Response: 'Antibody crossreactivity between the tumour suppressor PHLPP1 and the proto-oncogene β-catenin'

EMBO Reports (2013) 14 1, 10–11 doi:10.1038/embor.2012.188

Lobert and co-workers [1] reported a consequential flaw in the immunoreactivity of the antibodies A300-660A and IHC-00382 to PH domain leucine-rich repeat protein phosphatase 1 (PHLPP1) manufactured by Bethyl Laboratories, Inc. (Montgomery, Texas, USA). The flaw is the crossreactivity of these preparations with β-catenin. This response is to acknowledge the validity of the report of Lobert and co-workers, to report that testing at Bethyl Laboratories has confirmed their findings and to report production of antibodies to the carboxyl-terminus of PHLPP1 with less than 3% residual crossreactivity with β-catenin when compared with crossreactivity of A300-660A to β-catenin.

The antibodies A300-660A and IHC-00382 had a common means of production in which a peptide (LPDYYDTPL) representing the C-terminus of PHLPP1 was conjugated to a carrier protein and used as an antigen for the production of hyperimmune sera in rabbits. Subsequently, antibodies to the peptide were purified by using the peptide conjugated to beads of agarose. The affinity-purified antibodies were tested in immunoblotting and immunoprecipitation. Having found that the preparation of antibodies was useful for immunoblotting and immunoprecipitation, the preparation was offered as part of Bethyl's line of products as product number A300-660A. Subsequent to testing for usefulness in immunohistochemical staining, a portion of A300-660A was reformulated to produce product number IHC-00382. The reformulation consisted of adjusting the concentration of antibody to that which gave positive immunohistochemical staining when IHC-00382 was used at a final working dilution within the range of 1:100–1:500.

When using A300-660A sufficiently concentrated, we observed a more rapidly migrating band (approximately 90 kDa) when immunoblotting whole-cell lysate from HeLa and 293T cells. The band in question was reported by Lobert and co-workers [1] to be β-catenin. Under our conditions for immuno-blotting, the intensity of the β-catenin band was considerably less than the intensity of the PHLPP1 bands in HeLa and 293T cells. In addition, when using the antibody sufficiently dilute, PHLPP1 could be readily detected whilst the β-catenin band was not observed. Furthermore, the β-catenin band was not readily observed when using the antibody to perform immunoblot of immunoprecipitates produced using A300-660A. Thus, we had no evidence that the more rapidly migrating band was immunoprecipitated by A300-660A.

On becoming aware of the report by Lobert and co-workers [1], we took three actions. The first was to remove from the market the antibody Bethyl IHC-00382, which had previously been offered with an explicit recommendation for use in immunohistochemistry. Second, we removed the recommendation for use of A300-660A in immunohistochemistry and stated that the use of A300-660A in immunohistochemistry was strongly discouraged. We chose to continue to offer A300-660A with recommendations for its use in immunoblotting and immunoprecipitation. Third, we initiated testing of the crossreactivity of A300-660A with β-catenin.

We confirmed crossreactivity with β-catenin using A300-660A in immunoblot of immunoprecipitates produced by using a β-catenin antibody (Bethyl product number A302-012A; Fig 1A). In addition, β-catenin was observed by immunoblotting for it (Fig 1B,C) in the immunoprecipitate produced using A300-660A. Notably, β-catenin was not detected when immunoblotting the immunoprecipitates produced by using a different antibody to PHLPP1 (A300-659A, which recognizes an upstream epitope on PHLPP1).

Figure 1.

Assessment by immunoblot and immunoprecipitation of crossreactivity of antibodies to PHLPP1 with β-catenin. Antibodies to PHLPP1 and β-catenin were used in immunoblotting of the immunoprecipitates produced by using antibodies to PHLPP1 (A300-659A, A300-660A, A304-029A), antibodies to β-catenin (A302-012A) and antibodies purified from PHLLP1 antisera by adsorption to a β-catenin peptide (BL16200). (A) An immunoblot using antibody A300-660A. (B) Short exposure of an immunoblot for β-catenin using antibody A302-012A. (C) Long exposure of an immunoblot for β-catenin using antibody A302-012A. (D) Short exposure of an immunoblot using antibody A304-029A. (E) Long exposure of an immunoblot using antibody A304-029A. PHLPP1, PH domain leucine-rich repeat protein phosphatase 1.

Whilst β-catenin was not listed among the results from Blast analysis of the PHLPP1 peptide LPDYYDTPL, alignment of the C-terminal 50 residues of human PHLPP1 (accession number O60346) with the C-terminal 50 residues of β-catenin (accession number P35222) revealed sufficient similarity to presume the C-terminus of β-catenin (LAWFDTDL) to be the epitope bound by A300-660A (Fig 2). We elected to remove antibodies responsible for crossreactivity with the C-terminus of β-catenin by adsorption to a peptide containing the epitope.

Figure 2.

Alignment of the carboxy-terminal 50 residues of PHLPP1with the corresponding portion of β-catenin. The C-terminal residues of PHLPP1 were aligned with the C-terminal residues of β-catenin. The underlined portion indicates the residues used to generate the antibody A300-660A. Given the antibody A300-660A was shown to crossreact with β-catenin, the similarity of the C-terminus of β-catenin to that of PHLPP1 was sufficient to anticipate that binding of Bethyl A300-660A to β-catenin was at the C-terminus of β-catenin. PHLPP1, PH domain leucine-rich repeat protein phosphatase 1.

The peptide SNQLAWFDTDL was synthesized and conjugated to beads of agarose. The agarose containing the peptide was transferred to a 2.5 × 20 cm column, and hyperimmune sera from rabbits immunized with the PHLPP1 peptide PDYYDTPL were passed through the column to adsorb antibodies that bound to the β-catenin peptide. Adsorption of β-catenin crossreactive antibodies in the sera was repeated multiple times with regeneration of the column between each cycle of adsorption. Subsequently, antibodies to PHLPP1 remaining in the hyperimmune sera were affinity-purified following adsorption to, and elution from, beads of agarose conjugated to the PHLPP1 peptide.

The resulting affinity-purified β-catenin crossreactive antibodies (BL16200) and PHLPP1 antibodies (A304-029A) were tested for crossreactivity with β-catenin. Whereas β-catenin was immunoprecipitated by A300-660A and BL16200, the amount of β-catenin detected in the immunoprecipitate produced by using A304-029A was negligible (Fig 1B,C). Using immunoblot, there was no detectable crossreactivity of A304-029A with β-catenin in whole-cell lysate from HeLa cells or with β-catenin immunoprecipitated from HeLa cells (Fig 1D,E). When tested for reactivity to the β-catenin peptide by indirect ELISA, reactivity with β-catenin was less than 3% of the reactivity of A300-660A with β-catenin (Table 1). If thoroughly validated in immunohistochemistry, A304-029A might prove useful in determining cellular localization of PHLPP1.

Table 1. Assessment by ELISA of crossreactivity of antibodies to PHLPP1.

Peptide	Concentration of antibody (ng/ml)						Antibody
	333.33	100.00	33.33	10.00	3.33	1.00
PHLPP1	3.18	2.90	2.60	1.52	0.58	0.19	A300-660A
β-catenin	2.58	1.34	0.56	0.23	0.06	0.03
PHLPP1	3.02	2.69	2.02	1.11	0.38	0.08	A304-029A
β-catenin	0.01	0.04	0.00	0.00	0.00	0.00

Values are the mean absorbance for duplicate wells after subtraction of background. PHLPP1, PH domain leucine-rich repeat protein phosphatase 1.