| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online March 10, 2005 as doi:10.1096/fj.04-2812fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
School of Life and Health Sciences, Aston University, Birmingham, UK
1 Correspondence: E-mail: a.v.hine{at}aston.ac.uk
SPECIFIC AIMS
Affinity purification of plasmid DNA is an attractive option for the biomanufacture of therapeutic plasmids, which are strictly controlled for levels of host protein, DNA, RNA, and endotoxin. Plasmid vectors are considered to be a safer alternative than viruses for gene therapy, but milligram quantities of DNA are required per dose. We have exploited the lac operator, present in a wide diversity of plasmids, as a target sequence for affinity capture (Fig. 1
) and have purified plasmid DNA, in good yield, free from detectable RNA and protein, and with minimal genomic DNA contamination.
|
PRINCIPAL FINDINGS
1. A LacI fusion protein is an effective affinity ligand for plasmid purification
We wanted to investigate whether the protein-DNA interaction between LacI and the native operators in pUC19 would be strong enough to allow plasmid capture. Alternatively, would a low-affinity operator on its own be sufficient or would high-affinity, synthetic operator sequences be required to permit capture of a 1.8 MDa plasmid by a 274 kDa tetrameric protein? We engineered several variants of pUC19 with combinations of native and synthetic operators.
Since we anticipated using a His6 tag for protein immobilization and GFP as a sensitive monitor of protein leaching from the resin, it was necessary to confirm the DNA binding activity of LacI protein when fused to both a His6 tag and GFP. LacI is a tetrameric protein, and we questioned whether the bulk of the GFP fusion might sterically hinder protein tetramerization or DNA binding. Binding experiments between the LacI fusion protein (NH2-LacI-His6-GFP-COOH) and a single lacOs sequence (ds oligonucleotide) were performed as described previously. This experiment yielded a Kdapp of 5.1 ± 0.8 nM (data not shown) and demonstrated that DNA binding activity is retained by LacI in the presence of these fusion domains.
We anticipate that affinity purification will be best used as an early-stage process during plasmid biomanufacture. As large-scale recombinant protein purification of the affinity ligand is unlikely to prove cost effective, we used an Escherichia coli crude lysate that contained the LacI fusion protein rather than purified protein per se as the affinity ligand. Plasmid DNA was purified conventionally by alkaline lysis/ion exchange chromatography, which would enable quantification of loads and yields during purification. Affinity capture experiments were performed with native pUC19 and plasmid variants. After complex formation between the protein and DNA, the mixture was absorbed onto an IMAC resin, which is specific for the His6 tag. The resin was washed extensively and bound plasmid was eluted with a NaCl/IPTG solution, which should induce an allosteric change in LacI, allowing release of the plasmid, while the protein ligand remains bound to the resin. Eluted plasmid was desalted and analyzed. Samples throughout the process were quantified for DNA (Table 1
) and examined by electrophoresis (Fig. 2
) and protein contamination.
|
|
As might be predicted, the affinity between LacI and the operator sequence correlates with the degree of plasmid capture: the stronger the affinity between protein and DNA, the higher the efficiency of plasmid immobilization. Thus, the greatest immobilization is mediated by dual lacO3/lacOs operators, although satisfactory capture is also obtained with native pUC19 (lacO3/lacO1) and the plasmid containing a single lacOs operator (Table 1)
. Plasmid capture is possible with a single low-affinity operator (lacO3), although retention levels on the column with this operator are poor. Conversely, plasmid elution from the column is independent of LacI/operator affinity with a consistent yield (after desalting) of 
80–90% of that originally immobilized. This finding is predictable, as DNA release is affected by an allosteric mechanism. Allolactose and its analogs induce a conformational change in the LacI structure, rendering it incapable of binding to the operator sequence. Under such circumstances, it is unlikely that the affinity between protein and DNA would affect the efficiency of DNA release.
For therapeutic plasmid DNA, a lack of contamination by RNA, genomic DNA, host protein, and endotoxin is crucial. It is notable, given the crude nature of the affinity ligand used in these experiments, that neither the RNA nor sheared genomic DNA present in the crude protein preparation (Fig. 2A-E
, lanes 1–3) is visible in the affinity-purified fractions (Fig. 2A-E
, lanes 7 and 8). For more sensitive detection of genomic DNA, we performed PCR analysis of the eluted plasmid pUC19lacO3/lacOs for the gyrA gene of E. coli. Low level contamination of the first, but not the second elution (Fig. 2B
, lanes 7, 8) was detectable by PCR (data not shown). Conversely, analysis of the purified plasmid by fluorescence spectroscopy for GFP, by silver stained SDS-PAGE, and by BCA assay (data not shown) revealed no detectable levels of protein contamination either by host protein or by leaching of the affinity adsorbent. Silver staining typically detects as little as 0.25 ng of protein, whereas a BCA assay may detect >5 µg of protein. We have determined previously that the minimal protein concentration detectable by the fluorescence assay is 0.4 nM GFP.
2. Affinity purification of plasmid DNA directly from crude lysates
Analogous to the affinity purification of proteins, we wished to be able to isolate plasmid DNA directly from a bacterial crude lysate, as an initial purification step. As described above, we had already demonstrated that LacI fusion protein from within a crude E. coli lysate would isolate plasmid that had already been purified conventionally. We therefore reasoned that this procedure should work with dual crude cell lysates (both plasmid and protein). Accordingly, a 200 mL culture of E. coli carrying the pUC19lacO3/lacOs plasmid was lysed conventionally by alkaline lysis. The solution was neutralized, its pH adjusted and combined with a second crude lysate containing the LacI-His6-GFP protein. The mixture was incubated with IMAC resin, washed, eluted, and the eluate desalted. Examination of the resulting fractions demonstrated that affinity capture of plasmid DNA, directly from a neutralized alkaline lysate is indeed feasible. Given the crude nature of both starting materials, we believe that the degree of plasmid purification achieved by this one-step procedure is remarkable.
CONCLUSIONS AND SIGNIFICANCE
We have demonstrated a high-yielding generic, affinity-based plasmid purification process. Although most efficient (84–87% recovery of immobilized plasmid) with plasmids containing two lac operator sequences, like those found in native pUC-based vectors, the process also works well (81% recovery of immobilized plasmid) with a plasmid containing a single engineered operator sequence. The process is simple: crude bacterial lysate containing a LacI-His6-GFP fusion protein is mixed with plasmid that contains lac operator(s) and the resulting protein/DNA complex is isolated using an IMAC resin, which is specific for the His6 Tag. Plasmid DNA is then eluted with a NaCl/IPTG solution, and desalted. The process is exceptionally specific for plasmid DNA, excluding host cell protein, (leaching of the affinity adsorbent is also undetectable), RNA, and the majority of genomic DNA. The process is even effective as a dual crude lysate approach in which crude cell lysate (protein) is combined with crude, neutralized alkaline lysate (plasmid) to yield apparently pure plasmid DNA, although yields are currently low.
Even with the current, nonoptimized protocol, the high recovery rates mean that it should be possible to generate milligram quantities of purified plasmid, with 100 mL of resin and 2.5 mg of tetrameric fusion protein. However, the IMAC resin used in our experiments is far from ideal. The resin is more typically used in protein purifications and therefore is of high porosity. Use of alternatives such as a nonporous micropellicular support, a macroporous monolith, or even adsorptive membranes should vastly increase the surface area available for immobilization of the protein-plasmid complex and so reduce the quantity of support required.
Alternative supports may also reduce the potential for nonspecific interactions. Although no protein contamination of eluted plasmid was detectable, elution of plasmid pUC19
lacO (albeit at levels too low to be quantified, Fig. 2E
) suggest a potential for nonspecific interaction between nucleic acids and the TALONTM resin. We therefore recommend that PCR-based analysis of eluted fractions should be performed for all potential contaminating DNAs.
Although we anticipate application of an affinity process as an early stage procedure in plasmid purification, the high degree of purification may also make it attractive as a final "polishing" step. In such an application, it would be advisable to first purify the fusion protein from the crude bacterial lysate to prevent contamination of the partially purified plasmid DNA. Likewise, the final desalting stage should remove the IPTG from the eluate, but given the toxicity of this molecule, we determined that a simple salt elution (elution buffer lacking IPTG), can also elute the plasmid DNA, although at a much reduced yield. We therefore suggest that if this process were to be used as a polishing step, elution should either be carried out in the absence of IPTG, or by replacing IPTG with allolactose, a nontoxic, but more expensive, natural alternative.
We have developed an affinity purification strategy we believe is suitable for development for use in large-scale manufacturing of therapeutic grade plasmid DNA. The process uses affinity-based procedures that are synonymous with protein purification but that have so far eluded DNA manufacture.
FOOTNOTES
To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.04-2812fje;
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
