Unexpected lack of specificity of a rabbit polyclonal TAP-L (ABCB9) antibody

In this article, we describe the surprising non-specific reactivity in immunoblots of a rabbit polyclonal antibody (ref. Abcam 86222) expected to recognize the transporter associated with antigen processing like (TAP-L, ABCB9) protein. Although this antibody, according to company documentation, recognizes a band with the expected molecular weight of 84 kDa in HeLa, 293T and mouse NIH3T3 whole-cell lysates, we found that this band is also present in immunoblots of TAP-L deficient bone marrow-derived dendritic cell (BMDC) whole-cell lysates in three independent replicates. We performed extensive verification by multiple PCR tests to confirm the complete absence of the ABCB9 gene in our TAP-L deficient mice. We conclude that the antibody tested cross-reacts with an unidentified protein present in TAP-L knockout cells, which coincidentally runs at the same molecular weight as TAP-L. These findings underline the pitfalls of antibody specificity testing in the absence of cells lacking expression of the target protein.

Introduction TAP-L (TAP-Like), also known as ABCB9, is an ATP-dependent membrane half-transporter. It belongs, like TAP, the transporter associated with antigen processing, to the ABC transporter family, the members of which transport various molecules across membranes. TAP-L can form homodimers and is located primarily in lysosomes, presumably importing peptides from the cytosol. TAP-L has broad specificity for peptides ranging from a length of 6 to 59 amino acids, with an optimal activity for peptides of 23 residues (Wolters et al., 2005). TAP-L can transport two peptides at a time (Herget et al., 2009). Considering its similarity to the heterodimeric TAP transporters (ABCB2/3) importing MHC class I peptide ligands into the endoplasmic reticulum, TAP-L is a potential candidate involved in antigen presentation by MHC molecules (Bangert et al., 2011). Indeed, the length of the peptides transported by TAP-L (6-59 residues) is compatible with loading of both MHC class I and class II molecules. Moreover, TAP-L is highly expressed in lysosomes of professional antigen presenting cell (APC) lysosomes, and upregulated during differentiation of dendritic cells. However, such a function remains hypothetical, and the biological role of TAP-L is presently unknown.
In this article, we describe experimentation designed to specifically detect the ABCB9 protein in bone marrow-derived dendritic cells (BMDCs) by immunoblot. We purchased a rabbit polyclonal antibody generated by Abcam Company using a synthetic peptide as the immunogen, corresponding to a region between residues 475 and 525 of human ABCB9. This antibody is expected to recognize mouse and human ABCB9 and recommended for immunohistochemistry (IHC), immunoprecipitation (IP) and western blot (WB).

Materials and methods
Mice C57/BL6 TAP-L KO/WT heterozygous mice (ABCB9 tm1 (KOMP) Vlcg ) were purchased from The Komp Repository at the University of California at Davis, CA 95616 (see the results section for details). Heterozygous mice were bred in our laboratory and inter-crossed to obtain homozygous knock out (KO) mice (TAP-L KO/KO) along with their C57/BL6 wild type (WT) littermates.

Sample preparation
On day 6 of culture, WT and TAP-L KO BMDCs (Table 1) were lysed in a buffer containing 20mM Tris-HCl pH 7.4, 150mM NaCl, 5mM MgCl 2 , 1% NP40 and protease inhibitors (protease inhibitor cocktail, Roche) for 1 h at 4°C. Protein concentration was determined by Lowry's method, a biochemical assay for determining the total level of protein in a solution, using DC Protein Assay Reagents Package™ (BioRad).

Electrophoresis and western blot (WB)
Reagents are listed in Table 2 and Table 3 and the WB protocol is given in Table 4. The samples were loaded on a 10% acrylamide gel for electrophoresis at 80V. Separated proteins were transferred onto polyvinylidine fluoride (PVDF) membrane (pore size 0.4μm) for 1 h at 75V. The membrane was blocked with 5% BSA (Bovine Serum Albumin) in Tris-Buffered Saline (50mM Tris, 150mM NaCl) containing 0.5% Tween 20 (TBS-T) for 1h at room temperature, then incubated with the polyclonal rabbit ABCB9 antibody (Abcam, Catalog number 86222, Lot number: GR22408-1) diluted 1/2000 in TBS-T with 5% BSA for 1h at room temperature. The membrane was washed four times for 5 min with TBS-T then incubated with a goat polyclonal anti-Rabbit-HRP (Jackson ImmunoResearch Laboratory; Suffolk, UK) secondary antibody diluted 1/5000 in TBS-Tween 5% BSA for 1h at room temperature. An enhanced chemiluminescence (ECL) detection system, Immobilon Western HRP (Millipore, Guyancourt, France) was used for developing the membranes. Images were taken with a CCD camera (Fujifilm, Tokyo, Japan). Three independent experiments were performed.

Results
Seeking to detect the ABCB9 protein, we performed a series of WBs on whole-cell lysates obtained from BMDCs, thought to correspond to an inflammatory subtype of DCs. It has previously been shown that ABCB9 expression by monocyte-derived human DCs is increased under inflammatory conditions (Demirel et al., 2007). To validate specificity of antibody staining, we included TAP-L deficient BMDCs as a negative control. TAP-L KO/WT heterozygous mice (ABCB9 tm1 (KOMP) Vlcg ), in which the region located between nucleotides 5625 and 33216 of the TAP-L gene has been     Figure 2A). Heterozygous mice were bred in our laboratory and inter-crossed to obtain homozygous KO mice (TAP-L KO/KO). To our surprise, the ABCB9 antibody recognized a band, with an apparent molecular weight (84kDa) corresponding to that of ABCB9 protein, both in WT and TAP-L deficient BMDCs (Figure 1). Three different immunoblots were performed in three independent experiments.
Given these surprising results, we verified that the TAP-L KO mice were truly deficient for the target gene. We performed a series of polymerase chain reactions (PCRs). Different fragments of the WT allele (located in exons 2, 4, 8, 11 and introns 5, 6, 9) and the expected genomic region in KO mice (located between the upstream or downstream arm and within the Neo cassette) were amplified by PCR.
The following forward (F) and reverse (R) primers were used: PCRs were performed in a total volume of 25μl containing: 17.3μl H2O (DEPC treated water, pathogen free, DNase/RNase Figure 1. WB anti-ABCB9 on total cell lysates from WT and TAP-L KO BMDCs. 20-200μg of total BMDC cell lysate from WT and TAP-L KO BMDCs was loaded on 10% acrylamide gels. The proteins were transferred onto a PVDF membrane. The rabbit ABCB9 antibody was used to detect the TAP-L protein (84kDa), followed by incubation with an HRP-conjugated goat anti-rabbit secondary antibody. An ECL detection system was used for developing the membranes by chemoluminescence. Three immunoblots from three independent experiments are shown.
Free-Invitrogen), 5 μl 5x GoTaq Green Reaction Buffer (Promega), 0.5μl dNTP (10mM), 20 μM primers, 0.2μl polymerase (5 U/μl) (GoTaq-Promega polymerase) and 1μl DNA or water (negative control). The amplification reaction was performed as follows: for the WT allele, an initial denaturation at 94°C for 5 min, 10 cycles: final elongation 72°C for 5 min. The PCR products obtained were migrated on a 1.5% agarose gel containing 10 μg/ml of Ethidium Bromide. Migration was performed in a buffer tank filled with TAE buffer containing 40mM Tris, 20mM acetic acid, 1mM EDTA, pH=8 for 20 min at 120 V and visualization of the PCR products under a UV lamp connected to a photographic device.
The resulting PCR products from multiple KO mice confirmed the absence of the TAP-L gene and the presence of the Neo cassette ( Figure 2B), indicating that the TAP-L gene was deleted as expected and that the mice obtained were TAP-L KO/KO. Consequently, the band recognized by the ABCB9 antibody, even though A B running at the expected molecular weight, could not correspond to the TAP-L protein.

Conclusion
Collectively, these results show that the commercial ABCB9 antibody recognizes a protein with a molecular weight similar to that of TAP-L. It is impossible to know whether it also recognizes TAP-L. Our findings highlight the importance of verifying commercial antibody specificity using knockout cells. If such cells are not available, lentiviruses encoding target-specific shRNA, which are now readily available for an essentially complete range of proteins, can be used to produce cells that provide informative negative controls. In this article, the authors report the results of studies in which they attempt to validate the specificity of a commercially available antibody to TAP-Like (TAP-L), which is an endosomal peptide transporter from the same family at the prototypical ER TAP proteins. They report that, unexpectedly, the anti-TAP-L antibody exhibits immunoreactivity on samples prepared from TAP-L knock mice.

Suggestions for improvement:
It would be helpful to know the number of cell equivalents loaded on each lane of the gels. The gels were loaded for equal protein, so the number of cell equivalents is likely similar, but this point should be addressed. Figure 1 -Since the purpose of the experiment is to test the specificity of the primary western blot antibody (i.e., anti-TAP-L), it would be appropriate to include a control blot that was probed with secondary antibody only. It is possible that the unexpected reactivity on the western blot is due to the secondary antibody (and that the primary antibody is generating no signal). A blot probed with secondary antibody only would address this possibility. The company makes specific note of the peptide used to generate the reagent under analysis. Therefore, it would be interesting of the authors compared this sequence to the protein sequence in the available databases to see if they could identify candidates for the non-TAP-L protein being recognized by the antibody under analysis.
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
No competing interests were disclosed. Competing Interests: