluation of effects of the A J in reducing viral replication in lung tissue and nasal secretions arid reducing lung lafiammaiion is in progress. These data provide strong evidence that AB anti-flu minibinder mediates■ ^' rofound protection imm Influenza disease m the standardised preclinical ferret model for ft ens infection.

References

1. Kirtteiiig, J. R. & Cochran, J. R. Engineered Rnottin peptides as ^■ diagnostics, therapeutics, and drug delivery vehicles. Cwr, Opin. Chem. Biol 34, 143-150 (2016).

2. Gebauer, M. & Sfcerra. A. Engineered protein, scaffolds as next-generation antibody therapeutics.

Cur. Opm. Chem. Hfal, 13, 245-255 (2009).

3. Zahnd, C. ei ah Efficient tumor targeting with high-affmity designed atrkyrin repeat proteins; effects of affinity and molecular .size. Cancer R s. 7¾ 1595- 1605 (2010).

4. VazQuez-toinbardi R. i' l

Dbcev. IbdayM, 1271 -1.283 (2015).

5. Bhard aj„ Q. ei at Accurate de novo design of hyperstabie constrained peptides. Nature 538,

329-335 (2016).

6. Rocklin, G. J. ei l Global analysts of pr tein ^■ folding usin massively parallel design,, synthesis, and testing. Scietwa 3S7, 168-175 (2017).

7. Berger, S. ei el Computationally designed high specificity inhibitors delineate the roles of 8CL2 farnily proteins in cancer. Elt e 5, (20.16).

8. Procko, E. ettnl A computationally designed inhibitor of an Epsiein-Barr viral Bci-2 protein induces- apoptosis in infected ceils.-C / 157, 1644-1656 (201.4).

9. Cieary, M. A. el at Production, of comple nucleic acid libraries using highly parallel in situ oligonucleotide, synthesis. Nat. Methods 1, 2 1-248 (2004).

10, Sun, M. G. F,, Seo, M.-H,, ko, S„ Corbi-Verge, G. <¾ Kim, P, M. .Protein engineering by highly parallel screening of computationally designed variants, SciAd 2, -el 600692 (2036).

i 1. Fleishman, S. J, el at RosettaSeripis: a scripting iaftguage interface to the Rosetta

raacronioiecaiar modeling suite, FLoS One 6, e20161 (2 1 i ),

1 . Hurt, A. G. ei at Antiviral resistance during the 2009 influenza A HI Nl pandemic: public health, laboratory, and clinical perspecti ves, lancet Infect 7%. 12, 240-248 (2012),

13. Slteer, A- Spasmodic dysp onia and hottdinnm toxin: experience from the largest treatment series, Em-. J. Neurol 1% 28-3 (2010),

14, oday, M. T. iff ^' at A Computationally Designed Hemagglutinin Stein-Binding Protein Provides in Vivo Protection from Influenza independent of a Host Immune Response. PLo$ Paihog. 12,

15. Whitehead, T„ A. el ah Optimization 0? affinity, specificity and. function of designed influenza inhibitors using deep ^' sequencing. Nat. io echnoL 30, 543.-548 {2012}.

1.6. Fleishman, S. J- el at Computational design of proteins targeting the conserved stem, region of influenza eniagglutinin. Sciatce 332, 816-S21 (20! 1),

17. Benrtsson, R, P. A., Peng, L., .Dong, M. & Stenraark, P.. Structure of Botulinum neurotoxiit B binding domain in complex with both syrtapiotagmin H and GD I a, {2013).

doi 10.22 ί 0 pdb4kbb/pdb

IS, Corti,.f . vi ol.. A neutndizing. antibody selected from plasma cells■■mat binds to group: 1 and group 2 -i fluenza A hemagglutinins. Science 333, 850-856 {20 ! 1).

19. Cass, L. M. R., Efthymioponlos, C. & Bye, A. Pharmacokinetics of Zaaatmvir After Intravenous, Oral, Inhaled or Intranasal Administration to Healthy Volunteers.€¾«■ Pharmacaktitet 36, 1 - 1 6 (1999).

20. King,. C, el a .. Removing T-ce!l epitopes with Computational protein -design. Frm, al Acad.

M U. S. A. 11 L 8577--S582 (2014).

21. Huang _s P.-S. et ah RosettaReraotfc!: a generalized framework for flexible backbone protein

design. ?l.4)S One 6, «24109 (2011),

22. Lin, Y.-R. et al Control over overall shape and size in de novo designed proteins. Prpc, Nail Amd. Set. a Λ. 112, E547S-85 (20 ! 5).

23. .Koga, , a ah .Principles for designing ideal protein .structures. NatumAM, 222-227 (2 12).

24. Siiva, D.-A., Correia, B. B. & Procko, E. Motif-Driven Design of Protein-Protein interfaces, in Methods in Moiectdar Biology 285-304 (2016).

25. Hoover,, .D, M. U A Works: an automated method: for designing oligonucleotides for PCR-based gene synthesis, ftfu iefe Acids Res. 3% 43e--43 (2002),

26. Baworio, P. & Heririga, 1 .PRALINE: A Versatile Multiple Sequence Alignment Toolkit, i

Methods it? M leadar Bi gp 245-262 (2013),

27. Zhang,. J.,■Kobert, K.., Fiotiri., T. & - ^' Stamatakis, A. FEAR: a fast and accurate lUumina Paired-End reAd roergeR, imnfi rmatics 30, 61 -6 0 (2013).

28, askom, M. ei l seaborn; vO.7.1 (June 2 16).. dot; 10.528 l¾e»0do.5484

29, Chao, G. el at isolating and engineerin human antibodies using yeast surface display. ^' Ate

Pm oc. i, 755-768 (2006).

30, Benamil, L. ₍ Perez, 1 M, _» Beik, J, & Hsieh, C..--M. An improved yeast .-transformation, method for the generation of very large humaiti antibody libraries, Prolein Eng. Dm. Set 23, 155-1 9 (2010). 3 L Jacobs, T. M, , Yumerefendi,-H, _s Kuhlmati, B. & Leaver-Fay, A, SwiftLib: rapid degencrate- codon-library optimization through dynamic programming. Nucleic Acids- ties. 43, e34 (2015).

32. Jin, R., Rmnmei. A.., Binz, T. & Bmnger, A. T. Botulinum neurotoxin B recognizes its protein receptor with high affinity and specificity.. Niatw 444, 1092- 1095 (2006). 33. absch, W. DS. dem Crystal fogr:,D Biol. Crmmifogr, 66, 125-132 ^■ (201.0).

34. Emsiey, P. & Cowtan, K, Coot: model-building tools for molecular graphics. Acta Qy&taUojyr. D Biol Crystal!ogr. ίκ% 21.26-2132 (2004).

35. McCoy, A. J. et al. Phasercrystalipgraphsc software. J, Apjpl CrysiaUogr. 40, 658-674 (2007). 36. Btt iger, A, T, Free R value: a novel statistical quantity for assessing t e accuracy of crystal sfmctores.. Na re-SSS, 472-475 (1992).

37. Chen, V , B. et al. MolProbity: all-atom structure validation for macromoiecular crystallography.

Acta Cfyxtel!ogr. D Biol Ciystathgr. 66, 12- 1 (2010).

38;, Otwinowski, Z. Sc - Minor, W. 20j Processing of X-ray diffraction -data collected in oscillation mode . in Methods in Emymohgy 307--326 ( 1997),

39. McCoy, A. 1 et at Phaser cfystaUogfaphic software. J. Appl. Crytfalbgr. 40, 658-674 (2007).

40. Adams, P. D, et al. The Phemx software for automated- determination of acfomoieeular

structures. Methods 55, 94- 106 (201 }.}.

41. Gamblm, S. I. at al T¾e structure and receptor binding properties of the 1918: influenza

¾emagjgbtttmn. Sciance 303, ! 38 ^» 1842 (2004);

42. Van Der Spoel, D. et al iROMACS: fast,. flexible, and free. J. Comp l Chem. .26, 17 1- 1 7 ί .8 (20Θ5),

43. Lindorff-Larsen, . & al Improved side-chain, torsion potentials for the Amber ff99S.B protein force field. PwMm 78, 1 50-1958 (2050},

44. Jorgensen, W. L., Cnandwekhar,. i., Madura,, J. D. _v impey, R, W, ^' 8c Klein, M. 1... Comparison ef simple potential fimctions for simulating liquid water. J. kem. Pirn. 79, 926 (1.983).

45. Hess, B.„ Bekl er, H., Berendsen, H. J. C. & Johannes Q E. L1NCS: A linear constraint solver for molecular simulations.,/ Comput. Ckem, 18, 1 63-1472 (1 97).

6, Essmapn, U, et al A smooth particle mesh Ewaid method. Chem. Pkyss. 103, 8.577 {1095}. 47. Berendsen, H. J. C... Transport Properties Computed by Linear .Response through Weak Coupling to a Bath, in Computer Simtikttmn in Materials Science 139-155 (.1991).

4.8·. Bussi,.<3., Dooadio. P. Pamnello, M..Canonical sam lin through velocity reseaiing. J. Chem.

Fhys. 1.26, 14.1 1 (2007).

49, Nose ^" , S, & Klein, . L. Constant pressure molecular dynamics- for molecular systems. Mot.

Rhys. SO, 1055-1076 (1983).