dna vaccines, methods and protocols

1
From:
Methods in Molecular Medicine, vol. 29, DNA Vaccines: Methods and Protocols
Edited by: D. B. Lowrie and R. G. Whalen Humana Press Inc., Totowa, NJ
1
Purification of Supercoiled Plasmid
Anthony P. Green
1. Introduction
Current technologies for the purification of supercoiled plasmids are lim-
ited. The use of cesium chloride gradients in the presence of ethidium bromide
is time consuming, labor intensive, requires the use of known mutagens and is
not conducive to large scale. As a result, first-generation high-performance
liquid chromatography (HPLC) methods based on anion-exchange and size
exclusion have been developed but are difficult to accommodate production at
large scale and still result in compromised purity (1,2). The success of DNA
vaccines in animal models and the initiation of human trials (3,4) has led to a
need to increase the level of supercoiled plasmid purity as well as the method-
ology utilized to produce these plasmids at large scale. Several parameters of
the purification process need to be addressed:
• The ability to prepare supercoiled plasmid at purity levels acceptable for clini-
cal material.
• The ability to prepare clinical grade supercoiled plasmid that will be scalable in
order to produce gram quantities of product.
• The ability to prepare clinical grade supercoiled plasmid in accordance with
cGMP principles.
• The ability to develop validated assays to assess purity, yield, and contamina-
tion levels.
Challenges to the successful development of a purification process can be
divided into biological and practical. The biological challenge arises from the
spectrum of biomolecules that must be purified away from the supercoiled plas-

mid product ( Table 1). Additionally, the spectrum of nucleic acid contami-
nants and plasmid isoforms within that spectrum, as shown in Table 2, must be
removed. The removal of the relaxed DNA, DNA catenanes as well as endo-
toxins (5,6) are a particular problem requiring additional steps in the process.
2 Green
The practical challenge arises because the purification process that is devel-
oped must produce highly pure product at high yield and must be reproducible,
scalable and economical ( see Note 1).
We describe a new purification process that has been used to generate clini-
cal material using a proprietary non-porous polymer resin, PolyFlo
®
, which
uses principles of ion-pair reverse-phase chromatography to achieve separa-
tion based on size and charge density. The process can be performed using
either acetonitrile (ACN) or ethanol (EtOH). Simultaneous removal of con-
taminating endotoxins, chromosomal DNA, RNA, proteins, and plasmid iso-
forms during purification is a unique advantage of this resin. This process meets
the challenges for purity, yield, reproducibility, and scalability (1).
2. Materials
2.1. Crude Starting Material
The preparation of crude starting material from biomass is typically per-
formed using acid/base extraction (7). This classic alkaline-lysis process pro-
vides material significantly reduced in protein, lipid and chromosomal DNA.
Newer protocols have been adopted to improve the initial purity even further,
including the use of temperature shift during fermentation (8) or the addition
of a second acetate precipitation (NH
4
Ac) to reduce the RNA burden (9). We
describe two purification methods: an ACN process using starting material in
which minimal efforts have been made to reduce the RNA burden, and (2) an

EtOH process in which the starting material has been reduced for RNA by
anion-exchange chromatography and/or diafiltration. The protocols described
do not incorporate the use of RNase ( see Note 2).
2.2. ACN Purification Materials
1. Glass borosilicate chromatography column packed with PolyFlo resin (see
Note 3).
2. Starting material (ammonium acetate supernatant) (see Note 4).
Table 1
Constituents in Crude Lysate
Plasmid DNA
Chromosomal DNA
RNA
Lipids
Endotoxin
Proteins
Carbohydrates
Table 2
Plasmid DNA Forms
Monomer supercoiled
Nicked
Linear
Dimers
Catenanes
Purification of Supercoiled Plasmid 3
3. 1.0 M TEAA (triethylamine acetate) pH 7.0.
4. 0.5 M KPO
4
pH 7.0.
5. 1.0 M TBAP (tetrabutylammonium phosphate); Aldrich Chemicals (Milwaukee,
WI) no.# 26, 810-0; 1 M in H

2
O).
6. 100% Acetonitrile (ACN, American Chemical Society (ACS) grade or equivalent).
7. TES (20 mM Tris-HCl pH 8.0, 1.0 mM EDTA, 5.0 mM NaCl).
8. Equilibration buffer: 0.1 M TEAA pH 7.2, 6% ACN.
9. Wash buffer I: TES, 5% ACN.
10. Wash buffer II: 0.1 M KPO
4
, 2.0 mM TBAP, 5% ACN.
11. Wash buffer III: 0.1M KPO
4
, 2.0 mM TBAP, 15% ACN.
12. Elution buffer: 0.1 M KPO
4
, 2.0 mM TBAP, 25% ACN.
13. Sanitization buffer: 0.5 N NaOH.
2.3. Ethanol Purification Materials
1. Glass borosilicate chromatography column packed with PolyFlo resin.
2. RNA-reduced plasmid sample.
3. 0.5 M KPO
4
pH 7.0.
4. 1.0 M TBAP (tetrabutylammonium phosphate; Aldrich Chemicals # 26, 810-0;
1 M in H
2
O).
5. Ethanol (EtOH, ACS grade or equivalent).
6. TES (0.02 M Tris-HCl pH 8.0, 1.0 mM EDTA, 5.0 mM NaCl).
7. Equilibration buffer: 0.1 M KPO
4

pH 7.0, 2.0 mM TBAP, 1% ethanol.
8. Wash buffer I: TES, 7% ethanol.
9. Wash buffer II: 0.1 M KPO
4,
2.0 mM TBAP, 5% ethanol.
10. Elution buffer: 0.1 M KPO
4
, 2.0 mM TBAP, 25% ethanol.
11. Sanitization buffer: 0.5 N NaOH.
2.4. Post-Purification Materials
Millipore Pellicon II (Bedford, MA) or A/G Technology (Needham, MA)
diafiltration/ultrafiltration technologies are applicable for buffer exchange and
concentration.
3. Purification Protocols
A schematic of the two protocols is shown in Fig. 1.
3.1. ACN Protocol
1. Adjust to obtain a linear flow rate of 150 cm/h and equilibrate column in ≥3
column volumes of equilibration buffer.
2. Prepare sample by diluting 1/5 with TES and adjusting to 0.1 M TEAA using 1 M
TEAA stock solution. Sample load should be no more than 0.5 mg/mL of resin
(load concentration is based on total A
260nm
).
3. Load sample and wash with equilibration buffer until the monitor returns to
baseline (~2 column volumes). Collect wash.
4 Green
4. Wash with ~3 column volumes of wash buffer I (TES, 5% ACN). Make sure
monitor returns to baseline. Collect wash.
5. Wash with ~3 column volumes of wash buffer II (0.1 M KPO
4

, 2.0 mM TBAP,
5% ACN).
6. Wash with ~3 column volumes of wash buffer III (0.1 M KPO
4
, 2.0 mM TBAP,
15% ACN) until the monitor returns to baseline. Collect wash.
7. Elute product with a 10-column volumes linear gradient from 0.1 M KPO
4
, 2.0
mM TBAP, 15% ACN to 0.1 M KPO
4
, 2.0 mM TBAP, 25% ACN. Collect elution
fractions. This is the purified product.
8. Clean column by running 2 column volumes of sanitization buffer (0.5 N NaOH).
Turn pump off and let sit in 0.5 N NaOH for 1 h. Re-equilibrate column with ≥3
column volumes of equilibration buffer. Monitor pH to assure that all the NaOH
has been removed.
9. Purified sample from step 8 may be processed through concentration and/or
buffer exchange steps. It is recommended to diafilter against 0.5 M Na-acetate
pH 7.8 to remove residual TBAP.
3.2. ACN Protocol Results
The results of the ACN protocol are shown in Fig. 2. No difference in the
purity of the product is seen using starting material representing 1 g or 100 g of
biomass. The RNA is eliminated. Despite the significant quantities of relaxed
DNA, >50% of total plasmid, this contaminant is removed in the wash step.
The final purity is >90%.
3.3. Ethanol Protocol
1. Adjust pump speed to obtain a linear flow rate of 150 cm/hr and equilibrate col-
umn in ≥ 3 column volumes of equilibration buffer.
Fig.1. Schematic flow-chart of PolyFlo purification process.

Purification of Supercoiled Plasmid 5
2. Prepare sample for loading by adjusting to 2 mM TBAP and 1% ethanol. Sample
load should be no more than 0.5 mg/mL of resin (load concentration is based on
total A
260nm
). Sample should be <0.5 M NaCl.
3. Load sample and wash with equilibration buffer until the monitor returns to
baseline. Collect flow through.
4. Wash with ~3 column volumes of wash buffer I (TES, 6% ethanol). Make sure
monitor returns to baseline. Collect wash.
5. Wash with ~3 column volumes of wash buffer II (0.1 M KPO
4
, 2.0 mM TBAP,
5% ethanol).
6. Elute product with a 10-column volumes linear gradient from 0.1 M KPO
4
, 2.0 mM
TBAP, 5% Ethanol to 0.1 M KPO
4
, 2.0 mM TBAP, 25% ethanol. Collect elution
fractions. This is the purified product.
7. Clean column by running 2 column volumes of sanitization buffer (0.5 N NaOH).
Turn pump off and let sit in 0.5 N NaOH for 1 h. Re-equilibrate column with ≥3
column volumes of equilibration buffer. Monitor pH to assure that all the NaOH
has been removed.
Fig. 2. PolyFlo chromatography of plasmid using ACN process. (A) Chromato-
graphic tracing for application of nucleic acid sample extracted from 1.0 g E. coli
cells. (B,C) 1% Agarose gel analysis of resolved peaks from 1.0 g biomass (B) and
100 g biomass (C) . Lane 1 = 1 µg nucleic acid sample; lanes 4 and 5 = TES and ACN
wash; lanes 6 and 7 = 15% (v/v) ACN/TBAP wash; lanes 8 and 9 = 15-25% (v/v) ACN

gradient; and lane 10 = 50% ACN strip. Reprinted from (1).
6 Green
8. Purified sample from step 7 may be processed through concentration and/or
buffer exchange steps. It is recommended to diafilter against 0.5 M Na-acetate
pH 7.8 to remove residual TBAP.
3.4. Ethanol Protocol Results
The results of the Ethanol protocol are shown in Fig. 3. The residual RNA is
eliminated. Despite the significant quantities of relaxed DNA, >60% of plas-
mid, this contaminant is removed in the wash step. The final purity is >95%.
4. Notes
4.1. Organic Solvent
The choice of organic solvent for chromatography is predicated on the
amount of contaminating RNA. In general, if the RNA burden is less than 50%,
the ethanol process may be employed. This can be accomplished through anion-
exchange chromatography, diafiltration or RNase treatment. A rigorous test of
the amount of contaminating RNA below which the ethanol process can be
used has not been performed. If RNA reduction is achieved by RNase diges-
tion, the sample must be diafiltered or dialyzed to remove excess ribonucle-
otides prior to PolyFlo chromatography.
Fig. 3. PolyFlo Chromatography of plasmid using EtOH process. Crude lysate was
reduced for RNA by diafiltration against 10–20 vol TES using a Millipore XL mem-
brane (100,000 MWCO) prior to application onto a 1 × 4 cm PolyFlo column. (A)
Chromatographic tracing at 254 nm; (B) 1% Agarose gel of resolved peaks. Lane 1 =
starting material; lane 2 = 1% EtOH flow-through peak; lane 3 = 6% EtOH wash peak;
and lane 4 = 5–25% EtOH gradient.
Purification of Supercoiled Plasmid 7
4.2. Endotoxin Removal
PolyFlo has an extremely hydrophobic surface. As such, significant quanti-
ties of endotoxin are removed as part of the purification process and in a repro-
ducible manner ( Fig. 4).

4.3. Specifications
No specifications for the purity of supercoiled plasmid, the levels of residual
contaminants or even the methods for evaluating purity have been codified
(10). While many methods can be used to analyze plasmid DNA, it is only
recently that these methods have been applied to plasmid DNA as a potential
pharmaceutical product (11). Table 3 describes some of the target specifica-
tions and methods used within the industry.
4.4. Multiple Chromatography Runs
PolyFlo is a chemically inert polymer that withstands rigorous sanitization
procedures which allows for multiple runs. One hundred consecutive applica-
tions of crude plasmid with no change in purity or contamination levels have
Fig. 4. Endotoxin binding to a PolyFlo column (1 × 4 cm). (A) Total endotoxin
units (EU) were determined in the flow-through (solid squares) and gradient elution
(open circles) after loading sample buffer was spiked with increasing levels of endo-
toxin. (B) Analysis of endotoxin levels in purified plasmid preparations at defined
intervals during 100 consecutive applications of a single PolyFlo chromatography col-
umn. Reprinted from (1).
8 Green
been documented (1). The resin can be sanitized to remove any residual nucleic
acid, protein, lipid and endotoxin by exposure to 0.5 N NaOH alone or in com-
bination with 0.1 M HCl.
4.5. Potential Interferences
PolyFlo resin is sensitive to detergents but not chaotropes or salts. The use
of PEG in the lysis process does not affect PolyFlo performance (1). However,
detergents such as SDS, in concentrations >0.005% prevent binding to the resin
and should be avoided.
4.6. Process Optimization
As with all chromatographic procedures, there are several key steps that will
affect results. Using PolyFlo in the chromatography of plasmids, the key ele-
ments are the organic solvent concentration in the loading buffer and the col-

umn wash steps. For example, if the load concentration of organic is too high,
product will be lost in the flow-through fraction. If the load concentration is
too low, then the RNA will bind and will not be eliminated in the flow-through.
As a consequence, trace RNA levels may be seen throughout the chromatogra-
phy. The same considerations can be applied to the wash concentration.
References
1. Green, A. P., Prior, G. P., Helveston, N. M., Taittinger, B. E., Liu, X F., and
Thompson, J. A. (1997). Preparative purification of supercoiled plasmid DNA for
therapeutic applications. BioPharm 10, 52–62.
2. Horn, N. A., Meek, J. A., Budahazi, G., and Marquet, M. (1995). Cancer gene
therapy using plasmid DNA: purification and DNA for human clinical trials. Hum.
Gene Ther. 6, 565–573.
Table 3
Target Specifications
Target Testing
Parameter specification method
Purity
% Monomer supercoiled >95 1% Agarose gel
Purity 1.8–2.0 A260:A280
Contaminants
RNA <1% 1% Agarose gel
Genomic DNA <1% Slot-blot hybridization
Endotoxin <100EU/mg LAL gel clot
Protein Negative SDS-PAGE
Purification of Supercoiled Plasmid 9
3. Ulmer, J. B., Donnelly, J. J., and Liu, M. (1996). Toward the development of
DNA vaccines. Curr. Opin. Biotechnol. 7, 653–658.
4. Donnelly, J. J., Ulmer , J. B., Shiver, J. W., and Liu, M. (1997). DNA vaccines.
Ann. Rev. Immunol. 15, 617–648.
5. Weber, M., Möller, K., Welzeck, M., and Schoor, J. (1995). Effects of lipopolysac-

charide on transfection efficiency in eukaryotic cell. BioTechniques 19, 930–940.
6. Wicks, I. P., Howell, M. L., Hancock, T., Kohsaka, H., Olee, T. -W., and Carson,
D. (1995). Bacterial lipopolysaccharide copurifies with plasmid DNA: implica-
tions for animal models and human gene therapy. Hum. Gene Ther. 6, 317–323.
7. Birnbaum, H. C. and Doly, J. (1979). A rapid alkaline extraction procedure for
screening recombinant plasmid DNA. Nucleic Acids Res. 7, 1513–1523.
8. Lahijani, R., Hulley, G., Soriano, G., Horn, N., and Marquet, M. (1996). High-
yield production of pBR322-derived plasmids intended for human gene therapy
by employing a temperature-controllable point mutation. Hum. Gene Ther. 7,
1971–1980.
9. Thompson, and Blakesley, (1983). Purification of nucleic acids by RPC-5 analog
chromatography: peristaltic and gravity-flow applications. Meth. Enzymol. 110,
123–127.
10. Office of Vaccine Research and Review (1996). Points to consider on plasmid
DNA vaccines for preventive infectious disease indications, Food and Drug
Administration, Bethesda, MD.
11. Middaugh, C. R., Evans, R. K., Montgomery, D. L., and Casimiro, D. R. (1998).
Analysis of plasmid DNA from a pharmaceutical perspective. J. Pharm. Sci. 8,
130–146.

11
From:
Methods in Molecular Medicine, vol. 29, DNA Vaccines: Methods and Protocols
Edited by: D. B. Lowrie and R. G. Whalen Humana Press Inc., Totowa, NJ
2
Production of Plasmid DNA in Industrial Quantities
According to cGMP Guidelines
Joachim Schorr, Peter Moritz, and Martin Schleef
1. Introduction
Within the last five years, the exponential growth of research activities on

the development of genetic vaccination and gene therapy has made it neces-
sary to develop an easy, cost-effective, industrial scale process for production
of plasmid DNA (see Note 1). One main issue is that the process should con-
form to cGMP guidelines and be acceptable to the FDA or other national regu-
latory agencies. The cGMP environment should be implemented independently
of the intended use of the DNA product.
A typical application would be the supply of genetic information that is miss-
ing within the cell, e.g., because of a genetic disease like cystic fibrosis (CF).
In such cases the “therapy” has been performed by transferring liposome-plas-
mid DNA complexes to the lung epithelium to express the absent chloride chan-
nel gene (CFTR) for the restoration of the draining system of this tissue (1).
The more preventive approach of “gene medicine” could involve vaccination
using plasmid DNA either by subcutaneous or intramuscular injection (2–6) or
other techniques (for an overview see ref. 7, and this volume). The expression
of immunogenic epitopes can cause both humoral and CTL response (8,9), and
Chapter 6.
In all cases, it is essential to be able to use a therapeutic agent (the “bio-
logic,” usually a plasmid DNA) free of any other materials. Such contamina-
tion includes components used in the isolation process and coming from the
organism from which the plasmid is isolated, mainly residual proteins, RNA,
and genomic DNA of the host cell. In this chapter, we describe the develop-
ment of a pharmaceutical manufacturing process to isolate plasmid DNA start-
12 Schorr, Moritz, and Schleef
ing from a technology (Qiagen, Hilden, Germany) that has been shown over
the past years to be a tool in research and development that fulfills the require-
ments described in Subheading 1. (see Notes 2–4).
1.1. The Host Cell Selection
A single appropriate host strain for all research work or industrial scale phar-
maceutical manufacturing does not exist (see Note 5). An appropriate strain
should be a clone derived from a host strain stock that is completely character-

ized and free of any contamination. It should be safe for the environment, for
the isolated product, the exposed patients, and the employees doing the manu-
facturing work as well as health care personnel. Considerable experience in the
field of molecular cloning and DNA techniques has been obtained with
Escherichia coli (E. coli), and E. coli K12 fulfills the needs for a safe, well-
characterized host strain for DNA production.
Systematic analysis of over 20 different E. coli substrains demonstrated that
very large qualitative and quantitative differences exist among all the substrains
tested. These differences mainly concern particularly the amount of plasmid
DNA per gram of biomass and the plasmids isoform distribution.
Plasmid isoforms consist of supercoiled molecules, dimers, or concatemers
or catenanes (chains of two or more plasmids), as well as linear or nicked plas-
mids. The observed differences in isoforms depend on the plasmid as well as
on the host strain. This means that not only the genetic background of the host
is responsible for the differences of the plasmid isoform distribution, but the
plasmid itself also contributes to a certain extent.
1.2. Growth Conditions
Bacterial cultures for the purpose of plasmid isolation were performed in a
batch mode, using culture bottle volumes of up to a maximum of 2 L. Studies
were done to determine the growth medium and conditions for optimal bacte-
rial growth and plasmid yields, resulting in optical density (O.D. 600) values
of around 3–6 O.D. units in complex bacterial growth media. For purposes
such as research grade plasmid preparations for cloning, sequencing, and trans-
fection experiments, this procedure is adequate and the final analysis of the
prepared DNA is usually an agarose gel electrophoresis, DNA quantification
and identity test (restriction digestion).
These test criteria are not stringent enough for pharmaceutical purposes,
and the procedure of the manufacturing had to be drastically modified. The
first point to consider in the development of a pharmaceutical grade process is
that the batch culture method has no type of online monitoring or regulation.

The growth conditions are adjusted before inoculating the medium and left
unchanged, usually for between 16–20 h. No pH monitoring or adjustment is
Production of Plasmid DNA 13
performed, oxygen and carbon source are also neither monitored nor regulated.
Essential substrates are depleted and toxic products accumulate. Degraded cel-
lular components, including plasmids, accumulate in overgrown cultures and
cell death follows.
To overcome these problems for the isolation of recombinant proteins,
high performance fermentation technology has been developed over the last
few years.
Fermentation processes require different growth media than batch cultures.
The possibility of monitoring the growth conditions allows for the introduction
of essential media components before they are exhausted (feeding), and for the
maintenance of a constant pH control and oxygen supplies. Besides the effects
of such regulation, the culture process becomes more defined and the pharma-
ceutical requirements on documentation can be fulfilled. A further feature of
the fermentation technology for the large-scale plasmid production, is the
potential of high-density fermentation to yield large amounts of biomass.
Experimental work on the composition of bacterial growth media for bottle
cultures and fermentation demonstrates that the choice of fermentation condi-
tions and growth media strongly influence the yields of plasmid that are
obtained from E. coli cells. One main focus was put on the amount of plasmid
per cell (copy number), that can be monitored on-line by capillary gel electro-
phoresis (10).
1.3. Downstream Processing
The isolation of a biomolecule from the bacterial culture (usually referred to
as downstream processing or DSP) is performed to separate plasmid DNA from
other undesired components present within the source of material. These unde-
sirable components are genomic DNA, RNA, proteins, lipids, lipopoly-
saccharides (LPS) or endotoxins, components of the cell wall and intact

bacteria (see Note 6). The alkaline lysis (11) was modified and is reproducibly
performed in scales up to five liter bacterial culture (ultrapure 100 chromato-
graphy system, Qiagen). The most important feature of this technique is the
formation of a complex of most of the undesired components mentioned above,
which can easily be removed by centrifugation (research scale) or floating and
filtration (research and industrial scale).
The resulting “cleared lysate” is applied to an industrial scale process chro-
matography column with anion exchange resin, to specifically bind the nega-
tively charged plasmid DNA and (under appropriate buffer conditions) not to
bind residual undesired components (e.g., protein, RNA, nucleotides, LPS).
Such anion exchange chromatography is not limited in scale (compared to
approaches such as gel filtration). Moreover, in the case of specific types of resin
material, possessing dense, high surface charge, a one-step process can be used.
14 Schorr, Moritz, and Schleef
As an additional pharmaceutical requirement, a process for the complete, rapid
removal of LPS molecules was developed (see Subheading 2.5.). Endotoxins
such as E. coli LPS can have cytotoxic effects on mammalian cells in vitro and in
vivo (12–15) and if present in large enough amounts in vivo can cause symptoms
of toxic shock syndrome and activation of the complement cascade (16).
In our process development we focused on the use of only non-toxic sub-
stances, and in particular avoided any potentially carcinogenic or immuno-
genic reagents. Additionally, the environment was controlled and the resulting
liquid waste was biodegradable.
1.4. Quality Assurance and Quality Control
When we began our work on DNA manufacturing, the only criteria for the
quality of plasmid DNA were those of typical research work. Usually the qual-
ity for this research grade material was estimated using analytical gel electro-
phoresis, restriction enzyme digestion, and DNA sequence readings.
We therefore established a set of quality criteria (17) that is now well
accepted by the scientific community. Relevant issues from this work were

discussed at the Food and Drug Administration (FDA)/World Health Organi-
zation (WHO) conference on Nucleic Acid Vaccines, February 5–7, 1996 at
the National Institute of Allergy and Infectious Diseases/National Institutes of
Health (NIAID/NIH) (Bethesda, MD) (18). An overview of the regulatory as-
pects for design, manufacturing, quality assurance, and quality control of vac-
cination vectors are summarized in the WHO “Guidelines for Assuring the
Quality of DNA Vaccines” (WHO Technical Report, Jan. 17, 1997).
The design of the production process focused on its acceptance by national
and international authorities such as the FDA (Washington, DC), Medicines
Control Agency (UK) and others, and had to fulfill the appropriate cGMP regu-
lations (see Note 7).
2. Materials
2.1. Buffers
1. P1: 50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 100 µg/mL RNaseA.
2. P2: 200 mM NaOH, 1% (w/v) SDS.
3. P3: 3.0 M KAc, pH 5.5.
4. QBT: 750 mM NaCl, 50 mM MOPS, pH 7.0, 15% (v/v) isopropanol, 0.15% (v/v)
Triton-X-100.
5. QC: 1.0 M NaCl, 50 mM MOPS, pH 7.0, 15% (v/v) isopropanol.
6. QN: 1.6 M NaCl, 50 mM MOPS, pH 7.0, 15% (v/v) isopropanol.
2.2. Transformation and Host Cells
Prepare competent cells such as E. coli K12 DH5α (Life Technologies,
Eggenstein, Germany), DH10B (Life Technologies) or TG1 (# 6056; Deutsche
Production of Plasmid DNA 15
Sammlung von Mikroorganismen und Zell Culturen, Braunschweig, Ger-
many), transform the plasmid DNA, and select recipients on agar plates con-
taining the appropriate selection factor.
2.3. Fermentation
Cultivate cells using a suitable fermenter, such as a Biostate B bioreactor
(B. Braun Biotech, Melsungen, Germany) with a working volume of 5 L. Use,

e.g., the complex bacterial growth media LB with additional salt (19).
2.4. Cell Harvest
Cells can be harvested by batch centrifugation at 4600g for 15 min at 4°C.
1. Beckman J2-21 centrifuge with a JA-10E rotor.
2. 500 mL Polypropylene bottles (Nalgene, Rochester, NY).
2.5. The Anion Exchange Chromatography System
Perform anion exchange chromatography (Qiagen) to specifically bind double
stranded DNA. Single stranded DNA, RNA, nucleotides, proteins, LPS. and other
contaminants do not bind to the chromatographic resin under appropriate conditions.
1. For small-scale preparations (e.g., test runs using the produced biomass), 500 µg
batches of DNA use the Qiagen EndoFree Plasmid Kit (Ref. #12362).
2. For larger scale preparations, use an anion-exchange chromatography column for
the isolation of up to 100 mg plasmid DNA (e.g., ultrapure 100 column #11100,
Qiagen) and LPS-free processing buffers (#11910, Qiagen).
3. Methods
The complete process of plasmid DNA production is performed under well-
documented conditions and in the case of GMP manufacturing under controlled
environmental conditions. The following examples of the process we use will
give some insight into the steps performed (Fig. 1).
3.1. The Host Cell Selection
To obtain a pure and well-characterized production strain capable of high
yields of DNA, the selection of an appropriate E. coli K12 plasmid host cell
clone is essential. Besides good microbiological practices and the use of standard
operating procedures (SOPs), a well established quality assurance and quality
control system is of great relevance, since all further process steps depend on this.
1. Check the DNA received by the Qiagen DNA Production Facility for large-scale
manufacturing for its identity first (size, restriction pattern, sequence), and if it is
satisfactory, release it for further processing.
2. Transform the DNA to E. coli K12 host cells, and select individual colonies for
further cultivation.

16 Schorr, Moritz, and Schleef
Fig. 1. Flow chart of a cGMP plasmid manufacturing procedure. “QA” indi-
cates the types of quality assurance tests that must be performed at certain steps to
ensure consistent quality and reproducibility, and to fulfill the needs of process
documentation.
3. Use 3 mL of an overnight culture of cells for a small-scale plasmid isolation
(QIAprep, Qiagen). In case of large numbers of clones, use an automated device
for the isolation of DNA in a 96-well format (BioRobot 9600, Qiagen).
4. Identify appropriate cell clones by comparing them and selecting those with high
plasmid yield and a proper plasmid isoform distribution for further production
steps (Fig. 2).
Production of Plasmid DNA 17
5. Further purify the selected clone by two single-colony passages and check it for
identity and absence of microbiological contaminants. Use it subsequently for
the inoculation of a culture to prepare a glycerol stock of between 100–500 vials.
This stock is called Master Cell Bank (MCB); it is necessary to be able to repro-
ducibly inoculate culture media from the MCB in the following process step and
any future manufacturing run.
6. Perform an extensive quality assurance program to check the quality of this MCB.
Test the identity, plasmid content, as well as absence of microbiological con-
taminants before proceeding with the following step. An important additional
requirement is the complete sequencing of the DNA construct at this stage to
exclude any difference to the original plasmid and to have a data backup for post-
production sequencing.
7. Use vials of the MCB to inoculate a fresh culture to produce an equally large set
of stocks (100–500 vials), which are required for the reproducible inoculation of
the fermentation-precultures. This second glycerol stock is called Manufacturing
Working Cell Bank (MWCB). Perform the same tests for Quality Assurance (QA)
as with the MCB.
3.2. Fermentation

A fermentation process for E. coli cells carrying plasmids in a certain copy
number must be well characterized, reproducible, easy to monitor and regu-
late; If possible it should run automatically. The MCB and MWCB described
above are the backbones for any reproducibility. Further important issues are
Fig. 2: Comparison of plasmid pUC21 DNA produced in different E. coli K12 host
cells. The upper panel shows the undigested DNA with its different isoforms. The
lower panel shows as a control the same amount of DNA after an EcoRI digestion. The
molecular size marker (M) is HindIII digested λ-DNA. Gel: 1% (w/v) agarose in TAE,
pH 8.0, run at 5V/cm and stained with ethidium bromide.
18 Schorr, Moritz, and Schleef
the types of fermenter, regulation and growth medium used. Batches of 5–200
L are routinely run, and if required, further scaling-up is possible.
1. Use an appropriate amount of the MWCB to inoculate a pre-culture in E. coli
growth medium.
2. Use the pre-culture to inoculate the fermenter for an overnight run at 37°C with
controlled pH (7.5) at maximum aeration.
3. Harvest the cells by use of a flow-through centrifuge and determine the biomass
content (wet and dry weights).
3.3. Lysis of Bacteria
To isolate the plasmid DNA from the E. coli cells, a modified alkaline lysis
procedure (11) is used. This step is of critical importance to reduce contami-
nants such as protein, RNA, genomic DNA, and cell wall residues. Here we
describe, as a pilot scale example, the approach of isolating up to 100 mg plas-
mid DNA starting from 60 g wet weight biomass (see also the protocol sup-
plied with the Qiagen ultrapure100 kit).
1. Thoroughly resuspend 60 g biomass in 1000 mL buffer P1 in a 5-L glass bottle.
2. Add 1000 mL of buffer P2, mix the complete volume and incubate it at room
temperature for 5 min.
3. Add 1000 mL of buffer P3 and mix it carefully.
4. Incubate the lysate for 30 min at room temperature to allow the flaky white pre-

cipitate of SDS, protein, genomic DNA, and cell residue to rise to the surface.
5. Carefully pump the lysate out of the bottle.
6. Filter the lysate through a QIAfilter™ unit (Qiagen), mixed with 1/10 volume of
buffer ER (Qiagen), and collect the filtrate for subsequent chromatography.
3.4. Anion Exchange Chromatography
The anion exchange chromatography columns are loaded by pumping lysate
with a peristaltic pump or preferably with a process chromatography system
for better monitoring of the process.
1. Equilibrate the Qiagen ultrapure 100 column with 350 mL buffer QBT at a flow
rate of 10 mL/min.
2. Load the column at a flow rate of approximately 4 mL/min overnight.
3. Wash the charged column with 3 L of LPS-free buffer QC at a flow rate of
20 mL/min.
4. Elute plasmid DNA with 400 mL LPS-free buffer QN at a flow rate of 3 mL/min.
5. Precipitate the DNA with 0.7 volumes of isopropanol at 4°C and centrifuge at
20,000g for 30 min in LPS-free centrifuge bottles.
6. Wash the DNA pellet with LPS-free 70% EtOH and rinse.
7. Dry the DNA pellet and resuspend it in the appropriate buffer system for further
applications.
Production of Plasmid DNA 19
3.5. Quality Assurance
The following QA is performed within the manufacturing process as an
in-process control (IPC).
1. Restriction Analysis: Digest the plasmid DNA to completion by use of differ-
ent restriction enzymes, following the instructions of the suppliers. Use aga-
rose gel electrophoresis to confirm that the total DNA size and molecular
weight of fragments are consistent with those expected from the knowledge
of the sequence.
2. Sequencing: Determine the complete nucleotide sequence of both DNA strands
by DNA sequencing. Perform all steps following SOPs and document the data in

a sequencing report.
3. Plasmid Stability: Monitor the presence or absence of a plasmid containing an
antibiotic resistance marker by inoculating a defined amount of cells on both
antibiotic and non-selective agar plates. If cells are not able to grow on selective
plates, the percentage of clones growing on both media represents the “plasmid
stability.”
4. DNA Quality: In addition to the analyses of fragment identity and sequence, use
spectrophotometric scans between 220–320 nm for the detection of salt and
organic contamination (20) within the DNA. Inspect the appearance of a sample
in an agarose gel electrophoresis. Important features are the isoform distribution
(by agarose gel electrophoresis) and the DNA concentration. Also determine the
content of RNA, genomic DNA and LPS by HPLC, Southern blot and the kinetic
QCL test kit (BioWhittaker, Walkersville, MD) respectively.
5. DNA Quantity: Determine the DNA concentration by spectrophotometric analy-
sis and calculation from its absorbance at 260 nm.
4. Notes
1. For large scale DNA production, we focused on the development of a technology
for industrial-scale manufacturing of nucleic acids that combines cost effective-
ness with the flexibility to install the system in every research laboratory (pilot
scale) or GMP facility (industrial scale).
2. A major consideration in the development of this technology was to avoid time-
consuming centrifugation and multiple chromatographic column runs. Centrifu-
gation of large volumes to clear bacterial lysates can now be replaced by just one
passage through a filtration unit that makes it possible to filter large volumes of
bacterial lysate.
3. The process includes the establishment of Master Cell Banks and Master Work-
ing Cell Banks; fermentation and downstream processing are monitored at all
stages by extensive in-process controls.
4. The three most important factors which need to be considered in the process
development for plasmid DNA production are: selection of the optimal host strain,

optimization of growth conditions, and the nucleic acid preparation method.
20 Schorr, Moritz, and Schleef
5. A large set of different E. coli host strains has been studied to identify strains
producing large amounts of plasmid DNA per cell with the highest quality. Qual-
ity criteria for the selection of a host strain are the homogeneity of the plasmid
DNA isolated from the host strain (>90% covalently closed circle), and the
endotoxin content of the DNA purified from a specific host strain.
6. Endotoxins (LPS) are major contaminants of nucleic acids, especially plasmid
DNA preparations. Due to their negatively charged phosphate groups, endotoxins
tend to co-purify with nucleic acids. It has been demonstrated that LPS contami-
nation of DNA has a direct influence on transfection efficiency into many types
of cultured cells, and that different cells show variable sensitivity to this con-
tamination (13).
7. The Qiagen procedure has been approved to produce DNA for human clinical
Phase I studies in the UK (1) and other European countries, as well as in the
United States by the FDA (21). A drug master file (DMF) for the clinical grade
manufacturing process is filed with the FDA.
References
1. Caplen, N. J., Gao, X., Hayes, P., Elaswarapu, R., Fisher, G., Kinrade, E., et al.
(1994). Gene therapy for cystic fibrosis in humans by liposome-mediated DNA
transfer: UK regulatory process and production of resources. Gene Therapy 1,
139–147.
2. Davis, H. L., Whalen, R. G., and Demeneix, B. A. (1993). Direct gene transfer
into skeletal muscle in vivo: factors affecting efficiency of transfer and stability of
expression. Hum. Gene Ther. 4, 151–159.
3. Manthorpe, M., Cornefer-Jensen, F., Hartikka, J., Felgner, J., Rundell, A.,
Margalith, M., and Dwarki, V. (1993) Gene therapy by intramuscular injection of
plasmid DNA: studies on firefly luciferase gene expression in mice. Hum. Gene
Ther. 4, 411–418.
4. Michel, M L., Davis, H. L., Schleef, M., Mancini, M., Tiollais, P., and Whalen,

R. G. (1995) DNA-mediated immunization to the hepatitis B surface antigen in
mice: aspects of the humoral response mimic hepatitis B viral infection in humans.
Proc. Natl. Acad. Sci. USA 92, 5307–5311.
5. Davis, H. L., Michel, M L., Mancini, M., Schleef, M., and Whalen, R. G. (1994)
Direct gene transfer in skeletal muscle: plasmid DNA-based immunization against
the hepatitis B surface antigen. Vaccine 12, 1503–1509.
6. Wolff, J. A., Williams, P., Acsadi, G., Jiao, S., Jani, A., and Chong, W. (1991)
Conditions affecting direct gene transfer into rodent muscle in vivo.
BioTechniques 11, 474–485.
7. Wolff, A. J. (1994) Gene Therapeutics—Methods and Applications of Direct Gene
Transfer, Birkhäuser, Boston.
8. Schirmbeck, R., Böhm, W., Ando, K., Chisari, F. V., and Reimann, J. (1995)
Nucleic acid vaccination primes hepatitis B surface antigen-specific cytotoxic T
lymphocytes in nonresponder mice. J. Virol. 69, 5929–5934.
Production of Plasmid DNA 21
9. Davis, H. L., Schirmbeck, R., Reimann, J., and Whalen, R.G. (1995) DNA-medi-
ated immunization in mice induces a potent MHC class I-restricted cytotoxic T
lymphocyte response to hepatitis B virus surface antigen. Hum. Gene Ther. 6,
1447–1456.
10. Schmidt, T., Friehs, K., and Flaschel, E. (1996) Rapid determination of plasmid
copy number. J. Biotech. 49, 219–229.
11. Ish-Horowics, D., and Burke, J. F. (1981). Rapid and efficient cosmid cloning.
Nucleic Acid Res. 9, 2989–2998.
12. Cotten, M., Baker, A., Saltik, M., Wagner, E., and Buschle, M. (1994). Lipopolysac-
charide is a frequent contamination of plasmid DNA preparations and can be toxic
to primary cells in the presence of adenovirus. Gene Ther. 1, 239–246.
13. Weber, M., Möller, K., Welzeck, M., and Schorr, J. (1995) Effects of lipopolysac-
charide on transfection efficiency in eukaryotic cells. BioTechniques 19, 930–940.
14. Wicks, I. P., Howell, M. L., Hancock, T., Kohsaka, H., Olee, T., and Carson, D.
A. (1995) Bacterial lipopolysaccharide copurifies with plasmid DNA: implica-

tions for animal models and human gene therapy. Hum. Gene Ther. 6, 317–323.
15. Morrison, D. C. and Ryan, J. L. (1987) Endotoxins and disease mechanisms. Annu.
Rev. Med. 38, 417–432.
16. Vukajlovich, S. W., Hoffman, J., and Morrison, D. (1987) Activation of human
serum complement by bacterial lipopolysaccharides: Structural requirements for
antibody independent activation of the classical and alternative pathways. Mol.
Immunol. 24, 319–331.
17. Schorr, J., Moritz, P., Seddon, T., and Schleef, M. (1995) Plasmid DNA for hu-
man gene therapy and DNA vaccines. NY Acad. Sci. 772, 271–273.
18. Smith, H. A., Goldenthal, K. L., Vogel, F. R., Rabinovich, R., and Aguado, T.
(1997) Workshop on the control and standardization of nucleic acid vaccines.
Vaccine 15, 931–933.
19. Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Labo-
ratory Press, Cold Spring Harbor, NY, p. 443.
20. Wilfinger, W.W., Mackey, K., and Chomczynski, P. (1997) Effect of pH and ionic
strength on the spectrophotometric assessment of nucleic acid purification.
BioTechniques 22, 474–481.
21. Isner, J. M., Walsh, J., Symes, A., Pieczek, A., Takeshita, S., Lowry, J., et al.
(1995) Arterial gene therapy for therapeutic angiogenesis in patients with periph-
eral artery disease. Circulation 91, 2687–2692.

23
From:
Methods in Molecular Medicine, vol. 29, DNA Vaccines: Methods and Protocols
Edited by: D. B. Lowrie and R. G. Whalen Humana Press Inc., Totowa, NJ
3
Development and Characterization
of Lyophilized DNA Vaccine Formulations
Nancy L. Shen, Jukka Hartikka, Nancy A. Horn,
Marston Manthorpe, and Magda Marquet

1. Introduction
The potential applications of using plasmid DNA for immunization and other
gene therapy approaches have been discussed in an increasing number of pub-
lications in the past few years. Injection of mouse muscle with naked DNA
(plasmid DNA in saline) resulted in significant episomal expression from a
number of encoded reporter genes such as firefly luciferase, chloramphenicol
acetyltransferase, and β-galactosidase (1). DNA vaccination has been shown
to induce neutralizing antibodies against the gene product, helper T-cell
responses of the Th1 phenotype, and cytotoxic T lymphocyte responses (2).
Vaccination with plasmid DNA stimulates immunogenicity and provides pro-
tection against various infectious diseases in pre-clinical animal models.
Examples include hepatitis B in chimpanzees (3), bovine herpes virus in mice
(4), influenza A virus in ferrets (5), human immunodeficiency virus in rhesus
monkeys (6), Mycobacterium tuberculosis in mice (7,8), malaria in mice (9,10),
and genital herpes simplex virus in guinea pigs (11). Recently, DNA vaccines
for the protection against influenza (Merck Research Laboratories, Rahway,
NJ), malaria (Vical Inc., San Diego, CA), and HIV (Apollon Inc., Philadel-
phia, PA), have entered phase I human clinical trials. Rapid progress has been
made in the areas of adjuvants for DNA vaccines (12), route of immunization
(13), industrial scale fermentation and pharmaceutical grade purification (14).
One major interest in the commercial development of DNA vaccines, espe-
cially for developing countries, is to increase DNA vaccine stability at room
24 Shen et al.
temperature, to reduce the requirement for costly cold storage, and to extend
product shelf-life.
Freeze-drying, or lyophilization, has been used in pharmaceutical processes
to prolong product stability, particularly for protein products (15,16). Freeze-
drying is used for an attenuated virus vaccine against yellow fever (17) and a
live rinderpest virus vaccine for cattle (18). The freeze-drying process can be
divided into three successive stages: freezing, primary drying, and secondary

drying. After freezing the product, the primary drying process involves lower-
ing pressure and supplying heat for water vapor sublimation. During the sec-
ondary drying stage, the residual absorbed moisture evaporates from the dried
material. In this chapter, we describe a lyophilized DNA vaccine formulation
that provides acute protection during lyophilization and permits a full recovery
of product activity.
1.1. Screening Buffer and pH
To screen excipients used in lyophilized vaccine DNA formulations, we
evaluated buffers and pH using an in vivo reporter gene assay. Plasmid DNA
VR1223 encoding a gene for luciferase (19,20) was formulated and injected
intramuscularly into adult mouse rectus femoris muscle at 50 µg DNA in
50 µL volume. Injections were performed on 5 mice (10 muscles) for each
formulation. Luciferase enzyme activity was measured 7 d post-injection. The
results are shown in Table 1. There was no statistical difference in expression
from any of the tested formulations by non-parametric Mann-Whitney rank
sum test (p < 0.05). Either pH 6.0 or pH 7.0 was appropriate. Similarly, either
phosphate-buffered saline (PBS) or citrate-buffered saline permitted gene
expression. There were no adverse effects detected in injected mice.
1.2. Screening Lyoprotectants
Lyoprotectant, a required component in a lyophilized formulation, provides
protection of biological molecules from freezing and drying processes and
gives mechanical support to the finished product. To screen and select
lyoprotectants, various sugars and polymers were added to the liquid formula-
tion and tested in mouse muscles. The results are summarized in Table 2.
Considerable variation in luciferase enzyme expression was noted among vari-
ous concentrations of sugars or polymers, and among similar experiments per-
formed by using different batches of mice. There was no statistical difference in
expression from any of the tested lyoprotectants compared to PBS, pH 7.0, by
non-parametric Mann-Whitney rank sum test (p < 0.05). However, a 2- to 3-fold
enhancement was observed from formulations containing sugar lyoprotectant or

sugar/polymer lyoprotectant. The sugars included trehalose, mannitol, lactose,
sucrose, and sorbitol; and the polymers included polyvinyl pyrrolidone (PVP)