Note: Descriptions are shown in the official language in which they were submitted.
CA 02220867 1997-11-12
WO 96/36706 PCT/US96J07083
TITLE OF THE INVENTION
A METHOD FOR LARGE SCALE PLASMID PURIF~CATION
RELATED APPLICATION
S This is a continll~tion-in-part application of U.S. Serial
Number 08/446,1 18 filed May 19, 1995.
BACKGROUND OF THE INVENTION
The classical techniques for isolating plasmid DNA from
microbial fermentations are suitable for small or laboratory scale
plasmid preparations. One such procedure involves the ~lk~line lysis of
microbial host cells cont~inin~: the plasmid, followed by acetate
neutralization causing the precipitation of host cell genomic DNA and
proteins which are then removed by, for example, centrifugation. The
liquid phase contains the plasmid DNA which is alcohol precipitated and
then subjected to isopycnic centrifugation using CsCl in the presence of
ethidium bromide. The ethi~ m bromide is required in order to
separate the total plasmid DNA into the three dirrercllt forms,
supercoiled (foln, I), nicked circle (form II), and linearized (form m),
and the desired plasmid form is collected. Further extraction with
butanol is required to remove residual ethidium bromide followed by
DNA precipitation using alcohol. Additional purification steps follow
to remove host cell proteins. The removal of host proteins is
performed by repeated extractions using phenol or a mixture of phenol
and chloroform. The pl~mi~l DNA is alcohol precipitated and residual
phenol i~ removed by reFeated isoamyllc~oroform extractions. l~e
final alcohol precipitated plasmid DNA is dissolved in water or a
suitable buffer solution.
There are numerous drawbacks and limit~ions to this
process including:
a) this process requires the use of expensive and hazardous
chemicals (CsCl and EtBr, which are used within the
density gradient centrifugation; EtBr is a known mutagen
CA 02220867 1997-11-12
W O 96/36706 PCTrUS96/07083
and must be removed from products; also it is an
interc~l~ting agent which can nick the plasmid);
b) the density centrifugation step is not easily scaleable;
c) there is a need for organic solvent extraction to remove
residual EtBr;
d) phenol extraction is used to remove residual proteins and
DNase, a process that would require a cellLlir lge to break
phenol/water emulsion;
e) highly repetitive steps making it laborious and time
consuming (isolation requires several days);
f) scalability of the chemical lysis step is an obstacle i.e.,
lysozyme/~lk~line/KOAc treatment step is efflcient in lysing
cells on a small scale, however, the increase in viscosity
makes large scale processing very difficult; and
g) use of large quantities of lysozyme to enzymatically weaken
the microbial cell wall prior to lysis.
The mixture is then neutralized by addition of acid which
results in precipitation of the high molecular weight chromosomal
2~ DNA. The high molecular weight RNA and protein-SDS complexes
precipitate with the addition of high concentration of KOAc salt. The
plasmid product rem~in~ in the clarified supernatant following
centrifugation. T imit~tions here include the need to process quickly and
on ice in order to retard the activity of nucleases which are not removed
30 until phenol extraction. The main cont~min~nt rem~ining in the
supernatant with the product is RNA.
Another commonly utilized method for isolating and
purifying plasmid DNA from bacteria provides a rapid process suitable
for only very small scale preparations.
CA 02220867 1997-11-12
W 096136706 PCT/US96J070X3
Holmes and Quigley (1981, Analytical Biochem., 114, pp
193-197) reported a simple and rapid method for preparing plasmids
where the bacteria are treated with lysozyme, then boiled at about
100~C in an appropriate buffer (STET) for 20 - 40 seconds forming an
5 insoluble clot of genomic DNA, protein and debris leaving the plasmid
in solution with RNA as the main cont~min~nt Lysozyme is apparently
a re4u~ement for this technique to work, and as such adds a treatment
step which is less desirable for large scale manufacture of DNA for
hllm~n or veterinary use. However, the addition of lysozyme may
10 enhance plasmid release during lysis. An advantage is that heat
treatment of the cells would also denature the DNase. However, this
technique is not suitable for scale up to a high volume of microbial
fermentations and is meant for fermentations less than five liters.
Alternatives to isopycnic centrifugation using CsCl for
i5 plasmid purifica~on have ~een p~lished. ~hese alternatives are
suitable only for laboratory scale plasmid isolation and include:
a) size exclusion chromatography, which is inherently limited
in throughput,
b) hydroxyapatite chromatography, which has the
disadvantage of requiring high concentrations of urea for
efficiency,
c) reversed phase chromatography; and
d) ion exchange chromatography.
Large scale isolation and purification of plasmid DNA
30 from large volume microbial fermentations, therefore, requires the
~ development of an improved plasmid preparation process. An isolation
and purification process for large scale plasmid DNA production is
necessitated by recent developments in many areas of molecular
biology. In particular, recent advances in the field of polynucleotide-
CA 02220867 1997-11-12
W O 96/36706 PCTrUS96/07083
based vaccines for hllm~n use, and potentially hllm~n gene therapy,
requires the ability to produce large quantities of the polynucleotide
vaccine in a highly purified form.
Unprecedented technology is required for developing/-
S implementing a large scale commercially viable process forfermentation, isolation, purification and characterization of DNA as a
bioph~ ceutical.
SUMMARY OF THE INVENTION
The current laboratory method used to isolate and purify
plasmid DNA consists of a series of classical laboratory techniques that
are not suitable for a manufacturing process. For example, density
gradient centrifugations are not scaleable; the purification procedure
necessitates the use of hazardous and expensive chemicals/solvents such
15 as ethydium bromide, a known mutagen, and is labor intensive and time
consuming. Therefore, a scaleable alternative process was developed,
and is disclosed herein. In addition, an HPLC assay was established to
track the plasmid product through the process steps and to distinguish
between the plasmid forms. The microbial cells harboring the plasmid
20 are suspended and optionally incubated with lysozyme in a buffer
cont~inin~ detergent, heated using a flow-through heat exchanger to lyse
the cells, followed by centrifugation. After centrifugation the clari~ied
lysate, which contains predomin~tely RNA and the plasmid product, is
filtered through a 0.45 micron filter and then diafiltered, prior to
25 loading on the anion exchange column. The plasmid product may
optionally be treated with RNase before or after filtration, or at an
earlier or later step. The anion exchange product fraction cont~inin~
the plasmid is loaded onto the reversed phase column, and is eluted with
an appropriate buffer, providing highly pure plasmid DNA suitable for
30 hllm~n use.
BRIEF DESCRIPTION OF THE DRAVVINGS
Figure 1. A schematic of a suitable heat exchanger apparatus is
shown.
CA 02220867 1997-11-12
WO 96/36706 PCT/U~"/D7~B3
Figure 2. The relationship between the outlet temperature and the
flow rate is shown, graphically.
Figure 3. Comparative chromatograms of total plasmid in clarified
supernatant with 50 mMEDTA and l00 mMEDTA are
shown.
Figure 4. The yield of supercoiled pl~mi~l as a function of
exit temperature is shown.
Figure 5. The elution profiles of anion exchange columns run with
RNase treated (bold line) and untreated (thin line) clarified
lysates are shown.
15 Figure 6. The elution profiles of anion exchange chromatography
with clarified lysate that was diafiltered before the column
or not ~ filtered before the column are shown.
Figure 7. An elution profile of plasmid DNA from cell lysate is
shown.
Figure 8. An agarose gel electrophoresis analysis of the DNA product
obtained at various intermediate steps of purification is
shown.
Figure 9. A tracing of the anion exchange HPLC analysis of the DNA
product demonstrating the purity of the product is shown.
DETAILED DESCRIPTION OF THE LNVENTION
We have identi~led a novel, scaleable, alternative
0 lysis/debris removal process for large scale plasmid isolation and
purification which exploits a rapid heating method to induce cell lysis
and precipitate genomic DNA, proteins and other debris while keeping
the plasmid in solution. The utility of this process is the large scale
CA 02220867 1997-11-12
W 096/36706 PCTAUS~ 70a3
isolation and purification of plasmid DNA. We have found that
suspending the microbial cells in modified STET buffer (described
below) and then heating the suspension to about 70-100~C in a flow-
through heat exchanger results in excellent lysis. Continuous flow or
5 batch-wise centrifugation of the lysate effects a pellet that contains the
cell debris, protein and most of the genomic DNA while the plasmid
remains in the supern~t~nt This invention offers a number of
advantages including higher product recovery than by chemical lyses,
inactivation of DNases, operational simplicity and scaleability.
The present invention is drawn to a process for the large
scale isolation and purification of plasmid DNA from microbial
fermentations. Large scale microbial cell fermentations as used herein
are considered to be total cell fermentation volumes of greater than
about 5 liters, or the cells harvested from a fermentation volume
15 greater than about 5 liters.
The present invention is also drawn to providing pl~mi~l
DNA in a highly purified form suitable for human use. DNA for
hllm~n use includes, but is not limited to, polynucleotide vaccines and
DNA for human gene therapy. Polynucleotide vaccines are intended for
20 direct injections into hllm~n~ [Montgomery, D.L. et al., 1993, Cell
Biol., 169, pp. 244-247; Ulmer, J.B. et al., 1993, Science, 259, pp.
1745-1749] .
The present invention is also drawn to an in-line
monitoring process for the tracking of the various forms of plasmid
25 DNA through the isolation and purification steps. The various forms of
plasmid DNA referred to above which can be individually isolated by
the process of the present invention are form I (supercoiled plasmid),
form II (nicked or relaxed plasmid), and form III (linearized plasmid).
The process of the present invention is suitable for use with
30 microbial fermentations in general. It is readily apparent to those
skilled in the art that a wide variety of microbial cells are suitable for
use in the process of the present invention, including but not limited to,
fungal cells including yeast, and bacterial cells. A ~lefell~d microbial
fermentation is a bacterial fermentation of cells cont~ining the plasmid
CA 02220867 1997-11-12
WO ~6/36706 PCT~US96/07083
to be isolated and purified. A preferred bacterial fermentation is a
fermentation of E. coli cont~inin~? the plasmid to be isolated and
purified. It is readily apparent to those skilled in the art that bacterial
fermentations other than E. coli fermentations are suitable for use in the
5 present invention. The microbial ferrnentation may be grown in any
liquid medium which is suitable for growth of the bacteria being
utilized.
The plasmid to be isolated and purified by the process of
the present invention can be any extrachromosomal DNA molecule.
10 The plasmids can be high copy number per cell or low copy number per
cell. The plasmids can also be of virtually any size. It is readily
apparent to those skilled in the art that virtually any plasmid in the
microbial cells can be isolated by the process of the present invention.
Microbial cells cont~ining the plasmid are harvested from
1~ the fermentation medium to provide a cell paste, or slurry. Any
conventional means to harvest cells from a liquid medium is suitable,
including, but not limited to centrifugation or microfiltration.
Isolation of the plasmid DNA from harvested microbial
cells using the current lab scale procedures consist mainly of enzymatic
20 treatment of microbial cells to weaken the cell wall followed by cell
lysis. The purification steps include repetitive CsCl/EtBr
centrifugations followed by organic solvent extractions and precipitation
to remove tRNA, residual proteins, EtBr and other host Co~ nt~.
These steps are not scaleable and therefore not suitable for use in large-
25 scale processing. In contrast, preparative scale chromatography is apowerful purification tool that provides high resolution, operational
ease and increased productivity for purifying DNA plasmid products.
Two different modes of chromatography, reversed phase and anion
exchange, show suitability in purifying DNA plasmid to the stringent
30 levels required for hllm~n use. Separations based on reversed phase are
governed by hydrophobic interactions while those for anion exchange
are based on electrostatic interaction. These two orthogonal
chromatography steps achieve separations between various forms of
plasmid (supercoiled, open relaxed, linear and concatemers) and remove
CA 02220867 1997-11-12
W O 96/36706 PCTrUS96/07083
host col,t~-l,in~nt~ like LPS (endotoxin), RNA, DNA and residual
proteins.
In the process of the present invention, harvested microbial
cells are resuspended in modified STET buffer which is comprised of
5 about 50 mM TRIS, about 50-100 mM EDTA, about 8% Sucrose, about
2% TRITON X-100, and optionally sub-microgram concentrations of
lysozyme, at a pH in the range of 6.0-10Ø The concentration of
lysozyme optionally used in the process of the present invention is
subst~nti~lly less than the concentration of lysozyme used in the
10 procedures known in the art. It is readily apparent to those skilled in
the art that modifications of this basic buffer formula can be made and
are suitable for use in the present invention. Modifications to this basic
buffer formula that do not subst~nti~lly affect or alter the outcome of
the present process are intended to be within the scope of the process of
15 the present invention. The pH range may be adjusted according to the
best results provided for the particular strain of bacteria being used.
The preferred pH range is about 8.0-8.5. The suspension is then heated
to about 70-100~C, with about 70-77~C preferred, in a flow-through
heat exchanger. The lysate is centrifuged to pellet large cell debris,
20 protein and most genomic DNA.
A prototype heat exchanger was built to demonstrate the
feasibility of flow-through heat lysis of microbial cells cont~inin~;
plasmid. The particular heat exchanger consisted of a 10 ft. x 0.25 inch
O.D. stainless steel tube shaped into a coil. The coil was completely
25 immersed into a constant high temperature water bath. The hold-up
volume of the coil was about 50 mL. Thermocouples and a
thermometer were used to measure the inlet and exit temperatures, and
the water bath temperature, respectively. The product stream was
pumped into the heating coil using a Masterflex peristaltic pump with
30 silicone tubing. Cell lysate exited the coil and was then centrifuged in a
Beckman J-21 batch centrifuge for clarification. Figure 1 provides a
schematic of this particular apparatus, however other types of heat
exchanger construction are suitable for use in the present invention,
.
CA 02220867 1997-11-12
WO 96136706 PCT/US96/07083
including but not limited to a shell and tube construction, which is
preferrable.
After centrifugation, the clarified lysate can optionally be
treated with RNase, and the plasmid product can be filtered to further
5 remove small debris. A wide variety of filtration means are suitable for
use in this process, including but not limited to filtration through a
membrane having a small pore size. A preferred filtration method is
filtration through a 0.45 micron filter.
To further remove cont~min~nts from the DNA product,
10 the material can be diafiltered. Standard, commercially available
diafiltration materials are suitable for use in this process, according to
standard techniques known in the art. A ~lefe-l~d diafiltration method
is diafiltration using an ultrafilter membrane having a molecular weight
cutoff in the range of 30,000 to 500,000, depending on the plasmid size.
15 The DNA ~r~al~tion described above is tli~filtered using an
ultrafiltration membrane (about 100,000 molecular weight cutoff)
against column buffer prior to loading on the anion exchange column.
Diafiltration prior to the anion exchange column is L)lefe-,ed, and it
greatly increases the amount of lysate that can be loaded onto the
20 column.
A wide variety of commercially available anion exchange
matrices are suitable for use in the present invention, including but not
limited to those available from POROS Anion Exchange Resins, Qiagen,
Toso Haas, Sterogene, Spherodex, Nucleopac, and Pharmacia. The
25 column (Poros II PI/M, 4.5 mm x 100) is initi~lly equilibrated with 20
mM Bis/l~IS Propane at pH 7.5 and 0.7 M NaCl. The sample is loaded
and washed with the same initial buffer. An elution gradient of 0.5 M
to 0.85 M NaCl in about 25 column volumes is then applied and
~ fractions are collected. Anion exchange chromatography is an ideal
30 first polishing step because it provides excellent clearance of RNA,
genomic DNA and protein. Figure 5 (bold) shows a sample elution
profile of filtered clarified cell lysate from the anion exchange column.
Agarose gel analysis revealed that ~e second peak which appears after
the flow-through is composed of the plasmid product. The earlier large
CA 02220867 1997-11-12
W 096136706 PCTrUS96/07083
- 10 -
peak is due to RNA. This is confirmed by incubating the clarified cell
lysate with ribonuclease prior to loading on the column, which showed
that the large peak disappears and is replaced by several smaller more
rapidly eluting peaks, due to the degradation products of ribonuclease
5 digestion.
The anion exchange product fraction is loaded onto a
reversed phase column. A wide variety of commercially available
matrices are suitable for use in the present invention, including but not
limited to those available from POROS, Polymer Labs, Toso Haas,
10 Pharmacia, PQ Corp., Zorbax, BioSepra resins, BioSepra Hyper D
resins, BioSepra Q-Hyper D resins and Amicon. The matrices can also
be polymer based or silica based. The reversed phase column (Poros
R/H), is equilibrated with about 100 mM Ammonium Bicarbonate at pH
8.5. A gradient of 0-11% isopropanol is then used to elute bound
15 material. The three forms of plasmids, forms I, II and m described
above, can be separated by this method.
The eluted plasmid DNA can then be concentrated and/or
diafiltered to reduce the volume or to change the buffer. For DNA
intended for hllm~n use it may be useful to diafilter the DNA product
20 into a ph~rm~ceutically acceptable carrier or buffer solution.
Pharmaceutically acceptable carriers or buffer solutions are known in
the art and include those described in a variety of texts such as
Remington's Pharmaceutical Sciences. Any method suitable for
concentrating a DNA sample is suitable for use in the present invention.
25 Such methods includes diafiltration, alcohol precipitation, lyophilyzation
and the like, with diafiltration being preferred. Following diafiltration
the final plasmid DNA product may then be sterilized. Any method of
sterilization which does not affect the utility of the DNA product is
suitable, such as sterilization by passage through a membrane having a
30 sufficiently small pore size, for example 0.2 microns and smaller.
The following examples are provided to illustrate the
process of the present invention without, however, limitin~ the same
thereto.
CA 02220867 1997-11-12
W O 96/36706 PCTrUS96/07083
EXAMPLE 1
Growth of Microbial Cells. Cell Lysis and Clarification
One liter of frozen E. coli cell slurry was used to make 8
liters of cell suspension in STET buffer (8% sucrose, 0.5% TRITON,
50 mM TRIS buffer, pH 8.5 and 50 mM EDTA). The absorbance of
the cell suspension at 600 nm was about O.D. 30. The suspension was
stirred continuously to ensure homogeneity. The viscosity of the cell
suspension was measured to be about 1.94 cp at room temperature
(24~C). The cell suspension was pumped through the heat exchanger at
81 mL/min which corresponded to a residence time of the cell solution
in the heat exchanger of about 35 seconds. The bath temperature was
m~int~in~d at 92~C. The inlet and outlet temperatures of the cell
solution were measured to be about 24~C and about 89~C (average),
respectively. About 1 liter of sample was run through the heat
exchanger. No visible clogging of the tube was observed although the
lysate was somewhat thicker than the starting material. The lysate was
cooled to room temperature and its viscosity was measured to be about
40 cp. The cell lysate was clarified by batch centrifugation at 9000
RPM for 50 minlltes using the Beckman J-21. Analysis of the
supernatant confirmed effective cell lysis and product recovery. The
product yield produced by flow-through heat lysis was at least
comparable to that made by the Quigley & Holmes boiling method. The
latter method; however, must be carried out at the laboratory scale in
batch mode and is therefore unsuitable for large-scale (5 liters or
greater) processing. Since the heat exchanger process is flow-through,
there is no maximum limit to the volume of cell suspension that can be
processed. This process can therefore accomodate very large scale
fermentations of bacteria to produce large quantities of highly purified
plasmid DNA.
The clarified lysate was then filtered through a membrane
having a pore size of 0.45 microns to remove finer debris. The filtrate
was then diafiltered using a membrane having a molecular weight cutoff
of about 100,000.
CA 02220867 1997-11-12
W 096/36706 PCTrUS9G~7~83
- 12-
EXAMPLE 2
Control and Reproducibility of Cell Lysis with the Heat Exchanger
Adjusting the flowrate (i.e., residence time) at which the
S cell slurry is pumped through the heat exchanger permits tight control
of the temperature of Iysis, i.e., the outlet temperature. A cell slurry
solution was prepared as described in Example 1 and pumped through
the heat exchanger at flow rates ranging from 160 to 850 mL/min. The
corresponding outlet temperatures ranged between 93~C and 65~C,
10 respectively. Figure 2 illustrates the relationship between flow rate and
temperature. The initial temperature of the cell slurry was 24~C and
the bath temperature was kept constant at 96~C. In addition, a number
of runs were performed where an outlet temperature of 80~C was
targeted. Yields of 24 mg of circular DNA per L of clarified
15 supernatant were consistently obtained demonstrating the
reproducibility of the process.
EXAMPLE 3
20 Purification of Plasmid DNA
Microbial cells and lysates were prepared as described in
Examples 1 and 2, and the following analyses were performed.
To illustrate that the addition of 100 mM EDTA vs 50 mM
EDTA increased the percentage of supercoiled DNA, and to determine
25 an acceptable range of outlet temperatures (i.e., lysis temperature) with
respect to recovery of supercoiled DNA, the following analyses were
performed. The supercoiled form of plasmid DNA is desirable since it
is more stable than the relaxed circle form. One way that supercoiled
DNA can be converted to open circle is by nicking with DNase. We
30 found that the addition of 100 mM EDTA vs 50 mM in the STET buffer
i.-,i7ed the formation of open circle plasmid. Figure 3 shows
comparative chromatograms of the total plasmid in the clarified
supernatant with 50 mM EDTA vs 100 mM EDTA. The cell suspension
was prepared as described in Example 1. The operating flow rate for
CA 02220867 1997-11-12
WO 96136706 PCT~IJS96/07083
these runs was approximately 186 ml/min. The temperatures of the
inlet, outlet and bath are 24~C, 92~C and 96~C respectively.
An acceptable range of lysis temperatures was determined
by measuring the percentage of supercoiled plasmid generated for each
run. Figure 4 illustrates the concentration of supercoiled plasmid as a
function of exit temperature. An acceptable range of lysis temperatures
is between 75~C and 92~C. At temperatures below 75~C, more relaxed
circle plasmid was generated, most likely due to increased DNase
activity. Above 93~C, the levels of supercoiled plasmid appear to
10 ~limini~h, possibly due to heat denaturation.
Following continuous heat lysis and centrifugation, 1 mL of
clarified lysate was either incubated with S ,ug RNase for 2 hours, or
was used untreated. The RNase treated and untreated samples were then
loaded onto an anion exchange column (Poros Q/M 4.6 X 100) that had
15 been previously equilibrated with a 50-50 mixture of solvents A and B
[HPLC solvent A: 20 mM Tris/Bis Propane, pH 8.0; and solvent B:
M NaCl in 20 mM Tris/Bis Propane, pH 8.0]. The column was eluted
using a gradient of 50% to 85% B run over 100 column volumes. Open
circle plasmid elutes at approximately 68% B and supercoiled elutes at
20 72% B.
A comparison of the anion exchange column eluate from
clarified lysate treated with RNase (thin line) and untreated (thick line)
is shown in Figure 5. The peak at about 10 minutes is plasmid DNA,
and is followed by a large peak in the untreated sample which is RNA.
25 In the RNase treated sample the large RNA peak has been elimin~ted
and a gFe~d~er separa~ion ofthe plasmid peak from cont~min~nt peaks is
produced.
As described above, diafiltration prior to anion exchange
- chromatography greatly increases the amount of lysate that can be30 loaded onto the column. This is demonstrated in Figure 6 which shows
- a comparison of clarified lysate which was diafiltered and clarified
lysate which was not diafiltered prior to anion exchange chromato-
graphy. Samples were prepared as described above except that one
sample was diafiltered before loading onto the anion exchange column,
CA 02220867 1997-11-12
W 096/36706 PCTrUS96/07083
- 14 -
and the other sarnple was not diafiltered. The column was run and
eluted as described above. Figure 6 shows that the amount of
cont~min~nt material eluted from the column is vastly greater in the
sample that was not diafiltered. The large amount of Cont~min~tin~
material which binds the anion exchange column matrix can overwhelm
the ma~imlmm capacity of the column causing loss of DNA product
because of the unavailability of the matrix to bind any more material.
Therefore, diafiltration removes co~t~ nts and allows more of the
DNA product to bind the anion exchange matrix, and in turn allows a
10 greater volume of clarified lysate to be loaded onto the column.
The plasmid DNA eluted from the anion exchange column
was separated into the individual forms by reversed phase HPLC
analysis. The separation of supercoiled plasmid (form 1) from nicked
circle (form 2) is shown in Figure 7. The two forms were easily
15 separated and allowed the isolation of individual forms of the plasmid.
EXAMPLE 4
Highly Purified Plasmid DNA From a Chromatography-based Process
A fermentation cell paste was resuspended in modified
STET buffer and then thermally lysed in a batchwise manner.
Alternatively a fermentation cell paste is resuspended in modified STET
buffer and then thermally lysed in the flow-through process described
above. The lysate was centrifuged as described above. Twenty ml of
25 the supern~t~nt were filtered as described above and loaded onto an
anion exchange column (Poros Q/M 4.6xlO0) that was previously
equilibrated with a 50-50 mixture of buffers A and B described above.
A gradient of 50% to 85% B was run over 50 column volumes with a
flow rate of lOml/~ lle. Fractions of 2.5ml each were collected from
30 the column. The supercoiled plasmid DNA eluted from the column at
72% B.
The anion exchange product was then loaded onto a
reversed phase chromatography column (Poros R/H) which had been
previously equilibrated wit,h lOOmM ammonium bicarbonate at pH 8.0,
CA 02220867 1997-11-12
W 096/36706 PCT~US96/07083
- 15 -
and a gradient of 0% to 80% methanol was used to elute the bound
material. The highly purified supercoiled plasmid DN~ eluted at 22%
methanol.
An agarose gel of the product fractions from each of the
5 major steps of the purification processis shown in Figure 8. Based on
the agarose gels and the colorimetric and HPLC assays described in
Example 3, the final product, shown in Figure 9, is highly pure. The
product consists of greater than 90% supercioled and less than 10%
open circle plasmid. RNA was below the limits of detection of the assay
10 used. Genomic DNA and protein cont~min~nt levels were also below
the limits of detection in the assays used. The overall supercoiled
plasmid yield at the end of the process was approximately 60% of the
supercoiled plasmid in the clarified lysate.
EXA~k~E 5~
Multi-Gram Scale Purification of Plasmid DNA
4.5 L of frozen E. coli cell slurry was used to make 33.7 L
of cell suspension in STET buffer (8% sucrose, 2% Triton, 50 mM Tris
20 buffer, 50 mM EDTA, pH 8.5) with 2500 units/ml of lysozyme. The
absorbance of the suspension at 600 nm was about O.D. 30. The
suspension was stirred at room temperature for 15 ~ l-les to ensure
proper mixing and then was incubated for 45 minlltes with continuous
stirring at 37~C. Following incubation, mixing was continued at room
25 temperature and the cell suspension was pumped through the heat
exchanger at a flowrate of 500 ml/min. The batch temperature was
m~int~ined at 100~C and the inlet and outlet temperatures of the cell
suspension were measured to be about 24~C and between 70-77~C,
respectively. The cell lysate exiting the heat exchanger was collected in
30 Beckman centrifuge bottles (500 mls each) and the material was
centrifuged immediately in Beckm~n J-21 centrifuges for 50 mimltes at
9000 RPM. Following centrifugation, the supernatant was found to
contain 4-5 times more plasmid product than in the case without
lysozyme incubation. The supernatant product of the centrifugation was
35 in~rnediately diafiltered against 3 volumes of TE buffer (25 mM Tris-
CA 02220867 1997-11-12
WO 96/36706 PCTrUS96107083
- 16 -
EDTA at pH 8.0) and then incubated with 20x105 units of E. coli RNase
for 2-4 hours at room temperature. After completion of the incubation,
the product solution was then diafiltered an additional 6 volumes with
TE buffer using a 100 kD MWCO membrane and then filtered through
5 a 0.45 micron filter to remove residual debris. The filtered lysate was
diluted to 0.7 M NaCl with a 20 mM Bis~rris Propane-NaCl buffer at
pH 7.5, which prepares the diluted filtrate for loading onto the anion
exchange column. The anion exchange column (3.6 L of POROS PI/M)
was previously equilibrated with 20 mM Bis/Tris Propane and 0.7M
10 NaCl. The filtered lysate was loaded to column capacity. In this case 5
grams of supercoiled plasmid was loaded onto the anion exchange
column. After loading, the column was washed with 2-4 column
volumes of 20 mM Bis/Tris Propane and 0.7 M NaCl. A 10 column
volume gradient from 0.7 M NaCl to 2.0 M NaCl in 20 mM Bis/Tris
15 Propane was performed to clear most of the E. coli protein. RNA and
some endotoxin. The supercoiled plasmid fraction eluted between 1.4
M and 2.0 M NaCl. The supercoiled fraction from the anion exchange
column, which contained 4 grams of supercoiled plasmid was then
diluted 2-3 times with pyrogen free water, adjusted to 1.2% IPA and pH
20 adjusted to 8.5 with 1 N NaOH. The diluted anion exchange supercoiled
fraction was then loaded onto a 7 L reversed phase column (POROS
R2/M) which had been previously equilibrated with 100 mM
Ammonium Bicarbonate col~t~i"i"~ 1.2% IPA. In this case, 3.2 grams
of supercoiled plasmid were loaded onto the reversed phase column and
25 then the column was washed with 6-10 column volumes of 1.2% IPA in
100 mM Ammonium Bicarbonate. This extensive wash was performed
to clear impurities. Next, a gradient of 1.2% IPA to 11.2% IPA in 5
column volumes was performed. The supercoiled plasmid fraction
elutes at about 4% IPA. The supercoiled product fraction from the
30 reversed phase column was then concentrated and diafiltered into
normal saline using a 30 kD MWCO membrane. The final product buL~
was filtered through a 0.22 micron filter. Table 1 provides a
purification table describing clearance of impurities and yields across
each of the major process steps. The overall product yield of the
CA 02220867 1997-11-12
W096/36706 PCT~US96J~7~83
process was more ~an 50% of the supercoiled plasmid in the clarified
cell lysate as indicated by ~e anion exchange HPLC assay described in
EXAMPLE 3. The purity of ~e product was very high with less than
1% E. coli RNA and protein, and less than 2.9% E. coli genomic DNA.
TABLE 1
MULTI-GRAM PURIFICATION AND RECOVERY SUMMARY
STEP Plasmid% Step genomicProtein RNA LAL
Product Yield DNA (mg/mg) (mg/mg) Eu/mg
(mg)(mglmg)
cl~rifi~ lysate 6750 100 0.52 7.6 196 l.lx107
concentration/RNase/ 6500 93 0.50 1.6 2.21 3.4x106
diafil~ation/dead-end
~ filtration
anionexchange 4000 of 80 0.41 0.3 0.1 1.2x 104
5000
reversed phase 2300 of 77 0.029<O.Ol:j: <0.01:: 62
3200
concentration 2110 100 0.029 <0.01~ <0.01:: 2.8
diafiltration into final
buffer
final process yield 54
~ below detection limits of assay method
- 15
