Sprinkling in extra validation for high-value PTMs and therapeutic Abs with MILKSHAKE Western blots and Sundae ELISAs
Abbratech, 25 Business Park Drive, Branford, Connecticut, USA
Tango Biosciences, 2201 W. Campbell Park Drive, Chicago, IL, USA
ProteoWise, Inc., 85 Willow Street, New Haven, CT, USA
Abstract
Thoroughly validated antibodies (Abs) are crucial for the generation of meaningful scientific data. Abs for post translationally modified (PTM) protein targets in particular pose added validation challenges. The MILKSHAKE method employs surrogate proteins which are either modified or non-modified at a specific site. Western blot is used to observe the binding of PTM Abs to the surrogate proteins, indicating the specificity of the PTM Ab under test. In this study, we expand the utility of MILKSHAKE by validating acetyl and methyl specific Abs and by introducing another surrogate protein antigen based on cellulose binding domain (CBD) to evaluate Abs in a single western blot lane. This study also explores the use of ‘Sundae’ surrogate protein ELISA specifically for therapeutic Ab evaluation.
Untitled section
Keywords: Post translational modification, Antibody, Sortase ligation, Validation, Western blot, ELISA
Article notes
Untitled section
Issue date 2025 Nov 25.
Introduction
Post translational modifications (PTM) expand the functional diversity of proteins by altering their activity, interactions, and stability [1–3]. Once identified, PTMs are frequently studied to establish their impact in biological systems. Acetylation and methylation are among the most well-studied modifications in recent literature [4]. Acetylation of lysine residues has a wide range of physiological effects on histone and non-histone proteins [5,6] while protein methylation contributes significantly to cellular regulation, especially via transcription regulation [7]. In addition, the enzymes responsible for both methylation and acetylation are themselves emerging therapeutic targets, particularly for certain cancers [8,9].
Antibodies (Abs) are important tools for studying cellular processes but are often unreliable in terms of specificity [10–13]. Several groups have endeavored to make the selection and use of Abs more transparent and accessible, [14–17] but thorough validation of Abs remains complex and inconsistent [18–20]. Unfortunately the use of poorly validated antibodies waste time, funding, and biological samples. By improving Ab validation, MILKSHAKE and Sundae align with research principles which promote the responsible use of resources and minimizing unnecessary experimental replication. Studies of PTMs often rely upon Abs which are specific to the modification of interest. Validation of these Abs requires materials, such as treated cell lysates or peptide arrays. Such materials are not always reproducible or cost-effective for researchers to produce on their own, leaving them with fewer resources to properly validate their antibodies.
The MILKSHAKE method aims to make validation of PTM Abs simpler and more cost-effective, thereby enabling the generation of reliable data. The MILKSHAKE protein, a surrogate for the target protein, is a modified maltose binding protein (modMBP) conjugated (via sortase A) to a synthetic peptide containing the PTM of interest [21]. The (malted) MILKSHAKE name stems from the modMBP protein. We have previously demonstrated by western blot that MILKSHAKE binding can distinguish phospho-specific Abs from Abs that also bind the non-phosphorylated protein target [22,23]. In this study, we have further tested the method in western blot with acetylated and methylated target proteins. In addition, we tested a new surrogate antigen, using cellulose binding domain (CBD) instead of modified maltose binding protein which is attached to the target sequence via sortase. The CBD-based antigen can be used in combination with MILKSHAKE to test the specificity of PTM Abs in a single lane of a western blot, taking advantage of their different molecular weights.
Another surrogate protein assay, the ‘Sundae’ ELISA, can also be used to study Ab-epitope pairs. Instead of utilizing sortase, the Sundae protein, is composed of a genetically modified modMBP which can be used to test a specific residue of interest and its contribution to binding [21]. This method is similar to alanine scanning, however, here the residue of interest can be replaced by any of the twenty amino acids (different Sundae ‘flavors’) to determine the impact of each side chain [24,25]. We have previously established this method with residue-specific Abs binding a viral sequence [23]. In this study, we have tested therapeutic Abs and residue-specific Abs with Sundae antigens to determine the efficacy of this method.
2.Materials & methods
2.1.Plasmid construction-MILKSHAKE and Sundae
The vector pMAL-c6T [New England Biolabs (NEB) Ipswich, MA, USA] was modified via a sequence insert using NotI and EcoRI for both plasmids. MILKSHAKE insert: a sortase site (LPETG) and a GS linker as previously described [21]. Sundae inserts were: pAT155 {RTKQHGQFSLAVVSLNI} (amino acids 429–445 of the epidermal growth factor receptor protein); pAT158 {SELLSLINDMPITNDQKKLMSNNVQ} (amino acids 255–279 of the respiratory syncitial virus F protein); pAT253 {VVVGAGGVGK}, pAT254 {VVVGAVGVGK}, pAT255 {VVVGAAGVGK}, pAT256 {VVVGADGVGK}, pAT257 {VVVGASGVGK}, pAT258 {VVVGARGVGK}, and pAT259 {VVVGACGVGK}.
2.2.Plasmid constructionCBD-(LPETG)-6XHis protein
A plasmid encoding the Cellulose Binding Domain (CBD) with C-terminal Sortase site, His6-tag (LPETGHHHHHH-COOH) and β-lactamase was synthesized by GenScript Biotech [Piscataway, NJ, USA].
2.3.modMBP purification
As previously described [21], modMBP was isolated from E. coli induced with 0.2 mM IPTG [Teknova Hollister, CA, USA]. Pellets were lysed and modMBP was purified using amylose resin [NEB].
2.4.CBD-(LPETG)-6XHis purification
Four hundred mL of 2XYT medium, supplemented with 50 μg/mL carbenicillin, was inoculated with 2 mL of an overnight culture of the KRX strain of E. coli. The culture was grown to mid-log phase and was then induced with 0.2 % rhamnose for 3 h. Cells were centrifuged at 10,000 rpm for 20 min, and the cell pellet resuspended in 25 mL of lysis buffer (50 mM sodium phosphate, 300 mM NaCl, 20 mM β-mercaptoethanol). Resuspended cells were lysed using a Sonics Vibra-Cell sonicator, at 50 % amplitude with 10-second On/Off pulses, until clear. The lysate was then centrifuged at 20,000 rpm for 20 min, and the supernatant incubated with His60 Ni Superflow resin [Takara Bio, San Jose, CA, USA] that had been resuspended in 15 mL lysis buffer without β-mercaptoethanol. After 2 h of tumbling at 4°C, the lysate-resin mixture was loaded onto a gravity flow column, and the flowthrough discarded. The resin was washed three times with 25 mL of wash buffer (50 mM sodium phosphate, 300 mM NaCl, 10 mM Imidazole). The bound protein was eluted over 15 1 mL fractions of elution buffer (50 mM sodium phosphate, 300 mM NaCl, 300 mM imidazole). The fractions containing protein, as monitored by absorbance at 280 nm wavelength, were pooled, buffer-exchanged against PBS, and concentrated using Pierce™ Protein Concentrator PES, 3 K centrifugal filters. The final concentration of desalted and concentrated protein samples was determined by Bradford assay.
2.5.Target peptides
MILKSHAKE peptides [LifeTein Somerset, NJ, USA] were designed to include a target sequence that is either non-modified or post-translationally modified (i.e. non-acetylated or acetylated). In addition, the MILKSHAKE peptides include a C-terminal hemagglutinin tag {YPYDVPDYA} to be used for detection in western blot and an N-terminal sortase site {GGGSGSS} to allow for conjugation of the peptide to a modified maltose binding protein [21]. For CBD-conjugated proteins, four glycine-terminated peptides, corresponding to variants of a phospho site within the DNA repair protein 53BP1, and a biotinylated peptide were synthesized with > 90 % purity: 53BP1-pT543, GGGSGSSIDEDGEN-(pT)-QIEDTEP; 53BP1-T543, GGGSGSSIDEDGENTQIEDTEP; 53BP1-pS543, GGGSGSSIDEDGEN-(pS)-QIEDTEP; and GGGSGSSST-Lys(Biotin)-KK.
2.7.Cell lysates
As previously described [21], HEK293-T cells [Sigma] alone were washed and lysed with RIPA buffer, and the lysate clarified by centrifugation. HEK lysates were then mixed with MILKSHAKE proteins before loading in SDS-PAGE gels for use in the Strawberry version of the method. NIH/3T3 lysates were purchased from Rockland Immuno-chemicals [Pottstown, PA, USA] and were used as recommended by the manufacturer.
2.8.Protein gel and western blot
As previously described for traditional western blot [21], polyacrylamide gels, 4–20 % [Bio-Rad] were loaded as per figure legends. Proteins were transferred to nitrocellulose (MILKSHAKE) or PVDF (CBD-LPETGHis6) membranes. Blots were blocked with 3 % Milk-TBST (Tris Buffered Saline, 0.1 % Tween 20). Primary Abs were diluted to vendor recommendations in 5 % BSA-TBST (or with 10 mL of biotinylated engineered FHA domain, A2, at 5 ng/μL for CBD blot). Blots were imaged with chemiluminescence reagent [Bio-Rad and ThermoFisher].
Hyperblot automated multiplex western blot system was also used to analyze MILKSHAKE proteins [ProteoWise New Haven, CT, USA]. In short, 2.0 μL of MILKSHAKE proteins solubilized in Laemmli buffer (0.1 mg/mL) mixed 1:50 with PageRuler strep-tagged unstained molecular weight ladder [ThermoFisher] were loaded in parallel into the Hyperblot system cartridge, consisting of a 2.0 cm× 2.0 cm micro polyacrylamide gel, 3D-NPC (3-Dimensional Nanoscale Protein Capture) membrane and fluidic immunostaining chamber. After sample loading, the cartridge was inserted into a Hyperblot instrument, along with 2.5 μL of each primary Ab (1 μg/mL final concentration) contained in discrete bar-code vials received by the instrument. Samples were separated on the basis of molecular weight, transferred to 3D-NPC membrane for covalent capture, and subjected to repeated immunocycling with successive image capture at each immunocycle. Automated image processing, signal intensity measurement and molecular weight determination per intra-lane ladder for each protein were performed in an automated fashion by the Hyperblot system.
2.9.CBD and sundae ELISA
Plates were coated with Sundae proteins at 1 μg/mL and blocked with 1 % BSA/PBS as previously described [21]. Primary antibodies Ipilimumab, Cetuximab, Nivolumab, Motavizumab and Palivizumab [Selleck Chemicals LLC, Houston TX, USA] and anti-KRAS G12V and G12D [Abcam Cambridge, UK] were diluted in 1 % BSA/PBS. All washes were performed 3X with PBST. For the CBD-LPETG- His6 ELISA experiment, 50 μL of the sortase reactions, diluted 1:20 in PBS, were used to coat microtiter plate wells. Non-specific binding sites were blocked with 2 % milk PBS for 1 h at room temperature, followed by three washes with 100 μL of PBS. Next, 50 μL (10 ng/μL) of biotinylated engineered Forkhead-associated (FHA) domain, A2 [26], in PBST, was added and incubated for 1 h at room temperature. Afterward, 50 μL of HRP-conjugated streptavidin [Jackson ImmunoResearch, West Grove, PA, USA], at a 1:5000 dilution in PBST, was added and incubated for 1 h at room temperature. Wells were developed with 50 μL of TMB substrate and stopped with 50 μL of 1 N HCl, and the absorbance was measured at 450 nanometer wavelength in a microplate reader. Measurements were averaged across triplicate wells, and the standard deviation was calculated.
3.Results
3.1.Testing acetylated lysine Abs with MILKSHAKE
We produced MILKSHAKE proteins containing acetylated or non-acetylated epitopes from a variety of human proteins and compared the acetyl-specificity of Abs from multiple vendors (Fig. 1). The acetylated lysine specific Abs in Fig. 1 include those for acetyl-p53 (Lys379), acetyl-NFκβ (Lys310), and acetyl-Histone H4 (Lys12). The Abs were purchased from a variety of Ab companies, many of which have been widely cited in research. The identity of each company is anonymized (e. g., Vendor A). The acetyl Abs tested bind to the acetyl MILKSHAKE protein and do not bind the non-acetyl MILKSHAKE protein in western blot (Vanilla version: MILKSHAKE protein only loaded in each lane). The same result is also the case for Abs tested with acetyl-α-tubulin (Lys40) MILKSHAKE proteins (data not shown). In order to detect possible off-target binding to other cellular proteins, we increase the stringency of the method by spiking the MILKSHAKE proteins into an HEK lysate (Strawberry MILKSHAKE). Using this method, we visualized some bands which do not correspond to the expected size of the p53 protein for acetyl-p53 (Lys379) (Fig. 1 B, Vendor A and Vendor C). In the case of Vendor A, these other bands are much less intense than the primary band corresponding to the size of p53. When comparing the strength of binding to the MILKSHAKE protein, Vendor C’s Ab for acetyl-p53 (Lys379) shows weaker binding than Abs from other suppliers (Fig. 1 A and B). This weaker target binding elevates the concern for the off-target bands which are of similar intensity to the primary p53 target band.
3.2.Validating methyl-specific Abs with MILKSHAKE
We evaluated the performance of selected methyl-specific Abs using MILKSHAKE protein targets in Fig. 2. The Histone H3 (tri-methyl Lys4) Abs from all Vendors bind to the tri-methyl Lys4 MILKSHAKE protein, with varying signal intensity in the di-methyl MILKSHAKE lane (Fig. 2A). However, when these same Abs are tested in the HEK293-T lysate background (Strawberry version), the Vendor H Ab, recognizes both the di-methyl and tri-methyl versions with similar signal intensity while the other Abs tested bind more strongly to the tri-methyl MILKSHAKE protein (Fig. 2B). The Ab from Vendor A in Fig. 2B binds strongest and most specifically to the tri-methyl MILKSHAKE protein. Strong non-specific binding to other proteins was observed for Vendor G’s Histone H3 (tri-methyl Lys4) Ab and well was Vendor H’s Ab (Fig. 2B).
Comparison of the Vanilla and Strawberry MILKSHAKE methods for Histone H4 (mono-methyl Lys20) Abs revealed a similar phenomenon. Vendor A’s Ab shows some weak binding to the tri-methyl and non-methyl MILKSHAKE proteins which is eliminated once tested in a lysate background (Fig. 2C vs. 2D). However, Vendor C’s Ab binds to the tri-methyl and non-methyl MILKSHAKE proteins in the Vanilla version and the non-specificity persists in the Strawberry version, where significant non-specific binding to HEK293-T lysate proteins was found (Fig. 2C vs. 2D). The Abs from Vendors E and F bind specifically only to the mono-methylated Lys20 MILKSHAKE protein.
3.3.Testing specificity of Abs for dual sites: methylated lysine - phosphorylated serine
We tested Abs which recognize Histone H3 when tri-methylated at Lys9 and phosphorylated at Ser10 (Fig. 3). This dual PTM is important for research involving gene expression and gene silencing [27]. Antibodies that recognize Histone H3 when it is both methylated and phosphorylated at these sites can be used to study mitosis, specifically chromosome condensation [28]. Due to the complexity of this PTM, four different MILKSHAKE proteins were generated in order to test binding with and without phosphorylation and with and without tri-methylation at the two sites. All four Abs tested display strong binding to the tri-methylated and phosphorylated MILKSHAKE protein in both versions of the method. These Abs also exhibit only weak or absent binding to non-phospho Ser10 and non-methyl Lys9 (Fig. 3A and B). However, Abs from Vendor D and Vendor I do show binding to other cellular proteins which do not correspond to the 17 kDa molecular weight of Histone H3 (Fig. 3B). We also tested an Ab that recognizes Histone H3 when Lys9 is not methylated; this Ab had strong binding to non-methylated MILKSHAKE proteins, regardless of the phosphorylation status at the Ser10 site (data not shown).
3.4.For methylated target proteins, Strawberry MILKSHAKE is a more reliable validation technique than untreated cell lysates
Antibodies against PTM sites are commonly validated using cell lysates, sometimes after undergoing PTM-specific treatment. Cell lysate generation for methylated target proteins is difficult since some residues can be mono- di- or tri-methylated and more than one residue may be methylated on the same protein. Ab developers have therefore used cell lysates without any specific treatment in these cases.
In Fig. 4, we used commercially available untreated lysates to test methyl-specific Abs. In each case, the lysate tested was the same as is commonly used for validation by the (anonymized) Ab companies for product datasheets. In Fig. 4A, NIH/3T3 lysate was loaded into each lane, blotted, and probed with four Histone H4 mono-methyl Lys20 Abs. The expected size of Histone H4 is 11 kDa. The Ab from Vendor A has a strong band at the expected size and Abs from Vendors E and F have weaker bands at the expected size. However, the Ab from Vendor C has a strong band at the expected size accompanied by many additional off-target bands. This data closely matches the results from the Strawberry MILKSHAKE experiment in Fig. 2D including the Vendor C non-specific bands.
Next, we tested Abs which recognize Histone H3 tri-methyl Lys4 with NIH/3T3 lysates. These lysates are also commonly used for validation by the Ab vendors for this target. The expected size of Histone H3 is 17 kDa; all four Abs have bands at the expected size. However, the Ab from Vendor G has prominent, non-specific bands (Fig. 4B). Data for Vendors A, B and G here closely match the results from the Strawberry MILKSHAKE experiment in Fig. 2B. This includes the additional bands seen with Vendor G’s Ab. One glaring difference, however, is exposed by comparing the Vendor H data from Fig. 2B to the data in Fig. 4B (using NIH/3T3 lysate). The typically used NIH/3T3 validation method would suggest this Ab performs as expected, but Strawberry MILKSHAKE demonstrates that this Ab in fact binds to both the di-methyl and tri-methyl versions of the protein which may lead a researcher to erroneous conclusions.
Four Abs which recognize Histone H3 when tri-methylated at Lys9 and phosphorylated at Ser10 were tested using NIH/3T3 western blot. Abs from Vendors D, C and I have bands at the expected size while Vendor B does not produce a band at this exposure. These data are similar to those from the Strawberry MILKSHAKE experiment in Fig. 3B. Vendor B’s Ab produced a very weak signal in MILKSHAKE (Fig. 3B). The comparison of Figs. 3B and 4C also shows non-specific bands for Vendor D and I Abs in both experiments as well.
To further assess the MILKSHAKE method, we also tested these surrogate antigens using a high throughput western blot instrument, the Hyperblot. This device allows antigens to be separated and then probed with multiple antibodies in succession. The Hyperblot data with Histone H3 (tri-methyl Lys4) Abs show specific binding for the Ab from Vendor A (Fig. 5). However, the Ab from Vendor H binds to both the di-methyl and tri-methyl MILKSHAKE protein variants. These data are very similar to those obtained in traditional western blot experiments (Fig. 2A).
3.5.MILKSHAKE and CBD surrogate proteins can be used in combination to increase efficiency
To increase the efficiency of the MILKSHAKE method, we have developed a similar surrogate protein using cellulose binding domain (CBD) with a C-terminal sortase acceptor site. This CBD protein, once conjugated to a modified or unmodified target peptide is tested in western blot and ELISA to determine the binding specificity of PTM Abs or other affinity reagents. In Figs. A and B, the CBD protein is conjugated to a peptide sequence from the DNA repair protein 53BP1 which is either phosphorylated or non-phosphorylated at residue Thr543. The affinity reagent engineered Forkhead-associated (FHA) domain, A2, [26] was able to bind to the CBD protein containing the phosphorylated Thr543 and did not bind the non-phosphorylated Thr543 CBD or a phosphorylated Ser543 CBD control.
In Fig. 6C, two Histone H3 tri-methyl Lys4 Abs were tested using both MBP and CBD conjugated proteins in a single lane. MILKSHAKE has a molecular weight of 42 kDa and CBD is 25.2 kDa, making it possible to test one Ab on two different surrogate antigens at once. In both blots, Lane 1 was loaded with unconjugated CBD and MBP, Lane 2 was loaded with CBD conjugated to a tri-methyl Lys4 target peptide and MBP conjugated to a di-methyl Lys4 target peptide. Lane 3 was loaded with the opposite combination: CBD conjugated to a di-methyl Lys4 target peptide and MBP conjugated to a tri-methyl Lys4 target peptide. Vendor B’s Ab binds only to the tri-methyl antigen displaying one band at the expected size of CBD in Lane 2 and one band at the expected size of MBP in Lane 3 (Fig. 6C). These results correspond with those obtained for the same Ab in Fig. 2. Vendor H’s Ab binds to both the tri-methyl and di-methyl antigens in each lane (Fig. 6C). This binding profile matches the Vendor H Ab binding in Fig. 2.
3.6.Sundae ELISA: varied utility for therapeutic and residue-specific Abs
Sundae surrogate proteins are used in ELISA to determine binding specificity of Abs when a single residue is replaced by any one of up to twenty different amino acids. The epitopes for selected therapeutic Abs were used to design Sundae proteins, each of which contained between 17 and 24 amino acids of the drug epitope. Sundae ELISA tested binding to the corresponding therapeutic Abs.
A Sundae protein (pAT155) containing residues 429–445 of the epidermal growth factor receptor (EGFR) protein was designed to target Cetuximab binding. The residues chosen for this sundae protein are known to be important for Cetuximab binding and function [29]. In Fig. 7A, Cetuximab was unable to bind these residues as presented in the Sundae protein. Additional therapeutic antibodies that were also unable to bind Sundae proteins include Ipilimumab and Nivolumab (data not shown).
The Abs Palivizumab and Motavizumab both target respiratory syncytial virus (RSV). Both are known to bind residues between 255 and 279 of the viral fusion (F) glycoprotein [30,31]. A Sundae protein incorporating these residues was designed (pAT158) and tested in ELISA (Fig. 7B). Both Abs bind pAT158 and show minimal binding to the negative control, a modMBP protein which does not contain residues from the RSV fusion glycoprotein. Previous reports have shown that Motavizumab binds more strongly to the RSV epitope than Palivizumab [32]. That difference is also evident in the ELISA data with the pAT158 Sundae protein (Fig. 7B).
It is worth noting that Motavizumab was pulled from the market after initial excitement about its enhanced binding affinity compared to Palivizumab. Although it showed stronger binding to the RSV fusion glycoprotein, as demonstrated in studies similar to the one discussed here, the decision to withdraw it was likely influenced by its failure to significantly outperform Palivizumab in terms of clinical outcomes and safety [33,34].
To further demonstrate the use of the Sundae method, we selected Abs which can recognize a single amino acid difference between two proteins, specifically those that bind mutants of the KRAS protein at position G12, an important driver of many cancers [35]. We identified two antibodies that each bind a KRAS mutant and tested their binding to each of the seven most prevalent KRAS G12 mutants in the Sundae antigen format (Fig. 8) The Ab which recognizes KRAS G12V is highly specific in Sundae ELISA, binding strongly to the G12V Sundae antigen. The KRAS G12D Ab however binds to the KRAS G12C Sundae antigen as well as the KRAS G12D version at concentrations above 0.3 μg/mL (Fig. 8B)
4.Discussion
Effective Ab validation is critical for biological assays such as western blotting and ELISA. In this study, we have expanded the use of MILKSHAKE and introduced a CBD-conjugated method, both designed to streamline and improve PTM Ab validation. The issues of time and cost can both be considerable hurdles for scientists and these are important to consider as research moves toward more stringent antibody validation. Unlike current methods, MILKSHAKE and CBD allow for standardization and increased convenience of Ab validation using western blot, an assay accessible to most research laboratories. The single lane combination of CBD and MBP fusion proteins offers a quick and cost-effective shortcut to reliable data. Furthermore, the choice of surrogate protein may influence the outcomes of Ab validation. Continuing to explore alternative scaffolds beyond MBP and CBD could address these type of validation issues as they arise.
In our analysis of acetylated lysine Abs, all tested Abs (n = 23) showed specificity in MILKSHAKE western blot, binding exclusively to acetylated residues (Vanilla version). This result outperforms previous findings with phospho-specific antibodies [22,23]. These results also highlight the value of the stringent validation practices used by Ab developers for acetyl antibodies. It has long been known that some proteins, especially histones, are acetylated at multiple sites on the same protein which makes for a more challenging validation process [36]. Antibody developers have employed competition assays with acetylated peptides to ensure their products are specific [37,38]. We did, however, observe some off-target binding with the use of cell lysates (Strawberry version). This result emphasizes the need for antibodies to be validated in assays specific to each study since Ab performance can change depending on the assay or materials used.
Our experiments with methyl-specific Abs and combination PTMs produced more variable results. While Vanilla MILKSHAKE identified non-PTM binding in some cases, the Strawberry MILKSHAKE experiment reduced this liability in certain instances. This result emphasizes the importance of using Strawberry MILKSHAKE which more closely mimics researchers’ experimental conditions. Some Abs tested, such as those targeting di- and tri-methyl Histone H3 (Lys4), displayed non-specific binding across both MILKSHAKE methods, demonstrating the risk of false-positive or misleading data when using only traditional lysate methods (Fig. 4B). Complex PTMs require more thorough validation than a cell lysate can offer. Researchers must be able to determine if Abs bind to other PTMs or other proteins and the MILKSHAKE method affords that opportunity. Our data using the Hyperblot instrument supports our traditional western blot findings. The ability of the MILKSHAKE surrogate antigen to perform well in other systems and produce similar results highlights a key advantage of the method: a reproducible, standardized antigen. The Hyperblot instrument itself also reduces the effort involved in thorough Ab validation, allowing a researcher to load one set of antigens and efficiently test multiple Abs to find the best reagent.
While the focus of this study was on acetylation, methylation, and dual modifications, the MILKSHAKE platform could be further expanded to less common PTMs, such as O-linked glycosylation, ubiquitination, or sumoylation. Developing surrogate proteins with these modifications could provide a standardized approach to studying emerging PTM-related pathways. It is also important to note that both polyclonal antibodies (pAbs) and monoclonal antibodies (mAbs) were tested in this manuscript. Our data reveal that the clonality of the antibody does not appear to contribute to its performance. Each antibody must always be validated for PTM specificity and off-target binding regardless of monoclonal or polyclonal status.
Validation using untreated cell lysates can be challenging, as they lack the ability to distinguish between modified and non-modified residues. While alternative methods, such as peptide blocking or Luminex assays, can be useful, they are less accessible, and produce results that are not directly comparable to those obtained from western blotting. For comprehensive validation, researchers should assess all relevant PTM combinations within a protein and use a ‘same-assay’ standardized method such as Strawberry MILKSHAKE. While MILKSHAKE has shown robustness in a lysate background, it may still face challenges with yet untested target types or when applied to more complex samples, such as serum or tissue extracts. Future adaptations, such as optimized blocking protocols, might enhance performance in these contexts.
The Sundae surrogate protein can be used to assess a particular residue’s role in an Ab—epitope pair, utilizing the full range of naturally occurring amino acids. It is important however to evaluate if this antigen works well for each protein target. As we demonstrated, some Abs are not able to recognize Sundae proteins (Cetuximab) while others are (Palivizumab and Motavizumab). We attribute this difference to the epitope itself; those targets with linear epitopes are more amenable to Sundae interrogation than those that are discontinuous or rely on a three-dimensional structure. Once it is determined that the Sundae antigen is appropriate for testing the Ab of interest, amino acid-variant sets of these proteins can then be tested to determine the contribution of each site to binding. Sundae may be important for proteins which have sequence-similar counterparts either in the human proteome or in experimental animals used in therapeutic development. The Sundae method is also useful for epitopes with a single amino acid difference such as KRAS G12 mutants. A set of full-length proteins with all twenty amino acid variations may be costly or difficult to produce. However, with Sundae researchers can determine, if an Ab is truly specific to just one mutation and better evaluate the Ab concentration they plan to use in an experiment in a more time and cost efficient way.
Overall, MILKSHAKE and Sundae advance Ab validation by identifying specificity and off-target binding concerns. Beyond basic research, these methods have broader applications in drug development, clinical diagnostics, and improving reproducibility across labs. Their use in high-throughput screening and therapeutic Ab refinement demonstrates their potential to streamline Ab development and enhance binding precision. While MILKSHAKE and Sundae are cost-effective relative to some current methods, small laboratories with limited budgets might still find the initial setup (e.g., cloning, protein production) to be a barrier. Development of ready-to-use kits or collaboration with core facilities could help democratize access.
Acknowledgements
Research in this study was supported by awards, 2R44GM146473-02 and 1R43GM146514-01, from the National Institutes of Health.
Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Michael Weiner reports financial support was provided by Abbratech Inc. Brian Kay reports financial support was provided by Tango Biosciences. Michael Weiner has patent #63/655050 pending to Abbratech Inc. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Footnote Group
References
Untitled section
References
- [1]Chen J, Tsai YH. Applications of genetic code expansion in studying protein posttranslational modification. J Mol Biol 2022;434(8):167424. 10.1016/j.jmb.2021.167424. Epub 2021 Dec 28.
- [2]Ramazi S, Zahiri J. Posttranslational modifications in proteins: resources, tools and prediction methods. Database 2021;2021:baab012. 10.1093/database/baab012.
- [3]Wang S, Osgood AO, Chatterjee A. Uncovering post-translational modification-associated protein-protein interactions. Curr Opin Struct Biol 2022;74:102352. 10.1016/j.sbi.2022.102352. Epub 2022 Mar 22.
- [4]Bobalova J, Strouhalova D, Bobal P. Common post-translational modifications (ptms) of proteins: analysis by up-to-date analytical techniques with an emphasis on barley. J Agric Food Chem 2023;71(41):14825–37. 10.1021/acs.jafc.3c00886. Epub 2023 Oct 4.
- [5]Drazic A, Myklebust LM, Ree R, Arnesen T. The world of protein acetylation. Biochim Biophys Acta 2016;1864(10):1372–401. 10.1016/j.bbapap.2016.06.007. Epub 2016 Jun 11.
- [6]Narita T, Weinert BT, Choudhary C. Functions and mechanisms of non-histone protein acetylation. Nat Rev Mol Cell Biol, 20; 2019. Mar. p. 156–74. https://doi.org/10.1038/s41580-018-0081-3. Erratum in: Nat Rev Mol Cell Biol. 2019 Aug;20 (8):508. doi: 10.1038/s41580-019-0156-9.
- [7]Lee DY, Teyssier C, Strahl BD, Stallcup MR, 2005. Role of protein methylation in regulation of transcription. Endocr Rev. (2), 147–170. doi: 10.1210/er.2004-0008. Epub 2004 Oct 12.
- [8]Dilworth D, Barsyte-Lovejoy D. Targeting protein methylation: from chemical tools to precision medicines. Cell Mol Life Sci 2019;76(15):2967–85. 10.1007/s00018-019-03147-9. Epub 2019 May 18.
- [9]Yang J, Song C, Zhan X. The role of protein acetylation in carcinogenesis and targeted drug discovery. Front Endocrinol 2022;13:972312. 10.3389/fendo.2022.972312.
- [10]Baker M Reproducibility crisis: blame it on the Abs. Nature 2015;521(7552): 274–6. 10.1038/521274a.
- [11]Bordeaux J, Welsh A, Agarwal S, Killiam E, Baquero M, Hanna J, Anagnostou V, Rimm D. Ab validation. Biotechniques 2010;48(3):197–209. 10.2144/000113382. Erratum in: Biotechniques. 2010 May;48(5):351.
- [12]Bradbury A, Plückthun A. Reproducibility: standardize Abs used in research. Nature 2015;518(7537):27–9. 10.1038/518027a.
- [13]Couchman JR. Commercial Abs: the good, bad, and really ugly. J Histochem Cytochem 2009;57(1):7–8. 10.1369/jhc.2008.952820. Epub 2008 Oct 14.
- [14]Bandrowski A, Pairish M, Eckmann P, Grethe J, Martone ME. The Antibody Registry: ten years of registering antibodies. Nucleic Acids Res 2023;51(D1): D358–67. 10.1093/nar/gkac927.
- [15]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;9(1):4130. 10.1038/s41467-018-06642-y.
- [16]Biddle M, Stylianou P, Rekas M, Wright A, Sousa J, Ruddy D, Stefana MI, Kmiecik K, Bandrowski A, Kahn RA, Laflamme C, Krockow EM, Virk HS. Improving the integrity and reproducibility of research that uses antibodies: a technical, data sharing, behavioral and policy challenge. MAbs 2024;16(1):2323706. 10.1080/19420862.2024.2323706. Epub 2024 Mar 6.
- [17]Ayoubi R, Ryan J, Biddle MS, Alshafie W, Fotouhi M, Bolivar SG, Moleon VR, Eckmann P, Worrall D, McDowell I, Southern K, Reintsch W, Durcan TM, Brown CM, Bandrowski A, Virk HS, Edwards AM, McPherson PS, Laflamme C. Scaling of an antibody validation procedure enables quantification of antibody performance in major research applications. bioRxiv 2023. https://doi.org/10.1101/2023.06.01.543292. Update in: Elife. 2023 Nov 23;12:RP91645. doi: 10.7554/eLife.91645.
- [18]Taussig MJ, Fonseca C, Trimmer JS. Ab validation: a view from the mountains. N Biotechnol 2018;45:1–8. 10.1016/j.nbt.2018.08.002. Epub 2018 Aug 4.
- [19]Kahn RA, Virk H, Laflamme C, Houston DW, Polinski NK, Meijers R, Levey AI, Saper CB, Errington TM, Turn RE, Bandrowski A, Trimmer JS, Rego M, Freedman LP, Ferrara F, Bradbury ARM, Cable H, Longworth S. Antibody characterization is critical to enhance reproducibility in biomedical research. Elife 2024;13:e100211. 10.7554/eLife.100211.
- [20]Lithgow GJ, Driscoll M, Phillips P. A long journey to reproducible results. Nature 2017;548(7668):387–8. 10.1038/548387a.
- [21]Ferguson MR, Mendez QM, Acca FE, Chapados CD, Driscoll HA, Jones KS, Mirando G, Weiner MP, Li X. Validation and the determination of antibody bioactivity using MILKSHAKE and sundae protocols. Methods Mol Biol 2023;2702: 451–65. 10.1007/978-1-0716-3381-6_24.
- [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;72(1):11–20. 10.2144/btn-2021-0078. Epub 2021 Nov 29.
- [23]Mendez Q, Driscoll HA, Mirando GR, Acca F, Chapados CD, Jones KS, Weiner M, Li X, Ferguson MR. MILKSHAKE Western blot and Sundae ELISA: we all scream for better antibody validation. J Immunol Methods 2023;521:113540. 10.1016/j.jim.2023.113540. Epub 2023 Aug 18.
- [24]Morrison KL, Weiss GA. Combinatorial alanine-scanning. Curr Opin Chem Biol 2001;5(3):302–7. 10.1016/s1367-5931(00)00206-4.
- [25]Warren PD, Dodson MS, Smith MH, Landowski TH, Palting JD, Towne P. Highresolution epitope mapping and affinity binding analysis comparing a new antihuman LAG3 rabbit antibody clone to the commonly used mouse 17B4 clone Antibodies 2022;11(4):60. 10.3390/antib11040060.
- [26]Thota Srinivas S, Allen Grace L, Grahn Ashley K, Kay Brian K. Engineered FHA domains can bind to a variety of Phosphothreonine-containing peptides. Protein Eng Des Sel 2024. 10.1093/protein/gzae014.
- [27]Sabbattini P, Sjoberg M, Nikic S, Frangini A, Holmqvist PH, Kunowska N, Carroll T, Brookes E, Arthur SJ, Pombo A, Dillon N. An H3K9/S10 methyl-phospho switch modulates Polycomb and Pol II binding at repressed genes during differentiation. Mol Biol Cell 2014;25(6):904–15. 10.1091/mbc.E13-10-0628. Epub 2014 Jan 15.
- [28]Prigent C, Dimitrov S. Phosphorylation of serine 10 in histone H3, what for? J Cell Sci 2003;116(Pt 18):3677–85. 10.1242/jcs.00735.
- [29]Li S, Schmitz KR, Jeffrey PD, Wiltzius JJ, Kussie P, Ferguson KM. Structural basis for inhibition of the epidermal growth factor receptor by cetuximab. Cancer Cell 2005;7(4):301–11. 10.1016/j.ccr.2005.03.003. 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.
- [30]López JA, Andreu D, Carreño C, Whyte P, Taylor G, Melero JA. Conformational constraints of conserved neutralizing epitopes from a major antigenic area of human respiratory syncytial virus fusion glycoprotein. J Gen Virol 1993;74(Pt 12): 2567–77. 10.1099/0022-1317-74-12-2567.
- [31]McLellan JS, Chen M, Kim A, Yang Y, Graham BS, Kwong PD. Structural basis of respiratory syncytial virus neutralization by motavizumab. Nat Struct Mol Biol 2010;17(2):248–50. 10.1038/nsmb.1723. Epub 2010 Jan 24.
- [32]Wu H, Pfarr DS, Johnson S, Brewah YA, Woods RM, Patel NK, White WI, Young JF, Kiener PA. Development of motavizumab, an ultra-potent antibody for the prevention of respiratory syncytial virus infection in the upper and lower respiratory tract. J Mol Biol 2007;368(3):652–65. 10.1016/j.jmb.2007.02.024. Epub 2007 Feb 20.
- [33]Resch B. Product review on the monoclonal antibody palivizumab for prevention of respiratory syncytial virus infection. Hum Vaccin Immunother 2017;13(9): 2138–49. 10.1080/21645515.2017.1337614. Epub 2017 Jun 12.
- [34]Carbonell-Estrany X, Simões EA, Dagan R, Hall CB, Harris B, Hultquist M, Connor EM, Losonsky GA. Motavizumab study group. motavizumab for prophylaxis of respiratory syncytial virus in high-risk children: a noninferiority trial. Pediatrics 2010;125(1):e35–51. 10.1542/peds.2008-1036. Epub 2009 Dec 14.
- [35]Huang L, Guo Z, Wang F, Fu L. KRAS mutation: from undruggable to druggable in cancer. Signal Transduct Target Ther 2021;6(1):386. 10.1038/s41392-021-00780-4.
- [36]Sterner DE, Berger SL. Acetylation of histones and transcription-related factors. Microbiol Mol Biol Rev 2000;64(2):435–59. 10.1128/MMBR.64.2.435-459.2000.
- [37]Suka N, Suka Y, Carmen AA, Wu J, Grunstein M. Highly specific antibodies determine histone acetylation site usage in yeast heterochromatin and euchromatin. Mol Cell 2001;8(2):473–9. 10.1016/s1097-2765(01)00301-x.
- [38]Rothbart S, Lin S, Britton LM, et al. Poly-acetylated chromatin signatures are preferred epitopes for site-specific histone H4 acetyl antibodies. Sci Rep 2012;2: 489. 10.1038/srep00489.