As to purification of bispecific antibodies based on altering amino acids in the antibody see purificaiton of bispecific antibodies 

Structural Comparison between the SpA Domains

It is known that Protein A is comprised of domains called E, D, A, B and C each of which has about 60 amino acids and these domains show high homology to each other and can bind to the common Fc region of immunoglobulin independently. (Sato, US2010/0168395).

Domains A, B and C of spA are highly homologous, with no amino acid differences within the two alpha-helices. Domain D is slightly more diveresed with three amino acid substitutions in the alpha helical regions. Region E is strikingly different, in particular within the N-temrinal segment. The first eight residues have been substituted by six others. Despite these differences region E binds IgG. (Nilsson, Protein Engineering, 192), pp. 107-113 1987).

Organization of the Protein A gene. S=signal sequence, A-E=IgG binding domains and X=C-terminal with no IgG binding activity.

From US12/653888.  AA alignments of the wild type IgG binding domains of SpA domains (E, D, A, B and C).

Staphylococcus aureus (“SpA”) refers to a 42 Kda multi-domain protein isolated from the bacteriumStaphylococcus aureus. SpA is bound to the bacterial cells wall via its carboxy-terminal cell wall binding region, referred to as the X domain. At the amino-terminal region, it includes 5 immunoglobulin binding domains, referred to as E, D, A, B, and C. Each of these domains contains about 58 amino acid residues, and they share 65-90% amino acid sequence identity. 

Alignment of common Wild-type (WT) SpA domain sequences (sequences taken from Bian, 2010/0221844) 

--B/C

B domain: ADNKFNKEQQNAFYEILHLPNLNEEQRNGFIQSLKDDPSQSANLLAEAKKLNDAQAPK

C domain: ADNKFNKEQQNAFYEILHLPNLTEEQRNGFIQSLKDDPSVSKEILAEAKKLNDAQAPK

 91.4% identity in 58 residues overlap; Score: 272.0; Gap frequency: 0.0%

 1 ADNKFNKEQQNAFYEILHLPNLNEEQRNGFIQSLKDDPSQSANLLAEAKKLNDAQAPK

1 ADNKFNKEQQNAFYEILHLPNLTEEQRNGFIQSLKDDPSVSKEILAEAKKLNDAQAPK

 
--B/Z

B domain: ADNKFNKEQQNAFYEILHLPNLNEEQRNGFIQSLKDDPSQSANLLAEAKKLNDAQAPK

Z domain: VDNKFNKEQQNAFYEILHLPNLNEEQRNAFIQSLKDDPSQSANLLAEAKKLNDAQAPK

98.2% identity in 57 residues overlap; Score: 288.0; Gap frequency: 0.0%

2 DNKFNKEQQNAFYEILHLPNLNEEQRNGFIQSLKDDPSQSANLLAEAKKLNDAQAPK

2 DNKFNKEQQNAFYEILHLPNLNEEQRNAFIQSLKDDPSQSANLLAEAKKLNDAQAPK

--C/Z

C domain: ADNKFNKEQQNAFYEILHLPNLTEEQRNGFIQSLKDDPSVSKEILAEAKKLNDAQAP

Z domain: VDNKFNKEQQNAFYEILHLPNLNEEQRNAFIQSLKDDPSQSANLLAEAKKLNDAQAPK

89.3% identity in 56 residues overlap; Score: 257.0; Gap frequency: 0.0%

 2 DNKFNKEQQNAFYEILHLPNLTEEQRNGFIQSLKDDPSVSKEILAEAKKLNDAQAP

2 DNKFNKEQQNAFYEILHLPNLNEEQRNAFIQSLKDDPSQSANLLAEAKKLNDAQAP

                 ********************* ***** ********** *   *************

Residues on SpA involved in binding

Protein A (SpA) exists in both secreted and membrane associated forms possesses two distinct Ig binding activities: each of its domains E, D, A, B and C can bind Fcy (the constant region of IgG involved in effector functions) and Fab (the Ig fragment responsible for antigen recognition. (Graille, PNAS, 97(10), May 9, 2000, pp. 5399-5404). 

SPa binds human immunoglobulins at two target sites: one in the Fab portion of the immunoglobulins of the IgM, IgA, IgG and IgE isotypes and the other in the Fc portion of IgG1, IgG2, IgG3 allotype G3m(s,t) and IgG4 immunoglobulins. It was demsontrated that these two SPA binding sites are distinct when chemically modified SPA lost the ability to bind Fcy but retained ability to bind in the Fab porition and by inhibition analysis in which IgG F(ab’)2 fragments but not IgG Fc fragments, blocked SPA binding to IgE. (Potter “Staphylococcal Protein A binding to VH3 encoded immunoglobulins”, Intern. Rev. Immunol., Vol. 14, pp 291-308, 1996). 

While all five domains of Protein A (A, B, C, D, E) bind IgG via their Fc region, only domains D and E exhibit significant VH3 binding. (Slough, US Patent Application No: 15/566231, published as US 2018/010007). 

SpA possesses distinct Ig binding sites. One site is for Fcy (the constant region of IgG class of Ig) and the other is for the Fab portion of certain Ig molecules (the porition of the Ig that is responsible for antigen recognition). The Fcy binding site has been localized to the elbow region at the CH2 and CH3 interface of most IgG subclasses, and this binding property has been extensively used for purification of antibodies. Silverman (US2006/0205016).  

Protein A interacts with antibodies through two distinct binding evests: the “classical” binding site on the Fc portion of human IgG1, IgG2, and IgG4 and the “alternate” binding site found on the Fab portion of human IgG, IgM, IgA and IgE that contain heavy chains of the VH3 subfamily. (Starovasnik, Protein Science, 1999, 8; 1423-1431).

Fcy binding site: 

The Fcy binding site of SpA has been localized to the elbow region at the CH2 and CH3 interfact of most IgG subclasses and this binding property has been extensively used for the labeling and purificaiton of antibodies. (Graille, PNAS, 97(10), May 9, 2000, pp. 5399-5404).

The N-terminal part of the mature protein is responsible for the binding to the Fc-portion of IgG (this is the region that has the 5 hoologous domains each of about 58 amino acid residues which are individually IgG binding). Each of the domains have two alpha-helices involved in the IgG binding. The A, B and C domains are highly homologous, with no amino acid differences within the two alpha-helices. Region D is slighly more diverged with three amino acid substitutiosn in the alpha helical regions. Region E is strikingly different, in particular within the N-terminal segment. (Nilsson, “A synthetic IgG-binding domain based on staphylococcal protin A” Protein Engineering, 1(2), 1987, 107-13).

Protein A binds the Fc region of immunoglobulins. The suitability of protein A as an affinity ligand depends on the species and isotype of any immunoglobulin Fc domain that is present in the antibody. It binds with high affinity to human IgG1, IgG2 and IgG4 as well as mouse IgG2a, IgG2b and IgG3. Protein G is recommended for all mouse isotypes  (WO2008/091740). Protein A binds with moderate affinity to human IgD, IgM, IgA and IgE as well as mouse IgG1. One molecule of immunobilized protein A can bind at least two molecules of IgG. The advantage of Protein A compared to Protein G is that the binding of IgG is weaker and consequently milder conditions can be used to release IgG from Protein A (WO1997017361).

–Residues in the D domain:

Graille, (PNAS, 97(10), May 9, 2000, pp. 5399-5404) disclosed the cyrstal structure of domain D of SpA complexed with the Fab fragment of human IgM and the key contact residues from both partners in the interaction. The residues in domain D involved in the interaction with Fab are highly conserved in other Ig binding domains of SpA and are distinct from the residues that mediate binding to Fcy above. The interaction involves the following domain D residues: Gln26, Gy29, Phe-30, Gln32, Ser88 and Asp 36 of helix II; Asp37 and Gln 40 in the loop between helix II and helix III; and Asn43, Glu47 and Leu51 of helix III. In the Fab, the interaction is mediated by heavy chain residues GlyH15, SerH17 in the beta turn befroe strnd B; Arg H19 of strand B; LysH57 and Tyr H59 of strand C; lys-H64, Ly-H65 and Arg-H66 before strand D, Thr-H68 and Ser-H70 of strand D, Gln-H81 of strand E; Asn-H82a and Ser-H82b after strand E. Six of these residues are in framework region (FR) beta strands whereas the other seven residues are in VH region FR interstrand loops on the side farthest from the antigen binding pocket. 

Silverman (US2006/0205016) teaches that the Z domain binds to the Fc region of immunoglobulins as do the 5 homlogous SpA domains, but unlike the parental domain it does not bind to the Fab region (¶102).

Residues on the Antibody Important for Binding

Protein A binds to the framework regions of V domains of the human VH3 family and to certain V domains of the VH1 family. (Holliger, US 5,837,242).

SpA, which exists in both secreted and membrane-associated forms, possesses two distinct Ig-binding adtivities: each domain can bind Fcy (the constant region of IgG involved in effector functions) and Fab (the Ig fragment responsible for antigen recognition). SPA is a 42-kDa protein covalently anchored to the staphylococcal cell wall through its carboxyl temrinal end. The protein is comprised of five repeated domains (E, D, A, B, C) of 48 reisudes linked to teh cell surface by region Xr, which contains a variable number of short 8-residue repeats. Each SpA domain can bind with high affinity to the Fc region of IgG and to teh Fab region of immunoglobulin of the VH3 subclass. (Lydon, US 2016/0068589). 

Fab binding site:

A subset of antibodies bind Protein A in the variable region of the heavy chain. In particular, IgGs that contain heavy chains dervied from the H3 gene fmaily have been shown to exhibit this behavior. Nearly half of human VH germline genes belong to this family. Becasue of this, on SpA-based resins, many bsAbs exhibit poor resolution of the two binding species as well as retention of the non-binindg Fc&Fc& parental antibody. This is becasue VH binding reduces the avidity differences between the bispecific and the FcFc parental and Fc*Fc* parental is also retained by VH binding. This phenomenon can be resolved through the use of the alkli stablized MabSelect SuRe ligand. This is becasue binding binding studes between SpA and antibodies have shown that all five domains of SpA (E, D, A, B, and C) have antibodies via the Fc region, only domains D and E exhibit significant Fab binding. As the MabSelect SuRE lgiand is a treamer of the Z-domain, a protein-engineered version of the native SpA B domain, it would be expected to lack significnat Fab binding, and this has been demonstrated in multiple studies. (Tustian “Development of a novel affinity chromatography resin for platform purificaiton of bispecific antibodies with modified protein A binding avidity, 2018). 

–Framework Subgroup 3 (VH3) domain encoded antibodies:

The term “VH3 domain” refers to the framework subgroup 3 of human heavy chain variable region of an immunoglobulin. The heavy chain variable domains of antibodies are classified into distinct subfamilies (VH1 to VH6) on the basis of DNA sequence and protein homologies. (Slough, US Patent Application No: 15/566231, published as US 2018/010007)

The Fab specificity of SpA is less well characterized but it has been shown to involve a site on the variable region of the Ig heavy chain. Fab binding specificity is restricted to products of the human variable region of the Fab heavy chain VH3 family that represent nearly half of inherited VH genes and their homologues in other mammalian species. Graille, (PNAS, 97(10), May 9, 2000, pp. 5399-5404) In the context of IgG purification using Protein A, VH3 domain interactions have been considered undersirable given that they affect the elution profile of the antibody to be purified and alternative resins have been developed that contain only domain B of Protein such as Sure from GE Healthcare. However, many of the antibodies and antibody derived molecules currently available and in development such as antibody fragments like Fab do not contain Fc regions and binding of protein A by VH3 regions enables the use of protein A in the purification of these types of antibodies. (Slough, US Patent Application No: 15/566231, published as US 2018/010007).

An alternate binding site for binding protein A has been described in antibodies that contain the specific framework subgroup 3 of the human H chain variable region, also referred to as the VH3 domain and has often been classified as a secondary interaction.  (Sasso, J. Immunol 1991, 147: 1877-1883, Human IgA and IgG F(ab’)2 that bind to staphylococcal protein A belong to the VHIII subgroup). While all five domains of Protein A (A, B, C, D, and E) bind IgG via their Fc-region, only domains D and E exhibit significant VH3 binding. These VH3 domain interactions with Protein A have been considered undesirable given that they affect the elution profile of the antibody to be purified, and alternative resins have been developed and are available on the market that contain only domain B of Protein A such as SuRe® from GE Healthcare. However, many of antibodies in development do not contain Fc regions and binding of protein by VH3 regions enables the use of protein A for the purification of such antibodies. (Slough, US Patent Application No: 15/566231, published as US 2018/010007)

Immunoglobulins of human heavy chain subgroup III have a binding site for SpA on the heavy chain variable domain (VH), in addition to the well known binding site on the Fc portion of the antibody (Starovasnik, Protein Science, 1999, 8; 1423-1431). Protein A shows binding to some human VH domains of the VH3 family but not to VH1 and VH2 (Jansson (1998) FEMS Immunology and Medical Microbiology 20: 69-78). 

SPA binds human immunoglobulins in the variable region encoded only by gene segments belonging only to the VH 4amily. The simultaneous interaction of amino acid resiudes in FR1, CDRH2 and FR3 of VH3 encoded antibodies is required for binding to SPA. When any one of these regions is replaced by sequences from a non-SPA binding antibody, the ability to bind SPA is abrogated. (Potter “Staphylococcal Protein A binding to VH3 encoded immunoglobulins”, Intern. Rev. Immunol., Vol. 14, pp 291-308, 1996). 

Nearly half of human VH germline genes belong to the VH3 subfamily and pharmaceuticals containing IgG antibodies of the VH3 subfamily are under study and some of them are alaredy commercially available. (Yoshida, US 2012/0208234). 

Hillson discloses that the ability of human VH3 Ig to bind SPA via their Fab region is analogous to the binding of bacterial superantigens to T cell receptors. Hillson showed that binding to SPA in ELISA cocurred with 15 of 15 VH3 IgM, but none of 12 IgM from the VH1, VH4, VH5 or VH6 vamilies. Use of D, JH and Cl genes was similar among VH3 and non-VH3 IgM. A comparison of the corresponding VH protein sequences identified a probably site for SPA binding that includes VH3 residues in the framework region 3 (FR3) and perhaps FR1 and 3′ complementary determining region 2. The results thus demonstrate that among human IgM, specificity for SPA is encoded by at least 11 different VH3 germline genes and that SPA likely binds to residues in the VH framework region, outside the classical antigen binding site of the hypervariable loops. (J Exp. Med, 178, 1993, 331-336). 

SpA is a 45 kDa bacterial membrane prtoein that can interact with either Fcy, a constant reigon portion of IGG, or with the Fab portion that also mediates conventional Ag binding. SpA has been shown to specificlally interact with Fab dervied form the VH3 family, and is little aaffected by VH CDR3, JH or light chain usage. The 61 mino acid sequence of SpA that represents domain D exhibited both the FH3 Fab and Fcy binding specificites. It had the strongest binding with an IgM-k encoded by the germline configuation of the VH3 gene VH26C (Roben, “VH3 family antibodies bind domain D of Staphylococacal Protein A” J Immunology, 1995, 154: 6437-6445). 

Jansson (FEMS Immunology and Medical Microbiology 20(1) (1998) 69-78 discloses that all the domains of SpA (A, B, C, E and D) bind Fab with similar strenght, contrary to reported work that only domains E and/or D are responsible for Fab binding. Most amino acids reported to be important for good SPA binding are present in the DP_53 VH3 region. However the Z domain having only a 1 amino acid difference (glycine to alanine mutation at posiiton 29) shows very little or no binding activity. 

Potter, (Intern. Rev. Immunol., 14, pp. 291-308) discloses that replacement of a single involved region with the corresponding region from a nonbinding Ig resutls in loss of ability of VH3 encoded Ab to bind SPA.  SPA binds human immunoglobulins in the variable region encoded only by gene segments belowing to the VH3 family. 

Potter (J. Immunology, 1996, 157: 2982-2988) also disclsoes that at least 15 different VH germline genes have been shown to bind SPA. A single amino acid change at position 57 in the CDR2 of a human SPA nonbinding VH encoded rheumatoid factor converted it to an SPA binding, implicating CDR2 in SPA binding. When regions of the mutated binder were exchange with those form a mouse nonbinding antibody, the pattern of SPA binding indicated that residues in CDR1, CDR2 and FR3 are involved in the interaction between VH3 encoded antibodies and SPA. In addition, all three regions are simultaneously required for SPA binding to occur because when any one of the three regions are altered, SPA binding is disrupted. 

Yeung (WO2010/075548) discloses a variant IgG comprising a human IgGVH region having one or more amino acid substitutions relative to the WT IgGVH region which has altered binding to Staphylococcus aureus prtoein A. In some embodiments, the variant has increased binding to protein A by an amino acid substitution at residue(s) 70, 79, and 82 (e.g., S70A, Y79A or S82bA). In some embodiments, the variant has decreased binding to protein A as by substitution in the IgG VH region at residues 17, 19, 57 66 81 and 82a (e.g., S17A, RI9A, T57A, T57K, R66A, Q81A or N82aA). In some embodiments, the variant IgG is Ig1, IgG2, IgG3 or IgG4. 

–Particular substitutions in VH3 Germline/Subclass region: 

—-T57E:

Bodn (US 2005/0079574) discloses that in classical VHH3 domains, the mutation T57E abolishes affinity for protein. 

Igawa (EP2522724 and WO/2011/078332) disclsoes that the H chain variable region of the VH3 subclass has protein A binding activity and that accordingly to increase protein A binding, the amino acid sequences are preferably identical to those of the H chain variable region of the VH3 subclass and to reduce protein A binding, the amino acid sequences are preferably identical to those of the H chain variable region of another subclass. Such modifications can be part of purifying multimers/bispecific antibodies such that amino acid residues in either or both of the first/second polypeptides are modified so that there is a larger difference of protien A binding ability between the first and second polypeptides. 

As to the different isotypes of IgG

Protein A bind to the CH2-CH3 interface of human IgG1, human IgG2 and Human IgG4 (Fc) and as such cannot be used for the purification of human IgG3. Protein A shows binding to some human VH domains of the VH3 family but not to human VH1 and VH2 and thus Protein A cannot be used for purification of all human IgG derived Fab gragments. In addition A domain based Protein A variants (like MabSelect Sure) lack the ability to bind to human VH3 domains and as such can not be used for Fab purification. (Hermans, 13.982970). 

CH3 domain modifications:

Davis (WO/2010/151792) discloses a bispecific antibody comprising a H chain variable domain that are differentially modified in the CH3 domain wherein the differential modificaitons result in a differential affinity for the bispecific antibody for an affinity reagent such as Protein A. For exaple, the second CH3 region can include a 95R modification(by IMGT exon numbering or 435R by EU numbering). in another embodiment, the second CH3 region includes a 436F by EU modifiecation. In another embodiment,t he second CH3 region is from a modified human IgG1 and fruther includes a modificaiton selected from D16E, L8M, N44S, K52N, V57M and V821. In one emboeidment, the second CH3 region is from a modified human IgG2 and includes a modification selected from N44S, K52N and V82I. In anotehr embodiment, the second CH3 region is from a modified human IgG4 and includes Q15R, N4S, K52N, V57M, R69K, E79Q and V82I. 

Domains of SpA 

D domain is a 61 amino acid polypeptide that folds into a 3 helix bundle structure. It is capable of Fc bnding via residues on the surface of helices 1 and 2, or to Fab via residues on the surface of helices 2 and 3.

Silverman (WO 00/69457) discloses critical residues in domain D that interact with Ig-Fab to mediate the Ig-Fab binding site in SpA. (Roben, J. Immunogloy, 1995, 154: 6437-6445)

Binding studies demonstrate that domain D possesses the same VH3 restricted Fab binding specificity as the native SpA and also possesses Fc-y-binding (Fcy is a constant region porition of IgG) activity. Structurally domain D is in part distinguished form other domains of SpA because it has a three amino insertion between the second and third amino acids in the consensus sequence. 

A domain of Spa is a 58 amino acid polypeptide that also folds into a 3 helix structure. It is capable of Fc binding via residues on the surface of helices 1 and 2, or to Fab via residues on the surface of helices 2 and 3.

B domain is a 58 amino acid polyeptide that also folds into a 3 helix bundle structure. It is capable of Fc binding via residues on the surface of helices 1 and2, or to Fab via residues on the surface of helices 2 and 3.

C domain is a 58 amino acid polypeptide that also folds into a 3 helix structure and capable of Fc binding via residues on the surface of helices 1 and 2 or to Fab via residues on the surface of helices 2 and 3 (Bian US 12/653888). 

The C domain has superior resistance to alkaline conditions compared to the other domains. Thr-23, an amino acid found only in the C domain, is easily predicted to have a beneficial effect on alkaline stability. Val-40 one of the unique residues of the C domain appears to have a positive effect on its alkaline stability. The C domain also lacks a protease susceptible Gly-Glu sequence. Teh G29A mutation in the C domain increases alkaline stability, although smaller than that observed for teh B domain. (Yoshida “Remarkable alkline stability of an engineered protein A as immunoglobulin affinity ligand: C domain having only one amino acid substitution” (2013), Protein Science). 

Hall (WO 2008/039141A1) disclose Domain C from SpA and variants thereof. 

Z domain of SpA (see analogues of SpA) is an engineered analogue of the B domain of SpA and includes an alanine instead of glycine residue at position 29. The Z domain binds to the Fc region of immunoglobulines as do the 5 homologous SPA domains, but unlike the parental domain it does not bind to the Fab region. The term “affibody” is used to define a class of engineered proteins selected fro their specific binding activity towards a desired target and based on the Z domain.