Anal. Chem. 2005, 77, 6818-6827
Multidimensional High-Throughput Screening of Displacers Kaushal Rege,† Asif Ladiwala,† and Steven M. Cramer*
Department of Chemical and Biological Engineering, Rensselaer Polytechnic Institute, 110 8th Street, Troy, New York 12180
A multidimensional, batch high-throughput screening (MD-HTS) protocol was developed to investigate the effects of various parameters on the selectivity of ionexchange protein displacement systems. A variety of molecules were screened, and the results were employed to provided insights into the influence of displacer chemistry and concentration, resin chemistry, and mobilephase salt counterion on the efficacy and selectivity of these nonlinear chromatographic systems. These results open up the possibility of tailoring the selectivity of displacement separations by choosing appropriate combinations of operating conditions using the MD-HTS technique. The screens were also employed for the identification of displacers and conditions for the separation of a challenging protein mixture by selective displacement chromatography. Column displacements were carried out with potential lead compounds identified from the MD-HTS screens, and the results confirmed that selective displacement could indeed be achieved for this model mixture. Furthermore, the results indicated that this approach is particularly useful when the order of elution is not changed, but the inherent selectivity is increased in the presence of the displacer. The results presented in this paper demonstrate the utility of the MDHTS technique for rapid method development in protein ion-exchange displacement chromatography. Ion-exchange displacement chromatography has attracted significant attention as a powerful technique for the purification of biomolecules in biotherapeutic downstream processes.1-7 In particular, low molecular weight (30%) in the presence of calcium. These results are consistent with observations on FF Sepharose SP and indicate that the presence of calcium resulted in enhanced efficacies of displacers on both resin systems. Similar trends were observed with 10 mM input displacer concentration (data not shown).
Figure 7. (a) Selective displacement of B-CytC on Source 15S using 47 mM 1-(2-aminoethyl)piperidine. Column, 100 mm × 4.6 mm i.d. Source 15S; carrier, 50 mM sodium acetate pH 4.8; protein, 18 mg of H-CytC and B-CytC (1:1 w/w); flow rate, 0.2 mL/min. (b) Displacement of H-CytC and B-CytC on HP Sepharose SP using 90 mM N-methyl-1,3-propanediamine. Column, 100 mm × 5 mm i.d. HP Sepharose SP; carrier, 50 mM sodium acetate pH 4.8; protein, 18 mg of H-CytC and B-CytC (1:1 w/w); flow rate, 0.1 mL/min.
Validation of Screening Results. As seen from the above discussion, the MD-HTS approach can indeed provide valuable insights into the factors influencing the efficacy and selectivity of displacement separations. A more practical application of these screens, however, is for the identification of the appropriate conditions for the resolution of biological mixtures. For the model proteins examined in this study, the separation of the two CytCs by displacement is clearly the more challenging separation. Accordingly, this mixture was employed to validate the utility of this screening technique. The MD-HTS data were analyzed to identify conditions where the P values for the two CytCs differed by more than 10% in order to establish potential scenarios for the selective displacement of one of the CytCs. The selective displacer leads identified on different stationaryphase materials under different conditions are summarized in Table 3. As seen in the table, there were conditions that resulted in higher displacement of either H-CytC or B-CytC. Column displacement experiments were carried out with 1-(2-aminoethyl)piperidine, which was found to be selective toward B-CytC on the Source 15S material, and N-methyl-1,3-propanediamine, which showed a selectivity toward displacing H-CytC on HP Sepharose SP. Initial column experiments carried out at the same displacer Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
6825
Table 3. Summary of Potential Lead Displacer Candidates for the Selective Displacement of B-CytC and H-CytC conditions
displacer
% H-CytC
27.52 40.29
38.83 60.35
FF Sepharose SP [displacer]: 20 mM counterion: Na+
5-amino-1,3,3-trimethylcyclohexane N-1,3-propanediamine
FF Sepharose SP [displacer]: 10 mM counterion: Na+
(none)
Source 15S [displacer]: 20 mM counterion: Na+
piperazine 5-amino-1,3,3-trimethylcyclohexane 1-(2aminoethyl)piperadine
55.23 66.87 69.42
44.31 54.65 56.09
Source 15S [displacer]: 10 mM counterion: Na+
2,2-dimethyl-1,3-propanediamine
40.36
29.67
FF Sepharose SP [displacer]: 20 mM counterion: Ca2+
L-lysine methyl ester 2,2-dimethyl-1,3-propanediamine 1,2-diaminocyclohexane
43.53 61.65 73.71
33.18 47.81 60.44
FF Sepharose SP [displacer]: 10 mM counterion: Ca2+
N-(2-aminoethyl)-1,3-propanediamine
67.90
94.52
Source 15S [displacer]: 20 mM counterion: Ca2+
(none)
Source 15S [displacer]: 10 mM counterion: Ca2+
benzylamine N-methyl-1,3-propanediamine
31.98 81.75
43.53 68.20
concentrations as the batch screens (i.e., 20 mM) were not successful in displacing either protein. However, when the column experiments were repeated at higher displacer concentrations, the proteins were displaced to varying degrees. The chromatogram for the displacement of a 1:1 (w/w) B-CytC/H-CytC mixture on a Source 15S column using 1-(2aminoethyl)piperidine as the displacer is shown in Figure 7a. As seen in the figure, by optimizing the concentration of 1-(2aminoethyl)piperidine, we were indeed able to obtain the selective displacement of B-CytC while H-CytC was eluted after the breakthrough of the displacer. Previous results from our laboratory have demonstrated that “transient selective displacement” of the CytC mixture can be established (for example, ref 26). However, those displacements were not stable and occurred at a specific feed load under multicomponent conditions. Once those conditions changed, the proteins were no longer displaced. In contrast, displacement with 47 mM 1-(2-aminoethyl)piperidine was found to be stable. This was validated by carrying out the displacement using a single-component feed (B-CytC) as well as under varying feed load conditions (results not shown). Figure 7 shows the chromatogram obtained for the displacement of a 1:1 (w/w) B-CytC/H-CytC mixture on HP Sepharose SP using N-methyl-1,3-propanediamine as the displacer (note: these column experiments were carried out with HP Sepharose SP, which has ligand, spacer, and backbone chemistry very similar to that of the FF Sepharose SP material). As described above, this displacer showed selectivity toward displacing H-CytC on the Sepharose material under batch conditions. In contrast to the expected selective displacement of the H-CytC, both proteins were displaced by N-methyl-1,3-propanediamine. Despite several efforts (26) Kundu, A.; Barnthouse, K. A.; Cramer, S. M. Biotechnol. Bioeng. 1997, 56, 119-129.
6826
% B-CytC
Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
to optimize the separation, no resolution was obtained between the two proteins, although a higher concentration of the H-CytC was consistently displaced. To understand the observed results, it is important to note that, under both linear gradient and multicomponent frontal conditions, B-CytC always eluted prior to H-CytC. Accordingly, during the loading of the column, B-CytC will move further down the column than H-CytC due to multicomponent frontal chromatography. Upon its introduction into the column, the displacer first encounters the H-CytC, which it easily displaces. However, even though this displacer can displace H-CytC more effectively than B-CytC, the displaced H-CytC will in turn “displace” the adsorbed B-CytC moving it further down the column, making it difficult to establish a selective displacement of H-CytC. In summary, the above results demonstrate that it is indeed possible to employ the MD-HTS technique to identify both displacers and operating conditions where one can successfully carry out selective displacement chromatography of very challenging mixtures. Furthermore, from the results of the above column displacement experiments, it is evident that this approach is particularly useful when the order of elution is not changed, but the inherent selectivity is increased in the presence of the displacer. CONCLUSIONS A multidimensional batch displacement assay was employed as a high-throughput screen for the investigation of a variety of operating parameters in displacement chromatography. The screening results indicated that the displacer chemistry and concentration, stationary-phase resin chemistry, and nature of the salt counterion can have a major influence on the efficacy of displacement separations. These results open up the possibility of tailoring the selectivity of displacement separations by choosing
the appropriate combinations of operating conditions using the MD-HTS technique. In addition to examining the influence of various parameters on the selectivity of displacement separations, the screens were also employed for the identification of displacers and conditions for the separation of model protein mixtures. Column displacements were carried out with potential selective displacers identified from the MD-HTS screens for a challenging cytochrome c model separation. The results indicated that selective displacement of these two very similar proteins could indeed be achieved. Furthermore, the results indicated that this approach is particularly useful when the order of elution is not changed, but the inherent selectivity is increased in the presence of the displacer. The MD-HTS protocol enables the rapid evaluation of a large number of possible displacer candidates under a wide variety of conditions. This approach can be readily scaled up to include more operating parameters and wider ranges of operating conditions (e.g., other factors such as pH, salt concentration, mobile-phase modifiers). While model proteins were employed to demonstrate the utility of MD-HTS as a platform technology for displacement chromatography process development, future work will examine
a variety of high-throughput analytical techniques as well as appropriately labeled bioproducts and impurities to enable multicomponent MD-HTS screening of real bioprocessing mixtures. For such applications, MD-HTS will be employed to identify conditions that can produce selectivity between a bioproduct of interest and its key impurities. Finally, the implementation of this approach in a robotic system would allow for automated displacement methods development. ACKNOWLEDGMENT The authors acknowledge NSF Grant BES-0079436 and NIH Grant GM047372-07 for funding this research. SUPPORTING INFORMATION AVAILABLE Complete MD-HTS screening data. This material is available free of charge via the Internet at http://pubs.acs.org.
Received for review February 19, 2005. Accepted July 24, 2005. AC050314A
Analytical Chemistry, Vol. 77, No. 21, November 1, 2005
6827
