Methods of predicting and mitigating tryptophan oxidation

Abstract

Provided herein are methods of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, methods of modifying a polypeptide comprising a tryptophan residue (e.g., an unstable tryptophan residue) to reduce the likelihood of oxidation of the tryptophan residue, and polypeptides produced by such methods.

Description

[0001] METHODS OF PREDICTING AND MITIGATING TRYPTOPHAN OXIDATION

[0002] CROSS REFERENCE TO RELATED APPLICATIONS

[0003] This application claims priority to U.S. Patent Application No. 63 / 830,239, filed on June 25, 2025, and U.S. Patent Application No. 63 / 664,650, filed on June 26, 2024, the entire contents of which are incorporated herein by reference in their entirety.

[0004] SEQUENCE LISTING

[0005] The instant application contains a Sequence Listing which has been submitted electronically in XML format and is hereby incorporated by reference in its entirety. Said XML copy, created on June 24, 2025, is named 50474-346WO3_Sequence_Listing_6_24_25 and is 18,590 bytes in size.

[0006] FIELD OF THE INVENTION

[0007] Provided herein are methods of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, methods of modifying a polypeptide comprising a tryptophan residue (e.g., an unstable tryptophan residue) to reduce the likelihood of oxidation of the tryptophan residue, and polypeptides produced by such methods.

[0008] BACKGROUND

[0009] The amino acid residue tryptophan (Trp) is a key amino acid in protein-protein interactions. For example, tryptophan residues are frequently located (and may be strategically placed) in the antigenbinding sites of protein therapeutics, e.g., engineered antibodies, to enhance binding affinity and / or specificity.

[0010] However, tryptophan has a high susceptibility to oxidation under stressed conditions. Tryptophan oxidation has various undesirable consequences, including detrimental effects on protein stability, function, and efficacy, and presents a significant challenge in protein design. For example, oxidation is a major degradation pathway in monoclonal antibodies (mAbs), e.g., mAb therapeutics, where it can negatively impact binding affinity, stability, potency, viability and overall developability. For example, oxidation of tryptophan residues within mAb complementarity-determining regions (CDRs) can disrupt antigen binding affinity and specificity, and oxidation within the Fc domain may interfere with vital interactions with receptors like FcRn and Protein A. Tryptophan oxidation can also exacerbate the susceptibility of a protein (e.g., mAb) to degradation pathways, such as fragmentation and aggregation, thus impacting thermal stability, serum half-life, and immunogenicity, further complicating the stability landscape.

[0011] A major challenge regarding tryptophan oxidation in therapeutic proteins is the difficulty of predicting oxidation liability a priori. Consequently, such liability is often identified late in the drug discovery process, or even during development. Further, once a high-risk tryptophan residue is identified, mitigation strategies may not be straightforward, especially if the tryptophan residue is within a binding site of the protein. Existing mitigation strategies may also increase drug cost, decrease drug potency, and / or lead to formulation issues. While molecular redesign may be considered, the unique properties of the tryptophan side chain often render it irreplaceable, such that deletion or substitution of tryptophan is likely to negatively impact protein function (e.g., binding).

[0012] Accordingly, there is a need in the art for methods of identifying tryptophan residues of a polypeptide that are likely to be oxidized (e.g., in silico methods that allow identification of such residues early in drug discovery and molecule development) and methods of modifying polypeptides comprising a high-risk tryptophan residue to reduce the likelihood of tryptophan oxidation without deleting or replacing the tryptophan residue.

[0013] SUMMARY OF THE INVENTION

[0014] In one aspect, the disclosure features a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of (a) determining the likelihood of oxidation of a tryptophan residue of the polypeptide, the determination based on (i) the electrostatic potential of a region comprising the tryptophan residue, and (ii) the solvent-accessible surface area (SASA) of the tryptophan residue; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue; (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region. In some aspects, the method comprises identifying a tryptophan residue of the polypeptide that is likely to be oxidized.

[0015] In another aspect, the disclosure features a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of (a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises (i) determining the electrostatic potential of a region comprising the tryptophan residue; and (ii) determining the solvent-accessible surface area (SASA) of the tryptophan residue; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue; (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

[0016] In another aspect, the disclosure features a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of (a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises: (i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above 50 A2; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue;

[0017] (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or

[0018] (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue.

[0019] In another aspect, the disclosure features a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of: (a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises: (i) determining that the tryptophan residue is in a region having a negative electrostatic potential of -0.5 or less; and (ii) determining that the tryptophan residue has a SASA that is above 50 A2; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises: (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue.

[0020] In some aspects, the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced.

[0021] In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

[0022] In some aspects, the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH or VL comprising the tryptophan residue.

[0023] In some aspects, the polypeptide is an antibody or antibody fragment, the tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR loop comprising the tryptophan residue.

[0024] In some aspects, the method comprises determining that the tryptophan residue is in a region having a negative electrostatic potential. In some aspects, the tryptophan residue is identified as being in a region having the negative electrostatic potential using the Adaptive Poisson-Boltzmann Solver (APBS) software package.

[0025] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -0.5 or less. In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .0 or less. In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .5 or less.

[0026] In some aspects, the method comprises determining that the tryptophan residue is solvent- exposed.

[0027] In some aspects, the method comprises determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, the reference SASA level is 50 A2.

[0028] In some aspects, identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises (i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above a reference level.

[0029] In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues that contribute to a negative electrostatic potential of the region comprising the tryptophan residue.

[0030] In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 20 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 15 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 1 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are at a position in the amino acid sequence within 10 amino acid residues of the tryptophan residue.

[0031] In some aspects, the one or more candidate amino acid residues of the polypeptide comprise an aspartic acid (D) residue or a glutamic acid (E) residue at position n-1 relative to the tryptophan residue.

[0032] In some aspects, the modifying comprises introducing an amino acid substitution in the amino acid sequence. In some aspects, the amino acid substitution mutation comprises replacement of a negatively charged amino acid residue with a neutral or positively charged amino acid residue. In some aspects, the amino acid substitution mutation comprises replacement of a D residue or an E residue at position n-1 relative to the tryptophan residue. In some aspects, the amino acid substitution mutation comprises replacement of a neutral amino acid residue with a positively charged amino acid residue. In some aspects, the positively charged amino residue is a lysine (K) residue or an arginine (R) residue.

[0033] In some aspects, the modifying comprises deletion of one or more amino acid residues in the amino acid sequence.

[0034] In some aspects, the modifying comprises addition of one or more amino acid residues in the amino acid sequence.

[0035] In some aspects, the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a neutral electrostatic potential.

[0036] In some aspects, the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a positive electrostatic potential.

[0037] In some aspects, the tryptophan residue is at an interaction surface of the polypeptide. In some aspects, the interaction surface is a protein-protein interaction surface.

[0038] In some aspects, the modifying does not substantially affect a binding interaction of the polypeptide.

[0039] In some aspects, the polypeptide is an antibody or an antibody fragment. In some aspects, the tryptophan residue is in a Fab domain of the antibody or antibody fragment. In some aspects, the tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment. In some aspects, the tryptophan residue is in a framework region of the antibody fragment. In some aspects, the tryptophan residue is in a CDR loop of the antibody or antibody fragment. In some aspects, the modifying does not substantially affect the binding affinity of the antibody or antibody fragment.

[0040] In some aspects, the polypeptide is a cytokine.

[0041] In some aspects, the polypeptide is a VHH.

[0042] In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 85%.

[0043] In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 35%, as assessed in an oxidation assay using 2,2'-Azobis(2-amidinopropane) dihydrochloride (AAPH), wherein the length of the assay is about 16 hours. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 30%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 hours. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 25%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 hours. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 1%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 hours. In some aspects, the oxidation assay is performed at 37°C. In some aspects, the oxidation assay is performed at a pH of 5.5.

[0044] In some aspects, the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced.

[0045] In another aspect, the disclosure features a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of oxidation of a solvent-exposed tryptophan residue of the polypeptide, the method comprising modifying the amino acid sequence to increase the electrostatic potential of a region comprising the solvent-exposed tryptophan residue, wherein the modifying comprises (a) replacing one or more candidate amino acid residues within the region with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the solvent-exposed tryptophan residue; (b) deleting one or more candidate amino acid residues within the region, wherein the one or more candidate amino acid residues do not include the solvent-exposed tryptophan residue; or (c) inserting, within the region comprising the solvent-exposed tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

[0046] In some aspects, the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the solvent-exposed tryptophan residue is reduced.

[0047] In some aspects, the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the solvent- exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

[0048] In some aspects, the polypeptide is an antibody or antibody fragment, the solvent-exposed tryptophan residue is within the fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the Fab domain comprising the solvent-exposed tryptophan residue.

[0049] In some aspects, the polypeptide is an antibody or antibody fragment, the solvent-exposed tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the solvent- exposed tryptophan residue is the electrostatic potential of the VH or VL comprising the solvent-exposed tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment, the solvent-exposed tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the CDR loop comprising the solvent-exposed tryptophan residue.

[0050] In some aspects, the method comprises determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, the reference SASA level is 50 A2.

[0051] In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues that contribute to a negative electrostatic potential of the region comprising the solvent- exposed tryptophan residue.

[0052] In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 30 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 20 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 15 A of the solvent- exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 10 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 5 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 1 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

[0053] In some aspects, the one or more candidate amino acid residues of the polypeptide are at a position in the amino acid sequence within 10 amino acid residues of the solvent-exposed tryptophan residue. In some aspects, the one or more candidate amino acid residues of the polypeptide comprise an aspartic acid (D) residue or a glutamic acid (E) residue at position n-1 relative to the solvent-exposed tryptophan residue.

[0054] In some aspects, the modifying comprises introducing an amino acid substitution in the amino acid sequence. In some aspects, the amino acid substitution mutation comprises replacement of a negatively charged amino acid residue with a neutral or positively charged amino acid residue. In some aspects, the amino acid substitution mutation comprises replacement of a D residue or an E residue at position n-1 relative to the solvent-exposed tryptophan residue. In some aspects, the amino acid substitution mutation comprises replacement of a neutral amino acid residue with a positively charged amino acid residue. In some aspects, the positively charged amino residue is a lysine (K) residue or an arginine (R) residue.

[0055] In some aspects, the modifying comprises deletion of one or more amino acid residues in the amino acid sequence.

[0056] In some aspects, the modifying comprises addition of one or more amino acid residues in the amino acid sequence. In some aspects, the modifying converts the electrostatic potential of the region comprising the solvent-exposed tryptophan residue from a negative electrostatic potential to a neutral electrostatic potential.

[0057] In some aspects, the modifying converts the electrostatic potential of the region comprising the solvent-exposed tryptophan residue from a negative electrostatic potential to a positive electrostatic potential.

[0058] In some aspects, the solvent-exposed tryptophan residue is at an interaction surface of the polypeptide. In some aspects, the interaction surface is a protein-protein interaction surface. In some aspects, the modifying does not substantially affect a binding interaction of the polypeptide.

[0059] In some aspects, the polypeptide is an antibody or an antibody fragment. In some aspects, the solvent-exposed tryptophan residue is in a Fab domain of the antibody or antibody fragment. In some aspects, the solvent-exposed tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment. In some aspects, the solvent-exposed tryptophan residue is in a framework region of the antibody fragment. In some aspects, the solvent-exposed tryptophan residue is in a CDR loop of the antibody or antibody fragment. In some aspects, the modifying does not substantially affect the binding affinity of the antibody or antibody fragment.

[0060] In some aspects, the polypeptide is a cytokine.

[0061] In some aspects, the polypeptide is a VHH.

[0062] In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 85%.

[0063] In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 35%, as assessed in an oxidation assay using 2,2'-Azobis(2-amidinopropane) dihydrochloride (AAPH), wherein the length of the assay is about 16 hours. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 30%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 hours. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 25%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 hours. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 1%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 hours. In some aspects, the oxidation assay is performed at 37°C. In some aspects, the oxidation assay is performed at a pH of 5.5.

[0064] In some aspects, the disclosure features a polypeptide comprising a tryptophan residue having a low likelihood of oxidation, wherein the polypeptide is produced by any one of the methods provided herein.

[0065] In another aspect, the disclosure features a method of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, the method comprising the steps of: (a) measuring the electrostatic potential of a region comprising the tryptophan residue; and (b) measuring a solvent- accessible surface area (SASA) of the tryptophan residue; wherein the electrostatic potential of the region comprising the tryptophan residue and the SASA of the tryptophan residue together identify the tryptophan residue as likely to be oxidized. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

[0066] In some aspects, the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue.

[0067] In some aspects, the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH or VL comprising the tryptophan residue.

[0068] In some aspects, the polypeptide is an antibody or antibody fragment, the tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR loop comprising the tryptophan residue.

[0069] In some aspects, the method comprises determining that the tryptophan residue is in a region having a negative electrostatic potential. In some aspects, the tryptophan residue is identified as being in a region having the negative electrostatic potential using the APBS software package.

[0070] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -0.5 or less. In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .0 or less. In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .5 or less.

[0071] In some aspects, the method comprises determining that the tryptophan residue is solvent- exposed. In some aspects, the method comprises determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, the reference SASA level is 50 A2.

[0072] In some aspects, determining whether a tryptophan residue of a polypeptide is likely to be oxidized comprises: (i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above a reference level.

[0073] In some aspects, the tryptophan residue is at an interaction surface of the polypeptide. In some aspects, the interaction surface is a protein-protein interaction surface.

[0074] In some aspects, the polypeptide is an antibody or an antibody fragment. In some aspects, the tryptophan residue is in a Fab domain of the antibody or antibody fragment. In some aspects, the tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment. In some aspects, the tryptophan residue is in a framework region of the antibody fragment. In some aspects, the tryptophan residue is in a CDR loop of the antibody or antibody fragment.

[0075] In some aspects, the polypeptide is a cytokine.

[0076] In some aspects, the polypeptide is a VHH.

[0077] BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Fig. 1 is a schematic diagram showing oxidation products of tryptophan (Trp) generated using the reagent 2,2'-Azobis (2-amidinopropane) dihydrochloride (AAPH). AAPH generates peroxy radicals (ROO-) that react with Trp to form a Trp radical, which further forms the oxidation products 4, 5, 6, or 7-hydroxy tryptophan (hydroxy tryptophan, mass of tryptophan (M)+16 Daltons (Da) molecular weight (MW)), oxindolealanine (M+16 Da MW), N-formylkynurenine (M+32 Da MW), or kynurenine (M+4 Da MW).

[0079] Fig. 2 is a stacked bar graph showing Trp oxidation frequency in a set of tested antibodies at each of the indicated complementarity-determining region (CDR) loops under AAPH stress conditions. Trp sites in each CDR loop were identified as modified (reactive) (light gray) if the oxidation frequency at that site was at or above 35%, and were identified as unmodified (non-reactive) (dark gray) if the oxidation frequency at that site was below 35%.

[0080] Fig. 3 is a diagram showing various structural parameters and environmental features that may contribute to risk of Trp oxidation in a fragment antigen-binding (Fab) domain of an antibody. CDR = complementarity-determining region. Epot = the electrostatic potential of the region comprising the tryptophan residue. SASA = solvent accessible surface area. Hbond = hydrogen bond. WW= Wimley- White whole residue hydrophobicity scale. El = Eisenberg hydrophobicity scale. BM = Bandyopadhyay- Mehler hydrophobicity scale. KD = Kyte-Doolittle hydrophobicity scale.

[0081] Fig. 4A is a dendrogram showing features that may be associated with a Trp site. The features are described in Table 1 .

[0082] Fig. 4B is a chart showing the relative importance of each of the indicated features in determining likelihood of oxidation of the Trp, as determined using random forest (RF) regression and Shapley additive explanations (SHAP) values. The features are described in Table 1 .

[0083] Fig. 5 is a scatter plot showing Trp oxidation rate as a function of Trp solvent-accessible surface area (SASA) (in A2) for Trp sites in a set of tested antibodies. Circles with dots indicate Trp residues that were classified as at a high risk for oxidation, and open circles indicate Trp residues that were classified as at a low risk for oxidation. The horizontal dashed line shows an oxidation rate threshold of 30%; the vertical dashed line shows a SASA threshold of 50 A2. The top left quadrant represents Trp residues that were buried and oxidized (false negatives using SASA alone as a predictor of oxidation); the bottom right quadrant represents Trp residues that were exposed but not oxidized (false positives using SASA alone as a predictor of oxidation).

[0084] Fig. 6 is a scatter plot showing Trp oxidation rate (measured as the percent of Trp sites that were modified (oxidized) under AAPH stress; “% Trp oxidized”) as a function of Trp solvent accessible surface area (SASA) (in A2) for Trp sites in a set of tested antibodies. The horizontal line shows an oxidation frequency threshold of 35%; the vertical line shows a SASA threshold of 50 A2. The top left quadrant represents Trp residues that were buried and oxidized (false negatives using SASA alone as a predictor of oxidation); the bottom right quadrant represents Trp residues that were exposed but not oxidized (false positives using SASA alone as a predictor of oxidation).

[0085] Fig. 7 is a scatter plot showing Trp oxidation rate (“% Trp oxidized”) as a function of the electrostatic potential of the region comprising the tryptophan residue (Epot). The horizontal line shows an oxidation frequency threshold of 35%; the vertical line shows an Epot threshold of 0. The bottom left quadrant represents Trp residues having a negative Epot that were not oxidized (false positives using Epot alone as a predictor of oxidation); the top right quadrant represents Trp residues having a positive Epot that were oxidized (false negatives using Epot alone as a predictor of oxidation).

[0086] Fig. 8 is a scatter plot showing Epot as a function of SASA (in A2) for Trp sites in a set of tested antibodies. Heat map shading indicates level of Trp oxidization. The horizontal dashed line shows an Epot threshold of 0; the vertical dashed line shows a SASA threshold of 50 A2.

[0087] Fig. 9 is a box plot showing the rate of Trp oxidation for Trp sites in a set of tested antibodies, with data grouped by (1 ) a positive Epot (Epot>0) or a neutral or negative Epot (Epot<=0) and (2) a SASA at or less than 50 A2(SASA<=50), or greater than 50 A2(SASA>50).

[0088] Fig. 10 is a scatter plot showing Epot as a function of SASA for Trp sites in a set of clinical-stage antibodies. MabG H3, MabG L3, MabA L3, MabH, MabF, and MabE represent Trp residues at which an oxidation rate of less than 35% was observed upon AAPH stress. MabD, MabC, MabA H3, and MabB represent Trp residues in which an oxidation rate of more than 35% was observed upon AAPH stress.

[0089] Fig. 11 is a structural diagram showing features of the antibody Mab1 (wild-type; WT, left) and the antibody variant Mab1 .RRR (right). The side chains of Trp 91 , Asp 97, Asp 99, and Asp 101 are shown in ball-and-stick representation in the diagram of Mab1 . The side chains of Trp 91 , Arg 97, Arg 99, and Arg 101 are shown in ball-and-stick representation in the diagram of Mab1 RRR.

[0090] Fig. 12A is a structural diagram showing the presence of a negative electrostatic potential around a Trp (W) residue in the antibody Mab2 (wild-type; WT). Dashed lines indicate distance in A between the Trp residue (W58) and amino acid residues D100d, D100c, and D101 (Kabat numbering).

[0091] Fig. 12B is a structural diagram showing features of the Mab2 variant Mab2.RRR, which comprises three amino acid substitution mutations replacing aspartic acid (“D”) with arginine (“R”). Dashed lines indicate distance in A between the Trp residue (W58) and amino acid residues R100D, R100C, and R101 (Kabat numbering).

[0092] Fig. 13 is a set of graphs showing SASA, Epot (evaluated from MD trajectory), and Trp oxidation rate for Trp sites in the indicated engineered antibodies and wild-type comparators.

[0093] Fig. 14A is a structural diagram showing the electrostatic potential of the antibody Mab2 (left) and the antibody variant Mab2.W (right). R = arginine; W = tryptophan.

[0094] Fig. 14B is a pair of structural diagrams showing the antibody Mab2 (left) (showing a negative electrostatic potential in the H3 CDR loop) and the antibody variant Mab2.W (right). R = arginine; W = tryptophan.

[0095] Fig. 15 is a schematic diagram showing the redox potential (E°) in mV for Trp residue W100a (Kabat numbering) in Mab4 or Mab4.RR versus normal hydrogen electrode (NHE).

[0096] Fig. 16 is a graph showing a titration curve for the Trp residue at position 100 (Trp100a) of mab4 (light gray) or its engineered variant Mab4.RR (dark gray). Modeled data were fitted to the Nernst Equation and the oxidation / reduction (redox) potential was evaluated. Data are shown as fraction of reduced species versus redox potential (mV).

[0097] Fig. 17A is a structural diagram showing an anti-CD33 antibody in complex with CD33. The positions of Trp 96 (W96), Lys 52 (K52), and Asp 101 (D101 ) are indicated.

[0098] Fig. 17B is a structural diagram showing the electrostatic potential of the anti-CD33 antibody of Fig. 17A. Spheres indicate residues that were modified to reduce Trp oxidation and maintain binding.

[0099] Fig. 17C is a scatter plot showing the distribution of AAPH-stressed Trp oxidation (reduction in oxidation rate (%)), Epot (kT / e), and KD for the anti-CD33 antibody of Fig. 17A and thirteen variants thereof (S1 -S13). Relative KD is indicated by a heat map.

[0100] Fig. 17D is a chart showing the amino acid residues at select positions in the anti-CD33 antibody of Fig. 17A and thirteen variants thereof (S1 -S13) that showed reduction in Trp oxidation and relevant properties of the antibodies.

[0101] Fig. 18 is a graph showing the oxidation rates (percent of relevant Trp residues that were oxidized) measured in eight clinical-stage antibodies under stress induced by AAPH or an uncharged variant thereof (“AAPH variant”). The structures of AAPH and its variant are shown at right.

[0102] DETAILED DESCRIPTION OF THE INVENTION

[0103] The present invention is based, at least in part, on the discovery that the oxidation risk of a tryptophan residue of a polypeptide can reliably be predicted by two parameters: (1 ) the electrostatic potential of a region comprising the tryptophan residue; and (2) the solvent accessibility of the tryptophan residue, wherein a low electrostatic potential of a region comprising the tryptophan and a high solvent accessibility identify a tryptophan residue as one that is likely to be oxidized (i.e., identify a tryptophan as “high-risk” or “unstable”).

[0104] The present invention is also based, at least in part, on the discovery that the risk of oxidation of a tryptophan residue in a polypeptide can be reduced by modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, without modifying the polypeptide to replace or delete the tryptophan residue itself, rendering the long-range electrostatics unfavorable for tryptophan oxidation. In particular, it was discovered that alterations to the polypeptide that were distant from the tryptophan residue (e.g., were well outside the “contact” distance of 3-5 A) were successful in decreasing tryptophan oxidation by rendering the long-range electrostatics of the polypeptide unfavorable for tryptophan oxidation. The methods described herein provide a unique engineering strategy to reduce the risk of tryptophan oxidation without modifying the polypeptide to replace or delete the tryptophan residue itself, by introducing modifications to the polypeptide sequence that alter long-range electrostatics of tryptophan oxidation.

[0105] I. DEFINITIONS

[0106] Unless otherwise defined, all terms of art, notations, and other scientific terminology used herein are intended to have the meanings commonly understood by those of skill in the art to which this invention pertains. In some cases, terms with commonly understood meanings are defined herein for clarity and / or for ready reference, and the inclusion of such definitions herein should not necessarily be construed to represent a substantial difference over what is generally understood in the art.

[0107] The term “about” as used herein refers to the usual error range for the respective value readily known to the skilled person in this technical field. Reference to “about” a value or parameter herein includes (and describes) aspects that are directed to that value or parameter per se.

[0108] As used herein, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. For example, reference to “an isolated peptide” means one or more isolated peptides.

[0109] Throughout this specification and claims, the word “comprise,” or variations such as “comprises” or “comprising,” will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.

[0110] “Affinity” refers to the strength of the sum total of noncovalent interactions between a single binding site of a molecule (e.g., a receptor) and its binding partner (e.g., a ligand). Unless indicated otherwise, as used herein, “binding affinity” refers to intrinsic binding affinity, which reflects a 1 :1 interaction between members of a binding pair (e.g., receptor and ligand). The affinity of a molecule X for its partner Y can generally be represented by the dissociation constant (KD). Affinity can be measured by common methods known in the art, including those described herein. A preferred method for measuring affinity is Surface Plasmon Resonance (SPR).

[0111] The term “antibody” encompasses various antibody structures exhibiting the desired antigenbinding activity, including, but not limited to, monoclonal antibodies, multispecific antibodies (e.g., bispecific antibodies), and antibody fragments.

[0112] The term “monoclonal antibody” as used herein refers to an antibody obtained from a population of substantially homogeneous antibodies, i.e. the antibodies forming this population are essentially identical, except for possible post-translational modifications arising e.g. during manufacturing and / or storage. These antibodies are directed against the same epitope (or the same group of epitopes in the case of multispecific monoclonal antibodies, e.g. the same pair of epitopes in the case of bispecific monoclonal antibodies). This definition expressly excludes polyclonal antibody preparations which are mixtures of antibodies directed against different epitopes. Monoclonal antibodies in accordance with the present invention may be made by a variety of techniques, including but not limited to hybridoma methodology, recombinant DNA methods, phage-display methods, and methods utilizing transgenic animals containing all or part of the human immunoglobulin loci, such methods and other exemplary methods for making monoclonal antibodies being described herein.

[0113] The term “full-length antibody” refers to an antibody having the structure of an immunoglobulin comprising two light chains and two heavy chains, and comprising an Fc region as defined herein.

[0114] An “antibody fragment” refers to a molecule other than a full-length antibody that comprises a portion of a full-length antibody that binds the antigen to which the full-length antibody binds. Examples of antibody fragments include but are not limited to Fv molecules, Fab molecules, Fab' molecules, Fab’- SH molecules, F(ab')2 molecules, diabodies, linear antibody molecules, single-chain antibody molecules (e.g., scFv and scFab molecules), and multispecific (e.g. bispecific) antibodies formed from antibody fragments. The term “variable region” or “variable domain” refers to the domain of an antibody heavy or light chain that is involved in binding the antibody to antigen. The variable domains of the heavy chain and light chain (VH and VL, respectively) of a native antibody generally have similar structures, with each domain comprising four conserved framework regions (FRs) and three complementary determining regions (CDRs). (See, e.g., Kindt et al. Kuby Immunology, 6thed., W.H. Freeman and Co., page 91 (2007)). A single VH or VL domain may be sufficient to confer antigen-binding specificity. Furthermore, antibodies that bind a particular antigen may be isolated using a VH or VL domain from an antibody that binds the antigen to screen a library of complementary VL or VH domains, respectively. See, e.g., Portolano et al., J. Immunol. 150:880-887 (1993); Clarkson et al., Nature 352:624-628 (1991 ).

[0115] The term “complementarity determining region” or “CDR” as used herein refers to each of the regions of an antibody variable domain which are hypervariable in sequence and which determine antigen binding specificity. Generally, antibodies comprise six CDRs: three in the VH (CDR-H1 , CDR-H2, CDR- H3), and three in the VL (CDR-L1 , CDR-L2, CDR-L3). CDRs are defined by a variety of methods / systems by those skilled in the art. These systems and / or definitions have been developed and refined over a number of years and include Kabat, Chothia, IMGT, AbM, and Contact. The Kabat definition is based on sequence variability and generally is the most commonly used. The Chothia definition is based on the location of the structural loop regions. The IMGT system is based on sequence variability and location within the structure of the variable domain. The AbM definition is a compromise between Kabat and Chothia. The Contact definition is based on analyses of the available antibody crystal structures. Software programs (e.g., abYsis: http: / / www.abysis.org / abysis / sequence_input / key_annotation / key_annotation.cgi) are available and known to those of skill in the art for analysis of antibody sequences and determination of CDRs.

[0116] Exemplary CDRs herein include (numbering of amino acid residues according to the reference cited, i.e. Chothia numbering for the Chothia and Contact definition, Kabat numbering for the Kabat definition and IMGT numbering for the IMGT definition):

[0117] (a) hypervariable loops occurring at amino acid residues 26-32 (L1 ), 50-52 (L2), 91 -96 (L3), 26- 32 (H1 ), 53-55 (H2), and 96-101 (H3), according to Chothia and Lesk, J. Mol. Biol. 196:901 -917 (1987) (“Chothia definition”);

[0118] (b) CDRs occurring at amino acid residues 24-34 (L1 ), 50-56 (L2), 89-97 (L3), 31 -35B (H1 ), SO- 65 (H2), and 95-102 (H3), according to Kabat 1991 (“Kabat definition”);

[0119] (c) antigen contacts occurring at amino acid residues 30-36 (L1 ), 46-55 (L2), 89-96 (L3), 30-35 (H1 ), 47-58 (H2), and 93-101 (H3), according to MacCallum et al. J. Mol. Biol. 262: 732-745 (1996) (“Contact definition”); and

[0120] (d) CDRs occurring at amino acid residues residues 27-38 (L1 ), 56-65 (L2), 105-117 (L3), 27-38 (H1 ), 56-65 (H2), and 105-117 (H3), according to Lefranc et al. Dev. Comp. Immunol. 27: 55-77 (2003) (“IMGT definition”).

[0121] “Framework” or “FR” refers to variable domain residues other than complementary determining regions (CDRs). The FR of a variable domain generally consists of four FR domains: FR1 , FR2, FR3, and FR4. Accordingly, the CDR and FR sequences generally appear in the following sequence in VH (or VL): FR1 -CDR-H1 (CDR-L1 )-FR2- CDR-H2(CDR-L2)-FR3- CDR-H3(CDR-L3)-FR4. The term “Fc region” herein is used to define a C-terminal region of an immunoglobulin heavy chain that contains at least a portion of the constant region. The term includes native sequence Fc regions and variant Fc regions. In one aspect, a human IgG heavy chain Fc region extends from Cys226, or from Pro230, to the carboxyl-terminus of the heavy chain. However, antibodies produced by host cells may undergo post-translational cleavage of one or more, particularly one or two, amino acids from the C- terminus of the heavy chain. Therefore an antibody produced by a host cell by expression of a specific nucleic acid molecule encoding a full-length heavy chain may include the full-length heavy chain, or it may include a cleaved variant of the full-length heavy chain. This may be the case in particular where the final two C-terminal amino acids of the heavy chain are glycine (G446) and lysine (K447, Kabat EU numbering). Therefore, the C-terminal lysine (Lys447), or the C-terminal glycine (Gly446) and lysine (Lys447), of the Fc region may or may not be present. The corresponding sequence including a C- terminal lysine residue is also encompassed, however. Accordingly, in one aspect, a heavy chain including an Fc region as specified herein comprises an additional C-terminal lysine residue (K447, Kabat EU numbering). Also encompassed is the corresponding sequence without the C-terminal glycine residue. Accordingly, in one aspect, a heavy chain including an Fc region as specified herein lacks the C- terminal glycine residue (G446, Kabat EU numbering). In such heavy chain, the C-terminal amino acid residue may be proline (P445, Kabat EU numbering) or proline amide (P445-NH2, Kabat EU numbering). Unless otherwise specified herein, numbering of amino acid residues in the Fc region or heavy chain constant region is according to the EU numbering system, also called the EU index, as described in Kabat 1991.

[0122] As used herein, the term “region comprising the tryptophan residue” refers to an area of a polypeptide that comprises a tryptophan residue of interest. In some aspects, the region comprising the tryptophan residue may be a region of the polypeptide within a radial distance of 100 A of the tryptophan residue (e.g., within a radial distance of 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A of the tryptophan residue) in a structural model of the polypeptide. In some aspects, the region comprising the tryptophan residue is the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the region comprising the tryptophan residue is the Fab domain comprising the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an lgG1 isotype antibody or antibody fragment), the tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the region comprising the tryptophan residue is the VH or VL comprising the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the region comprising the tryptophan residue is the CDR loop comprising the tryptophan residue.

[0123] As used herein, the terms “electrostatic potential of the region comprising the tryptophan residue” and “Epot” are used interchangeably and refer to the electrostatic potential of an area of a polypeptide that comprises a tryptophan residue of interest. In some aspects, the electrostatic potential of the region comprising the tryptophan residue may be the electrostatic potential of a region of the polypeptide within a radial distance of 100 A of the tryptophan residue (e.g., within a radial distance of 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A of the tryptophan residue) in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH or VL comprising the tryptophan residue. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR loop comprising the tryptophan residue. Methods of assessing electrostatic potential of a region of a polypeptide are known in the art and include, without limitation, use of the Adaptive Poisson-Boltzmann Solver (APBS) software package (Jurrus et al. 2018. Protein Science 27 (1 ): 112-28).

[0124] As used herein, the terms “solvent-accessible surface area” and “SASA” are used interchangeably and refer to a standard measurement of the surface area of a molecule that is accessible to a solvent (typically water). A skilled artisan may readily calculate both SASA and the change in SASA upon alteration of an amino acid sequence for a given protein structure. For example, SASA can be calculated by using the ‘rolling ball’ algorithm developed by Shrake & Rupley in 1973 (Shrake, A; Rupley, J A. (1973). J Mol Biol 79 (2): 351 -71 ), which models a sphere approximating the size of the solvent molecule to ‘probe’ the surface of the molecule.

[0125] II. METHODS OF IDENTIFYING HIGH-RISK TRYPTOPHAN RESIDUES

[0126] The present disclosure provides methods (e.g., in silico methods) of identifying tryptophan (Trp) residues within a polypeptide sequence that have a high risk of oxidation, e.g., oxidation under stressed conditions and / or oxidation under any condition relevant to the manufacture, storage, or use of the polypeptide. In particular, the disclosure provides methods of assessing the oxidation risk of a tryptophan residue based on assessment of the electrostatic potential and solvent-accessible surface area (SASA) of the tryptophan residue and / or the region comprising the tryptophan residue.

[0127] For example, in one aspect, the disclosure provides a method of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, the method comprising the steps of (a) measuring the electrostatic potential of a region comprising the tryptophan residue; and (b) measuring a solvent-accessible surface area (SASA) of the tryptophan residue, wherein the electrostatic potential of the region comprising the tryptophan residue and the SASA of the tryptophan residue together identify the tryptophan residue as likely to be oxidized. In another aspect, the disclosure provides a method of identifying a tryptophan residue of a polypeptide as one that is likely to be oxidized, the method comprising the steps of (a) measuring the electrostatic potential of a region comprising the tryptophan residue; and (b) measuring a SASA of the tryptophan residue; wherein the electrostatic potential of the region comprising the tryptophan residue and the SASA of the tryptophan residue together identify the tryptophan residue as likely to be oxidized.

[0128] Exemplary polypeptides that may be assessed according to these methods include, but are not limited to antibodies (e.g., lgG1 isotype antibodies), antibody fragments (e.g., a VHH), cytokines, enzymes, and / or therapeutic polypeptides.

[0129] Methods of assessing electrostatic potential and solvent accessibility, as well as exemplary values that identify a tryptophan residue as one that is likely to be oxidized, are described in further detail below.

[0130] A. Identification of tryptophan residues that are likely to be oxidized

[0131] Electrostatic potential and SASA

[0132] The present methods are based, at least in part, on the discovery that oxidation risk of a tryptophan residue of a polypeptide can reliably be predicted by two parameters: (1 ) the electrostatic potential of a region comprising the tryptophan residue; and (2) the solvent accessibility of the tryptophan residue, wherein a low electrostatic potential of a region comprising the tryptophan and a high solvent accessibility identify a tryptophan residue as one that is likely to be oxidized (i.e., identify a tryptophan as “high-risk” or “unstable”). Accordingly, the disclosure provides methods of determining whether a tryptophan residue of a polypeptide is likely to be oxidized (or identifying a tryptophan residue as one that is likely to be oxidized) based on assessment of electrostatic potential and solvent accessibility.

[0133] For example, in some aspects, identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises (i) determining that the tryptophan residue is in a region having a low electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above a reference level.

[0134] In some aspects, identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises (i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above a reference level (e.g., is above 50 A2).

[0135] Accordingly, in some aspects, the disclosure provides a method of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, the method comprising the steps of (a) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (b) determining that the tryptophan residue has a SASA that is above 50 A2. In some aspects, the disclosure provides a method of identifying a tryptophan residue of a polypeptide as one that is likely to be oxidized, the method comprising the steps of (a) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (b) determining that the tryptophan residue has a SASA that is above 50 A2.

[0136] In some aspects, a tryptophan residue is not likely to be oxidized if the tryptophan residue is in a region having an electrostatic potential above 0.5. In some aspects, a tryptophan residue is not likely to be oxidized if the tryptophan residue has a SASA below 50 A2. In some aspects, a tryptophan residue is not likely to be oxidized if (a) the tryptophan residue is in a region having an electrostatic potential above 0.5, and (b) the tryptophan residue has a SASA below 50 A2. In some aspects, a tryptophan residue is likely to be oxidized if (a) the tryptophan residue is in a region having an electrostatic potential of 0.5 or less and (b) the tryptophan residue has a SASA above 50 A2. In some aspects, a tryptophan residue is likely to be oxidized if (a) the tryptophan residue is in a region having an electrostatic potential of -0.5 or less, and (b) the tryptophan residue has a SASA above 50 A2.

[0137] In some aspects, the identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises use of a model or equation (e.g., a two-parameter regression model) in which electrostatic potential and SASA are factors.

[0138] Level of oxidation risk

[0139] As used herein, a tryptophan residue of a polypeptide that is “likely to be oxidized” (i.e. , a “high- risk” or “unstable” tryptophan residue) is a residue that is likely to undergo oxidation under stressed conditions and / or is likely to undergo oxidation under any condition relevant to the manufacture, storage, or use of the polypeptide at a rate that is expected or has been determined to negatively impact the efficacy (e.g., therapeutic efficacy) and / or stability (e.g., shelf stability) of the polypeptide.

[0140] Methods for empirically assessing the oxidation rate of a tryptophan residue of a polypeptide are known in the art and include use of chemical stressors (e.g., hydrogen peroxide (H2O2), 2,2-azobis(2- amidinopropane) or dihydrochloride (AAPH)) and photooxidation. For example, in some aspects, the oxidation rate of a polypeptide is assessed in an oxidation assay using AAPH. An exemplary AAPH assay is performed for about 16 hours (e.g., oxidation rate is measured following treatment with AAPH for 16 hours, e.g., at pH 5.5, 37°C).

[0141] In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that has an oxidation rate of at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, or 95% or has an oxidation rate of more than 95% (e.g., has an oxidation rate of 5-10%, 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%, or 90- 100%), as measured using an appropriate assay for oxidation (e.g., an AAPH assay).

[0142] In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that has an oxidation rate of at least 35%, as measured using an appropriate assay for oxidation (e.g., an AAPH assay). In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that is predicted to have an oxidation rate of at least 35%, as determined using an in silico assay.

[0143] In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that has an oxidation rate of at least 30%, as measured using an appropriate assay for oxidation (e.g., an AAPH assay). In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that is predicted to have an oxidation rate of at least 30%, as determined using an in silico assay.

[0144] In any of the methods provided herein, the method may further comprise empirically determining the oxidation rate of a polypeptide comprising a tryptophan residue. Electrostatic potential

[0145] The present methods involve assessment of the electrostatic potential of a region of a polypeptide comprising a tryptophan residue, wherein a low or negative electrostatic potential is a factor in determining whether the tryptophan residue is likely to be oxidized. The methods are based, at least in part, on the finding that modulation of electrostatic potential by mutating distant residues can impact tryptophan oxidation: thus, the presently claimed methods include assessment of electrostatic potential across large regions of the polypeptide. In some aspects, electrostatic potential is assessed in silico, e.g., is assessed in a structural model of the polypeptide. Methods of assessing electrostatic potential of a region of a polypeptide are known in the art and include, without limitation, use of the Adaptive Poisson- Boltzmann Solver (APBS) (Jurrus et al. 2018. Protein Science 27 (1 ): 112-28). Exemplary methods for assessing electrostatic potential using the APBS software package are provided, e.g., in Park et al. 2023. mAbs 16 (1 ): 2362788. Briefly, in some aspects, to analyze the electrostatic potential of a region comprising a tryptophan residue, the APBS software package is used to compute Poisson-Boltzmann electrostatics for the region using the corresponding PQR file. Nanoshaper may then be used to generate a triangulated mesh representing the molecular surface. The solvent-excluded surface (SES) (Park et al. 2023. mAbs 16 (1 ): 2362788) may be constructed using a probe size of 1 .5 A and a density of 5 points per A2. The charge at each vertex of the meshed surface can be assigned using Multivalue, a tool within the APBS suite. Next, a Python script is used to map the electrostatic potential from the vertices to the corresponding atoms. To obtain per residue values, the sum of all electrostatic potentials can be calculated, as well as the sum of all the positive or negative values, yielding three values for each residue: APBS_neg, APBS_pos, and APBS_sum (Epot).

[0146] For example, in some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 200 A (e.g., within a radial distance of 150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60- 70 A, 70-80 A, 80-90 A, 90-100 A, 100-110 A, 110-120 A, 120-130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A)) of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region of the polypeptide within a radial distance of 200 A, 150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60-70 A, 70-80 A, 80-90 A, 90-100 A, 100-110 A, 110-120 A, 120-130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A)). In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 100 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region within a radial distance of 30 A of the Trp residue in a structural model of the polypeptide). In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide). In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide). Thus, in some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 200 A ,150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60-70 A, 70-80 A, 80-90 A, 90-100 A, 100-1 10 A, 1 10-120 A, 120- 130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A) of the tryptophan residue in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 100 A of the tryptophan in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan in a structural model of the polypeptide is negative.

[0147] In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of a particular domain or sub-domain of the polypeptide that comprises the tryptophan residue (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of a particular domain or sub-domain of the polypeptide that comprises the tryptophan residue). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the particular domain or sub-domain of the polypeptide that comprises the tryptophan residue is negative.

[0148] For example, in some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of a fragment antigen-binding (Fab) domain, variable domain, complementarity-determining region (CDR), framework region, or Fc region of the antibody or antibody fragment that comprises the tryptophan residue. Thus, in some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a Fab domain, variable domain, CDR, framework region, or Fc region that comprises the tryptophan residue is negative.

[0149] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue.

[0150] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within the heavy chain variable (VH) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a VH domain that comprises the tryptophan residue is negative.

[0151] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), and the tryptophan residue is within the light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VL comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a VL domain that comprises the tryptophan residue is negative,

[0152] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR that comprises the tryptophan residue is negative.

[0153] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a complementarity-determining region heavy chain 1 (CDR-H1 ) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-H1 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-H1 loop that comprises the tryptophan residue is negative.

[0154] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-H2 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-H2 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-H2 loop that comprises the tryptophan residue is negative. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-H3 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-H3 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-H3 loop that comprises the tryptophan residue is negative.

[0155] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-L1 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-L1 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-L1 loop that comprises the tryptophan residue is negative.

[0156] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-L2 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-L2 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-L2 loop that comprises the tryptophan residue is negative.

[0157] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-L3 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-L3 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-L3 loop that comprises the tryptophan residue is negative.

[0158] In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the of the tryptophan residue side chain. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the tryptophan residue side chain is negative.

[0159] In some aspects, the method comprises determining that the tryptophan residue is in a region having a low electrostatic potential. For example, in some aspects, the tryptophan residue is in a region having an electrostatic potential that is less than 0.5 (e.g., having an electrostatic potential of -10.0 to - 9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, - 1 .0 to 0, 0 to 0.1 , 0.1 to 0.2, 0.2 to 0.3, 0.3 to 0.4, or 0.4 to 0.5). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is less than 0.5 (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, -1 .0 to 0, 0 to 0.1 , 0.1 to 0.2, 0.2 to 0.3, 0.3 to 0.4, or 0.4 to 0.5).

[0160] In some aspects, the method comprises determining that the tryptophan residue is in a region having a negative electrostatic potential. For example, in some aspects, the tryptophan residue is in a region having an electrostatic potential that is less than 0 (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to 0). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is less than 0 (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to - 6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to 0).

[0161] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -0.5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to - 0.5). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is -0.5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to - 6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to -0.5).

[0162] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .0 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .0). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is -1 .0 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, - 5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .0).

[0163] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .5). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is -1 .5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, - 5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .5).

[0164] Solvent accessibility

[0165] The present methods involve assessment of the solvent accessibility (e.g., solvent-accessible surface area (SASA)) of a tryptophan residue, wherein a high level of solvent accessibility (e.g., high SASA) is a factor in determining whether the tryptophan residue is likely to be oxidized. In some aspects, solvent accessibility (e.g., SASA) is assessed in silico, e.g., is assessed in a structural model of the polypeptide. Methods of assessing SASA in a polypeptide are known in the art and include, without limitation, use of ptraj (Roe et al. 2013. Journal of Chemical Theory and Computation 9 (7): 3084-95) and mdtraj (McGibbon et al. 2015. Biophysical Journal 109 (8): 1528-32). SASA may be assessed at a tryptophan residue and / or at a region surrounding the tryptophan residue. For example, in some aspects, SASA of a tryptophan residue refers to SASA of a single position (the tryptophan residue). In other aspects, the SASA of the tryptophan residue is the SASA of n-1 to n+1 residues, wherein n is the tryptophan residue (3 residues total). In some aspects, the SASA of the tryptophan residue is the SASA of n-2 to n+2 residues, wherein n is the tryptophan residue (5 residues total). In some aspects, the SASA of the tryptophan residue is the SASA of n-3 to n+3 residues, wherein n is the tryptophan residue (7 residues total). In some aspects, the SASA of the tryptophan residue is the SASA of n-4 to n+4 residues, wherein n is the tryptophan residue (9 residues total).

[0166] In some aspects, the method comprises determining that the tryptophan residue is solvent- exposed. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is solvent-exposed.

[0167] In some aspects, the method comprises determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, the reference SASA level is 30 A2, 35 A2, 40 A2, 45 A2, 50 A2, 55 A2, 60 A2, 65 A2, 70 A2, 75 A2, or 80 A2. In some aspects, the reference SASA level is 50 A2. In some aspects, the reference SASA level is 80 A2. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue has a SASA that is above 50 A2. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue has a SASA that is above 80 A2.

[0168] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR of the antibody or antibody fragment, and the SASA of the tryptophan residue is the SASA of the CDR containing the tryptophan residue.

[0169] III. METHODS OF MODIFYING A POLYPEPTIDE TO REDUCE THE LIKELIHOOD OF TRYPTOPHAN OXIDATION

[0170] The present disclosure provides methods of producing optimized polypeptides in which the risk of oxidation of at least one tryptophan residue of the polypeptide is decreased. For example, in some aspects, the method comprises identifying a polypeptide in which at least one tryptophan residue is at a high risk of oxidation and modifying the polypeptide to reduce the likelihood of oxidation of the tryptophan residue, thereby producing a modified, optimized polypeptide.

[0171] In particular, the methods provided herein comprise (1 ) determining whether a tryptophan residue of a polypeptide is likely to be oxidized (or identifying a tryptophan residue as one that is likely to be oxidized) based on assessment of electrostatic potential and solvent accessibility, wherein a low electrostatic potential of a region comprising the tryptophan and a high solvent accessibility identify a tryptophan residue as one that is likely to be oxidized, and (2) modifying the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue and / or decrease the solvent accessibility of the tryptophan residue, thereby reducing the likelihood of oxidation of the tryptophan residue. In some aspects, the modifications are distant from the tryptophan residue (e.g., are outside the “contact” distance of 3-5 A, e.g., are within a radial distance of more than 10, more than 50, more than 100, or more than 200 A of the tryptophan residue) and decrease tryptophan oxidation by rendering the long-range electrostatics of the polypeptide unfavorable for tryptophan oxidation.

[0172] For example, in one aspect, the disclosure provides a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of (a) determining the likelihood of oxidation of a tryptophan residue of the polypeptide, the determination based on (i) the electrostatic potential of a region comprising the tryptophan residue, and (ii) the solvent-accessible surface area (SASA) of the tryptophan residue; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises: (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue; (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region. In some aspects, the method comprises identifying a tryptophan residue of the polypeptide that is likely to be oxidized. In some aspects, the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced.

[0173] In another aspect, the disclosure provides a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of: (a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises: (i) determining the electrostatic potential of a region comprising the tryptophan residue; and (ii) determining the solvent-accessible surface area (SASA) of the tryptophan residue; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises: (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue; (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region. In some aspects, the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced.

[0174] In another aspect, the disclosure provides a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of: (a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises: (i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above 50 A2; and (b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises: (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue;

[0175] (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or

[0176] (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue.

[0177] In another aspect, the disclosure provides a method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of oxidation of a solvent-exposed tryptophan residue of the polypeptide, the method comprising modifying the amino acid sequence to increase the electrostatic potential of a region comprising the solvent-exposed tryptophan residue, wherein the modifying comprises: (a) replacing one or more candidate amino acid residues within the region with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the solvent-exposed tryptophan residue; (b) deleting one or more candidate amino acid residues within the region, wherein the one or more candidate amino acid residues do not include the solvent-exposed tryptophan residue; or (c) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

[0178] A modified amino acid sequence, as used herein, may be an amino acid sequence that comprises one or more alterations (e.g., amino acid substitutions, insertions, or deletions) that mitigate or are intended to mitigate the risk of oxidation of a tryptophan residue relative to a reference amino acid sequence (e.g., the amino acid sequence of a protein in which a high-risk tryptophan was identified).

[0179] In some aspects of any of the methods provided herein, the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced (e.g., reduced below a threshold that identifies a high-risk tryptophan) or eliminated. In some aspects, the disclosure provides a polypeptide comprising a tryptophan residue having a low likelihood of oxidation, wherein the polypeptide is produced by any method described herein.

[0180] Exemplary polypeptides that may be produced according to these methods include antibodies (e.g., IgG 1 isotype antibodies), antibody fragments (e.g., a VHH), cytokines, enzymes, and / or therapeutic polypeptides.

[0181] Methods of assessing electrostatic potential and solvent accessibility, as well as exemplary values that identify a tryptophan residue as one that is likely to be oxidized, are described in further detail below.

[0182] A. Identification of tryptophan residues that are likely to be oxidized

[0183] Electrostatic potential and SASA

[0184] The present methods are based, at least in part, on the discovery that oxidation risk of a tryptophan residue of a polypeptide can reliably be predicted by two parameters: (1 ) the electrostatic potential of a region comprising the tryptophan residue; and (2) the solvent accessibility of the tryptophan residue, wherein a low electrostatic potential of a region comprising the tryptophan and a high solvent accessibility identify a tryptophan residue as one that is likely to be oxidized (i.e., identify a tryptophan as “high-risk” or “unstable”). Accordingly, the disclosure provides methods of determining whether a tryptophan residue of a polypeptide is likely to be oxidized (or identifying a tryptophan residue as one that is likely to be oxidized) based on assessment of electrostatic potential and solvent accessibility.

[0185] For example, in some aspects, identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises (i) determining that the tryptophan residue is in a region having a low electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above a reference level.

[0186] In some aspects, identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises (i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (ii) determining that the tryptophan residue has a SASA that is above a reference level (e.g., is above 50 A2).

[0187] Accordingly, in some aspects, the disclosure provides a method of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, the method comprising the steps of (a) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (b) determining that the tryptophan residue has a SASA that is above 50 A2. In some aspects, the disclosure provides a method of identifying a tryptophan residue of a polypeptide as one that is likely to be oxidized, the method comprising the steps of (a) determining that the tryptophan residue is in a region having a negative electrostatic potential; and (b) determining that the tryptophan residue has a SASA that is above 50 A2.

[0188] In some aspects, a tryptophan residue is not likely to be oxidized if the tryptophan residue is in a region having an electrostatic potential above 0.5. In some aspects, a tryptophan residue is not likely to be oxidized if the tryptophan residue has a SASA below 50 A2. In some aspects, a tryptophan residue is not likely to be oxidized if (a) the tryptophan residue is in a region having an electrostatic potential above 0.5, and (b) the tryptophan residue has a SASA below 50 A2. In some aspects, a tryptophan residue is likely to be oxidized if (a) the tryptophan residue is in a region having an electrostatic potential of 0.5 or less and (b) the tryptophan residue has a SASA above 50 A2. In some aspects, a tryptophan residue is likely to be oxidized if (a) the tryptophan residue is in a region having an electrostatic potential of -0.5 or less, and (b) the tryptophan residue has a SASA above 50 A2. In some aspects, identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises use of a model or equation (e.g., a two-parameter regression model) in which electrostatic potential and SASA are factors.

[0189] Level of oxidation risk

[0190] As used herein, a tryptophan residue of a polypeptide that is “likely to be oxidized” (i.e. , a “high- risk” or “unstable” tryptophan residue) is a residue that is likely to undergo oxidation under stressed conditions and / or is likely to undergo oxidation under any condition relevant to the manufacture, storage, or use of the polypeptide at a rate that is expected or has been determined to negatively impact the efficacy (e.g., therapeutic efficacy) and / or stability (e.g., shelf stability) of the polypeptide.

[0191] Methods for empirically assessing the oxidation rate of a tryptophan residue of a polypeptide are known in the art and include use of chemical stressors (e.g., hydrogen peroxide (H2O2), 2,2-azobis(2- amidinopropane) or dihydrochloride (AAPH)) and photooxidation. For example, in some aspects, the oxidation rate of a polypeptide is assessed in an oxidation assay using AAPH. An exemplary AAPH assay is performed for about 16 hours (e.g., oxidation rate is measured following treatment with AAPH for 16 hours, e.g., at pH 5.5, 37°C).

[0192] In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that has an oxidation rate of at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, or 95%, or has an oxidation rate of more than 95% (e.g., has an oxidation rate of 5-10%, 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%, or 90- 100%), as measured using an appropriate assay for oxidation (e.g., an AAPH assay).

[0193] In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that has an oxidation rate of at least 35%, as measured using an appropriate assay for oxidation (e.g., an AAPH assay). In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that is predicted to have an oxidation rate of at least 35%, as determined using an in silico assay.

[0194] In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that has an oxidation rate of at least 30%, as measured using an appropriate assay for oxidation (e.g., an AAPH assay). In some aspects, a tryptophan residue of a polypeptide that is likely to be oxidized is a residue that is predicted to have an oxidation rate of at least 30%, as determined using an in silico assay.

[0195] In any of the methods provided herein, the method may further comprise empirically determining the oxidation rate of a polypeptide comprising a tryptophan residue.

[0196] Electrostatic potential

[0197] The present methods involve assessment of the electrostatic potential of a region of a polypeptide comprising a tryptophan residue, wherein a low or negative electrostatic potential is a factor in determining whether the tryptophan residue is likely to be oxidized. The methods are based, at least in part, on the finding that modulation of electrostatic potential by mutating distant residues can impact tryptophan oxidation: thus, the presently claimed methods include assessment of electrostatic potential across large regions of the polypeptide. In some aspects, electrostatic potential is assessed in silico, e.g., is assessed in a structural model of the polypeptide. Methods of assessing electrostatic potential of a region of a polypeptide are known in the art and include, without limitation, use of the Adaptive Poisson- Boltzmann Solver (APBS) (Jurrus et al. 2018. Protein Science 27 (1 ): 112-28). Exemplary methods for assessing electrostatic potential using the APBS software package are provided, e.g., in Park et al. 2023. mAbs 16 (1 ): 2362788. Briefly, in some aspects, to analyze the electrostatic potential of a region comprising a tryptophan residue, the APBS software package is used to compute Poisson-Boltzmann electrostatics for the region using the corresponding PQR file. Nanoshaper may then be used to generate a triangulated mesh representing the molecular surface. The solvent-excluded surface (SES) (Park et al. 2023. mAbs 16 (1 ): 2362788) may be constructed using a probe size of 1 .5 A and a density of 5 points per A2. The charge at each vertex of the meshed surface can be assigned using Multivalue, a tool within the APBS suite. Next, a Python script is used to map the electrostatic potential from the vertices to the corresponding atoms. To obtain per residue values, the sum of all electrostatic potentials can be calculated, as well as the sum of all the positive or negative values, yielding three values for each residue: APBS_neg, APBS_pos, and APBS_sum (Epot). For example, in some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 200 A (e.g., within a radial distance of 150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60- 70 A, 70-80 A, 80-90 A, 90-100 A, 100-110 A, 110-120 A, 120-130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A)) of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region of the polypeptide within a radial distance of 200 A, 150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60-70 A, 70-80 A, 80-90 A, 90-100 A, 100-110 A, 110-120 A, 120-130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A)). In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 100 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide). In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide). In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of the region within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide). Thus, in some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 200 A ,150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60-70 A, 70-80 A, 80-90 A, 90-100 A, 100-110 A, 110-120 A, 120- 130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A) of the tryptophan residue in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 100 A of the tryptophan in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan in a structural model of the polypeptide is negative. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan in a structural model of the polypeptide is negative.

[0198] In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of a particular domain or sub-domain of the polypeptide that comprises the tryptophan residue (e.g., determining the electrostatic potential of a region comprising the tryptophan residue comprises determining the electrostatic potential of a particular domain or sub-domain of the polypeptide that comprises the tryptophan residue). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the particular domain or sub-domain of the polypeptide that comprises the tryptophan residue is negative.

[0199] For example, in some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of a fragment antigen-binding (Fab) domain, variable domain, complementarity-determining region (CDR), framework region, or Fc region of the antibody or antibody fragment that comprises the tryptophan residue. Thus, in some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a Fab domain, variable domain, CDR, framework region, or Fc region that comprises the tryptophan residue is negative.

[0200] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue.

[0201] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within the heavy chain variable (VH) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a VH domain that comprises the tryptophan residue is negative.

[0202] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), and the tryptophan residue is within the light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VL comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a VL domain that comprises the tryptophan residue is negative. In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR that comprises the tryptophan residue is negative,

[0203] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a complementarity-determining region heavy chain 1 (CDR-H1 ) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-H1 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-H1 loop that comprises the tryptophan residue is negative.

[0204] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-H2 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-H2 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-H2 loop that comprises the tryptophan residue is negative.

[0205] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-H3 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-H3 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-H3 loop that comprises the tryptophan residue is negative.

[0206] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-L1 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-L1 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-L1 loop that comprises the tryptophan residue is negative.

[0207] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-L2 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-L2 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-L2 loop that comprises the tryptophan residue is negative.

[0208] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR-L3 loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR-L3 loop comprising the tryptophan residue. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of a CDR-L3 loop that comprises the tryptophan residue is negative.

[0209] In some aspects, the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the tryptophan residue side chain. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the electrostatic potential of the tryptophan residue side chain is negative.

[0210] In some aspects, the method comprises determining that the tryptophan residue is in a region having a low electrostatic potential. For example, in some aspects, the tryptophan residue is in a region having an electrostatic potential that is less than 0.5 (e.g., having an electrostatic potential of -10.0 to - 9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, - 1 .0 to 0, 0 to 0.1 , 0.1 to 0.2, 0.2 to 0.3, 0.3 to 0.4, or 0.4 to 0.5). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is less than 0.5 (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, -1 .0 to 0, 0 to 0.1 , 0.1 to 0.2, 0.2 to 0.3, 0.3 to 0.4, or 0.4 to 0.5).

[0211] In some aspects, the method comprises determining that the tryptophan residue is in a region having a negative electrostatic potential. For example, in some aspects, the tryptophan residue is in a region having an electrostatic potential that is less than 0 (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to 0). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is less than 0 (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to - 6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to 0).

[0212] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -0.5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to - 0.5). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is -0.5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to - 6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, -2.0 to -1 .0, or -1 .0 to -0.5).

[0213] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .0 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .0). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is -1 .0 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, - 5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .0).

[0214] In some aspects, the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, -5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .5). In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is in a region having an electrostatic potential that is -1 .5 or less (e.g., having an electrostatic potential of -10.0 to -9.0, -9.0 to -8.0, -8.0 to -7.0, -7.0 to -6.0, -6.0 to -5.0, - 5.0 to -4.0, -4.0 to -3.0, -3.0 to -2.0, or -2.0 to -1 .5).

[0215] Solvent accessibility

[0216] The present methods involve assessment of the solvent accessibility (e.g., solvent-accessible surface area (SASA)) of a tryptophan residue, wherein a high level of solvent accessibility (e.g., high SASA) is a factor in determining whether the tryptophan residue is likely to be oxidized. In some aspects, solvent accessibility (e.g., SASA) is assessed in silico, e.g., is assessed in a structural model of the polypeptide. Methods of assessing SASA in a polypeptide are known in the art and include, without limitation, use of ptraj (Roe et al. 2013. Journal of Chemical Theory and Computation 9 (7): 3084-95) and mdtraj (McGibbon et al. 2015. Biophysical Journal 109 (8): 1528-32). SASA may be assessed at a tryptophan residue and / or at a region surrounding the tryptophan residue. For example, in some aspects, SASA of a tryptophan residue refers to SASA of a single position (the tryptophan residue). In otheraspects, the SASA of the tryptophan residue is the SASA of n-1 to n+1 residues, wherein n is the tryptophan residue (3 residues total). In some aspects, the SASA of the tryptophan residue is the SASA of n-2 to n+2 residues, wherein n is the tryptophan residue (5 residues total). In some aspects, the SASA of the tryptophan residue is the SASA of n-3 to n+3 residues, wherein n is the tryptophan residue (7 residues total). In some aspects, the SASA of the tryptophan residue is the SASA of n-4 to n+4 residues, wherein n is the tryptophan residue (9 residues total).

[0217] In some aspects, the method comprises determining that the tryptophan residue is solvent- exposed. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue is solvent-exposed.

[0218] In some aspects, the method comprises determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue has a SASA that is above a reference level. In some aspects, the reference SASA level is 30 A2, 35 A2, 40 A2, 45 A2, 50 A2, 55 A2, 60 A2, 65 A2, 70 A2, 75 A2, or 80 A2. In some aspects, the reference SASA level is 50 A2. In some aspects, the reference SASA level is 80 A2. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue has a SASA that is above 50 A2. In some aspects, identifying a tryptophan residue of a polypeptide that is likely to be oxidized comprises, in part, determining that the tryptophan residue has a SASA that is above 80 A2.

[0219] In some aspects, the polypeptide is an antibody or antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is within a CDR of the antibody or antibody fragment, and the SASA of the tryptophan residue is the SASA of the CDR containing the tryptophan residue. B. Modification of the amino acid sequence of the polypeptide

[0220] The disclosure provides methods of modifying the amino acid sequence of a polypeptide to reduce the likelihood of oxidation of a tryptophan residue therein (e.g., a tryptophan residue that has been identified as having a high risk of oxidation according to the above-described methods). In particular, the methods comprise modifying the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue and / or decrease the solvent accessibility of the tryptophan residue, thereby reducing the likelihood of oxidation of the tryptophan residue. Surprisingly, the modifications to the amino acid sequence of the polypeptide can be introduced at a distance from the tryptophan residue, rendering the long-range electrostatics unfavorable for tryptophan oxidation.

[0221] Modifying the amino acid sequence of a polypeptide encompasses both in silico modification of the amino acid sequence (e.g., modifying an in silico sequence that is used to create a structural model of the polypeptide) and ex silico modification of the amino acid, e.g., in vitro or in vivo production of a polypeptide having a modified amino acid sequence. A skilled artisan would be readily able to produce a desired modified amino acid sequence, e.g., by producing a nucleotide sequence encoding the desired modified amino acid sequence and expressing the nucleotide sequence in an appropriate production system (e.g., a host cell).

[0222] Modifications contemplated by the invention include, without limitation, (i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue; (ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or (iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

[0223] A skilled artisan would be able to identify insertion, deletion, and / or substitution mutations that would be likely to increase the electrostatic potential of the region. In some aspects, the methods further comprise determining the likelihood of oxidation of the tryptophan residue of the modified polypeptide, e.g., by assessing electrostatic potential and solvent accessibility as provided herein (e.g., in an in silico structural model).

[0224] In some aspects, the modifications comprise only a single amino acid replacement, insertion, or deletion. In other aspects, the modifications comprise 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, or more than 25 (e.g., 1 -5, 5-10, 10-15, 15-20, or 20-25) amino acid replacements, insertions, or deletions. For example, the modifications may comprise 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25 or more than 25 (e.g., 1 -5, 5-10, 10-15, 15- 20, or 20-25) amino acid replacements; 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25 or more than 25 (e.g., 1 -5, 5-10, 10-15, 15-20, or 20-25) amino acid insertions (e.g., consecutive insertions of two or more amino acid residues or non-consecutive insertions of amino acid residues at least two discrete sites); and / or 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25 or more than 25 (e.g., 1 -5, 5-10, 10-15, 15-20, or 20-25) amino acid deletions (e.g., consecutive deletions of two or more amino acid residues or non-consecutive deletions of amino acid residues at least two discrete sites). In some aspects, the one or more candidate amino acid residues of the polypeptide (e.g., the one or more amino acid residues that are deleted or replaced to generate the modified polypeptide) are amino acid residues that contribute to a negative electrostatic potential of the region comprising the tryptophan residue.

[0225] In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 200 A (e.g., within a radial distance of 150 A, 125 A, 100 A, 90 A, 80 A, 70 A, 60 A, 50 A, 40 A, 30 A, 25 A, 20 A, 15 A, 10 A, or 5 A (e.g., within a radial distance of 5-10 A, 10-20 A, 20-30 A, 30-40 A, 40-50 A, 50-60 A, 60-70 A, 70-80 A, 80-90 A, 90-100 A, 100-110 A, 110-120 A, 120-130 A, 130-140 A, 140-150 A, 150-160 A, 160-170 A, 170-180 A, 180-190 A, or 190-200 A)) of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide. In some aspects, the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

[0226] In some aspects, one or more of the modifications that increases the electrostatic potential of the region surrounding the tryptophan residue is distant from the tryptophan residue (e.g., is outside of the “contact” distance of 3-5 A). In some aspects, the modifications comprise one or more amino acid replacements, insertions, or deletions that are more than 5 A, more than 10 A, more than 20 A, more than 30 A, more than 40 A, more than 50 A, more than 60 A, more than 70 A, more than 80 A, more than 90 A, more than 100 A, more than 110 A, more than 120 A, more than 130 A, more than 140 A, more than 150 A, more than 160 A, more than 170 A, more than 180 A , more than 190 A away, or more than 200 A from the tryptophan residue (e.g., 5-10 A away, 10-20 A away, 20-30 A away, 30-40 A away, 40-50 A away, 50-60 A away, 60-70 A away, 70-80 A away, 80-90 A away, 90-100 A away, 100-110 A away, 110-120 A away, 120-130 A away, 130-140 A away, 140-150 A aways150-160 A away, 160-170 A away, 170-180 A away, 180-190 A away, or 190-200 A away from the tryptophan residue). In some aspects, the disclosure provides methods of modifying a polypeptide to reduce the likelihood of oxidation of a tryptophan residue therein, wherein no modifications are made within a radial distance of 3 A of the tryptophan residue; no modifications are made within a radial distance of 5 A of the tryptophan residue; no modifications are made within a radial distance of 10 A of the tryptophan residue; no modifications are made within a radial distance of 20 A of the tryptophan residue; no modifications are made within a radial distance of 30 A of the tryptophan residue; no modifications are made within a radial distance of 40 A of the tryptophan residue; no modifications are made within a radial distance of 50 A of the tryptophan residue; no modifications are made within a radial distance of 60 A of the tryptophan residue; no modifications are made within a radial distance of 70 A of the tryptophan residue; no modifications are made within a radial distance of 80 A of the tryptophan residue; no modifications are made within a radial distance of 90 A of the tryptophan residue; no modifications are made within a radial distance of 100 A of the tryptophan residue; no modifications are made within a radial distance of 110 A of the tryptophan residue; no modifications are made within a radial distance of 120 A of the tryptophan residue; no modifications are made within a radial distance of 130 A of the tryptophan residue; no modifications are made within a radial distance of 140 A of the tryptophan residue; no modifications are made within a radial distance of 150 A of the tryptophan residue; no modifications are made within a radial distance of 160 A of the tryptophan residue; no modifications are made within a radial distance of 170 A of the tryptophan residue; no modifications are made within a radial distance of 180 A of the tryptophan residue; or no modifications are made within a radial distance of 190 A of the tryptophan residue; and / or wherein no modifications are made at a position within 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 150, 200, 250, or 300 amino acid residues of the tryptophan residue in the amino acid sequence (e.g., all of the modifications are distant from the tryptophan residue). In some aspects, no modifications are made within a radial distance of 10 A of the tryptophan residue. In some aspects, no modifications are made within a radial distance of 15 A of the tryptophan residue. In some aspects, no modifications are made within a radial distance of 20 A of the tryptophan residue. In some aspects, no modifications are made within a radial distance of 50 A of the tryptophan residue.

[0227] In some aspects, all of the modifications that increases the electrostatic potential of the region surrounding the tryptophan residue are distant from the tryptophan residue (e.g., have a radial distance of at least 15 A or at least 20 A from the tryptophan residue), and a relatively large number of modifications are made. For example, in some aspects, all of the modifications that increases the electrostatic potential of the region surrounding the tryptophan residue have a radial distance of at least 15 A or at least 20 A from the tryptophan residue, and at least 5, at least 10, or at least 20 modifications are made.

[0228] In other aspects, one or more of the modifications that increases the electrostatic potential of the region surrounding the tryptophan residue is proximal to the tryptophan residue (e.g., is within the “contact” distance of 3-5 A), and only one or a few modifications (e.g., 1 -3 modifications) are made. In some aspects, only one modification that increases the electrostatic potential of the region surrounding the tryptophan residue is made, and the modification is proximal to the tryptophan residue.

[0229] In some aspects, the structural model of the polypeptide is determined using Molecular Operating Environment (MOE) homology modeling or the DeepAb modeling technique (Ruffolo et al. 2022. Patterns 3 (2). In some aspects, the structural model of the polypeptide is determined using x-ray diffraction. In some aspects, the structural model of the polypeptide is determined using cryogenic electron microscopy.

[0230] In some aspects, the one or more candidate amino acid residues of the polypeptide are at a position in the amino acid sequence within 100 amino acid residues (e.g., within 95, 90, 85, 80, 75, 70, 65, 60, 55, 50, 45, 40, 35, 30, 25, 20, 15, 10, 5, 4, 3, 2, or 1 amino acid residues, e.g., within 1 -5, 5-10, 10-15, 15-20, 20-25, 25-30, 30-35, 35-40, 40-45, 45-50, 50-55, 55-60, 60-65, 65-70, 70-75, 75-80, 80-85, 85-90, 90-95, or 95-100 amino acid residues) of the tryptophan residue.

[0231] In some aspects, the one or more candidate amino acid residues of the polypeptide that are deleted or replaced to generate the modified polypeptide are negatively charged amino acid residues. In some aspects, the one or more candidate amino acid residues of the polypeptide that are deleted or replaced to generate the modified polypeptide are glutamic acid (E) or aspartic acid (D) residues. In some aspects, the polypeptide comprises a glutamic acid (E) residue at position n-1 relative to the tryptophan residue, and the modifications include replacement or deletion of the E residue. In some aspects, the polypeptide comprises an aspartic acid (D) residue at position n-1 relative to the tryptophan residue, and the modifications include replacement or deletion of the D residue.

[0232] In some aspects, the modifying comprises introducing an amino acid substitution mutation in the amino acid sequence, and the amino acid substitution mutation comprises replacement of a negatively charged amino acid residue (e.g., an E or D residue) with a neutral or positively charged amino acid residue (e.g., a lysine (Lys; K), arginine (Arg; R), histidine (His; H), glycine (Gly; G), alanine (Ala; A), cysteine (Cys; C), valine (Vai; V), proline (Pro; P), leucine (Leu; L), isoleucine (lie; I), methionine (Met; M), phenylalanine (Phe; F), tryptophan (Trp; W), asparagine (Asn; N), glutamine (Gin; Q), serine (Ser; S), threonine (Thr; T), or tyrosine (Tyr; Y) amino acid residue). In some aspects, the amino acid substitution mutation comprises replacement of a D residue or an E residue at position n-1 relative to the tryptophan residue with a neutral or positively charged amino acid residue (e.g., a lysine (Lys; K), arginine (Arg; R), histidine (His; H), glycine (Gly; G), alanine (Ala; A), cysteine (Cys; C), valine (Vai; V), proline (Pro; P), leucine (Leu; L), isoleucine (lie; I), methionine (Met; M), phenylalanine (Phe; F), tryptophan (Trp; W), asparagine (Asn; N), glutamine (Gin; Q), serine (Ser; S), threonine (Thr; T), or tyrosine (Tyr; Y) amino acid residue).

[0233] In some aspects, the amino acid substitution mutation comprises replacement of a neutral amino acid residue with a positively charged amino acid residue. For example, the positively charged amino residue may be a lysine (K) residue or an arginine (R) residue.

[0234] In some aspects, the amino acid substitution mutation comprises replacement of an amino acid residue (e.g., a positive or neutral amino acid residue) with an amino acid residue having a hydrophobic side chain. For example, the amino acid residue having a hydrophobic side chain may be an alanine (A) residue.

[0235] In some aspects, the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a neutral electrostatic potential (e.g., the tryptophan residue that is likely to be oxidized is in a region having a negative electrostatic potential in an unmodified polypeptide, and the modifying results in a modified polypeptide in which the tryptophan residue is in a region having a neutral electrostatic potential).

[0236] In some aspects, the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a positive electrostatic potential (e.g., the tryptophan residue that is likely to be oxidized is in a region having a negative electrostatic potential in an unmodified polypeptide, and the modifying results in a modified polypeptide in which the tryptophan residue is in a region having a positive electrostatic potential).

[0237] C. Mitigation of tryptophan oxidation liability

[0238] A skilled artisan would be readily able to determine whether a modified polypeptide (e.g., a polypeptide having an amino acid sequence that has been modified to increase the electrostatic potential of a region comprising a high-risk tryptophan residue) has a reduced likelihood of tryptophan oxidation. A reduced likelihood of tryptophan oxidation may be a likelihood that is reduced relative to an appropriate reference polypeptide, e.g., a polypeptide having an amino acid sequence that has not been modified to increase the electrostatic potential of the region comprising a high-risk tryptophan residue. In some aspects, the reference polypeptide is the polypeptide in which the high-risk tryptophan residue was identified.

[0239] In some aspects, the modified polypeptide is compared to the reference polypeptide in an in silico assay. For example, in some aspects, (1 ) the electrostatic potential of a region comprising the tryptophan residue; and (2) the solvent accessibility of the tryptophan residue are assessed in a structural model of the modified polypeptide according to the methods provided herein, wherein these parameters predict a lower likelihood of tryptophan oxidation than in the reference polypeptide. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide as predicted by an in silico assay is reduced by at least 1% (e.g., at least 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100% (e.g., is reduced by 1 -10%, 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%, or 90-100%). In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 3%. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide as predicted by an in silico assay is reduced by at least 85%. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide as predicted by an in silico assay is reduced by at least 1 .1 -fold (e.g., at least 1.5-fold, 2-fold, 3-fold, 4-fold, 5-fold, 10-fold, 20-fold, 30-fold, 40-fold, 50-fold, 60-fold, 70-fold, 80-fold, 90- fold, 100-fold, 110-fold, 120-fold, 130-fold, 140-fold, 150-fold, 160-fold, 170-fold, 180-fold, 190-fold, 200- fold (e.g., is reduced by 1 .1 -fold to 5-fold, 5-fold to 10-fold, 10-fold to 20-fold, 20-fold to 30-fold, 30-fold to 40-fold, 40-fold to 50-fold, 50-fold to 60-fold, 60-fold to 70-fold, 70-fold to 80-fold, 80-fold to 90-fold, 90- fold to 100-fold, 100-fold to 110-fold, 110-fold to 120-fold, 120-fold to 130-fold, 130-fold to 140-fold, 140- fold to 150-fold, 150-fold to 160-fold, 160-fold to 170-fold, 170-fold to 180-fold, 180-fold to 190-fold, or 190-fold to 200-fold). In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 5-fold. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 20-fold. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide as predicted by an in silico assay is reduced by 66-fold. In some aspects, the likelihood of oxidation of the tryptophan residue as predicted by an in silico assay is reduced below a threshold that identifies a high-risk tryptophan, e.g., the modification abrogates the tryptophan liability of the polypeptide.

[0240] In some aspects, the modified polypeptide is compared to the reference polypeptide in an in vitro assay, e.g., an assay that includes use of a chemical stressor (e.g., hydrogen peroxide (H2O2), 2,2- azobis(2-amidinopropane) or dihydrochloride (AAPH)) or photooxidation. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 1% (e.g., at least 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100% (e.g., is reduced by 1 -10%, 10-20%, 20-30%, 30- 40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%, or 90-100%)), as determined using an in vitro assay. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 3%, as determined using an in vitro assay. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%, as determined using an in vitro assay. In some aspects, the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 85%, as determined using an in vitro assay. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 50% (e.g., less than 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, or 5% (e.g., is 0-5%, 5-10%, 10-15%, 15-20%, 20-25%, 25-30%, 30-35%, 35-40%, 40-45%, or 45-50%)), as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days (e.g., wherein oxidation rate is measured following treatment with AAPH for 16 days, e.g., at pH 5.5, 37°C). In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 35%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 30%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 25%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days. In some aspects, the frequency of tryptophan oxidation of the modified polypeptide is less than 1%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

[0241] In some aspects, modification of the amino acid sequence of the polypeptide does not substantially impact a desired property of the polypeptide. For example, in some aspects in which the polypeptide binds to a target, modification of the amino acid sequence decreases the affinity of the polypeptide for the target by less than 20%, less than 15%, less than 10%, less than 5%, less than 2%, or less than 1% (e.g., by 0-1%, 1 -5%, 5-10%, 10-15%, or 15-20%) or does not decrease the affinity of the polypeptide for the target.

[0242] For example, in some aspects, the tryptophan residue is at an interaction surface of the polypeptide (for example, a protein-protein interaction surface), and the modification of the amino acid sequence of the polypeptide does not substantially disrupt the interaction.

[0243] In some aspects, the polypeptide is an antibody or an antibody fragment (e.g., an IgG 1 isotype antibody or antibody fragment), the tryptophan residue is in a Fab domain, VH domain, a VL domain, a framework region, or a CDR region of the antibody or antibody fragment, and the modifying does not substantially affect the binding affinity of the antibody or antibody fragment.

[0244] IV. EXAMPLES

[0245] Example 1. Evaluation of AAPH-induced Trp oxidation in a diverse range of antibodies

[0246] Oxidation is a major degradation pathway in proteins with far-reaching implications in physiological functions and pathological conditions (Jung et al., Archives of Biochemistry and Biophysics, 462: 231 -237, 2007; Stadtman and Berlett, Chemical Research in Toxicology, 10: 485-494, 1997; Butterfield and Sultana, Journal of Amino Acids, 2011 : 198430, 2011 ; Cheignon et al., Redox Biology, 14: 450-464, 2018; Rakhit et al., Journal of Biological Chemistry, 277: 47551 -47556, 2002; Zigman in Progress in Tryptophan and Serotonin Research; Schlossberger, H. G., Steinhart, H., Kochen, W., Linzen, B., for Tryptophan Research, I. S. G., Eds.; De Gruyter: Berlin, Boston, 1984; p. 449-468; Spector, FASEB Journal, 9: 1173-1182, 1995; Stadtman, Free Radical Biology and Medicine, 9: 315-325, 1990; Berlett and Stadtman, Journal of Biological Chemistry, 272: 20313-20316, 1997). Oxidative damage is thought to contribute to aging and the development of diseases such as Alzheimer’s (Butterfield and Sultana, Journal of Amino Acids, 2011 : 198430, 2011 ; Cheignon et al., Redox Biology, 14: 450-464, 2018), amyotrophic lateral sclerosis (ALS) (Rakhit et al., Journal of Biological Chemistry, 277: 47551 -47556, 2002), and cataracts by increasing protein susceptibility to proteolysis (Zigman in Progress in Tryptophan and Serotonin Research; Schlossberger, H. G., Steinhart, H., Kochen, W., Linzen, B., for Tryptophan Research, I. S. G., Eds.; De Gruyter: Berlin, Boston, 1984; p. 449-468; Spector, FASEB Journal, 9: 1173-1182, 1995) and reducing enzymatic activity (Stadtman, Free Radical Biology and Medicine, 9: 315-325, 1990; Berlett and Stadtman, Journal of Biological Chemistry, 272: 20313-20316, 1997). In therapeutic antibodies, oxidative damage can compromise structural integrity and function, potentially leading to adverse outcomes including loss of binding affinity (Dashivets et al., MAbs, 1525-1535, 2016; Sreedhara et al., Molecular Pharmaceutics, 10: 278-288, 2013; Wei et al., Analytical Chemistry, 79: 2797-2805, 2007; Hensel et al., PLoS One, 6: e17708, 2011 ; Hageman et al., Pharmaceutical Research, 36: 1 -10, 2019), aggregation (Hensel et al., PLoS One, 6: e17708, 2011 ; Luo et al., Journal of Biological Chemistry, 286: 25134-25144, 2011 ; Joubert et al., Journal of Biological Chemistry, 286: 25118-25133, 2011 ), fragmentation (Wecksler et al., Molecular Pharmaceutics, 15: 1598- 1606, 2018; I Hyes et al., International Journal of Biochemistry Research & Review, 4(5): 367-385, 2014), color change of drug substance (Li et al., Analytical Chemistry, 86: 6850-6857, 2014; Bellmaine et al., Free Radical Biology and Medicine, 160: 696-718, 2020), and altered immunogenicity (Hermeling et al., Pharmaceutical Research, 21 : 897-903, 2004). Understanding the determinants and underlying mechanisms of oxidation and the associated risks are critical for assessing its role in disease progression and ensuring the long-term chemical and physical stability of therapeutic antibodies (Torosantucci et al., Pharmaceutical Research, 31 : 541 -553, 2014; Krause et al., Current Opinion in Biotechnology, 60: 159- 167, 2019).

[0247] Among the amino acids prone to oxidation, methionine (Met) and tryptophan (Trp) stand out as particularly susceptible in antibody therapeutics (Hensel et al., PLoS One, 6: e17708, 2011 ; Krause et al., Current Opinion in Biotechnology, 60: 159-167, 2019; Jarasch et al., Journal of Pharmaceutical Sciences, 104: 1885-1898, 2015; Gaza-Bulseco et al., Journal of Chromatography B, 870: 55-62, 2008). Met oxidation has been extensively studied, with its susceptibility closely tied to solvent accessibility (Sankar et al., mAbs, 10: 1281 -1290, 2018; Yang et al., mAbs, 9: 646-653, 2017; Agrawal et al., Journal of Pharmaceutical Sciences, 107: 1282-1289, 2018). In contrast, Trp oxidation is less well understood, yet arguably more consequential in therapeutic contexts (Wei et al., Analytical Chemistry, 79: 2797-2805, 2007; Hageman et al., Pharmaceutical Research, 36: 1 -10, 2019; Jarasch et al., Journal of Pharmaceutical Sciences, 104: 1885-1898, 2015) for two key reasons. First, Trp oxidation involves multiple reaction pathways and produces a variety of oxidative products, including oxindolealanine, hydroxytryptophan, kynurenine, and N-formylkynurenine (Fig. 1 ). This complexity makes predicting and characterizing Trp oxidation far more challenging (Hageman et al., Pharmaceutical Research, 36: 1 -10, 2019; Li et al., Analytical Chemistry, 86: 6850-6857, 2014). Second, Trp plays a critical role in proteinprotein and hydrophobic core interactions (Ramaraj et al., Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics, 1824: 520-532, 2012; Khemaissa et al., Crystals, 11 : 1032, 2021 ; Barik, International Journal of Molecular Sciences, 21 : 8776, 2020), owing to its unique structural properties.

[0248] Tryptophan stands out among the 20 amino acids due to its aromatic nature, characterized by a six-membered benzene ring fused to a five-membered pyrrole ring with an integrated NH group, forming the indole side chain. Trp’s unique molecular properties arises from a combination of structural features, including (i) a highly accessible polar surface (e.g., the largest non-polar (hydrophobic) area on its two- sided ir-electron face, which is both polarizable and highly accessible due to its planar topology); (ii) the strongest electrostatic potential for cation-n interactions, and (iii) an indole N-H group that is capable of hydrogen bonding. The binuclear aromatic structure of the indole ring enables Trp to navigate large hydrophobic pockets, facilitating interactions requiring a substantial surface area. Such properties make Trp prevalent in membrane proteins, membrane-active toxins, and channels. Redox-active Trp side chains play essential roles in electron transfer and protein function through regulated protonation state Flbonding. Further, these unique properties makes Trp more prevalent (Robin et al., Journal of Molecular Biology, 426: 3729-3743, 2014) and highly surface exposed (Jarasch et al., Journal of Pharmaceutical Sciences, 104: 1885-1898, 2015; Osajima and Hoshino, Computational Biology and Chemistry, 64: 368- 383, 2016) in the complementarity-determining regions (CDRs) of antibodies compared to other proteins.

[0249] Due to its unique features, Trp is frequently strategically placed within the binding sites of protein therapeutics, particularly therapeutic monoclonal antibodies (mAbs), to enhance binding affinity. However, the susceptibility of Trp to oxidation presents a significant challenge, leading to various undesirable consequences. For mAbs, oxidation of Trp residues within crucial regions such as complementarity determining regions (CDRs) can disrupt antigen binding affinity and specificity. Oxidation within the Fc domain may also interfere with interactions with vital receptors like FcRn and Protein A ligand. Trp oxidation can exacerbate susceptibility to other degradation pathways, such as fragmentation and aggregation, impacting thermal stability, serum half-life, and immunogenicity, further complicating the stability landscape of therapeutic proteins.

[0250] A current challenge with Trp oxidation in therapeutic proteins lies in the lack of a complete understanding of the reaction pathway and the difficulty in predicting this liability a priori. Consequently, such liability is often identified late in the drug discovery process or even during development.

[0251] Computational models ranging from quantum mechanical calculations (Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016) to molecular dynamics simulations (Agrawal et al., Journal of Pharmaceutical Sciences, 107: 1282-1289, 2018; Barnett et al., Journal of Pharmaceutical Sciences, 108: 1944-1952, 2019) and machine learning approaches (Sankar et al., mAbs, 10: 1281 -1290, 2018; Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ), have been used to probe the mechanisms of protein oxidation and to develop predictive models.

[0252] Among the key factors influencing oxidation susceptibility, the solvent-accessible surface area (SASA) of Trp and Met has been consistently identified as a critical determinant (Sharma et al., Proceedings of the National Academy of Sciences 111 (52): 18601-6, 2014; Delmar et al., Molecular Therapy Methods & Clinical Development 21 : 466-77, 2021 ; Jacobitz et al., Journal of Pharmaceutical Sciences 110 (2): 719-26, 2021 ; Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016; Barnett et al., Journal of Pharmaceutical Sciences, 108: 1944-1952, 2019; Pavon et al., Analytical Chemistry, 91 : 2192-2200, 2019). Additionally, structural parameters such as dihedral angles {(p and < / / ) (Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ), torsional properties (Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ), loop flexibility (Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016), and proximity to aromatic residues (Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ; Aledo et al., Scientific Reports, 5: 16955. 2015; Aledo et al., BMC Bioinformatics, 18: 430, 2017) have also been investigated, albeit to a lesser extent.

[0253] Design strategies to mitigate or prevent Trp oxidation are as critical as predictive efforts; however, they remain relatively limited. Once a liable Trp degradation site (high-risk oxidation hotspot) is identified (often in later stages of discovery and development), especially for Trp residues within binding sites, one common approach is to substitute oxidation-prone Trp residues with other aromatic amino acids, such as phenylalanine or tyrosine, to reduce oxidation susceptibility (Wei et al., Analytical Chemistry, 79: 2797-2805, 2007; Taylor et al., J. Biol. Chem., 282: 16329-16335, 2007). However, due to the distinct physicochemical properties of Trp, its complete removal (e.g., by substitution) is not always viable and can negatively impact antigen binding or specificity (Luz et al., J. Lipid Res., 61 : 1347-1359, 2020; Kobayashi et al., Protein Engineering, Design and Selection, 12: 879-884, 1999; Saleh et al., Biotechnol. Bioeng., 118: 2923-2933, 2021 ). The unique properties of the Trp side-chain often render it irreplaceable. Mitigation strategies such as the addition of antioxidants such as N-acetyl tryptophan (NAT), lyophilization, or molecular redesign may also be considered. However, the addition of antioxidants is not always effective and can introduce challenges during analytics and development.

[0254] Since avoiding Trp residues in binding sites is not always feasible, rational engineering strategies that can (1 ) predict the oxidation risk a priori and (2)preserve Trp within the binding site while mitigating its susceptibility to oxidation are highly desirable.

[0255] The present Examples address these challenges by leveraging a large dataset of antibody AAPH-driven Trp oxidation profiles of therapeutic proteins containing Trp and employing machine learning to identify novel structural factors that influence oxidation susceptibility. Specifically, the critical role of local electrostatics is emphasized and, as described in the results below, electrostatic potential of the region comprising the tryptophan residue was identified as a structural descriptor which, when used in combination with the solvent accessible surface area, robustly predicted oxidation rate. Furthermore, the studies below demonstrate how modulation of the electrostatic potential be harnessed to rationally design oxidation-resistant antibodies and effectively mitigate oxidation liability. In particular, it is shown in several case studies that by introducing point mutations distant from the binding sites, oxidative stability can be significantly improved while maintaining binding affinity.

[0256] Furthermore, the present Examples introduce a multi-parameter optimization approach that integrates these insights to systematically enhance oxidation resistance while minimizing trade-offs typically associated with Trp substitution (e.g., loss of binding affinity) — an elusive goal that traditional Trp substitutions often fail to achieve.

[0257] The results from this work not only provide a computational predictive model to assess oxidation risk a priori, but also offer a practical strategy to engineer oxidation resisting protein drugs by design, optimizing protein stability and functionality in biopharmaceutical development. Methods

[0258] Data set for tryptophan oxidation

[0259] A data set of about 187 unique monoclonal antibodies (mAbs) was curated. These antibodies were candidates targeting various antigens, including variants designed for the same target. The data set includes approximately 264 tryptophan (Trp) residues that fall within the complementarity-determining regions (CDRs) of the antibodies (“CDR Trps”). Each of the antibodies was subjected to oxidative stress induced by 2-2’-azobis(2-amidinopropane) dihydrochloride (AAPH) as a radical generator. The antibodies were oxidized at a pH of 5.5 using AAPH for 16 hours at 37°C.

[0260] Oxidation rates, measured as the percentage of modifications post-stress, were obtained for each CDR Trp and are reported with respect to a non-spiked control. A historically determined cutoff threshold of 35% oxidation rate was used to classify Trp sites as susceptible (high risk, n=55) or resistant (low risk, n=209) to oxidation. The clinical-stage antibodies and engineered mAbs discussed in the below Examples were similarly oxidized using AAPH. The clinical-stage antibodies were additionally oxidized in a separate set of experiments using an uncharged AAPH variant under the conditions described above for AAPH.

[0261] Results

[0262] A total of 264 CDR Trp data points were generated for the 187 antibodies. 205 CDR Trp residues were identified as “not oxidized” (rate of oxidization of 35% or less), and 59 Trp residues were identified as being “oxidized” (rate of oxidation greater than 35%).

[0263] Fig. 2 shows the frequency of oxidation risks across different antibody CDR loops. Oxidation risk varied significantly among different CDR loops. Analysis revealed that the H3 loop exhibited the highest ratio of high-risk to low-risk Trp sites, followed by H2 and L3. Approximately 40% of Trp sites within the H3 loop underwent oxidation, compared to 20% within the H2 loops, 16% within the L3 and H1 loops, and 10% in the L1 and L2 loops. Given the critical role of the H3 loop for antigen binding (Shirai et al., FEBS Lett., 455: 188-197, 1999; Sela-Culang et al., J. Immunol., 189: 4890-4899, 2012; Tsuchiya and Mizuguchi, Protein Sci., 25: 815-825, 2016), the high frequency of Trp oxidation in this loop is particularly noteworthy, highlighting the associated impact on overall stability and function of antibodies.

[0264] Overall, across all of the CDR loops, 20% of the Trp sites were identified as susceptible to oxidation. These data underscore the high risk of oxidation across antibody candidates, emphasizing the importance of understanding and mitigating oxidative damage in antibody engineering and development.

[0265] Example 2. Analysis of structural features contributing to Trp oxidation using machine learning Introduction

[0266] This example describes the analysis of structural features contributing to Trp oxidation using machine learning. A sizable dataset of AAPH-induced Trp oxidation data spanning a diverse range of antibodies was collected and machine learning classification models were used to pinpoint key structural and molecular descriptors associated with oxidation risk. Structural feature calculation and selection

[0267] Feature selection was grounded in mechanistic insights from previous studies (Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016; Barnett et al., Journal of Pharmaceutical Sciences, 108: 1944-1952, 2019; Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ; Sharma et al., Proceedings of the National Academy of Sciences 111 (52): 18601-6, 2014; Jacobitz et al., Journal of Pharmaceutical Sciences 110 (2): 719-26, 2021 ; Pavon et al., Analytical Chemistry, 91 : 2192-2200, 2019; Aledo et al., Scientific Reports, 5: 16955. 2015; Aledo et al., BMC Bioinformatics, 18: 430, 2017) and hypotheses specific to Trp oxidation. Features such as exposure, backbone dihedral angles, loop flexibility, neighboring amino acid types that have been previously investigated in oxidation contexts were included (Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016; Barnett et al., Journal of Pharmaceutical Sciences, 108: 1944-1952, 2019; Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ; Sharma et al., Proceedings of the National Academy of Sciences 111 (52): 18601-6, 2014; Jacobitz et al., Journal of Pharmaceutical Sciences 110 (2): 719-26, 2021 ; Pavon et al., Analytical Chemistry, 91 : 2192-2200, 2019). New features were also introduced, including electrostatic potential (Epot) and residue-level spatial aggregation propensity (SAP) scores using different hydrophobicity scales, often studied in the context of developability risks related to physical instability such as aggregation and viscosity (Chennamsetty et al., Proceedings of the National Academy of Sciences, 106: 11937-11942, 2009; Park and Izadi, bioRxiv, 2023-07, 2023). SAP scores quantify clustering of hydrophobic amino-acids on the surface, and Epot is the electrostatic potential on the residues for solvated antibodies, described using the Poisson-Boltzmann equation. Here, the potential effects of these descriptors on chemical degradation of antibodies was studied.

[0268] For each antibody in the dataset, 50 residue level structural features were computed, capturing various characteristics such as solvent accessibility (e.g., SASA, SASAJoop, n_loop_water_5), flexibility (e.g., Ioop_rmsd, res_rmsd), local hydrophobicity (e.g., Spatial Aggregation Propensity (SAP) scores), backbone dihedral angles (< / / , (p, and x angles of neighboring residues), local electrostatic potentials (e.g., Epot), and interactions with neighboring polar, non-polar, and aromatic residues (Fig. 3). These structural features were derived through the analysis of a conformational ensemble obtained from Molecular Dynamics (MD) simulations in explicit solvent, starting from a homology model structure of the Fab domains (see Detailed Methods below). A detailed list of features is provided in Table 1 .

[0269] Table 1. Features used to describe Trp oxidation for machine learning Detailed Methods

[0270] Structure modeling and molecular dynamics

[0271] Variable domain (Fv) structures were constructed from antibody sequences using the DeepAb (Ruffolo et al., Patterns, 3: 2022) modeling technique. To generate the full Fab structures, an internal VMD (Humphrey et al., Journal of Molecular Graphics, 14: 33-38, 1996) script was used to graft the Fvs onto the constant region of the herceptin Fab by aligning their elbow loops. The decision to generate the Fab domain, rather than the Fv alone, was made to account for any potential electrostatic contributions from the Fab’s constant region. The resulting Fab structures were then subjected to molecular dynamics (MD) simulations. The Fab structure was solvated and parameterized using the ff14SB force field (Maier et al., Journal of Chemical Theory and Computation, 11 : 3696-3713, 2015) and TIP3P (Mark and Nilsson, The Journal of Physical Chemistry A, 105: 9954-9960, 2001 ) water model in a 10 A truncated octahedron box. Following a relaxation protocol, 500 ns of MD simulation was performed. Further, 1000 snapshots were extracted every 0.5 ns for feature analysis.

[0272] Feature calculations

[0273] For each MD simulation frame, 50 features were extracted to describe the tryptophan oxidation propensity, as shown in Table 1 . For each antibody in the dataset, 50 residue-level structural features were computed, capturing various characteristics such as solvent accessibility (e.g., SASA, SASAJoop, n_loop_water_5), flexibility (e.g., Ioop_rmsd, res_rmsd), local hydrophobicity (e.g., Spatial Aggregation Propensity (SAP) scores), backbone dihedral angles (ip , <p, and x angles of neighboring residues), local electrostatic potentials (e.g., Epot), and interactions with neighboring polar, non-polar, and aromatic residues. The electrostatics Epot features in Table 1 were calculated using the Adaptive Poisson- Boltzmann Solver (APBS) software package (Jurrus et al. 2018. Protein Science 27 (1 ): 112-28) and the Moldesk tool (Park and Izadi, Molecular Surface Descriptors to Predict Antibody Developability, bioRxiv, 2023-07, 2023). The features SASA and Exposure in Table 1 were evaluated using the MSMS tool (Sanner et al., Biopolymers, 38.3: 305-320, 1996).

[0274] Machine learning models and feature analysis

[0275] Feature relationships were assessed by hierarchical clustering. The dendrogram plot was generated using Spearman rank correlation matrix and single linkage clustering to visualize feature relationships. For quantifying feature importance to tryptophan oxidation rate, a Random Forest regressor (1000 estimators) was trained with standardized features using a 5-fold cross-validation split, with SHAP (SHapley Additive exPlanations) values computed on the test set using shap.TreeExplainer. Further Gradient Boosting (GB) and Random Forest (RF) classifiers were used to predict oxidation rates as a binary classification problem. For classification, oxidation rates larger than 35% were used to create two classes (reactive vs non-reactive). Results

[0276] To examine inter-feature relationships (e.g., elucidate cross-correlation and hierarchical relationships among features), a dendrogram diagram was constructed (Fig. 4A). Eleven distinct clusters were identified, which could be grouped in six broad biophysical categories: solvent exposure, local electrostatics, hydrophobic cluster, flexibility, backbone, and side-chain dihedral angles.

[0277] Random Forest (RF) feature importance analysis and SHAP (SHapley Additive exPlanations) values (Fig. 4B) were used to identify key feature categories relevant to oxidation susceptibility. Consistent with prior findings, solvent accessibility (e.g., n_res_water_5, SASA_Loop, SASA_res (see Table 1 )) emerged as the top factor influencing oxidation risk. It is hypothesized that the solvent exposure is essential as buried Trp side-chains are protected from interacting with free radicals, thus are less prone to oxidation.

[0278] These analyses revealed that the second most important feature category was electrostatic potential (e.g., Epot_NEG, Epot_LOOP_SUMPOS, Epot_SUM, Epot_LOOP_SUM (see Table 1 )); this feature had not been identified in previous analyses in the context of oxidation susceptibility. Shapley additive explanations (SHAP) value analysis further revealed that higher Epot_NEG values, which indicate a stronger negative local electrostatic potential around the Trp side-chain, increase the risk of oxidation (Fig. 4B) while a positive potential (Epot_POS) tends to lower the oxidation risk.

[0279] Several additional features were also found to be important, although to a lesser degree. The local clustering of hydrophobic and aromatic groups (e.g., nb_seq_score_2, SAP scores (see Table 1 )) suggest an elevated oxidation risk. Further, greater local flexibility (e.g., higher res_rmsd) also correlated with increased oxidation susceptibility. These findings align with previous studies highlighting the effect of loop flexibility (Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016), and the role of neighboring aromatic residues (Delmar et al., Molecular Therapy Methods & Clinical Development, 21 : 466-477, 2021 ; Aledo et al., Scientific Reports, 5: 16955. 2015; Aledo et al., BMC Bioinformatics, 18: 430, 2017).

[0280] Example 3. A minimalistic two-parameter classifier for identifying Trp residues likely to have oxidation liabilities

[0281] Next, a mechanistic analysis of the role of the top two features the model described in Example 2 identified as driving Trp oxidation was performed. A negative electrostatic potential of the region comprising the tryptophan residue was found to substantially increase the susceptibility to oxidation. Importantly, as discussed in further detail below, a simple two-parameter model based on these features achieved 79% accuracy in categorical oxidation risk prediction, compared to 84% accuracy using a machine learning (ML) classifier trained on the full feature set.

[0282] Results

[0283] Deeper insights into the predictive power of the top features and their correlation with the dataset were gained by evaluating the performance of a simplified two-parameter model based on the top features, solvent accessibility and local electrostatics. The performance of this two-feature model was compared with that of Random Forest (RF) and Gradient Boosting (GB) classifiers that were trained on the full feature set (high-feature models) without a traditional train-test split, instead using 10-fold cross- validation (Table 2). The high-feature models achieved approximately 84% accuracy, with RF and GB models demonstrating high specificity (96% and 94%, respectively), but low sensitivity (40% for RF and 45% for GB), indicating a high rate of false negatives.

[0284] The predictive power of the top features was also assessed individually (Figs. 6 and 7, and Table 2).

[0285] The correlation plot between solvent-accessible surface area (SASA) and the oxidation rate (Fig. 6) indicates a positive correlation, suggesting that higher SASA values are associated with increased oxidation rates. To further investigate this, Trp residues were categorized as (i) buried or (ii) exposed. Based on the correlation plot (Fig. 6), Trp residues with SASA <50 A2were empirically defined as buried, which aligns with the relative solvent accessibility (RSA) threshold of <20-25% used in previous work (Wu et al., BioData Mining, 10: 1 , 2017). As shown in Table 2, the Solvent Accessible Surface Area (SASA) feature alone achieved a very high sensitivity of 91 %, indicating that almost all buried Trp residues (SASA < 50 A2) were resistant to oxidation. Buried Trp residues are linked to a high true negative rate, with 96 out of 101 instances correctly classified as non-reactive. This is expected, as the burial of Trp side-chains protects them from interacting with free radicals, thereby reducing oxidation risk. In contrast, exposure (SASA >50 A2) does not necessarily lead to oxidation. Only 50 out of 163 exposed Trp sites were found to be oxidation-prone, resulting in a specificity of just 46%, suggesting that high solvent accessibility was a necessary attribute, but not sufficient on its own to explain Trp oxidation.

[0286] Table 2. Summary of Classification Results

[0287] As is shown in Fig. 5, using SASA alone as a predictor of oxidation results in false negatives (top left quadrant; Trp residues that were buried and oxidized) and false positives (bottom right quadrant; Trp residues that were exposed but not oxidized).

[0288] The second top feature provided by the model, the electrostatic potential of the region comprising the tryptophan residue (Epot), shows a negative correlation to the oxidation rate, with approximately 85% (146 out of 171 ) of Trp residues in a positively charged potential (Epot > 0) being correctly classified as non-reactive (Fig. 7 and Table 2). Epot achieved a higher accuracy compared to SASA when used as a single classifier (67% vs. 55%) (Table 2). While specificity was improved relative to SASA, indicating that a positively charged electrostatic environment (positive Epot) correlates with a low risk of oxidation, the sensitivity was lower than that of SASA. Fig. 7 shows the Trp oxidation rate in the set of tested antibodies as a function of Epot alone.

[0289] Importantly, it was discovered that combining SASA and Epot significantly enhances the accuracy of the low-feature model compared to single-parameter models. As shown in Fig. 9, 85 out of 109 (79%) exposed T rp residues in a positively charged environment were below the oxidation rate cutoff, compared to only 28 out of 54 (52%) of the exposed sites in a negatively charged environment. The two- parameter classification model using SASA and Epot achieves an overall accuracy of 79%, which is only slightly (5%) lower than the performance of the RF and GB models that incorporate the full set of structural descriptors (Table 2). The result highlights that the solvent accessibility and electrostatic potential are major determinants of Trp oxidation rates, capturing much of the predictive power offered by the more complex, high-feature models.

[0290] As is shown in Figs. 8 and 9, Trp residues that have a high SASA (e.g., a SASA above a threshold of 50 A2) and are in a region having a low electrostatic potential (e.g., an electrostatic potential at or below 0) have a greater rate of oxidation.

[0291] Detailed Methods

[0292] Antigen binding kinetics

[0293] All surface plasmon resonance (SPR) measurements were conducted on a BIACORE™ T200 instrument (Cytiva). Anti-human IgG antibody was immobilized on a CM5 chip according to the manufacturer’s recommendations. The mAbs were then captured on the chip (1 pg / mL, 60 seconds) and affinity for the respective antigen was determined using single cycle kinetics. 0.01 M HEPES pH 7.4, 0.15 M NaCI, 0.005% v / v Surfactant P20 was used as the running buffer. Antigen concentrations were varied to bracket the KD, and dissociation was monitored for 600-1800 seconds after injection. The sensograms were corrected by subtracting from the blank flow cell as well as a zero point run (0 nM antigen). The corrected sensograms were analyzed using the manufacturer’s software and fit to a 1 :1 Langmuir binding model to calculate the kinetic and binding parameters.

[0294] Example 4. Validation of the two-parameter classifier on an independent set of clinical antibodies

[0295] To assess the generalizability of the low-feature model of Example 3 and ensure that the accuracy of SASA and Epot in identifying Trp residues likely to have oxidation liabilities is transferable to other datasets, blind predictions of Trp oxidation were conducted for a separate subset of eight clinical- stage antibodies, using SASA and Epot as key descriptors for Trp oxidation risk.

[0296] Methods

[0297] To design the validation set, variable domain sequences were collected for a panel of 629 clinical-stage IgG 1 molecules sourced from a 2022 snapshot of TheraSAbDab (Raybould et al., Nucleic Acids Research, 48: D383-D388, 2019). Static DeepAb (Ruffolo et al., Patterns, 3: 2022) structure models of their Fab domains were generated (see Detailed Methods in Example 2), and SASA and Epot values were calculated for all Trp residues appearing in CDR regions. A two-dimensional scatter plot showing these properties of the antibodies is provided in Fig. 10. The graph was divided into four quadrants based on SASA (with a cutoff of 50 A2) and Epot (with a cutoff of 0 kT / e). Several antibodies were randomly selected from each quadrant, with ten cases (Trp residues) from eight molecules selected in total. Four of the selected Trp residues had a large SASA and a negatively charged environment (negative Epot) (indicated by red data points in Fig. 10). These four CDR Trp residues were predicted to be susceptible to oxidation, while the Trp residues in low SASA and / or positively charged settings (indicated by blue data points) were expected to have be stable (i.e. , to be oxidation-resistant).

[0298] The VH and VL sequences of the eight selected molecules are provided in Table 3. All eight molecules used the same IgG 1 framework, but differed in the variable domain. The eight molecules were expressed and stressed using AAPH, and Trp oxidation rates were measured as described above.

[0299] Table 3. Antibody Sequences

[0300] Antibodies with Trp residues having positive Epot and low SASA (MabG H3 and MabG L3) positive Epot and high SASA (MabE, MabF, and MabH), negative Epot and low SASA (MabA L3), and negative Epot and high SASA (MabA H3, MabB, MabC, and MabD) represented all four quadrants of Fig. 10.

[0301] Results

[0302] Table 4 presents the SASA and Epot descriptors, blinded Trp stability predictions (based on SASA and Epot), and the experimentally measured Trp oxidation rates for each of the eight antibodies selected from the validation set. Remarkably, the model’s predictions agreed with the measured oxidation data for all ten Trp residues within the eight clinical antibodies. The four molecules with tryptophan residues with SASA of 50 A2or larger and negative electrostatic potential (<0 kT / e) showed significant oxidation (between 51 .34% and 93.35% modifications after AAPH stress) (Table 4). These findings demonstrate the predictivity of these newly identified descriptors.

[0303] Table 4. Properties of antibodies in the validation set

[0304] * SASA and Epot above the cutoff threshold for Trp oxidation.

[0305] ** Oxidation rate above the cutoff threshold (>35%).

[0306] SASA and Epot values are averaged over molecular dynamics (MD) trajectories.

[0307] Interestingly, in two molecules (Mab E and Mab F) that had exposed Trp residues (SASA > 50 A2), but were in a positively charged environment (had a positive Epot), the CDR Trp residues were oxidation-resistant (did not undergo oxidation). In addition, Mab A, which had a Trp with a slightly negative electrostatic potential (Epot) but borderline SASA (44.9 A2), also remained stable (did not undergo oxidation). These results confirm that SASA and Epot descriptors individually are not reliably predictive of Trp oxidation rate.

[0308] This external validation set highlights the generalizability and effectiveness of combining SASA and Epot for predicting oxidation susceptibility and identifying high-risk oxidation sites.

[0309] Example 5. Rational design engineering strategies to mitigate Trp oxidation risk

[0310] Antibody complementarity-determining regions (CDRs) often contain Trp residues that are important for binding of the antibody to its target. In many cases, these Trp residues need to be fully exposed and solvent-accessible to enable antigen binding. As a result, attempting to resolve issues with Trp oxidation by eliminating a Trp residue that is at risk of oxidation, or by making the Trp residue less solvent-exposed (a non-trivial endeavor) can negatively impact binding. In contrast, the long-range nature of the electrostatic potential offers a unique engineering opportunity: one can attempt to modulate electrostatic potential by introducing point mutations that are far from the Trp residue, rendering the long- range electrostatics unfavorable for oxidation reactions. The present example explores this strategy of reducing oxidation risk through targeted distant electrostatic mutations (modulating the strength and sign of the electrostatic potential around susceptible Trp residues).

[0311] Methods

[0312] Leveraging the long-range nature of electrostatics, a rational engineering strategy involving point mutations distant to susceptible Trp residues was used to successfully mitigate oxidation risks in several IgG 1 isotype antibodies, while preserving binding affinity in specific cases. The engineering strategy was tested in a set of five antibodies with high-risk Trp residues in a CDR region. All five antibodies had a CDR comprising a Trp residue within a region of negative electrostatic potential and underwent oxidation at a rate of more than 35%, as determined using AAPH-induced stress (Table 5). The Trp residues susceptible to oxidation were in the L3 (mab1 ), H2 (mab2) and H3 (mab3 and mab4) CDR loops.

[0313] Table 5. Summary of mutations, electrostatic potential, SASA, oxidation rate, and change in antigen binding kinetics upon point mutations (relative to wild-type; measured by SPR at pH 7.4) for the Trp residues in the CDRs of engineered lgG1 isotype antibodies

[0314] NB: No binding observed

[0315] *KD could not be accurately determined as off rate was reaching the limit of the instrument. To change the electrostatic potential of the CDR comprising the tryptophan residue and mitigate oxidation risk, point mutations substituting nearby acidic residues with basic residues were made. In particular, acidic amino acid residues (Asp residues) were identified and mutated to basic residues (Arg or Lys residues) or to Ala. For each antibody, engineering began with a single mutation of Asp to Arg (D to R), to Lys (D to K), or to Ala (D to A); if necessary, further D to R, D to K, or D to A mutations were then added until the electrostatic potential of the region comprising the tryptophan residue changed from negative to positive. Single, double and triple D R / K / A mutations were made. It is noted that these designs focused exclusively on removing or mitigating the oxidation risk, without necessarily optimizing or considering the impact on binding affinity.

[0316] Each of the new variants was expressed, and oxidation rates were measured under AAPH- stressed conditions.

[0317] Results

[0318] Table 5 and Fig. 13 show the calculated descriptors for the original antibodies (Mab1 , Mab2, Mab3, Mab4, and Mab5) and their engineered variants designed to mitigate oxidation rate (Mab1 .RRR, Mab2.RRR, Mab3.R, Mab4.K, Mab4.RR, Mab4.KR, Mab5.A, and Mab5.AA). The oxidation rate was reduced in most of the engineered variants relative to the original antibodies. As shown in Table 4 and Figs. 12A and 12B, the mutation of acidic residues (negative electrostatic potential) to basic residues (positive electrostatic potential) decreased the oxidation rate of Trp, further supporting the influence of the negative electrostatic potential on Trp oxidation.

[0319] For Mab1 , the Trp residue of interest (W91 in the L3 CDR loop) had a highly negative electrostatic potential (-1 .55 kT / e) and a SASA of 58.38 A2, resulting in a high oxidation rate of 70.6%. To counteract this, three Asp residues in the H3 loop were mutated to Arg (D97R, D99R, D101 R;

[0320] “Mab1 .RRR”). This triple mutation shifted the side-chain electrostatic potential to a mildly positive value (0.69 kT / e) and reduced SASA (28.5 A2). The impact on SASA was likely due to increased packing and conformational adjustments. Through these mutations, the oxidation rate decreased by 46%, lowering to 38.1 % in the triple variant Mab1 .RRR (see Table 5 and Figs. 1 1 and 13).

[0321] In Mab2, a similar approach was taken to reduce the oxidation rate of W58 in the H2 loop (initially 45.6%). Three Asp residues within the H3 loop were replaced with with 3 Arg residues (D100cR,D1 OOdR, D101 R; generation of Mab2.RRR), shifting the local electrostatic potential significantly from -1 .52 kT / e to 1 .65 kT / e. Unlike in Mab1 , the SASA for this variant remained high (66.04 A2compared to 80.53 A2for the wild type), and the mutations were positioned at distances ranging from 10.5 A to 18.6 A from the Trp site (Figs. 12A and 12B). Despite the high SASA in the variant Mab2.RRR, the oxidation rate decreased from 45.6% to 28.3%. The modified antibody had a Trp oxidation rate below the above-described threshold of 35%; thus, the mutations converted the reactive Trp to a non-reactive Trp.

[0322] The Mab3 variant was an outlier in the engineered set. For this molecule, a single D98R mutation was introduced to generate Mab3.R. This mutation significantly altered the electrostatic potential, but had minimal impact on SASA. The reduction in the oxidation rate of the Trp residue of interest (W100b) in H3 was marginal, with only a 2% decrease, suggesting that altering electrostatics alone was insufficient in this case. It is hypothesized that the extremely high SASA (close to RSA of 90% or 197 A2) may be the dominant factor, overshadowing the effects of changes in electrostatic potential and limiting the impact on oxidation.

[0323] The Mab4 variants exhibited a reduction in Trp oxidation similar to that observed in the Mab1 and Mab2 variants. Mab4 contains a reactive Trp residue (W100a) in the H3 loop. As shown in Fig. 13, when an Asp residue at position n-1 relative to the Trp was mutated to a Lys residue (thereby generating Mab4.K), the oxidation rate of the reactive Trp residue was decreased by more than 50% (from 86.2% to 41 .7%). Additionally, modifying Mab4 to replace two Asp residues with Arg residues in the H3 loop (thereby generating Mab4.RR) nearly eliminated the oxidation liability of Trp in the H3 loop, with Trp oxidation reduced to 1 .3%. Moreover, in the case of Mab4.RR, the mutation of these two residues increased the solvent accessibility (SASA) by 78% compared to the wild type (RSA = 49% or SASA = 107.85 A2), suggesting that for intermediate RSA values, electrostatic potential becomes the dominant factor in oxidation susceptibility.

[0324] The Mab4.KR was designed by mutating an Asp residue at position n-1 relative to the Trp to Lys, and mutating an Asp residue in the H3 loop to Arg. Interestingly, the Mab4.KR double mutation shifted the side-chain Epot to a small negative values (-0.24 kT / e), compared to a positive state. This modification was still highly effective in lowering the oxidation rate by approximately 80%, reducing it to 16.6%, though the oxidation is higher than mab4.RR (where the side chain Epot is positive).

[0325] In Mab5, a single D31 A mutation reduced the oxidation rate of the Trp residue of interest (W33 (H1 )) from slightly above the threshold (40.7%) to just below it (34.7%). It was hypothesized that the presence of Y32 in the n-1 position might be shielding the effect of the D31 A mutation on W33. To further investigate this, a Y32A mutation was introduced in combination with D31 A: this resulted in an increase in the positive potential of the W33 site. The double D31 A and Y32A mutation further reduced the oxidation rate by 54%, bringing it down to 18.7%. This combination underscores the synergistic effects of aromatic clustering (third important feature in the ML model) and charge modulation in mitigating oxidation risk.

[0326] These results demonstrate that Trp oxidation can be modulated by modulating the negative electrostatic potential of the region comprising the Trp, and further demonstrate that the mutation of distant residues (targeted distant electrostatic mutations) can impact Trp oxidation, providing an engineering strategy to mitigate Trp oxidation.

[0327] Example 6. Intentional design of oxidation-prone Trp residue in a case study

[0328] In addition to efforts aimed at reducing oxidation rates, experiments were performed to explore whether an oxidation-prone Trp residue could be intentionally created by introducing Trp into a negatively charged surface patch.

[0329] It was observed that the H3 loop of Mab2 had a substantial negative electrostatic potential (Figs. 14A and 14B), and a variant of Mab2 comprising replacement of an Arg residue (R100A) within the H3 loop with Trp (mab2.W) was designed (see Table 2 and Fig. 12A). The Arg residue was selected for replacement due to its bulky size (to resemble Trp) and positive charge. It was predicted that this introduced Trp in Mab2 would oxidize due to its location in a region with negative local electrostatic potential.

[0330] The SASA for W100a in Mab2.W was 220.81 A2, and the electrostatic potential of the Trp sidechain (Epot) was 2.22 kT / e (Table 2). Experimental measurement of the oxidation rate after AAPH stress confirmed that W100a in Mab2.W is susceptible to oxidation, exhibiting an extremely high oxidation rate of 96.8% (Table 2). This high rate of oxidation provides further direct evidence that local electrostatics and solvent accessibility together, independent of the identity or type of neighboring structures or sequences, significantly influence oxidation risk in proteins. Example 7. Impact of mutations on antigen binding and multi-parameter optimization strategy to balance affinity and oxidation risk in anti-CD33

[0331] Impact on Antigen Binding

[0332] In view of the successful reduction in oxidation rates achieved through point charge mutations (Example 6), the impact of the variants on binding was tested. Although the original designs were focused solely on minimizing oxidation (without considering affinity), binding was preserved in 4 out of 8 designs (50%), as shown in Table 5. Notably, Mab4.K, Mab5.A, and Mab5.AA retained strong binding while having reduced oxidation. Mab3.R also maintained binding, but did not have reduced oxidation risk. Overall, since no specific strategy was employed to preserve binding, the fact that binding was maintained in several cases is promising and suggests that the present approach holds potential for simultaneously optimizing both binding affinity and stability. This is further tested below.

[0333] Multi-parameter optimization strategy to balance affinity and oxidation risk in anti-CD33

[0334] The above-described approach was extended to simultaneously optimize both affinity and oxidation resistance for a critical Trp residue (W96) in an anti-CD33 mAb. W96 is located at the VH and VL interface and participates in a cation-n interaction with K52, as illustrated by a Fab-antigen complex CryoEM structure (see Fig. 17A). The folding structure and orientation of W96 are further stabilized through intermolecular hydrogen bonding between D101 and W96, as well as an interaction between R94 and D101 . The ability of the side-chain to form both cation-n and hydrogen bonds simultaneously appears to be essential for binding in this context. Fig. 17B illustrates that W96 is located near a negatively charged surface, making it prone to oxidation. In fact, W96 oxidizes at a rate of 97% after AAPH stress and severely impairs antigen binding affinity, reducing it by over 500-fold. A W96F mutation also led to an over 1000-fold reduction in binding affinity, rendering this traditional substitution approach impractical.

[0335] Building on the insights from the present study (see Examples 1 -6), a dual-optimization strategy was developed to target both Trp stability and binding affinity. It is noteworthy that the CD33 Fab-antigen CryoEM structure was not available during this optimization phase of molecule development. As such, the present approach did not rely on the complex structure, which is commonly used for optimization during discovery campaigns.

[0336] Over the course of three iterations, variants of the lead sequence were generated to explore (1 ) sequence stability and increase in the overall isoelectric point (pl) of the molecule, (2) locally targeted affinity and charge optimization to those residues within 10A around the W96 to maximize the impact on local electrostatics, and (3) a combinatorial exploration based on affinity and expression results of iterations 1 and 2.

[0337] Fig. 17C shows that the charge-altering point mutations are effective in increasing side-chain Epot and reducing the oxidation rate. A strong positive correlation between calculated Epot on W96 and percentage of oxidation reduction is observed. Fig. 17D describes the top thirteen variants and their electrostatic feature (Epot), oxidation reduction, and binding impact compared to the starting lead candidate (relative KD). The impact on binding affinity varied, with more mutations typically resulting in greater binding impact. Two candidates stood out as most promising based on their ability to reduce oxidation risk while maintaining sufficient binding. Variant S11 , with a single point mutation, achieved a 27.2% reduction in oxidation and a modest 1 .4-fold reduction in binding. Meanwhile, variant S13 showed a 49.9% oxidation reduction, albeit with a 20.7-fold decrease in binding affinity — still superior to the W96F variant, which resulted in a >1 OOO-fold binding loss. These findings highlight the potential of the dualoptimization approach for developing oxidation-resistant, high-affinity antibodies.

[0338] Detailed Methods

[0339] Affinity Determination of multi-property optimization clones using SPR

[0340] The binding affinity of protein and protein interaction was determined by surface plasmon resonance (SPR) technology (BIACORE™-8K+, Cytiva). Briefly, each antibody variant was captured by Protein A sensor chip (Series S) on a different flow cell to achieve approximately 150 response units (RU), followed by the injection of five-fold serial dilutions of human CD33 protein (R&D Systems; 0.16 nM to 100 nM) in HBS-EP buffer. Sensorgrams were processed with reference and blank subtraction and analyzed using a 1 :1 Langmuir binding model to calculate association rate constant (kon), dissociation rate constant (koff) and the equilibrium constant (KD).

[0341] Multi-property optimization of affinity and tryptophan oxidation

[0342] Variants of the internal lead candidates were engineered in three iterative steps with distinct optimization objectives. Initially, a structurally-aware sequence optimization approach from Hie et al. {Nature Biotechnology, 42: 275-283, 2024) was adapted to identify acidic framework residues (Asp / Glu) for substitution with basic residues (Arg / Lys) in tolerated positions, generating 24 candidate sequences for small-scale expression and affinity measurements.

[0343] Next, mutations were introduced to A, K, and R around the oxidizing Trp residue to increase affinity, identifying single mutations, light chain double mutations, and combinations of heavy chain double and triple mutations.

[0344] In the final step, beneficial mutations from both previous stages were combined in silico, producing 23,915 candidate sequences. Active learning (Daulton et al., Differentiable Expected Hypervolume Improvement for Parallel Multi-Objective Bayesian Optimization. Advances in Neural Information Processing Systems. 9851-9864, 2020) was used to select 90 sequences with desired properties, which were tested for improved affinity and oxidation resistance, demonstrating the effectiveness of multi-property optimization via active learning-assisted combinatorial mutagenesis.

[0345] Cryogenic Electron Microscopy Data Collection, Processing, and Model Refinement

[0346] Cryogenic electron microscopy (cryo-EM) data were collected using a Titan Krios G3i with a K3 camera, capturing over 18,000 movies at varying tilts. The data were processed with cryoSPARC, including motion correction, CTF estimation, and 2D classification, resulting in a 3D reconstruction at 3.5 A resolution. Model building was done using UCSF Chimera and the human CD33 PDB structure, followed by refinement in COOT and Phenix. The final model was validated and refined to 3.5 A resolution. Example 8. Mechanistic insight into the effect of electrostatic potential on Trp oxidation

[0347] To elucidate the mechanisms underlying the differences in oxidation rates between Trp residues in regions having negative vs. positive electrostatic potentials, oxidation / reduction (redox) potential molecular dynamics (MD) calculations were performed (Cruziero et al., The Journal of Chemical Physics, 149: 2018).

[0348] Methods and Results

[0349] Mab4 (an antibody comprising a Trp site in a region of negative electrostatic potential) and the engineered variant mab4.RR (comprising a Trp site in a region of positive electrostatic potential) were selected as exemplary models, and a redox replica exchange simulation was performed. Mab4 and Mab4.RR differ in that Mab4.RR comprises two aspartic acid (Asp; D) to arginine (Arg; R) substitution mutations. As shown in Example 6, these two mutations in mab4 decreased the rate of Trp oxidation by 85%, making it an ideal candidate for further exploration.

[0350] The oxidized state of Trp considered in this Example has the radical on the nitrogen atom of the indole ring instead of the carbon atom position 2 and 3 (Wood et al., The Journal of Physical Chemistry A, 120: 2926-2939, 2016). Previous reports have shown that Trp radical formation during oxidation occurs in the carbon atom close to the nitrogen of the indole ring. However, the nitrogen radical Trp was considered instead because the available electrochemical potential values and Gibbs free energy ( 1G) data corresponded to this particular oxidation transformation (Tyson et al., The Journal of Physical Chemistry Letters, 11 : 2408-2413, 2020).

[0351] Redox replica exchange was performed to determine whether any changes were associated with the electrochemical potential for any type of Trp radical formation. The Trp residue (TRX 100a) was titrated over a range of redox potentials between 900 mV and 1250 mV, at intervals of 50 mV. As shown in Fig. 16, the fraction of reduced species from the redox runs was evaluated, and the reduction potential (E°) of TRX 100a (a titratable Trp) of Mab4 and Mab4.RR was derived therefrom. The redox potential was obtained by fitting the fraction of reduced species with the Nernst equation, ultimately leading to a titration curve and a redox potential (E°).

[0352] As shown in Figs. 15 and 16, the Trx 100a of Mab4 (which is in a region having a negative electrostatic potential) and of Mab4.RR (which is in a region having a positive electrostatic potential) have different behaviors: the Trx 100a of Mab4 had an E0that was 88 mV lower than that of Mab4.RR, and the Mab4.RR had an E ° similar to a capped Trp control. The observed Trp oxidation may be a proton- coupled electron transfer process. The difference of 88 mV corresponds to approximately a 1 .5 pH unit drop, and thus a significant shift in the redox potential (decrease in E °).

[0353] To conclude, the lower reduction potential observed in Mab4 reveals that Trp residues in an area of negative electrostatic potential are more prone to oxidation (have lower E °) than Trp residues in an area of positive electrostatic potential, providing a mechanistic understanding of Trp oxidation. These results indicate that being in a region having a negative electrostatic potential affects the electrochemical potential of the Trp, making it more prone to oxidation than a Trp in a region having a neutral or positive electrostatic potential. Detailed Methods

[0354] For performing redox MD, the Trp (TRX) residue was parametrized to be titratable. Two states for Trp were defined: State 1 (Trp) and State 2 (Trp-+). The charges for each atom were obtained using the Gaussian 09 program with HF / 6-31 G* level capped Trp dipeptide (ACE-TRX-NME), as shown in Table 5. Constant redox potential MD was performed on capped Trp dipeptide (ACE-TRX-NME) and using Metropolis Monte Carlo redox state change attempt similar to previous publications (Cruzeiro et al. 2018. The Journal of Chemical Physics 149 (7)), the 21G_elec,ref term was fitted to predict E°. Next, the Trp 321 residue in Mab4 was defined as titratable Trp residue (TRX) and redox runs were conducted. Similar to previous MD setups, minimization, heating and equilibration were done for the Mab4 and Mab4.RR with the titratable TRX residue. The last equilibration coordinates were used as the starting points for all redox windows. A total of 8 windows ranging from 900 mV to 1250 mV in intervals of 50 mV were used for redox run. For each potential, an initial equilibration of 20 nanoseconds (ns) with a redox change attempt every 500 steps at constant volume were conducted. The equilibrated structures at each potential were used for the redox replica exchange run. Two replicates of 60 ns redox replica exchange runs were conducted with 10000 number of times exchange between redox potentials attempted and 10 ps of solvent relaxation after every exchange. The first 10 ns were considered as equilibration, and the last 50 ns were used to evaluate the redox potential using cestats.

[0355] Table 5. Charge distribution for the Redox Active Tryptophan (TRX) Residue

[0356] Discussion

[0357] In the study captured by these Examples, oxidation rates that were experimentally measured for a diverse panel of 187 antibodies were analyzed and ML classifiers using expert-guided features were developed to identify key structural determinants of Trp oxidation risk. Leveraging a comprehensive set of molecular descriptors, it was confirmed that solvent accessibility is a critical factor in oxidation susceptibility.

[0358] Additionally, it was found that a negative electrostatic potential surrounding Trp side chains significantly increases oxidation risk. A simplified two-parameter model incorporating these top predictors (SASA and Epot), achieved 79% accuracy in classifying oxidation risk, closely approaching the 84% accuracy of full-feature ML models (models that use the full feature set and / or are trained on the entire data set).

[0359] While the impact of solvent accessibility on oxidation has been well-studied, the strong correlation between local electrostatic potential and oxidation rate is particularly intriguing.

[0360] To further validate these findings, the model was applied to a blind test set of eight clinical-stage antibodies with 10 Trp residues within their CDRs. Despite variations in sequence, target, and other molecular attributes, the model accurately predicted oxidation susceptibility, reinforcing the strong association between local electrostatic environment, solvent accessibility, and Trp oxidation risk.

[0361] The long-range impact of electrostatics on oxidation risk enables a rational design strategy for mitigating oxidation through targeted point mutations. The present Examples demonstrate that introducing 1 -3 charge mutations near high-risk Trp residues reduced oxidation by an average of about 50% (in four of the five tested cases), while preserving binding affinity in half of them. Notably, in several case studies, effective mutations were introduced in CDR loops distinct from the one containing the oxidation-prone Trp, or even up to 10 A away from the Trp, underscoring the long-range impact of electrostatic modulation on oxidation susceptibility. Further, these observations provide a simplistic way of flagging high-risk Trp residues based on their SASA and the electrostatic potential of the region, which is useful in early-stage molecule development.

[0362] These insights were integrated into a multi-parameter optimization strategy designed to simultaneously enhance both oxidation resistance and binding affinity in an exemplary anti-CD33 antibody. By shifting the isoelectric point (pl) of the Fv region toward higher values (creating a more positively charged Fv) and employing a sequence-based affinity maturation approach, several variants of an anti-CD33 antibody containing an oxidation-prone Trp96 site (oxidizing at a 97% rate upon AAPH stress) were engineered. These optimized variants exhibited a significant reduction in oxidation while maintaining strong binding affinity. Notably, the binding loss was substantially lower than that observed when substituting Trp with Phe, a commonly used engineering alternative for mitigating oxidation.

[0363] To better understand how local electrostatics influence the energetics of oxidation, redox potential MD simulations were conducted on two closely related variants of an oxidation-prone antibody. These variants were generated through double D^R mutations in Mab4. The simulations revealed a higher population of the reduced (Trp) species compared to the oxidized (Trp.+) species at the Trp site when situated in a positively charged electrostatic environment. This finding confirms that local electrostatics alone can significantly modulate the Gibbs free energy of the electron transfer process.

[0364] Mechanistically, oxidation (electron loss) is stabilized in electron-rich environments due to electrostatic repulsion, which simultaneously makes electron regain unfavorable. Consequently, a positive electrostatic potential increases the energetic cost of oxidation.

[0365] While prior studies relating to Trp redox chemistry have primarily focused on short-range, water- mediated interactions and local hydrogen bonding, the results presented herein provide direct evidence that the effects of local electrostatics on oxidation can extend over long distances, as evidenced by Poisson-Boltzmann calculations. Furthermore, the results provided herein demonstrate that point mutations not directly interacting with the oxidation-prone Trp residue can still significantly reduce its oxidation rate in an antibody. These insights expand the scope of sequence-based modifications, offering a more efficient method for protein optimization, particularly in antibody discovery and development.

[0366] While the present study is focused on oxidation induced by AAPH, oxidation can also be triggered by other stressors, such as light exposure, metal-catalyzed oxidation, or thermal stress. The relevance of the electrostatically driven mechanism presented in these Examples to other oxidation pathways is supported by the redox potential MD simulations presented herein, which, independent of specific stressors, explain how changes in the electrostatic environment alter the energetics of oxidation.

[0367] To further assess the generalizability of the present findings, a neutral variant of AAPH was tested. It was noted that AAPH, being positively charged as an excipient, might be preferentially attracted to negatively charged environments, potentially increasing the local oxidation risk. To assess this, oxidation measurements for eight clinical antibodies were repeated using a neutral variant of AAPH. As shown in Fig. 18, the experimentally determined oxidation rates remained consistent across both conditions, indicating that AAPH primarily serves as a radical generator in solution rather than requiring direct interaction with oxidation-prone sites.

[0368] Overall, these findings suggest that the electrostatic descriptors provided herein are broadly applicable to other types of oxidation.

[0369] In conclusion, the study presented in these Examples establishes a strong link between local electrostatics and the Trp oxidation rate, demonstrating the feasibility of predicting Trp oxidation risk and rationally designing oxidation-resistant antibodies by analyzing and modulating the electrostatic environment around Trp residues. These findings provide a practical and effective method for optimizing both oxidation resistance and binding affinity, and provide a framework for designing oxidation-resistant biotherapeutics, enabling multi-parameter optimization of antibody stability and function and advancing antibody engineering efforts. These results enable early detection of Trp oxidation risks based solely on antibody structures and offer rational design engineering to optimize antibodies with minimal vulnerability to chemical degradation. Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, the descriptions and examples should not be construed as limiting the scope of the invention. The disclosures of all patent and scientific literature cited herein are expressly incorporated in their entirety by reference.

Claims

1. WHAT IS CLAIMED IS:1 . A method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of:(a) determining the likelihood of oxidation of a tryptophan residue of the polypeptide, the determination based on (i) the electrostatic potential of a region comprising the tryptophan residue, and (ii) the solvent-accessible surface area (SASA) of the tryptophan residue; and(b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises:(i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue;(ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or(iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

2. A method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of:(a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises:(i) determining the electrostatic potential of a region comprising the tryptophan residue; and(ii) determining the solvent-accessible surface area (SASA) of the tryptophan residue; and(b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises:(i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the tryptophan residue;(ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or(iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

3. The method of claim 1 , wherein the method comprises identifying a tryptophan residue of the polypeptide that is likely to be oxidized.

4. The method of any one of claims 1 -3, wherein the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced.

5. The method of any one of claims 1 -4, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide.

6. The method of any one of claims 1 -5, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide.

7. The method of any one of claims 1 -6, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

8. The method of any one of claims 1 -4, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue.

9. The method of any one of claims 1 -4, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH or VL comprising the tryptophan residue.

10. The method of any one of claims 1 -4, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR loop comprising the tryptophan residue.11 . The method of any one of claims 1 -10, wherein the method comprises determining that the tryptophan residue is in a region having a negative electrostatic potential.

12. The method of claim 11 , wherein the tryptophan residue is identified as being in a region having the negative electrostatic potential using the Adaptive Poisson-Boltzmann Solver (APBS) software package.

13. The method of any one of claims 1 -12, wherein the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -0.5 or less.

14. The method of claim 13, wherein the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .0 or less.

15. The method of claim 13, wherein the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .5 or less.

16. The method of any one of claims 1 -15, wherein the method comprises determining that the tryptophan residue is solvent-exposed.

17. The method of any one of claims 1 -16, wherein the method comprises determining that the tryptophan residue has a SASA that is above a reference level.

18. The method of claim 17, wherein the reference SASA level is 50 A2.

19. The method of any one of claims 2-18, wherein identifying a tryptophan residue of the polypeptide that is likely to be oxidized comprises:(i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and(ii) determining that the tryptophan residue has a SASA that is above a reference level.

20. The method of any one of claims 1 -19, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues that contribute to a negative electrostatic potential of the region comprising the tryptophan residue.21 . The method of any one of claims 1 -4 and 8-20, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide.

22. The method of claim 21 , wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide.

23. The method of claim 22, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

24. The method of any one of claims 1 -23, wherein the one or more candidate amino acid residues of the polypeptide are at a position in the amino acid sequence within 10 amino acid residues of the tryptophan residue.

25. The method of any one of claims 1 -24, wherein the one or more candidate amino acid residues of the polypeptide comprise an aspartic acid (D) residue or a glutamic acid (E) residue at position n-1 relative to the tryptophan residue.

26. The method of any one of claims 1 -25, wherein the modifying comprises introducing an amino acid substitution in the amino acid sequence.

27. The method of claim 26, wherein the amino acid substitution mutation comprises replacement of a negatively charged amino acid residue with a neutral or positively charged amino acid residue.

28. The method of claim 27, wherein the amino acid substitution mutation comprises replacement of a D residue or an E residue at position n-1 relative to the tryptophan residue.

29. The method of claim 26, wherein the amino acid substitution mutation comprises replacement of a neutral amino acid residue with a positively charged amino acid residue.

30. The method of any one of claims 27-29, wherein the positively charged amino residue is a lysine (K) residue or an arginine (R) residue.31 . The method of any one of claims 1 -25, wherein the modifying comprises deletion of one or more amino acid residues in the amino acid sequence.

32. The method of any one of claims 1 -25, wherein the modifying comprises addition of one or more amino acid residues in the amino acid sequence.

33. The method of any one of claims 1 -32, wherein the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a neutral electrostatic potential.

34. The method of any one of claims 1 -32, wherein the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a positive electrostatic potential.

35. The method of any one of claims 1 -34, wherein the tryptophan residue is at an interaction surface of the polypeptide.

36. The method of claim 35, wherein the interaction surface is a protein-protein interaction surface.

37. The method of any one of claims 1 -36, wherein the modifying does not substantially affect a binding interaction of the polypeptide.

38. The method of any one of claims 1 -7 and 11 -37, wherein the polypeptide is an antibody or an antibody fragment.

39. The method of claim 38, wherein the tryptophan residue is in a Fab domain of the antibody or antibody fragment.

40. The method of claim 39, wherein the tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment.41 . The method of claim 39 or 40, wherein the tryptophan residue is in a framework region of the antibody fragment.

42. The method of claim 39 or 40, wherein the tryptophan residue is in a CDR loop of the antibody or antibody fragment.

43. The method of any one of claims 8-10 and 38-42, wherein the modifying does not substantially affect the binding affinity of the antibody or antibody fragment.

44. The method of any one of claims 1 -7 and 11 -37, wherein the polypeptide is a cytokine.

45. The method of any one of claims 1 -7 and 11 -37, wherein the polypeptide is a VHH.

46. The method of any one of claims 1 -45, wherein the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%.

47. The method of claim 46, wherein the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 85%.

48. The method of any one of claims 1 -47, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 35%, as assessed in an oxidation assay using 2,2'-Azobis(2- amidinopropane) dihydrochloride (AAPH), wherein the length of the assay is about 16 days.

49. The method of claim 48, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 30%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

50. The method of claim 49, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 25%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.51 . The method of claim 50, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 1%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

52. A method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of:(a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises:(i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and(ii) determining that the tryptophan residue has a SASA that is above 50 A2; and(b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises:(i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue;(ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or(iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue.

53. A method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of tryptophan oxidation of the polypeptide, the method comprising the steps of:(a) identifying a tryptophan residue of the polypeptide that is likely to be oxidized, wherein the identifying comprises:(i) determining that the tryptophan residue is in a region having a negative electrostatic potential of -0.5 or less; and(ii) determining that the tryptophan residue has a SASA that is above 50 A2; and(b) modifying the amino acid sequence of the polypeptide to increase the electrostatic potential of the region comprising the tryptophan residue, wherein the modifying comprises:(i) replacing one or more candidate amino acid residues within the region comprising the tryptophan residue with one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue;(ii) deleting one or more candidate amino acid residues within the region comprising the tryptophan residue, wherein the one or more candidate amino acid residues do not include the tryptophan residue; or(iii) inserting, within the region comprising the tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region comprising the tryptophan residue.

54. The method of claim 52 or 53, wherein the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the tryptophan residue is reduced.

55. The method of any one of claims 52-54, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide.

56. The method of any one of claims 52-55, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide.

57. The method of any one of claims 52-56, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

58. The method of any one of claims 52-55, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the Fab domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue.

59. The method of any one of claims 52-55, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the VH domain or VL domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH or VL comprising the tryptophan residue.

60. The method of any one of claims 52-55, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within a CDR loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR loop comprising the tryptophan residue.61 . The method of any one of claims 52-60, wherein the tryptophan residue is identified as being in a region having the negative electrostatic potential using the APBS software package.

62. The method of any one of claims 52-61 , wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues that contribute to a negative electrostatic potential of the region comprising the tryptophan residue.

63. The method of any one of claims 52-54 and 58-62, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide.

64. The method of claim 63, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide.

65. The method of claim 64, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.

66. The method of any one of claims 52-65, wherein the one or more candidate amino acid residues of the polypeptide are at a position in the amino acid sequence within 10 amino acid residues of the tryptophan residue.

67. The method of any one of claims 52-66, wherein the one or more candidate amino acid residues of the polypeptide comprise an aspartic acid (D) residue or a glutamic acid (E) residue at position n-1 relative to the tryptophan residue.

68. The method of any one of claims 52-67, wherein the modifying comprises introducing an amino acid substitution in the amino acid sequence.

69. The method of claim 68, wherein the amino acid substitution mutation comprises replacement of a negatively charged amino acid residue with a neutral or positively charged amino acid residue.

70. The method of claim 69, wherein the amino acid substitution mutation comprises replacement of a D residue or an E residue at position n-1 relative to the tryptophan residue.71 . The method of claim 68, wherein the amino acid substitution mutation comprises replacement of a neutral amino acid residue with a positively charged amino acid residue.

72. The method of any one of claims 69-71 , wherein the positively charged amino residue is a lysine (K) residue or an arginine (R) residue.

73. The method of any one of claims 52-67, wherein the modifying comprises deletion of one or more amino acid residues in the amino acid sequence.

74. The method of any one of claims 52-67, wherein the modifying comprises addition of one or more amino acid residues in the amino acid sequence.

75. The method of any one of claims 52-74, wherein the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a neutral electrostatic potential.

76. The method of any one of claims 52-74, wherein the modifying converts the electrostatic potential of the region comprising the tryptophan residue from a negative electrostatic potential to a positive electrostatic potential.

77. The method of any one of claims 52-76, wherein the tryptophan residue is at an interaction surface of the polypeptide.

78. The method of claim 77, wherein the interaction surface is a protein-protein interaction surface.

79. The method of any one of claims 52-78, wherein the modifying does not substantially affect a binding interaction of the polypeptide.

80. The method of any one of claims 52-57 and 61 -79, wherein the polypeptide is an antibody or an antibody fragment.81 . The method of claim 80, wherein the tryptophan residue is in a Fab domain of the antibody or antibody fragment.

82. The method of claim 81 , wherein the tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment.

83. The method of claim 82, wherein the tryptophan residue is in a framework region of the antibody fragment.

84. The method of claim 82, wherein the tryptophan residue is in a CDR loop of the antibody or antibody fragment.

85. The method of any one of claims 58-60 and 80-84, wherein the modifying does not substantially affect the binding affinity of the antibody or antibody fragment.

86. The method of any one of claims 52-57 and 61 -79, wherein the polypeptide is a cytokine.

87. The method of any one of claims 52-57 and 61 -79, wherein the polypeptide is a VHH.

88. The method of any one of claims 52-87, wherein the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%.

89. The method of claim 88, wherein the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 85%.

90. The method of any one of claims 52-89, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 35%, as assessed in an oxidation assay using 2,2'-Azobis(2- amidinopropane) dihydrochloride (AAPH), wherein the length of the assay is about 16 days.91 . The method of claim 90, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 30%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

92. The method of claim 91 , wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 25%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

93. The method of claim 92, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 1%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

94. A method of modifying the amino acid sequence of a polypeptide to reduce the likelihood of oxidation of a solvent-exposed tryptophan residue of the polypeptide, the method comprising modifying the amino acid sequence to increase the electrostatic potential of a region comprising the solvent- exposed tryptophan residue, wherein the modifying comprises:(a) replacing one or more candidate amino acid residues within the region with one or more amino acid residues that increase the electrostatic potential of the region, wherein the one or more candidate amino acid residues do not include the solvent-exposed tryptophan residue;(b) deleting one or more candidate amino acid residues within the region, wherein the one or more candidate amino acid residues do not include the solvent-exposed tryptophan residue; or(c) inserting, within the region comprising the solvent-exposed tryptophan residue, one or more amino acid residues that increase the electrostatic potential of the region.

95. The method of claim 94, wherein the method further comprises producing a modified polypeptide having the modified amino acid sequence, thereby producing a polypeptide for which the likelihood of oxidation of the solvent-exposed tryptophan residue is reduced.

96. The method of claim 94 or 95, wherein the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

97. The method of any one of claims 94-96, wherein the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

98. The method of any one of claims 94-97, wherein the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

99. The method of claim 94 or 95, wherein the polypeptide is an antibody or antibody fragment, the solvent-exposed tryptophan residue is within the fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the Fab domain comprising the solvent-exposed tryptophan residue.

100. The method of claim 94 or 95, wherein the polypeptide is an antibody or antibody fragment, the solvent-exposed tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the solvent-exposed tryptophan residue is the electrostatic potential of the VH or VL comprising the solvent-exposed tryptophan residue.101 . The method of claim 94 or 95, wherein the polypeptide is an antibody or antibody fragment, the solvent-exposed tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the solvent- exposed tryptophan residue is the electrostatic potential of the CDR loop comprising the solvent-exposed tryptophan residue.

102. The method of any one of claims 92-101 , wherein the method comprises determining that the tryptophan residue has a SASA that is above a reference level.

103. The method of claim 102, wherein the reference SASA level is 50 A2.

104. The method of any one of claims 92-103, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues that contribute to a negative electrostatic potential of the region comprising the solvent-exposed tryptophan residue.

105. The method of any one of claims 94, 95, and 99-104, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 30 A of the solvent- exposed tryptophan residue in a structural model of the polypeptide.

106. The method of claim 105, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 10 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

107. The method of claim 106, wherein the one or more candidate amino acid residues of the polypeptide are amino acid residues within a radial distance of 5 A of the solvent-exposed tryptophan residue in a structural model of the polypeptide.

108. The method of any one of claims 92-107, wherein the one or more candidate amino acid residues of the polypeptide are at a position in the amino acid sequence within 10 amino acid residues of the solvent-exposed tryptophan residue.

109. The method of any one of claims 92-108, wherein the one or more candidate amino acid residues of the polypeptide comprise an aspartic acid (D) residue or a glutamic acid (E) residue at position n-1 relative to the solvent-exposed tryptophan residue.

110. The method of any one of claims 92-109, wherein the modifying comprises introducing an amino acid substitution in the amino acid sequence.

111. The method of claim 110, wherein the amino acid substitution mutation comprises replacement of a negatively charged amino acid residue with a neutral or positively charged amino acid residue.

112. The method of claim 111 , wherein the amino acid substitution mutation comprises replacement of a D residue or an E residue at position n-1 relative to the solvent-exposed tryptophan residue.

113. The method of claim 110, wherein the amino acid substitution mutation comprises replacement of a neutral amino acid residue with a positively charged amino acid residue.

114. The method of any one of claims 111 -113, wherein the positively charged amino residue is a lysine (K) residue or an arginine (R) residue.

115. The method of any one of claims 94-109, wherein the modifying comprises deletion of one or more amino acid residues in the amino acid sequence.

116. The method of any one of claims 94-109, wherein the modifying comprises addition of one or more amino acid residues in the amino acid sequence.

117. The method of any one of claims 94-116, wherein the modifying converts the electrostatic potential of the region comprising the solvent-exposed tryptophan residue from a negative electrostatic potential to a neutral electrostatic potential.

118. The method of any one of claims 94-116, wherein the modifying converts the electrostatic potential of the region comprising the solvent-exposed tryptophan residue from a negative electrostatic potential to a positive electrostatic potential.

119. The method of any one of claims 94-118, wherein the solvent-exposed tryptophan residue is at an interaction surface of the polypeptide.

120. The method of claim 119, wherein the interaction surface is a protein-protein interaction surface.121 . The method of any one of claims 94-120, wherein the modifying does not substantially affect a binding interaction of the polypeptide.

122. The method of any one of claims 94-98 and 102-121 , wherein the polypeptide is an antibody or an antibody fragment.

123. The method of claim 122, wherein the solvent-exposed tryptophan residue is in a Fab domain of the antibody or antibody fragment.

124. The method of claim 123, wherein the solvent-exposed tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment.

125. The method of claim 124, wherein the solvent-exposed tryptophan residue is in a framework region of the antibody fragment.

126. The method of claim 124, wherein the solvent-exposed tryptophan residue is in a CDR loop of the antibody or antibody fragment.

127. The method of any one of claims 99-101 and 122-126, wherein the modifying does not substantially affect the binding affinity of the antibody or antibody fragment.

128. The method of any one of claims 94-98 and 102-121 , wherein the polypeptide is a cytokine.

129. The method of any one of claims 94-98 and 102-121 , wherein the polypeptide is a VHH.

130. The method of any one of claims 94-129, wherein the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 40%.131 . The method of claim 130, wherein the likelihood of tryptophan oxidation of the modified polypeptide is reduced by at least 85%.

132. The method of any one of claims 94-131 , wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 35%, as assessed in an oxidation assay using 2,2'-Azobis(2- amidinopropane) dihydrochloride (AAPH), wherein the length of the assay is about 16 days.

133. The method of claim 132, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 30%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

134. The method of claim 133, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 25%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

135. The method of claim 134, wherein the frequency of tryptophan oxidation of the modified polypeptide is less than 1%, as assessed in an oxidation assay using AAPH, wherein the length of the assay is about 16 days.

136. A polypeptide comprising a tryptophan residue having a low likelihood of oxidation, wherein the polypeptide is produced by the method of any one of claims 1 -135.

137. A method of determining whether a tryptophan residue of a polypeptide is likely to be oxidized, the method comprising the steps of:(a) measuring the electrostatic potential of a region comprising the tryptophan residue; and(b) measuring a solvent-accessible surface area (SASA) of the tryptophan residue; wherein the electrostatic potential of the region comprising the tryptophan residue and the SASA of the tryptophan residue together identify the tryptophan residue as likely to be oxidized.

138. The method of claim 137, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 30 A of the tryptophan residue in a structural model of the polypeptide.

139. The method of claim 137 or 138, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 10 A of the tryptophan residue in a structural model of the polypeptide.

140. The method of any one of claims 137-139, wherein the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the region of the polypeptide within a radial distance of 5 A of the tryptophan residue in a structural model of the polypeptide.141 . The method of claim 137, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the fragment antigen-binding (Fab) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the Fab domain comprising the tryptophan residue.

142. The method of claim 137, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within the heavy chain variable (VH) domain or light chain variable (VL) domain of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the VH or VL comprising the tryptophan residue.

143. The method of claim 137, wherein the polypeptide is an antibody or antibody fragment, the tryptophan residue is within a complementarity-determining region (CDR) loop of the antibody or antibody fragment, and the electrostatic potential of the region comprising the tryptophan residue is the electrostatic potential of the CDR loop comprising the tryptophan residue.

144. The method of any one of claims 137-143, wherein the method comprises determining that the tryptophan residue is in a region having a negative electrostatic potential.

145. The method of claim 144, wherein the tryptophan residue is identified as being in a region having the negative electrostatic potential using the APBS software package.

146. The method of any one of claims 137-145, wherein the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -0.5 or less.

147. The method of claim 146, wherein the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .0 or less.

148. The method of claim 147, wherein the method comprises determining that the tryptophan residue is in a region having an electrostatic potential of -1 .5 or less.

149. The method of any one of claims 137-148, wherein the method comprises determining that the tryptophan residue is solvent-exposed.

150. The method of any one of claims 137-149, wherein the method comprises determining that the tryptophan residue has a SASA that is above a reference level.151 . The method of claim 150, wherein the reference SASA level is 50 A2.

152. The method of any one of claims 137-151 , wherein determining whether a tryptophan residue of a polypeptide is likely to be oxidized comprises:(i) determining that the tryptophan residue is in a region having a negative electrostatic potential; and(ii) determining that the tryptophan residue has a SASA that is above a reference level.

153. The method of any one of claims 137-152, wherein the tryptophan residue is at an interaction surface of the polypeptide.

154. The method of claim 153, wherein the interaction surface is a protein-protein interaction surface.

155. The method of any one of claims 137-140 and 144-154, wherein the polypeptide is an antibody or an antibody fragment.

156. The method of claim 155, wherein the tryptophan residue is in a Fab domain of the antibody or antibody fragment.

157. The method of claim 156, wherein the tryptophan residue is in a VH domain or a VL domain of the antibody or antibody fragment.

158. The method of claim 157, wherein the tryptophan residue is in a framework region of the antibody fragment.

159. The method of claim 157, wherein the tryptophan residue is in a CDR loop of the antibody or antibody fragment.

160. The method of any one of claims 137-140 and 144-154, wherein the polypeptide is a cytokine.161 . The method of any one of claims 137-140 and 144-154, wherein the polypeptide is a VHH.

Images