Improved Elution Conditions for Native Co-Immunoprecipitation

Background Native immunoprecipitation followed by protein A-mediated recovery of the immuno-complex is a powerful tool to study protein-protein interactions. A limitation of this technique is the concomitant recovery of large amounts of immunoglobulin, which interferes with down-stream applications such as mass spectrometric analysis and Western blotting. Here we report a detergent-based “soft” elution protocol that allows effective recovery of immunoprecipitated antigen and binding partners, yet avoids elution of the bulk of the immunoglobulin. Methodology/Principal Findings We assessed the performance of the soft elution protocol using immunoprecipitation of Adaptor protein complex 1 (AP-1) and associated proteins as a test case. Relative to conventional elution conditions, the novel protocol substantially improved the sensitivity of mass spectrometric identification of immunoprecipitated proteins from unfractionated solution digests. Averaging over three independent experiments, Mascot scores of identified AP-1 binding partners were increased by 39%. Conversely, the estimated amount of recovered immunoglobulin was reduced by 44%. We tested the protocol with five further antibodies derived from rabbit, mouse and goat. In each case we observed a significant reduction of co-eluting immunoglobulin. Conclusions/Significance The soft elution protocol presented here shows superior performance compared to standard elution conditions for subsequent protein identification by mass spectrometry from solution digests. The method was developed for rabbit polyclonal antibodies, but also performed well with the tested goat and mouse antibodies. Hence we expect the soft elution protocol to be widely applicable.


Introduction
Native immunoprecipitation (IP) is a widely-used tool in the study of protein-protein interactions. Standard protocols follow a regime of solubilizing cells in mild detergents, incubation with an antibody cross-reacting with the protein of interest, and recovery of the antibody-protein complex with protein A conjugated to an inert matrix, such as sepharose beads. The immunoprecipitated protein and any potential binding partners are eluted by boiling the matrix in SDS-containing buffer, which disrupts the protein Aimmunoglobulin (Ig) interaction [1].
An inherent complication of this method is the co-elution of large amounts of Ig, which usually accounts for the majority of the recovered material. The Ig can cause problems for down-stream applications such as SDS-PAGE, where it may mask whole regions of the gel, or cause high background on Western blots. Furthermore, it can interfere with mass-spectrometric identification of low-abundance co-precipitated proteins, especially when the immunoprecipitated material is analyzed directly from unfractionated solution digests.
Co-elution of Ig can be avoided by covalent coupling of antibodies to the matrix, but this requires individual optimization of the coupling conditions to prevent loss of antibody-antigen binding, and is hence time-consuming and expensive. Here we report an alternative approach based on a ''soft'' elution protocol that allows recovery of the immunoprecipitated material, whilst leaving a substantial proportion of the Ig bound to the protein A matrix.

Results and Discussion
Developing improved elution conditions SDS is a potent anionic detergent that very effectively disrupts protein-protein interactions [2]. We observed that at a concentration of 0.2% SDS (50 mM Tris, pH = 8.0) incubation at 25uC is already sufficient to elute rabbit polyclonal Ig from protein A sepharose beads (data not shown). To reduce the potency of the SDS, we titrated in small amounts of the nonionic detergent Tween-20. We established that at 25uC, a mixture of 0.2% SDS and 0.1% Tween-20 leaves the protein A-Ig interaction mostly intact, but allows effective elution of immunoprecipitated antigens as well as antigen-associated proteins.
Based on these findings, we recommend the following elution conditions: The elution protocol assumes that a native immunoprecipitation in PBS-T (phosphate buffered saline + 1% Triton X-100) has been performed, using protein A sepharose beads to recover the immuno-complex (see File S1 for the complete protocol). A diagrammatic overview of the procedure is shown in Figure 1.
The AP-1c-subunit was immunoprecipitated from detergent lysates of HeLa cells using a rabbit polyclonal antibody [7]. Protein A sepharose beads with bound immunoprecipitated material were split into two equal aliquots, and eluted in parallel under standard or improved conditions ( Figure 1). Recovered protein was analyzed by SDS-PAGE ( Figure 2). Soft elution allowed highly effective recovery of the antigen and associated proteins, whilst it substantially reduced the amount of co-eluting Ig compared to standard elution conditions.
To determine whether the reduction of Ig facilitates the identification of co-immunoprecipitated proteins, we again eluted AP-1c IPs under standard or improved conditions in parallel. Eluates were subjected to tryptic solution digest, and analyzed by mass spectrometry. In three independent repeats, we identified ten AP-1 constituent and associated proteins with both elution methods. However, the AP-1s1c subunit was detected in only one experiment under standard elution conditions, whereas it was identified in all three repeats using soft elution (Table 1). Hence, soft elution allowed the consistent identification of a minor coprecipitant that was largely missed under standard elution conditions.
Next we investigated if the quality of mass spectrometric protein identification also benefited from the soft elution protocol. We compared Mascot scores, which reflect the confidence of protein identification [8], of the seven most abundant co-precipitants ( Table 2). On average, scores were almost 40% higher in softeluted samples, demonstrating that the soft elution protocol significantly improved down-stream mass spectrometric analysis.
Although mass spectrometry is not intrinsically quantitative, the number of identified peptides can be converted into ''emPAI'' values (exponentially modified protein abundance index; [9]), an approximate measure of absolute protein abundance. We used emPAI values calculated by Mascot to estimate the amount of Ig present in standard and soft eluted samples. For each sample, we summed all Ig-related emPAI values to obtain a measure of total Ig present ( Figure 3). Averaging across the three experiments, we observed a 44% reduction in Ig through soft elution, consistent with our SDS-PAGE analysis ( Figure 2).
Next we investigated if the soft elution protocol also works for other rabbit polyclonal antibodies. We performed immunoprecipitations with antibodies against AP-2a [10], CVAK104 [11], and GGA1 [12], and as before eluted in parallel under standard or improved conditions (Figure 4). In all three cases, soft elution resulted in the effective recovery of co-precipitating proteins, as well as in a substantial reduction of co-eluting Ig.
Antibodies derived from different species vary in their affinity for protein A [13], and this may affect the performance of the soft elution protocol. To test this, we performed immunoprecipitations with a mouse monoclonal antibody against AP-1c (mAb100/3, an IgG 2a ), and a goat polyclonal antibody against CALM (clathrin assembly lymphoid myeloid leukemia protein [14]). While protein A has relatively high affinity for mouse IgG 2a , it binds only weakly to polyclonal goat Ig [13]. Hence, we used protein G sepharose to recover CALM immuno-complexes. The results are shown in Figure 5. In the case of the mouse antibody, soft elution allowed some reduction of recovered Ig; the relative difference to standard elution was however not as pronounced as that observed for rabbit antibodies (Figures 2 and  4). Remarkably, for the goat polyclonal antibody soft elution allowed Figure 2. SDS-PAGE analysis of standard and improved elution protocols. Native immunoprecipitation of the AP-1c subunit from HeLa cell lysates was performed as described in File S1. Immunocomplexes were recovered by addition of protein A sepharose beads. Prior to elution, beads were split into two equal aliquots. One was subjected to typical high SDS/high heat elution conditions (''standard elution'', lane 1), the other to our ''soft'' elution protocol (lane 2). Following soft elution, beads were re-eluted using the standard protocol (lane 3), to recover protein still bound to the beads. Thus, lane 3 shows the proportion of immunoglobulin (IgG) that is avoided through soft elution. Proteins in selected bands were identified by mass spectrometry (small arrows). For each band only the top scoring hit is shown. AP-1b, AP-1c and AP-1m are components of the AP-1 complex; p200 and c-Synergin are known AP-1 associated proteins. IgG: rabbit Ig gamma chain C region. Approximate molecular weights are indicated (MW, in kD). Gels were stained with Coomassie G-250. doi:10.1371/journal.pone.0018218.g002 almost complete avoidance of Ig co-elution; nearly all the Ig was left on the protein G beads, while the antigen was effectively recovered. These data suggest that the soft elution protocol is suitable for antibodies derived from species other then rabbit.

Conclusion and Perspective
The IP soft elution protocol presented here shows superior performance compared to standard elution conditions for subsequent protein identification by mass spectrometry from solution digests. We have formally demonstrated this for AP-1 IPs ( Figures 2&3,  Tables 1&2). Since we used a polyclonal antibody (ie a diverse mixture of antibodies with a range of affinities), our method will be widely applicable. Indeed, we have tested the protocol with further rabbit polyclonal antibodies (Figure 4), and in all cases observed a similar reduction of Ig relative to standard elution conditions as seen in Figure 2. Furthermore, soft elution appears to be suitable for use with mouse IgG 2a monoclonal antibodies, and with a minor modification (using protein G instead of protein A), it shows excellent performance with goat polyclonal antibodies ( Figure 5). Hence we expect our method to work for most rabbit polyclonal antibodies, as well as other antibodies with high affinity for protein A or protein G.

Materials and Methods
A detailed version of the IP-soft elution protocol as well as a full description of the mass spectrometric analyses performed in this study can be found in File S1. All LC-MSMS data presented here were generated using a nanoACQUITY LC (Waters) coupled to an LTQ OrbiTrap XL mass spectrometer (Thermo). This table shows AP-1 constituents and associated proteins identified in three independent AP1-c immunoprecipitations. In each experiment, samples (ie Protein A sepharose beads with bound antigen) were split into two equal aliquots prior to elution. Aliquots were either subjected to the soft elution protocol (''Soft'') reported in this study, or to standard elution conditions (''Std''). Eluates were subjected to tryptic solution digest without further fractionation, and analyzed by LC-MSMS. Mascot scores reflecting the confidence of identification [8], and the number of peptides identified for each protein are indicated. Please note that the AP-1s1c subunit was not identified in experiments 1 and 2 using standard elution conditions. Experiments 1 and 2 were performed in MES-D buffer; experiment 3 was performed in PBS-T buffer (see File S1 for details). doi:10.1371/journal.pone.0018218.t001 For a meaningful comparison only the seven most abundant co-precipitants of AP-1c were included (see Table 1). Ratios are the averages of three independent experiments. doi:10.1371/journal.pone.0018218.t002
Gel analysis SDS-PAGE was performed according to a standard protocol [13]. Gels were stained with Coomassie G-250 SimplyBlue SafeStain (Invitrogen, #LC6060), and scanned with an Odyssey Infrared Imager (LI-COR Biosciences). Scans were exported as TIFFs. Adjustment of brightness and contrast as well as despeckling was performed in Adobe Photoshop.

Application notes
The soft elution protocol has been thoroughly tested with rabbit, mouse (IgG 2A ), and goat antibodies. It is likely that it will perform well with antibodies derived from other species, provided they interact strongly with protein A or protein G. When using the soft elution protocol with an untested class of antibodies, it is recommended to select the protein with the highest predicted affinity (see for example http://www.piercenet.com/files/ TR0034-Ab-binding-proteins.pdf).
The IP soft elution protocol was primarily designed to facilitate the mass spectrometric identification of proteins co-precipitating with the primary antigen. We have tested the protocol with a wide range of antibodies, and in most cases observed a small reduction in the amount of recovered primary antigen (eg Figure 2, faint AP-1c band in lane 3). Our data suggest that this reduction is not detrimental for the purpose of protein identification (Table 1 average Mascot scores for AP-1c are still higher in soft eluted samples). Nevertheless, in cases where complete recovery of the primary antigen is desired, standard elution protocols may be more suitable. For pilot experiments, soft and standard elutions can be performed sequentially (as shown in lanes 2 and 3 of Figure 2), to ensure recovery of all immunoprecipitated material.

Supporting Information
File S1 Complete protocol for native immunoprecipitation and soft elution. (DOC)