Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Oct 1.
Published in final edited form as: J Immunol Methods. 2023 Aug 18;521:113540. doi: 10.1016/j.jim.2023.113540

MILKSHAKE Western Blot and Sundae ELISA: We All Scream for Better Antibody Validation

Qiana Mendez 1, Holland A Driscoll 1, Gregory R Mirando 1, Felicity Acca 1, Cassandra D Chapados 1, Kezzia S Jones 1, Michael Weiner 1, Xiaofeng Li 1,2, Mary R Ferguson 1,*
PMCID: PMC10568614  NIHMSID: NIHMS1927917  PMID: 37597727

Abstract

Knowing that an antibody’s sensitivity and specificity is accurate is crucial for reliable data collection. This certainty is especially difficult to achieve for antibodies (Abs) which bind post-translationally modified proteins. Here we describe two validation methods using surrogate proteins in western blot and ELISA. The first method, which we termed “MILKSHAKE” is a modified maltose binding protein, hence the name, that is enzymatically conjugated to a peptide from the chosen target which is either modified or non-modified at the residue of interest. The surety of the residue’s modification status can be used to confirm Ab specificity to the target’s post-translational modification (PTM). The second method uses a set of surrogate proteins, which we termed “Sundae”. Sundae consists of a set of modified maltose binding proteins with a genetically encoded target sequence, each of which contains a single amino acid substitution at one position of interest. With Sundae, Abs can be evaluated for binding specificities to all twenty amino acids at a single position. Combining MILKSHAKE and Sundae methods, Ab specificity can be determined at a single-residue resolution. These data improve evaluation of commercially available Abs and identify off-target effects for Research-Use-Only and therapeutic Abs.

Keywords: Post translational modification, antibody, validation, western blot, ELISA

1.1. INTRODUCTION

Antibodies (Abs) are essential tools in biomedical research for the detection and quantification of target proteins (1–3). There is great concern in scientific literature about the reliability and specificity of Abs used in a variety of applications (4–10). These concerns are magnified particularly when applied to post-translationally modified (PTM) proteins (11–15).

Post-translational modifications such as phosphorylation, acetylation, and glycosylation, are critical for regulating protein function and are involved in many disease processes (16–18). Post-translational modifications can make it challenging to develop specific Abs that differentiate between the modified or the native isoform of the protein, leading to false positive or negative results. The challenges associated with validating Abs against PTM proteins are numerous (19). Firstly, PTMs can be highly dynamic and may only be present in certain tissues, cell types, or developmental stages, making it difficult to generate Abs that work consistently across different samples. Secondly, while there have been several means established to validate most Abs, there is a lack of standardized methods for validating Abs against PTM proteins. When ELISA or traditional western blot is used for validation, the percent of modification at the site of interest is unknown. Knockout validation isn’t informative for PTM proteins and while mass spectrophotometry is useful it requires a level of expertise not found in all research teams. These issues have led to significant variability in results across different laboratories (20–21).

To help address these challenges, our solution involves the development of two Ab validation methods: MILKSHAKE which focuses on PTM-specific epitopes and Sundae which thoroughly explores the Ab-antigen interface at a one-residue resolution.

In the MILKSHAKE method, a modified maltose binding protein (modMBP) is fused to a peptide containing the moiety of interest using sortase A conjugation (22). The resulting protein is used in western blot to validate the specificity of Abs to the modified or non-modified epitope. The work in this study will be focused on phosphoproteins which account for the majority of PTM Abs cited in literature (23). To enhance the data produced in the MILKSHAKE method, we also incorporate cell lysates to increase the assay’s stringency and visualize off-target binding (strawberry MILKSHAKE). Specific binding to MILKSHAKE in western blot correlates with binding to full length target proteins found in treated cell lysates. (22) In addition, comparing the binding of an Ab to MILKSHAKE and to a treated cell lysate in western blot can verify the quality of the lysate treatment itself.

In addition to PTM Ab validation, the development of therapeutic Abs has hastened a need for more thorough evaluation of the antibody-antigen interaction. This evaluation is important to ensure specific Ab binding to the target of interest and reduce off-target effects. The Sundae method is similar to -- but we believe -- a superior means of understanding the paratope-epitope interaction at one-residue resolution compared with simpler methods such as alanine scanning. Alanine scanning is a technique that is often used to test the contribution of a specific amino acid residue to the function of a given protein. (24–25) Alanine is used because it is composed of the shortest sidechain among the 20 amino acids, showing the least electronic potential, and chemically inert for contacting with other residues. The alanine methyl group can be used to mimic the secondary structure preferences that the other amino acids possess. Sometimes bulky amino acids such as the branched chain amino acids valine and leucine are used where the size conservation of mutated residues is needed. (26–27) Often these tests can yield a quantitative measurement of point-mutation effects in bioactivities. However, one limitation to alanine scanning that is rarely discussed and underappreciated is that it only reveals whether alanine as a replacement at a specific site retains or loses bioactivity. It reveals nothing about the effect of the other 18 amino acids (i.e., the set of 20 amino acids minus the native and alanine residues).

The Sundae method consists of a modified maltose binding protein (modMBP) containing a sequence of interest genetically encoded in the C-terminus. The Sundae term stems from the fact that we are using multiple ‘flavors’ of residues at the position of interest on the epitope being tested fused to modMBP. Twenty different Sundae ‘flavors’ are produced, each of which contain the same peptide sequence except the residue of interest replaced with a different amino acid. The resulting proteins can be used in ELISA to validate the binding of an Ab to each residue substitution. To demonstrate Sundae, this manuscript focuses on evaluating the binding specificity profile of 2F5, a broadly neutralizing monoclonal antibody (bnAb) that targets Human Immunodeficiency Virus (HIV) (28).

1.2. MATERIALS & METHODS

1.2.1. Plasmid DNA construction-MILKSHAKE.

A plasmid expressing modified maltose binding protein (modMBP) was constructed by inserting a C-terminal sortase conjugation site with a preceding GlySer (GS) linker into the maltose binding protein expression vector pMAL-c6T [New England Biolabs (NEB) Ipswich, MA] {inserted sequence: GGTGGCGGTGGCTCGTTACCGGAAACTGGT which translates to GGGGSLPETG } using the NotI and EcoRI restriction sites [NEB].

1.2.2. Plasmid DNA construction-Sundae.

A plasmid expressing modified maltose binding protein (modMBP) was constructed by inserting a sequence from the target of interest (10–14 amino acids) preceded by a (GlySer)2 linker into the maltose binding protein expression vector pMAL-c6T [NEB] using the Not I and EcoR I restriction sites [NEB].

1.2.3. Protein Purification.

Modified MBP was purified from E. coli NEBExpress [NEB] cultures. Cultures grew at 37 °C until an OD600 = 0.5 was reached. Cultures were then induced with 0.2 mM IPTG (AmericanBio, Inc., Canton, MA) for an additional 1 hour. Cultures were centrifuged at 4,000 x g for 20 minutes at 4 °C. Pellets were frozen overnight at −80 °C. The following day the cells were thawed at room temperature and lysed using 2 mL of B-PER II Reagent [ThermoFisher, Waltham, MA]. Protein was purified via gravity flow using amylose resin [NEB] for immobilization and 10 mM maltose [MilliporeSigma, Burlington, MA] for elution. Protein was stored at 4 °C for short term and −80 °C for long term in 15% glycerol.

1.2.4. Peptide Design.

Peptides [Biopeptek Pharmaceuticals LLC, Malvern, PA] were designed as previously described (22) and contain an N-terminal sortase conjugation site {GlyGlyGlySerGlySerSer (GGGSGSS)}, a target-specific sequence (either phosphorylated or non-phosphorylated at a single amino acid) and a C-terminal hemagglutinin (HA) tag {TyrProTyrAspValProAspTyrAla (YPYDVPDYA)}.

1.2.5. Sortase Conjugation Reaction.

The following components (per reaction) were combined in a 96-well deep well plate: 0.145 mg purified modMBP, 10x Sortase Buffer [200 mM Tris-HCl, 1.5 M NaCl, 50 mM CaCl2, 2 mM beta-mercaptoethanol], 0.04 mg target peptide (design described above), 3 μg sortase enzyme [Moradec, San Diego, CA] and 1x PBS to final volume of 113 μL. One well without target peptide or sortase enzyme was used as a negative “unconjugated” control. The plate was incubated with shaking at 250 rpm at 37 °C for 2 hours. Reactions were then mixed with equal volume of 2x Laemmli Sample Buffer [Bio-Rad Laboratories, Hercules, CA] containing 200 mM dithiothreitol [AmericanBio Inc] and denatured at 95 °C for 10 minutes.

1.2.6. Preparation of Cell Lysates.

5 × 106 HEK293-F cells [ThermoFisher] were pelleted by centrifugation at 2,500 x g for 5 minutes at room temperature and washed twice with 10 mL chilled 1x PBS. The pellet was resuspended in 1 mL lysis buffer {RIPA buffer [MilliporeSigma] containing 3.2 mM sodium orthovanadate [NEB] and 10 μg each leupeptin, pepstatin, aprotinin and chymostatin [MilliporeSigma]}. The lysate was transferred to a microcentrifuge tube and incubated on ice for 15 minutes with shaking. The resulting lysate was centrifuged at 14,000 x g for 15 minutes at 4 °C. Supernatant was collected and quantified using an A280 microplate reader. Lysates were then mixed with equal volume of 2x Laemmli Sample Buffer [Bio-Rad Laboratories] containing 200 mM dithiothreitol [AmericanBio, Inc] and denatured at 95 °C for 10 minutes.

1.2.7. Polyacrylamide gel electrophoresis.

Each lane of a 4–20% polyacrylamide gel [Bio-Rad Laboratories] was loaded with MILKSHAKE (modMBP conjugated to a target peptide) and prepared HEK lysate. Gels were run for 40 minutes at 180V in SDS buffer.

1.2.8. Western blot.

After electrophoresis, proteins were transferred to nitrocellulose membranes using the Trans-Blot Turbo system [Bio-Rad Laboratories]. Membranes were washed with 1x Tris Buffered Saline (TBS) for 5 minutes at room temperature. Membranes were incubated in blocking buffer (3% Milk-TBST) for 1 hour at room temperature followed by washing three times for 5 minutes each with 1x TBST (Tris Buffered Saline with 0.1% Tween 20). Primary Abs were diluted according to vendor recommended concentrations for each individual Ab (unless otherwise specified) in 5% BSA-TBST and incubated shaking overnight at 4 °C. The following day, membranes were washed with 1x TBST three times for 5 minutes each. Secondary HRP-conjugated Abs were diluted 1:5,000 in 3% Milk-TBST and incubated at room temperature for 1 hour. Membranes were washed with 1x TBST three times for 5 minutes each at room temperature. Blots were visualized using chemiluminescence reagent [Bio-Rad Laboratories] with exposure times ranging from 0.4 to 16 seconds. Western blots were performed 2–4 times each per antigen-antibody pair tested.

1.2.9. ELISA.

Wells of a 96-well microtiter plate were coated with Sundae protein(s), gp41 protein [Abcam, Cambridge, UK] or biotinylated peptides via Neutravidin [ThermoFisher]. Antigens were then blocked with 2% BSA. Primary antibody 2F5 [Polymun Scientific, Klosterneuburg, Austria] or rabbit-derived IgG was diluted in 2% BSA and incubated in antigen-coated wells for 1 hour. Wells were washed three times with PBST (Phosphate Buffered Saline with 0.1% Tween 20). Secondary Abs: anti-Protein A-HRP [ThermoFisher] or goat anti-rabbit HRP [Jackson ImmunoResearch, West Grove, PA] were diluted 1:5,000 and incubated in wells for 1 hour at room temperature. Wells were washed again 3x with PBST. Wells were developed with TMB [ThermoFisher] and the reaction was quenched with 0.16 M sulfuric acid. Well absorbance was measured in a microplate reader at 450 nm. ELISAs were performed in triplicate for each antigen-antibody pair tested.

1.3. RESULTS

1.3.1. MILKSHAKE and Sundae surrogate proteins used for Western and ELISA Ab validation.

We took advantage of previous work showing sortase A can accept substrates of varying structure (29,30). We have developed a system in which a modified maltose binding protein (modMBP) with a sortase acceptor site is fused to a peptide containing our moiety of interest and a sortase donor site. The resulting protein, which we call MILKSHAKE, can be used in western blot to validate the specificity of phospho-specific binders to the modified or non-modified epitope of interest (Figure 1A). Positive binding to MILKSHAKE in western blot correlates with binding to full length target proteins found in treated cell lysates (22). We typically use MILKSHAKE proteins in two different experiment types: western blots in which lanes are loaded only with the MILKSHAKE protein as a test of the Ab’s specificity (which we have termed “vanilla”) and western blots in which lanes are loaded with a mixture of MILKSHAKE protein and HEK cell lysate as a more stringent test of the Ab’s sensitivity (which we have termed “strawberry”).

Figure 1. MILKSHAKE and SUNDAE methods for generating analyte material for western blot and ELISA.

Figure 1.

Open in a new tab

A) Modified maltose binding protein (modMBP) was purified from E. coli cultures and conjugated via sortase enzyme to peptides containing an N-terminal sortase conjugation site {GGGSGSS}, a target-specific sequence [either modified (e.g., phosphorylated) or non-modified at a single amino acid target site] and a C-terminal hemagglutinin (HA) tag {YPYDVPDYA}. Western blot analysis reveals binding specificity of the Ab under testing. The expected size of MILKSHAKE is 42 kDa. The anti-HA control blot was incubated with anti-HA HRP Ab [Abcam] at 1:5,000 dilution. Exposure 51 seconds. Each lane contains 500 ng MILKSHAKE and 15 μg HEK lysate. Lanes: M, Precision Plus Protein Standard [BioRad]; MBP, unconjugated modMBP negative control; 100, phospho conjugated MILKSHAKE Target A; 0, non-phospho conjugated MILKSHAKE Target A B) Sundae proteins contain genetically encoded epitope target sequences fused to the C-terminus of modMBP in a set of twenty different constructs. In each construct, one residue has been replaced by each of the 20 amino acids. These proteins are used in single-well or titration ELISA to determine binding affinity of Abs when this single residue is replaced.

Sundae proteins contain a genetically-encoded epitope from the target sequence fused to the C-terminus of modMBP. In each Sundae protein, one residue has been replaced by each of the 20 amino acids residues resulting in a full set of twenty Sundae proteins (Figure 1B). These surrogate proteins are used in single-well or titration ELISA to determine binding specificity of Abs when this single residue is replaced. Sundae proteins shed light on which residues are important for Ab binding as well as what physical and chemical attributes of the residue of interest may contribute to binding.

1.3.2. Vanilla MILKSHAKE demonstrates PTM Ab binding specificity.

Using modMBP and specifically designed peptides, we generated MILKSHAKE proteins containing the phosphorylated (single or dual depending on the target) or non-phosphorylated epitopes from desired proteins. With Abs from multiple vendors, we were able to detect binding to our phospho-peptide MILKSHAKE vs. non-phospho-peptide MILKSHAKE and not to a negative control: unconjugated modMBP (Figure 2).

Figure 2. Commercially acquired phospho-specific Abs tested using Vanilla MILKSHAKE.

Figure 2.

Open in a new tab

Blots were incubated with primary Abs from seven vendors diluted to vendor recommended concentrations for four different desired target sequences A) phospho-AKT (Ser473) B) phospho-Stat3 (Tyr705) C) phospho-ERK1 (Thr202/Tyr204) D) phospho-JNK1 (Thr183/Tyr185). Each IgG Ab binds the phosphorylated residues listed for each target which are present in the phospho conjugated modMBP (A and B) or the dual phosphorylated MILKSHAKE in Lane 2P or the single-phosphorylated MILKSHAKE proteins (in C and D). In some cases, Abs also bind non-phospho conjugated modMBP (see yellow arrows in A and Lane 0P in D (qualitative data). Each lane in A and B was loaded with 3 μg MILKSHAKE protein only, and each lane in C and D was loaded with 10 μg MILKSHAKE protein only. Exposure times are listed on each blot. Lane M: Precision Plus Protein Standard [BioRad] Lane MBP: unconjugated modMBP, negative control, Lane 2P: dual phosphorylated, Lane 0P: dual non-phosphorylated

The Abs for this study were chosen based upon recent Ab market data which identified them as being among the most cited phospho-specific Abs in literature (31). These include Abs important in cancer and diabetes research which bind phospho-AKT (Ser473), phospho-Stat3 (Tyr705), phospho-ERK1 (Thr202/Tyr204) and phospho-JNK1 (Thr183/Tyr185). Their prevalence in published research further supports the need to properly validate these Abs which are sold as phospho-specific to these sites. We chose Abs from top selling vendors but have chosen to keep the company names anonymous (e.g., Vendor A). Mainly because our results are meant to be illustrative of our methods and not a conclusive investigation of the quality of the manufacturers’ antibodies. For example, there may be lot-to-lot variation in these Abs or a performance difference if used under a different set of conditions than those used by the original manufacturer. Most of the Abs tested bind as expected to the phospho- MILKSHAKE and not to the non-phospho MILKSHAKE (13 out of 16 Abs tested). However, three Abs do show some reactivity to the non-phosphorylated MILKSHAKE protein in this vanilla version where only the MILKSHAKE protein is loaded in each lane (see Figure 2A and D, yellow arrows).

1.3.3. Strawberry MILKSHAKE demonstrates specificity and sensitivity in a cell lysate background.

In Figure 3, we demonstrate the range of sensitivity of PTM Abs for their MILKSHAKE protein targets when the surrogate protein is presented in a cell lysate background. In this strawberry version of the MILKSHAKE method, we are able to visualize, for some Abs, binding to the true target protein present in the HEK lysate as well as non-specific binding to other cellular proteins present in the HEK cell lysate. For example, the AKT protein has an expected size of approximately 60 kDa. Bands corresponding to this size can be seen in all four panels of Figure 3A.

Figure 3. Commercially acquired phospho-specific antibodies show varying degrees of specificity in strawberry MILKSHAKE.

Figure 3.

Open in a new tab

Blots were incubated with primary Abs from seven vendors at vendor recommended concentrations for four different desired target sequences A) phospho-AKT (Ser473) B) phospho-Stat3 (Tyr705) C) phospho-ERK1 (Thr202/Tyr204) D) phospho-JNK1 (Thr183/Tyr185). Each IgG Ab binds the phospho conjugated modMBP in A and B or the dual phosphorylated MILKSHAKE in Lane 2P or the single-phosphorylated MILKSHAKE proteins in C and D (qualitative data). In some cases, the Abs also bind non-specifically to other cellular proteins which do not match the expected size of the target protein (see orange arrows in A, C, D). In the dual PTM blots, some Abs fail to recognize both phospho sites individually (see pink arrows C and D). Each lane contains 500 ng MILKSHAKE and 15 μg of HEK cell lysate. Exposure times are listed on each blot. Lane M: Precision Plus Protein Standard [BioRad] Lane MBP: unconjugated modMBP, negative control Lane 2P: dual phosphorylated, Lane 0P: dual non-phosphorylated.

Some vendors’ Abs react to a variety of other cellular proteins whereas Abs from competitors are highly specific (Figure 3). For example, among the four Abs which bind phospho-AKT (Serine 473) each bind the phosphorylated MILKSHAKE and not the non-phosphorylated in this strawberry version of the MILKSHAKE method. However, the Ab from Vendor C shows reactivity to a variety of cellular proteins (Figure 3A, orange arrows), most of which are not the expected size of the AKT protein itself (approximately 60 kDa). In addition, Abs from Vendor G for phospho-JNK1 (Thr183/Tyr185) and for phospho-ERK1 (Thr202/Tyr204) show reactivity to other bands which do not correlate with the true sizes of those proteins either (Figure 3C, 3D, orange arrows).

1.3.4. Strawberry MILKSHAKE is a superior validation method for dual PTM site Abs compared to treated cell lysates.

Post-translational modification Abs to be used in western analysis are most commonly validated using mammalian cell lysates which have undergone treatment (i.e. chemical exposure, UV radiation etc.) to induce (or reduce) modification of residues on a protein of interest. This material can be extremely challenging to reproduce reliably. These challenges are especially evident when producing material for dual-phosphorylation proteins. Many proteins which undergo phosphorylation contain dual phospho sites which are typically in very close proximity in the amino acid sequence. Using only a treated cell lysate for Ab validation leads to uncertainty as to which site or sites the Ab is binding and which site or sites have been altered appropriately in the lysate.

In Figure 4, we used commercially available treated lysates [Cell Signaling Technology, Danvers, MA] to test phospho-specific Abs and compared these results to the strawberry MILKSHAKE results (Figure 3). For some targets, the data are compatible. For phospho-AKT (Ser473), Abs from Vendors A and D are phospho-specific in both assays and have very limited off-target binding. Antibodies from Vendors B and C demonstrate non-specific binding to other cellular proteins in both strawberry MILKSHAKE as well as treated cell lysates (Figure 3A, 4A).

Figure 4. Phospho-specific Abs tested with commercially available treated cell lysates.

Figure 4.

Open in a new tab

Lanes were loaded with cell lysates specifically treated for each PTM site under testing [Cell Signaling Technology]. Lane M: Precision Plus Protein Standard [BioRad], Lane Phospho lysate: cell lysate treated to produce phosphorylated site specific to the primary Ab tested, Lane Non-phospho lysate: cell lysate treated to produce non-phosphorylated site specific to the primary Ab tested A) Phospho lysate (20 μl): Jurkat cells, serum starved overnight followed by treatment with Calyculin A. Non-phospho lysate (20 μl): Jurkat cells, serum starved overnight followed by treatment with 50 μM LY294002. The expected size of AKT is 60 kDa. B) Phospho lysate (10 μl): serum-starved HeLa cells treated with 100 ng/ml interferon-alpha for 5 minutes. Non-phospho lysate (10 μl): serum-starved HeLa cell lysate. The expected size of Stat3 is 86 kDa. C) Phospho lysate (15 μl): Jurkat cells treated with 200 nM TPA for 20 minutes. Non-phospho lysate (15 μl): Jurkat cells treated with 10 μM U0126 for 1 hour. The expected size of ERK1 is 42 kDa. D) Phospho lysate (15 μl): 293 cells, treated with 50 mJ UV light followed by a 30-minute recovery. Non-phospho lysate (15 μl): 293 cells. The expected sizes of JNK1 are 46 and 54 kDa.

All four phospho-Stat3 (Tyr705) Abs tested are phospho specific and demonstrate limited off target binding in both assays (Figure 3B and 4B). The weakest binder (Vendor E) to phosphor-Stat3 (Tyr705) in treated lysates is also confirmed as the weakest binder in strawberry MILKSHAKE.

The targets containing dual phosphorylation sites however, tell a different story. In treated cell lysates, all four vendor Abs for phospho-ERK1 (Thr202/Tyr204) appear to correctly recognize only the phosphorylated cell condition (Figure 4C). However, when those same Abs were tested in strawberry MILKSHAKE, those from Vendors F and D are not able to recognize the MILKSHAKE protein on which only Tyr204 is phosphorylated (Figure 3C, pink arrows). These results suggest that these two Abs are only able to bind when Thr202 is phosphorylated.

Antibodies to phospho-JNK1 (Thr183/Tyr185) show vendor-specific results. The Ab from Vendor F does not bind in the treated lysate even after a prolonged exposure (Figure 4D). Antibodies from Vendors A, C and G appear phospho-specific in treated lysates for one or both of the expected sizes of the JNK1 protein, 46 and 54 kDa (Figure 4D). However, in strawberry MILKSHAKE these same Abs bind in three distinct ways: 1) to phospho-Tyr185 only (Vendors A and G) 2) to phospho-Thr183 only (Vendor F), or 3) to both phospho-Thr183 and phospho-Tyr185, albeit very weakly (Vendor C Figure 3D). In addition, off-target binding can be observed in both assays for Vendor G’s Ab to phospho-JNK1 (Thr183/Tyr185) (Figures 3D and 4D).

1.3.5. Sundae ELISA data demonstrate varied binding profiles for residue-specific Abs.

The epitope for 2F5, a human bnAb, is located on the HIV envelope glycoprotein gp41, specifically on a region of the protein known as the membrane-proximal external region (MPER) (28). The 2F5 epitope is composed of a conserved linear sequence of amino acids in MPER {662-ELDKWA-667}. This region of the HIV envelope protein is involved in the fusion of the virus with host cells during infection, making this region an important target for neutralizing antibodies. By binding to this region, 2F5 can block the fusion of the virus with host cells, preventing infection. Previous research has shown that the aspartic acid residue at the 664 position is important for 2F5 binding (32).

To test 2F5 binding, we generated 20 Sundae proteins, each with a different amino acid residue substituted at the D664 position (See Figure 1B). Titration ELISAs were performed to test the binding of the 2F5 antibody to each of the Sundae proteins. As expected 2F5 binds strongly to the gp41 protein, a peptide and the Sundae-D protein, all of which contain the MPER sequence with an aspartic acid residue at the 664 position (Figure 5A). These data confirm the importance of the aspartic acid residue (D) at this site. We also tested the binding of three IgGs generated from rabbits immunized with the MPER sequence using our site-directed Epivolve technology (33). Two of these IgGs (IGT-0035 and IGT-0037) show strong binding only when an aspartic acid is present at the 664 position in the Sundae-D protein (Figure 5A and Figure 5B). However, IGT-0034 is able to bind when a variety of other amino acids are present at this position, notably alanine, asparagine and proline, which implies its pan-variant binding capabilities (Figure 5B). The IGT-0034 antibody does also show some reactivity to the empty control protein however that binding is less strong than to the non-native sequence Sundae-N or Sundae-P proteins.

Figure 5. ELISA analysis of anti-MPER antibodies using panel of Sundae proteins.

Figure 5.

Open in a new tab

A) Sundae titration ELISA analysis (n = 3) of two IgGs: bnAb 2F5 and rabbit derived IGT-0037. These data suggest that the ‘D’ at position 664 is critical for binding. B) Single point Sundae ELISA analysis (n = 3) of four IgGs including 2F5 (measured at 7.5 μg mL−1). Ab reactivity varies depending on which residue is present at position 664 in the variant Sundae proteins. The three different rabbit IgG clones show different reactivity when a single residue in the Ab’s epitope (an aspartic acid in the native sequence) is replaced with other amino acids (Sundae-A, Sundae-E etc). In addition to Sundae proteins, positive controls (gp41 full length protein and D664 peptide) and a negative control (Sundae-Empty, modMBP without MPER sequence encoded) were also tested.

1.4. DISCUSSION

With the increasing importance of PTMs in regulating protein function and in many disease processes, it is essential to develop reliable and specific Abs that can detect and quantify PTM proteins accurately. The MILKSHAKE method aims to provide a standardized and comprehensive approach to PTM Ab validation, enabling researchers to generate more accurate and reproducible results and advance our understanding of complex biological systems. We believe this method will also work with other types of PTMs, including as examples: acetylation, methylation and glycosylation. An additional advantage of sortase is that it can be used to attach small molecules to proteins (Xiaofeng Li, personal communication).

Our study found that 13 out of 16 (81%) top selling PTM Abs tested performed as expected by binding the phosphorylated residue and not the non-phosphorylated residue in vanilla MILKSHAKE. However, the remaining 3 Abs (19%) tested which bind the non-phosphorylated MILKSHAKE protein are cause for concern. These data demonstrate that even among commonly used Ab vendors, the quality of Abs on offer are not always comparable or even valid. More concerning is that these findings apply to top selling Abs which likely have undergone more testing and scrutiny than less studied PTM sites. Antibodies to these lesser known sites may have more significant quality concerns. Researchers with access to MILKSHAKE data for PTM Abs could make a more informed decision when selecting an Ab for their experiments.

Dual phospho-sites have an even higher bar for validation. Traditional treated lysates as a method for PTM validation are not specific enough for dual or multi PTM sites because the treatments may affect each site unequally. With MILKSHAKE, the modification status of each individual residue is certain, and the binding specificity is more clearly defined. Out of the 8 dual phospho site Abs tested, we identified 5 Abs (62%) which appear dual-phospho-specific in a treated lysate experiment but only truly bind one phospho site in the strawberry MILKSHAKE method. Being able to validate Abs to this level of specificity may enable researchers to conduct more rigorous experiments. One assay could employ multiple Abs for these dual sites and dissect exactly which residues’ modifications are important for a particular pathway. Strawberry MILKSHAKE makes this possible, even when those modified sites are just one residue apart.

The strawberry MILKSHAKE data goes a step further than PTM specificity and reveals which Abs may have reactivity to other cellular proteins. The two Abs available from Vendor G for phosphor-JNK1 (Thr183/Tyr185) and for phospho-ERK1 (Thr202/Tyr204) demonstrate that an Ab may be PTM-specific but can also show off-target binding. The cell lysate used in the strawberry MILKSHAKE method could also be expanded to include other cell types besides HEK. This option would enable researchers to capture the full range of potential off-target binding in the cell types they plan to use in their own experiments. We are also planning two additional elements of the MILKSHAKE method to expand its application. The first is using purified MILKSHAKE protein to ensure the amount loaded in each Western blot lane is quantified. Currently, this method is qualitative and meant to help the end user quickly and easily evaluate the specificity of their purchased PTM antibody. A purified and quantified version will help researchers determine the lower limit of detection of the target protein for each Ab in western blot. The second is another version of MILKSHAKE which attaches peptides of interest to a mammalian cell (such as a CHO cell) surface via a sortase acceptor site on a membrane protein (which we have termed “chocolate”). This method would allow for testing in whole cell applications such as flow cytometry.

Additional evidence for the importance of Ab validation methods can be found in our own data when comparing results of the strawberry and vanilla tests. Antibodies for phospho-AKT (Ser 473) from Vendors B and C showed binding to the non-phospho MILKSHAKE protein in the vanilla experiment (Figure 2A). However, these same Abs have significantly reduced binding to the non-phospho MILKSHAKE protein when they are tested in a lysate background where many more epitopes, including the AKT protein itself, are present (Figure 3A). The results of both the vanilla and strawberry versions may help inform Ab selection depending on the type of analyte and how much target protein might be present in the planned experiment.

The Sundae method we present here can be used to better understand the relationship of the Ab paratope to the immunogen epitope. Sundae replaces the targeted site with up to 20 different amino acid residues and allows for retesting each mutated site for bioactivity. Similar peptide-based scanning methods have been used to explore Ab binding to a range of amino acid residues (34). The Sundae method allows for thorough Ab binding evaluation in a full-length surrogate protein antigen format. Our data for the broadly neutralizing 2F5 Ab correlates with previous 2F5 research which identifies the aspartic acid (D) at position 664 as crucial for binding to the MPER sequence in HIV (32). We were able to develop another Ab (IGT-0034) which relies less heavily on aspartic acid and is capable of binding to other residues at that same position. These residues include alanine (Sundae-A), proline (Sundae-P) and asparagine (Sundae-N) each of which have side chains that differ from the native aspartic acid both in size and electrostatic charge. Researchers could use the Sundae method to ‘walk down’ their target sequence and determine which Ab candidates have the appropriate binding profile for each residue along the target. This type of experiment would be especially helpful for identifying Abs with ‘pan-variant’ capabilities which can bind multiple amino acid residues at the same site. These may be useful when there are multiple known amino acid residue substitutions, and one Ab is desired for binding to all possible residues at that site. This method would also be helpful to evaluate therapeutic and diagnostic Abs for reduced ‘off-target’ effects which stem from binding to very similar sequences in non-target proteins.

Current methods for Ab validation are often time-consuming, labor-intensive, and lack standardization, leading to variability in results. We have developed two methods to address these limitations and better standardize Ab validation efforts: the MILKSHAKE method aimed at PTM Ab validation and the Sundae method aimed at fully exploring Ab-epitope interactions.

  • Surrogate protein Western blot validates post translational modification antibodies

  • Post translational modification antibodies for two sites may bind only one site

  • Surrogate protein ELISA reveals amino acid side chains’ impact on antibody binding

  • Surrogate protein ELISA may be used to reduce off-target antibody binding

1.5. ACKNOWLEDGEMENTS

Research reported in this study was supported by the National Institutes of Health under award number R01GM987654.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1.6 REFERENCES

  • 1.Goldman RD. Abs: indispensable tools for biomedical research. Trends Biochem Sci. 2000. Dec;25(12):593–5. doi: 10.1016/s0968-0004(00)01725-4. PMID: 11116184. [DOI] [PubMed] [Google Scholar]
  • 2.Groff K, Brown J, Clippinger AJ. Modern affinity reagents: Recombinant Abs and aptamers. Biotechnol Adv. 2015. Dec;33(8):1787–98. doi: 10.1016/j.biotechadv.2015.10.004. Epub 2015 Oct 19. PMID: 26482034. [DOI] [PubMed] [Google Scholar]
  • 3.Nelson PN, Reynolds GM, Waldron EE, Ward E, Giannopoulos K, Murray PG. Monoclonal Abs. Mol Pathol. 2000. Jun;53(3):111–7. doi: 10.1136/mp.53.3.111. PMID: 10897328; PMCID: PMC1186915. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Baker M. Reproducibility crisis: Blame it on the Abs. Nature. 2015. May 21;521(7552):274–6. doi: 10.1038/521274a. PMID: 25993940. [DOI] [PubMed] [Google Scholar]
  • 5.Bordeaux J, Welsh A, Agarwal S, Killiam E, Baquero M, Hanna J, Anagnostou V, Rimm D. Ab validation. Biotechniques. 2010. Mar;48(3):197–209. doi: 10.2144/000113382. Erratum in: Biotechniques. 2010 May;48(5):351. PMID: 20359301; PMCID: PMC3891910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Bradbury A, Plückthun A. Reproducibility: Standardize Abs used in research. Nature. 2015. Feb 5;518(7537):27–9. doi: 10.1038/518027a. PMID: 25652980. [DOI] [PubMed] [Google Scholar]
  • 7.Couchman JR. Commercial Abs: the good, bad, and really ugly. J Histochem Cytochem. 2009. Jan;57(1):7–8. doi: 10.1369/jhc.2008.952820. Epub 2008 Oct 14. PMID: 18854593; PMCID: PMC2605718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Edfors F, Hober A, Linderbäck K, Maddalo G, Azimi A, Sivertsson Å, Tegel H, Hober S, Szigyarto CA, Fagerberg L, von Feilitzen K, Oksvold P, Lindskog C, Forsström B, Uhlen M. Enhanced validation of Abs for research applications. Nat Commun. 2018. Oct 8;9(1):4130. doi: 10.1038/s41467-018-06642-y. PMID: 30297845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Taussig MJ, Fonseca C, Trimmer JS. Ab validation: a view from the mountains. N Biotechnol. 2018. Oct 25;45:1–8. doi: 10.1016/j.nbt.2018.08.002. Epub 2018 Aug 4. PMID: 30086383; PMCID: PMC6128662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Uhlen M, Bandrowski A, Carr S, Edwards A, Ellenberg J, Lundberg E, Rimm DL, Rodriguez H, Hiltke T, Snyder M, Yamamoto T. A proposal for validation of Abs. Nat Methods. 2016. Oct;13(10):823–7. doi: 10.1038/nmeth.3995. Epub 2016 Sep 5. PMID: 27595404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Zhou H, Watts JD, Aebersold R. A systematic approach to the analysis of protein phosphorylation. Nat Biotechnol. 2001. Apr;19(4):375–8. doi: 10.1038/86777. PMID: 11283598. [DOI] [PubMed] [Google Scholar]
  • 12.Kusebauch U, et al. Human SRMAtlas: A Resource of Targeted Assays to Quantify the Complete Human Proteome. Cell. 2016. Jul 28;166(3):766–778. doi: 10.1016/j.cell.2016.06.041. Epub 2016 Jul 21. PMID: 27453469; PMCID: PMC5245710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Maas MN, Hintzen JCJ, Mecinović J. Probing lysine posttranslational modifications by unnatural amino acids. Chem Commun (Camb). 2022. Jun 28;58(52):7216–7231. doi: 10.1039/d2cc00708h. PMID: 35678513. [DOI] [PubMed] [Google Scholar]
  • 14.Brandi J, Noberini R, Bonaldi T, Cecconi D. Advances in enrichment methods for mass spectrometry-based proteomics analysis of post-translational modifications. J Chromatogr A. 2022. Aug 16;1678:463352. doi: 10.1016/j.chroma.2022.463352. Epub 2022 Jul 19. PMID: 35896048. [DOI] [PubMed] [Google Scholar]
  • 15.Srivastava AK, Guadagnin G, Cappello P, Novelli F. Post-Translational Modifications in Tumor-Associated Antigens as a Platform for Novel Immuno-Oncology Therapies. Cancers (Basel). 2022. Dec 26;15(1):138. doi: 10.3390/cancers15010138. PMID: 36612133; PMCID: PMC9817968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chen J, Tsai YH. Applications of Genetic Code Expansion in Studying Protein Posttranslational Modification. J Mol Biol. 2022. Apr 30;434(8):167424. doi: 10.1016/j.jmb.2021.167424. Epub 2021 Dec 28. PMID: 34971673. [DOI] [PubMed] [Google Scholar]
  • 17.Ramazi S, Zahiri J. Posttranslational modifications in proteins: resources, tools and prediction methods. Database (Oxford). 2021. Apr 7;2021:baab012. doi: 10.1093/database/baab012. PMID: 33826699; PMCID: PMC8040245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Wang S, Osgood AO, Chatterjee A. Uncovering post-translational modification-associated protein-protein interactions. Curr Opin Struct Biol. 2022. Jun;74:102352. doi: 10.1016/j.sbi.2022.102352. Epub 2022 Mar 22. PMID: 35334254; PMCID: PMC9464464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Hattori T, Koide S. Next-generation antibodies for post-translational modifications. Curr Opin Struct Biol. 2018. Aug;51:141–148. doi: 10.1016/j.sbi.2018.04.006. Epub 2018 May 10. PMID: 29753204; PMCID: PMC6686900. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Voskuil JL. The challenges with the validation of research antibodies. F1000Res. 2017. Feb 17;6:161. doi: 10.12688/f1000research.10851.1. PMID: 28357047; PMCID: PMC5333605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Lithgow GJ, Driscoll M, Phillips P. A long journey to reproducible results. Nature. 2017. Aug 22;548(7668):387–388. doi: 10.1038/548387a. PMID: 28836615; PMCID: PMC5762131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Jones KS, Chapman AE, Driscoll HA, Fuller EP, Kelly M, Li X, Mansour S, McBride SL, Zhao Q, Weiner M, Ferguson MR. MILKSHAKE: novel validation method for Abs to post-translationally modified targets by surrogate Western blot. Biotechniques. 2022. Jan;72[1]:11–20. doi: 10.2144/btn-2021-0078. Epub 2021 Nov 29. PMID: 34841898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Helsby M. Who’s best supporting post-translational modifications? [Internet]. Bath, UK. 2016. Nov 16 – [cited 2023 Jun 21]. Available from https://www.citeab.com/ [Google Scholar]
  • 24.Isoda Y, Tanaka T, Suzuki H, Asano T, Nakamura T, Yanaka M, Handa S, Komatsu Y, Okuno S, Takahashi N, Okada Y, Kobayashi H, Li G, Nanamiya R, Goto N, Tateyama N, Yoshikawa T, Kaneko MK, Kato Y. Epitope Mapping of an Anti-Mouse CXCR6 Monoclonal Antibody (Cx6Mab-1) Using the 2 × Alanine Scanning Method. Monoclon Antib Immunodiagn Immunother. 2022. Oct;41(5):275–278. doi: 10.1089/mab.2022.0019. Epub 2022 Oct 25. PMID: 36301196. [DOI] [PubMed] [Google Scholar]
  • 25.Peng HP, Lee KH, Jian JW, Yang AS. Origins of specificity and affinity in antibody-protein interactions. Proc Natl Acad Sci U S A. 2014. Jul 1;111(26):E2656–65. doi: 10.1073/pnas.1401131111. Epub 2014 Jun 17. PMID: 24938786; PMCID: PMC4084487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Boersma MD, Sadowsky JD, Tomita YA, Gellman SH. Hydrophile scanning as a complement to alanine scanning for exploring and manipulating protein-protein recognition: application to the Bim BH3 domain. Protein Sci. 2008. Jul;17(7):1232–40. doi: 10.1110/ps.032896.107. Epub 2008 May 8. PMID: 18467496; PMCID: PMC2442000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Macri V, Nazzari H, McDonald E, Accili EA. Alanine scanning of the S6 segment reveals a unique and cAMP-sensitive association between the pore and voltage-dependent opening in HCN channels. J Biol Chem. 2009. Jun 5;284(23):15659–67. doi: 10.1074/jbc.M809164200. Epub 2009 Mar 6. PMID: 19269964; PMCID: PMC2708862. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Sánchez-Martínez S, Lorizate M, Katinger H, Kunert R, Nieva JL. Membrane association and epitope recognition by HIV-1 neutralizing anti-gp41 2F5 and 4E10 antibodies. AIDS Res Hum Retroviruses. 2006. Oct;22(10):998–1006. doi: 10.1089/aid.2006.22.998. PMID: 17067270. [DOI] [PubMed] [Google Scholar]
  • 29.Mao H, Hart SA, Schink A, Pollok BA. Sortase-mediated protein ligation: a new method for protein engineering. J Am Chem Soc. 2004. Mar 10;126(9):2670–1. doi: 10.1021/ja039915e. PMID: 14995162. [DOI] [PubMed] [Google Scholar]
  • 30.Guimaraes CP, Witte MD, Theile CS, Bozkurt G, Kundrat L, Blom AE, Ploegh HL. Site-specific C-terminal and internal loop labeling of proteins using sortase-mediated reactions. Nat Protoc. 2013. Sep;8(9):1787–99. doi: 10.1038/nprot.2013.101. Epub 2013 Aug 29. PMID: 23989673; PMCID: PMC3943461. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Antibody Market Dataset: Top 100 cited antibodies in 2021 [Internet]. Bath (UK) 2022 Sep - [cited 2023 Jun 21]. Available from: https://www.citeab.com/ [Google Scholar]
  • 32.Zwick MB, Jensen R, Church S, Wang M, Stiegler G, Kunert R, Katinger H, Burton DR. Anti-human immunodeficiency virus type 1 (HIV-1) antibodies 2F5 and 4E10 require surprisingly few crucial residues in the membrane-proximal external region of glycoprotein gp41 to neutralize HIV-1. J Virol. 2005. Jan;79(2):1252–61. doi: 10.1128/JVI.79.2.1252-1261.2005. PMID: 15613352; PMCID: PMC538539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Fuller EP, O’Neill RJ, Weiner MP. Derivation of splice junction-specific antibodies using a unique hapten targeting strategy and directed evolution. N Biotechnol. 2022. Nov 25;71:1–10. doi: 10.1016/j.nbt.2022.06.003. Epub 2022 Jun 22. PMID: 35750288; PMCID: PMC9464090. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Douglass J, Hsiue EH, Mog BJ, Hwang MS, DiNapoli SR, Pearlman AH, Miller MS, Wright KM, Azurmendi PA, Wang Q, Paul S, Schaefer A, Skora AD, Molin MD, Konig MF, Liu Q, Watson E, Li Y, Murphy MB, Pardoll DM, Bettegowda C, Papadopoulos N, Gabelli SB, Kinzler KW, Vogelstein B, Zhou S. Bispecific antibodies targeting mutant RAS neoantigens. Sci Immunol. 2021. Mar 1;6(57):eabd5515. doi: 10.1126/sciimmunol.abd5515. PMID: 33649101; PMCID: PMC8141259. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES