PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Gene Ther. Author manuscript; available in PMC 2012 October 8.
Published in final edited form as:
PMCID: PMC3465957
NIHMSID: NIHMS407518

Identification and removal of colanic acid from plasmid DNA preparations: implications for gene therapy

Abstract

Polysaccharide contaminants in plasmid DNA, including current good manufacturing practices (cGMP) clinical preparations, must be removed to provide the greatest safety and efficacy for use in gene therapy and other clinical applications. We developed assays and methods for the detection and removal of these polysaccharides, our Super Clean DNA (SC-DNA) process, and have shown that these contaminants in plasmid DNA preparations are responsible for toxicity observed post-injection in animals. Furthermore, these contaminants limit the efficacy of low and high doses of plasmid DNA administered by numerous delivery routes. In particular, colanic acid (CA) that is mainly long-chained, branched and has high molecular weight (MW) is most refractory when complexed to cationic delivery vehicles and injected intravenously (IV). Because CA is often extremely large and tightly intertwined with DNA, it must be degraded, in order, to be effectively removed. We have produced a recombinant, truncated colanic acid degrading enzyme (CAE) that successfully accomplishes this task. Initially, we isolated a newly identified CAE from a bacteriophage that required truncation for proper folding while retaining its full enzymatic activity during production. Any plasmid DNA preparation can be digested with CAE and further purified, providing a critical advance to non-viral gene therapy.

Keywords: plasmid DNA purification, polysaccharides, colanic acid, colanic acid degrading enzyme, boronate chromatography, non-viral delivery

Introduction

Investigators have observed toxicity leading to the death of animals following intravenous (IV) injections of plasmid DNA–liposome complexes at high doses, that is, greater than 80 µg of plasmid DNA in a mouse weighing 20 g.1 Mistakenly, some investigators believe that all cationic liposomes are toxic. However, some investigators, including our group, observed that administration of ‘next-generation’ liposomes, alone, at high doses without plasmid DNA does not cause toxicity or inflammation.1,2 Most investigators assume that plasmid DNA preparations, particularly clinical grade current good manufacturing practices (cGMP) plasmid preparations, are pure. However, optimized assays and methods to detect and remove contaminants from plasmid DNA preparations have not been reported. We have developed these assays and methods, and have shown that contaminants in plasmid DNA preparations are responsible for toxicity observed post-injection in animals. IV injections of high doses of DNA–liposome complexes may be required to treat certain diseases, for example, metastatic cancer. Furthermore, removal of contaminants from clinical grade DNA preparations most likely will increase safety and efficacy for patients receiving low doses of complexes because of immune responses produced from repeated administration of these contaminants. Currently, our liposomes,3 bi-lamellar invaginated vesicles (BIVs), are used in Phase I clinical trials to treat the late stage metastatic lung cancer4,5 and to treat hereditary inclusion body myopathy, a neuromuscular disorder (BBIND 13744). The lowest dose of IV injections (for example, 0.01 mg of DNA/kg body weight and less) require pretreatment of patients with the steroids, dexamethasone and diphenhydramine, to avoid grade 2 drug-related toxicities.4,5 Therefore, contaminants in plasmid DNA preparations must be identified and removed.

Cell wall polysaccharides are well-known contaminants of DNA that can be isolated from bacteria, yeast, plants, blue-green algae, protozoa, fungi, insects and mammals.68 In plasmid DNA, isolated from Escherichia coli (E. coli), the major known contaminating polysaccharide to date is lipopolysaccharide (LPS). This polysaccharide has attracted much attention because the lipid A moiety of LPS has an endotoxin activity and elicits a strong inflammatory response in mammals involving fever, decreased blood pressure, local inflammation and septic shock. Lipid A induces these responses through binding to serum lipopolysaccharide-binding protein and triggering signaling through the CD14 receptor, which is expressed on monocytes, endothelial cells and polymorphonuclear cells.9

Several methods are available to reduce the levels of LPS in plasmid DNA isolated from E.coli, including commercially available kits.10 These methods generally decrease the levels of polysaccharides linked to lipid A. However, lipid A is present in only a small fraction of the capsular polysaccharides of E. coli comprising group I.11,12 Several other major classes of polysaccharides are expressed by E. coli, including many that express capsular antigens O and K not bound to lipid A, colanic acid (CA) and enterobacterial common antigen. Some of these polysaccharides have phosphatidic acid as a lipid moiety, but this lipid is typically hydrolyzed during plasmid isolation procedures. Therefore, these polysaccharides cannot be removed from DNA by all currently available procedures that deplete endotoxin on the basis of hydrophobicity. Furthermore, endotoxin tests, for example, LAL, only measure the activity of lipid A and not that of polysaccharides.

Currently available DNA purification methods cannot quantitatively remove polysaccharides from DNA. Both DNA and polysaccharides are precipitated by organic solvents such as ethanol (EtOH) and by polyethylene glycol. Because polysaccharides are anionic, they co-purify with DNA by anion-exchange chromatography. The commercially available laboratory scale purification methods, including Qiagen, use anion-exchange chromatography for at least one major step in the purification process. Finally, on the basis of their similar buoyant densities, many polysaccharides copurify with DNA on cesium chloride density gradient centrifugation.6 Interestingly, investigators have had to purify CA using CA-overproducing bacteria that spew large amounts of CA into the growth media.13 These investigators chose not to isolate CA from the bacterial cell wall because they found it impossible to completely purify CA away from DNA because they are tightly intertwined.13 We have also observed the same result. Therefore, special purification methods are required to remove CA from plasmid DNA preparations. Furthermore, our data suggest that because CA is mainly long-chained, branched, and has high molecular weight (MW), this polysaccharide is most refractory to animals. CA is found in high abundance and comprises approximately 25% of the bacterial cell wall of gram-negative bacteria. Surprisingly, plant polysaccharides that contaminate plant genomic DNA preparations, inhibit both restriction endonuclease treatments and the PCR.8 Lectin-affinity chromatography has been used to remove polysaccharide contaminants from DNA isolated from fungi and plants,7 but the sugars recognized by this lectin are not present in most polysaccharides from E. coli. Therefore, new methods are needed to accomplish the separation of DNA plasmids from contaminating polysaccharides prepared in E. coli for use in gene therapy and other clinical applications.

We have produced several assays to assess polysaccharide contamination in plasmid DNA and have produced new methods for their removal. These contaminating polysaccharides are also present in cGMP, clinical grade plasmid DNA preparations and produce the following adverse effects when they are present at significant levels in DNA:

  1. contribute to poor complex formation when mixed with liposomes
  2. produce reduced gene expression in vitro and in vivo
  3. cause severe toxicity, including death, when complexed to liposomes and IV injected at high doses into animals
  4. cause immune responses when complexed to liposomes and IV injected at lower doses into animals

These polysaccharides inhibit DNA polymerase14 and RNA polymerase activities,15 and inhibit RNA synthesis.16 Several of these polysaccharides are also antigenic and have been shown to cause significant immune responses in animals.17 Investigators have reported inflammation and significant immune responses in animals after intramuscular injections of naked DNA,18 and the levels of these immune responses were identical to those produced after intramuscular injections of adenoviral vectors. We believe that these adverse effects produced by naked DNA, containing low levels of LPS and currently approved by the Food and Drug Administration for use in humans, are caused by the contaminating polysaccharides. Therefore, removal of these polysaccharides, particularly CA, is crucial to produce effective and safe gene therapy.

High-Performance liquid chromatography (HPLC) is used as a standard by the Food and Drug Administration to validate plasmid DNA purity for use in patients. However, HPLC cannot detect these contaminating polysaccharides. The MWs of the residual polysaccharides vary over too wide a range. Therefore, not enough polysaccharide at any given MW is present in a measurable quantity for detection by HPLC. In addition, the wavelength used for HPLC would need to be optimized for every different polysaccharide, and these wavelengths are not the optimized wavelength used for the detection of DNA. Optimization of HPLC wavelengths for polysaccharide detection is not ultimately useful because of the broad MW variation for any specific type of polysaccharide. Therefore, no significant peaks for any specific polysaccharide could ever be obtained by HPLC analysis. Furthermore, the high MW CA and other extremely large polysaccharides most likely cannot load onto the HPLC columns used for assessing plasmid DNA preparations.

Some investigators have attributed the observed plasmid toxicities solely to the presence of CpG sequences within the plasmid.19 They removed the majority of CpG sequences from their plasmids and reported reduced toxicity after IV injections of cationic liposomes complexed to these plasmids.19 However, only low doses containing up to 16.5 µg of DNA per injection into each mouse, approximately weighing around 20 g, were shown to reduce toxicity. To achieve efficacy for cancer metastases, particularly in mice bearing aggressive tumors, most investigators are interested in injecting higher doses in the range of 50–150 µg of DNA per 20 g mouse. Therefore, removal of CpG sequences from plasmid-based gene therapy vectors will not be useful for these applications because no difference in toxicity was shown after IV injections of these higher doses of plasmids, with or without reduced CpG sequences, complexed to liposomes.19 We believe that lack of removal of the contaminating polysaccharides in current DNA preparations, discussed above, is the major block to the safe IV injection of high doses of DNA–liposome complexes. Furthermore, use of Super Clean DNA (SC-DNA) provides greater efficacy and gene expression at lower doses compared with use of other plasmid DNA preparations of the identical plasmid.

CA is present at significant levels in all plasmid DNA, including cGMP clinical grade preparations. CA is also required for bacterial growth in large-scale shake flasks or fermentors because it protects gram-negative bacteria from stress.20 Therefore, large-scale plasmid production in CA-mutant bacterial strains is not possible, and only small 5 ml cultures can be effectively grown. Thus, CA must be removed in order to provide the greatest safety, especially when mixed with cationic carriers for delivery in animals and in humans. Removal of CA also increases gene expression from each plasmid because CA is an inhibitor of RNA polymerase activity and RNA synthesis. We have observed in the range of 2.2- to 4.4-fold increased reporter gene expression (for example, chloramphenicol acetyltransferase) in the organs of Balb/c mice post-IV injection of BIV–SC-DNAliposome complexes. We also observe no toxicity after IV injections of high doses of BIV–SC-DNA complexes. Because CA is often extremely large and is branched-chain, it must be degraded, in order, to be effectively removed. Initially, we produced a naturally occurring CA-degrading enzyme (CAE) from bacteriophage that is a newly identified protein; however, only small amounts could be produced, approximately 206 µg from a 4 l phage+ bacterial growth. Earlier attempts to produce full-length recombinant CAE in yeast, baculovirus and bacteria failed because of improper protein folding. We report production of a truncated, recombinant form of CAE for use in further purifying plasmid DNA preparations.

Results

Assays for the detection of polysaccharides in plasmid DNA preparations

We developed six different assays for the detection of contaminating polysaccharides in plasmid DNA preparations that include the following:

  1. uronic acid assay
  2. fucose assay
  3. DTAF ((4,6-dichloro- triazinyl)aminofluorescein) assay
  4. viscosity assay
  5. bicinchoninic acid (BCA) assay
  6. in vivo toxicity assay

For any purification process, multiple assays, each offering unique insights into purity, are useful. We believe that the lack of these assays to date is responsible for lesser quality of plasmid DNA that is currently produced including that used in clinical trials.

Uronic acid assay

Most polysaccharides present in bacteria, particularly the long-chain and branched polysaccharides found in plasmid DNA preparations, contain uronic acid. Furthermore, uronic acid is the major sugar found in bacterial polysaccharides and alternates with other sugars in the chain. Polysaccharides containing uronic acid have further modifications; therefore, uronic acid comprises approximately 10–35% of the total weight of the polysaccharide. Specifically, uronic acid comprises 11% of CA, 25% of capsular antigens O and K and 33% of enterobacterial common antigen. We modified a highly sensitive carbazole analysis assay that specifically quantifies the amount of uronic acid contained in polysaccharides (see Materials and methods for details).21 This assay detects sub-micro amounts of uronic acid, for example, 5 µg/ml of uronic acid. We adapted this assay for use in analyzing plasmid DNA. The total weight of polysaccharide contained in plasmid DNA preparations can be quantified by including standard curves using commercially available heparin sulfate or glucuronic acid. Heparin sulfate resembles the polysaccharide contaminant because uronic acid comprises about 25% of the total weight of heparin sulfate. Heparin sulfate consists of 50% sugars by weight. Half of the sugars are glucosamine, and the other half are iduronic acid and glucuronic acid. The other 50% of total weight of heparin sulfate is contributed by modifications including sulfates and acetylamides. Alternatively, glucuronic acid can be used to generate the standard curve. The total weight of polysaccharide in the unknown DNA sample can be conservatively estimated by multiplying the weight obtained for uronic acid in the unknown by four. We carried out our assays using both methods for creating the standard curves, and our results showed that both methods for quantifying the total weight of polysaccharide in plasmid DNA preparations are similar. We have assessed the weight of polysaccharide contamination in numerous plasmid DNA preparations including those obtained from companies that produce cGMP grade plasmid DNA for use in clinical trials. Table 1 shows the results obtained from our uronic acid assays. The endotoxin levels in the cGMP grade plasmid preparations were 0.02 EU/mg.

Table 1
Measurement of uronic acid in plasmid DNA preparations

Fucose assay

CA consists of 22% fucose in the ratio of 2:2:1:1:3 of fucose:galactose:glucose:uronic acid:other modifications; whereas, the other contaminating polysaccharides in plasmid DNA do not contain fucose. We modified a sensitive assay for the detection of fucose (see Materials and methods for details)22 for use with plasmid DNA. The results of our fucose assays showed that the vast majority of polysaccharides, greater than 90%, measured by the uronic acid assay are CA. For example, a plasmid DNA preparation containing about 0.7 mg polysaccharide per mg of DNA had 0.14 mg of fucose per mg of DNA, approximately 0.64 mg of CA. Therefore, it was generally found that the primary polysaccharide contaminant in plasmid DNA preparations was CA.

DTAF assay

This assay was used to detect lower MW polysaccharides in plasmid DNA preparations. We found that polysaccharides in plasmid DNA preparations are specifically labeled using DTAF obtained from Molecular Probes (Invitrogen Corporation, Carlsbad, CA, USA). DTAF specifically labels all polysaccharides whether or not they contain uronic acid (see Materials and methods for details). This fluorescence probe reacts with hydroxyl groups present in polysaccharides or carbohydrates. Therefore, DTAF does not label DNA because there are no free hydroxyl groups available on DNA. All hydroxyl groups in DNA are phosphorylated. We visualized DNA or polysaccharides in identical samples on agarose gel after electrophoresis was carried out. Ethidium bromide (EtBr), that labels DNA and not polysaccharides, was added only to the DNA sample that was loaded onto the gel for visualization of DNA. No EtBr was placed in the gel or in the gel running buffer. In addition, all plasmid DNA samples were overloaded in the lanes, loading 10 µg of DNA per lane, to further show that DTAF did not label DNA and that EtBr did not label the polysaccharides. Our results in Figure 1 show that DNA migrates at a different location than those of polysaccharides. The polysaccharide controls, LPS or detoxified LPS (Sigma-Aldrich, St Louis, MO, USA), were also loaded at 10 µg into lanes 10 and 8, respectively. Figure 1 also shows the cGMP plasmid DNA preparations that had the lowest amounts of uronic acid assessed in the uronic acid assay (Table 1). DTAF does not label SC-DNA.

Figure 1
Visualization of polysaccharides in plasmid DNA preparations by DTAF labeling. DTAF-labeled or EtBr-labeled plasmid DNA preparations were electrophoresed on a 1% agarose gel prepared in 1 × TAE that did not contain EtBr in the gel or in the 1 ...

Figure 1 shows that all the DNA samples contain polysaccharides that are also found in LPS (lane 10) and in detoxified LPS (D-LPS, lane 8). Furthermore, D-LPS is LPS from which lipid A has been removed. Therefore, removal of lipid A and endotoxin, does not remove the majority of polysaccharides. The only polysaccharides that cannot be visualized on the gel (Figure 1) are those that are large MW or highly branched and cannot migrate into the gel pores when subjected to electrophoresis. Therefore, the longchained, branched and high MW CA in all samples is either trapped in or moves out of the gel wells.

Production of CAE

Polysaccharide-degrading enzyme activities called polysaccharide depolymerases have been reported2326 although the enzymes were not purified to homogeneity and not sequenced. This enzymatic activity reduced the viscosity of polysaccharides in solution.25 CA is highly viscous in nature. We isolated a phage that lyses and produced plaques on CA-overproducing E. coli K-12 bacteria, SC12078.20 We also produced CA as previously described13 from the growth medium of SC12078 bacteria (see Materials and methods for details). We purified a CAE, of 84 354 MW, to homogeneity from our phage by ion-exchange chromatography followed by sizeexclusion chromatography (see Materials and methods for details). The fractions from the ion-exchange column (Figure 2a) that reduced the viscosity of CA were pooled and size fractionated (Figure 2b), and we isolated a single protein band that reduced the viscosity of CA. The isolated protein was subjected to analyses by mass spectrometry and Edman degradation to obtain amino-acid sequence data for nine different peptide fragments (see Materials and methods for details). Database searching using the peptide fragment amino acid data showed no significant hits and showed that our CAE is a newly identified phage protein (see Materials and methods for details). Degenerated oligos were prepared on the basis of the protein sequence data and were used to sequence the phage genomic DNA (see Materials and methods for details). The entire CAE coding region plus approximately 2 kb of flanking sequences at the 5′ and 3′ ends were sequenced on both strands, and the open reading frame (ORF) has a total of 2373 bp nucleotides (Figure 3; GenBank Accession Number HM214492). Protein structure homology searches using InterProScan (EMBL-EBI, European Bioinformatics Institute) showed some homology to virulence factor, pectin lyase fold at amino acids from 125–542; and showed high homology to a protein involved in CA synthesis, WCAM_- SHIFL_P37775, at amino acids 231–464.

Figure 2
Isolation of the colanic acid (CA)-degrading enzyme. (a) The processed supernatant from the NST1 phage+SC12078 bacterial growth was fractionated by ion-exchange chromatography on a Q Sepharose Fast Flow column (GE Healthcare, Piscataway, NJ, USA). Fractions ...
Figure 3
DNA sequence of the open reading frame (ORF) for the CA-degrading enzyme (CAE). The complete ORF for the CAE was sequenced on both strands obtained from the NST1 genomic DNA. The initial sequencing was performed using degenerate oligos designed from amino-acid ...

Our earlier attempts to produce full-length recombinant CAE in yeast (Pichia pastoris), baculovirus (SF9) and bacteria (E. coli) failed because of improper protein folding. We then made a truncated recombinant CAE, 73 197 MW (see Materials and methods for details). After examining the predicted structure of our CAE and chymotrypsin digestion of the full-length protein, we determined that 107 amino acids could be removed from the N-terminus without loss of activity. Chymotrypsin was the only protease that cleaved the natural full-length protein at this one location, namely, amino acids 106–107. Our CAE was not cleaved by other proteases including elastase, endoproteinase GluC, thermolysin, trypsin and proteinase K. Therefore, the recombinant, truncated CAE, that is devoid of the N-terminal 107 amino acids, is not cleaved by any protease and is extremely stable (to date tested out for 3 years, stored at 4 °C). CAE contains an N-terminal 6 His-tag that does not reduce its activity; therefore, CAE can be easily isolated on a Ni-NTA column (see Materials and methods for details; Invitrogen). We cloned CAE using the bacterial expression vector, pET-28a-c(+) vector (Novagen, EMD4Biosciences, San Diego, CA, USA). CAE is produced in the E. coli host, BL21 (DE3) grown in hyper broth medium containing kanamycin, and then purified (see Materials and methods for details). We produced approximately 10 mg of CAE from 1 l of growth in shake flasks; whereas only 206 µg of the natural, full-length enzyme was produced from a 4 l bacteriophage+bacterial growth.

Viscosity assay

We used the viscosity assay for many purposes, including checking the activity of CAE, and assessing the viscosity of plasmid DNA before and after CA digestion and removal. We used a Wells-Brookfield Cone Plate viscometer (Cat # LVDV-1+CP; Brookfield, Middleboro, MA, USA) with a CPE-40 cone (see Materials and methods for details). This set-up provides the most sensitive measurement of changes in viscosity in the smallest volume, 500 µl, and samples can be retrieved if needed. Viscometer accuracy is checked by measuring the viscosity of the standard, namely mineral oil. Measurements are taken at 100 r.p.m. for 30 s for a 500 µl sample. Viscosity readings are recorded in CentiPoise (cP), and subsequent calculations are based on the fact that water has low viscosity and has a cP value of 1.00. CAE at 100% activity would reduce the viscosity of CA to a cP of 1.00. Generally, samples vary from 1.57 to 1.00 cP, an approximate 64% difference in viscosity between these values. A test sample that would read 1.57 cP pre-digestion with CAE to 1.29 postdigestion indicates about 50% decreased viscosity. The Sutherland group measured 50% decreased viscosity in polysaccharides after incubation with their partially purified polysaccharide depolymerase over time out to 2 h.25 Our test samples were incubated with CAE at 37 °C for 1 h. Our data showed that concentrations of CAE at 0.5 µg and greater fully-digested 545 µg of CA; whereas 0.1 µg of CAE showed 55% activity in digesting this amount of CA. Incubation with CAE also reduces viscosity in plasmid DNA preparations.

Bicinchoninic acid (BCA) assay

Use of a highly sensitive BCA assay to measure the number of reducing ends of carbohydrates has been published.26,27 This assay can be used to assess the degradation of any polysaccharide. We used this assay primarily to detect extremely small CA or other polysaccharide fragments remaining in plasmid DNA preparations through the SC-DNA process that may not be detected by our other assays (see Materials and methods for details). However, this assay was also used to assess CAE activity and the amount of CA degradation in plasmid DNA post-digestion with CAE. Assay data were generated by measuring absorbance at 550 nm in a plate reader. If no polysaccharide ends were present, the reading was 0.0. Previous publications reported absorbance readings that ranged from 0.0–0.427. An increase in absorbance readings was observed post-digestion with CAE, and prior to removal of digested CA and other polysaccharides, because CA was cleaved into many fragments. Therefore, far more reducing ends were present post-digestion of CA. Readings in the low range indicated that either few polysaccharide ends were present (post-digestion and partial removal of polysaccharides) or that the CAE digestion was not complete (post-digestion). SC-DNA has an absorbance reading of 0.0 in this assay.

Production of SC-DNA

Our Super Clean process removes all contaminating polysaccharides including CA from plasmid DNA preparations. For all of our studies shown in Tables 15, we used good laboratories procedures with statement of purpose to produce SC-DNA, and bacterial growth was performed in shake flasks (see Materials and methods for details). We started by optimizing growth of bacteria in order to obtain the maximum plasmid DNA:biomass ratio. By doing this, we reduced the overall amount of polysaccharide present upfront in the purification process. It was noted that bacteria reach an OD600 of 4.0 at 11– 12 h post-inoculation of the starter culture (Qiagen Plasmid Purification Handbook 07/99, page 60). Furthermore, the maximal amount of plasmid DNA is produced from growth to an OD600 of 4.0. However, bacteria continue to grow to an OD600 of 5.0, if left in the incubator longer, with no further production of plasmid DNA. Therefore, we inoculated the ‘overnight’ growth late in the evening and started monitoring the OD600 in the morning, about 10 h post-inoculation. All bacterial growth was stopped at an OD600 of 4.0 measured accurately using a turbidity cell holder in a DU 600 spectrophotometer (Beckman Coulter, Brea, CA, USA).

Table 5
Toxicity of Qiagen-endofree purified DNA versus SC-DNA–BIV complexes (in human PANC-1 tumor bearing SCID mice)

We prepared plasmid DNA initially using anion-exchange chromatography (Qiagen EndoFree, Qiagen, Valencia, CA, USA). The plasmid DNA with contaminating polysaccharides was then digested with CAE at a weight ratio of 10:1 plasmid DNA:CAE, and the majority of polysaccharide fragments including digested CA were removed using boronate-chromatography (Pierce, Thermo Scientific, Rockford, IL, USA) (see Materials and methods for details). Any contaminant containing vicinal diols binds to the boronic acid column including small MW RNA, thymine glycol-containing DNA, and benzo(a)pyrene.2831 Polysaccharides also contain vicinal diols, and we found that polysaccharides of a certain size range including digested CA bound to boronic acid columns. The plasmid DNA eluted in the void volume and does not bind the boronic acid column. We found that extremely small fragments of CA were not able to bind the boronic acid column. Therefore, we concentrated the plasmid DNA and performed the final cleanup step using a Macrosep 100 Centrifugal Concentrator Unit (Pall Life Sciences Corporation, Port Washington, NY, USA) with a 300 k cutoff in the presence of 0.1% zwittergent (see Materials and methods for details) to remove small fragments of digested CA or other small polysaccharides.

In vivo toxicity assay

We performed in vivo studies in mice to test our SC-DNA preparations. These studies were performed in normal mice (Balb/c), and in severe combined immunodeficiency (SCID) mice with or without human pancreatic PANC-1 tumors (see Materials and methods for details). Our SCID mice are more sensitive to CA and die after IV injections containing 40 µg of commercially produced plasmid DNA complexed to liposomes and other cationic carriers, whereas Balb/c mice die at levels just above 50 µg of commercially produced plasmid DNA. In Tables 25, we showed that a total of 120 mice survived high doses (100 µg) of SC-DNA–BIV complexes post-IV injections in comparison with universal mortality with commercially produced DNA–BIV complexes at 100 µg dosing, regardless of methods used for plasmid DNA purification. The Company 1 and Company 2 cGMP plasmid DNA preparations (Table 2) are those tested in the uronic acid assay (Table 1) and DTAF assay (Figure 1). Adding contaminants back to SC-DNA preparations restores the universal mortality seen using non-SC-DNA–BIV complexes (Table 3). Furthermore, mice IV injected with SC-DNA–BIV complexes at all doses appeared healthy at all times, post-injection, that is, no ruffled or raised fur was observed. Use of commercially prepared plasmid DNA caused ruffled fur in mice that survived post-IV injections of BIV–non-SC-DNA complexes and that were injected at lower doses, 50 µg of DNA in Balb/c and 35 µg of DNA in SCID mice.

Table 2
Toxicity of third party cGMP and Qiagen-endofree purified DNA versus SC-DNA–BIV complexes (in Balb/c mice)
Table 3
Toxicity of contaminants added back versus SC-DNA–BIV complexes (in Balb/c mice)

Discussion

We have identified, purified and sequenced the ORF for a newly identified CAE. Other CA degrading activities have been identified; however, these enzymes have never been purified to homogeneity and have not been sequenced.2326 Furthermore, one of these hydrolases was shown to be part of a high MW multi-protein complex consisting of at least six different proteins, all of which were required for activity.26 Enzyme complexes of this type would be extremely difficult to use in the process of further purifying plasmid DNA, particularly for cGMP manufacture, or for studies of CA structure– function relationships. Generally, CAE requiring just a single protein have only been isolated from bacteriophages2325 and such autonomous activity is clearly established in our recombinant protein. To our knowledge, our study is the first report of a fully-characterized CAE that has been cloned and produced large-scale using a recombinant form of the enzyme (CAE). We found that the full-length enzyme was impossible to express and purify in a soluble form. Designing CAE was particularly challenging and required the creation of a truncated enzyme that could be properly folded during large-scale production in E. coli without loss of activity. This enzyme has multiple potential uses for a wide variety of molecular studies and in other industrial applications.

The sequence homology search of the protein databases did not yield any significant candidate protein. The secondary structure prediction of the CAE indicated that it has a short β-sheet repeat across the entire protein in conjunction with a few short helices. Moreover, some bacteriophages appear to have a homotrimeric tailspike protein that recognizes cell surface oligosaccharides as receptors. These tailspike proteins generally appear to adopt a right-handed β-helix fold32 with characteristic short β-sheets on their secondary structures. The initial protein structure homology searches using InterProScan showed some homology to virulence factor, pectin lyasefold, which is a typical example of such a β-helix fold.33 After close examination of the same class of β-helix proteins with 3D-structures available, we found that the core structure of the CAE (amino-acids 118–472) has better sequence homology to the phage P22 tailspike protein (amino-acids 125–542) as compared with other proteins.34 The predicted secondary structures of the two proteins also showed higher secondary structure similarity. Thus, we hypothesized that the core domain of CAE (amino acids 118–472) may adopt a similar fold and quaternary structure to the phage P22 tailspike protein. Figure 4 shows the modeling of the core domain of the CAE with the phage P22 tailspike protein as template (PDB code: 1TYX). The possible substrate recognition site is indicated as a ball model shown in yellow.

Figure 4
Predicted core structure for the CA-degrading enzyme (CAE). (a) The side-view for the modeling of the core domain of the CAE with the phage P22 tailspike protein as the template (PDB code: 1TYX). The possible substrate recognition site is indicated as ...

CAE was suitable for our application in providing for the efficient degradation and subsequent effective removal of CA from plasmid DNA preparations. CA is tightly intertwined with DNA and cannot be removed by standard chemical or biophysical methods, and therefore, highly purified CA must be isolated from the supernatants of overproducing CA strains grown in culture that spew CA into the growth medium.13 Conversely, removal of the ubiquitous CA contamination from plasmids produced in E. coli has proven extremely difficult. We tried a variety of different chemical methods, including gel filtration in the presence of zwitter-ion; however, none of these methods effectively removed CA from plasmid DNA preparations. CA is particularly refractory when complexed to cationic carriers and IV injected into animals because it is of high MW, is long-chained and branched. Therefore, it must be removed to provide for the greatest safety and efficacy. However, after digestion with CAE, boronate chromatography can be used to remove digested fragments of CA, while simultaneously removing other residual polysaccharides, as well as low MW RNA present in plasmid DNA preparations. Removal of CA and other contaminating polysaccharides is a critical advance to non-viral gene therapy. In addition, DNA vaccines could perhaps be more effective with the removal of CA. Although these naked plasmid DNA-based vaccines do not cause death in animals, CA could act as an adjuvant that is not controlled. The field of non-viral gene therapy has been plagued with toxicities associated with plasmid DNA for greater than 10 years,1,2 particularly for liposomal and other cationic plasmid-based delivery systems IV injected at higher doses. Many investigators, including our laboratory, have shown that IV delivery of high doses of ‘next-generation’ liposomes alone produce no toxicity, no adverse effects and no inflammation.1,2

Toxicity of liposomal plasmid DNA complexes has been considered to be an intrinsic property of the DNA and attributed to the presence of ‘CpG islands’ in bacterial DNA sequences, which can trigger TLR9- dependent and TLR9-independent immune recognition. 35 This led to efforts to delete or mutate such sequences in plasmids intended for therapeutic use. However, removal of CpG islands in plasmid DNA sequences did not solve this problem for high-dose IV injections of plasmid DNA delivered by cationic liposomal delivery systems.19 We propose that CA and other contaminating polysaccharides in plasmid DNA are delivered efficiently into cells by cationic carriers; whereas, the same contaminants in plasmid DNA injected IV with no carrier are not introduced into cells as efficiently, thereby increasing the amount of plasmid DNA that can be injected IV without causing acute toxicity. BIV complexes efficiently encapsulate plasmid DNA and enter cells through a fusogenic pathway,3,36 and therefore, plasmid DNA could avoid inflammation mediated through toll-like receptors on the cell surface. CA associated with plasmid DNA encapsulated in BIVs may induce specific inflammatory responses through an intracellular pattern recognition molecule-mediated pathway when efficiently delivered into cells.3743 These pattern recognition molecules include intracellular toll-like receptors, nucleotide-binding domain and leucinerich repeat containing receptors, intracellular glycan binding receptors (lectins) and intracellular retinoic acid-inducible gene-1-like receptors (reviewed in4143). Binding of CA to intracellular pattern recognition molecules could trigger a cytoplasmic adaptor molecule to induce an innate immune response by causing migration of factors, for example, nuclear factor kappa B, to the nucleus that regulate expression of proinflammatory cytokines and adhesion molecules, and can also induce immune cell maturation for generating effective immune responses (reviewed in38,40,41). For example, macrophages and dendritic cells in mammals and humans are activated by microbial components considered to be non-self.38 Interestingly, heparin sulfate induces increased levels of nuclear factor kappa B in the nucleus in a dose-dependent manner and causes maturation of dendritic cells. The signal transduction pathway for potential CA-mediated inflammatory and immune responses remains to be elucidated.

We showed by using six different assays that contaminating polysaccharides including CA are present in plasmid DNA preparations, including cGMP manufactured DNA used in clinical trials. Removal of these contaminants allowed for high-dose IV injections of BIV–DNA complexes without acute toxicities or any adverse effects (Tables 25). In fact, all mice including Balb/c, SCID and SCID tumor-bearing mice appeared as non-injected mice at all times post-IV injections. This indicates that CA contamination is a major contributor to the observed toxicity of cationic liposomal plasmid DNA complexes. Currently, we are including SC-DNA processing into our cGMP manufacture of plasmid DNA for animal studies and for use in clinical trials. CAE can be placed on a solid support, thus, precluding presence of CAE in the final preparation and allowing repeated use of the CAE solid support for the production of identical plasmids. In addition, we have produced peptide-based CAE antibodies that can be used to detect any residual CAE by western blotting or by ELISA in the final plasmid DNA preparations. Removal of CA and other contaminating polysaccharides is a major advance that should allow for a broader based use of plasmid DNA in non-viral delivery and gene therapy, providing greater safety and efficacy.

Materials and methods

Uronic acid assay

We modified a highly sensitive carbazole analysis assay that specifically quantifies the amount of uronic acid contained in polysaccharides.21 Briefly, 40 µg of each plasmid DNA was diluted in sterile water for irrigation USP (sterile water; B.Braun Medical, Irvine, CA, USA). The standard, for example, glucuronic acid, was diluted from a 1 mg/ml stock solution in sterile water to concentrations ranging from 0.005–0.5 mg/ml, and water as control was also prepared. All samples and controls were placed into 13 × 100 mm glass test tubes. A 0.025 m sodium tetraborate-decahydrate solution was prepared in concentrated sulfuric acid (specific gravity, 1.84). The borate was added to the sulfuric acid slowly to prevent clumps of fused borax from foaming. A 0.125% carbazole solution was also prepared by dissolving carbozole in absolute EtOH. Just before heating, 3 ml of the borate/sulfuric acid solution followed by 100 µl of the carbazole reagent were added to each tube. Glass disposable pipettes were used to dispense the borate/ sulfuric acid solution. Each tube was vortexed well, covered with a glass marble and placed into a suitable wire rack in a water bath. The water bath was set to boil, and the tubes were boiled for 10 min while maintaining constant boiling. The tubes were then placed at room temperature (RT) for 10 min. Each tube was vortexed and placed at RT for 5 min to allow any air bubbles to dissipate. The samples and the standards were measured for absorbance at 530 nm. All samples and standards were assayed in duplicate. The concentrations of polysaccharides in the samples were calculated on the basis of the linear regression data generated from the glucuronic acid standard curve.

Fucose assay

We modified an assay for measuring sub-nanomole amounts of fucose using enzymatic cycling.22 Each test sample was first prepared for analysis by lyophilizing 450 µg of plasmid DNA in a glass vial. In total, 200 µl of 5.5 m trifluoroacetic acid was added to each DNA sample, and the vials were tightly sealed with Teflonlined caps. The samples were then hydrolyzed by heating for 4 h at 100 °C. The samples were cooled to RT and the trifluoroacetic acid was removed with a passage of stream of argon gas under the fume hood. The DNA residue in each vial was dissolved in 200 µl of sterile water. Fucose standards were prepared ranging from 0.005–1 mg fucose/ml sterile water. Sterile water was used as a standard for the negative control. 200 µl of each test sample and the standards were aliquoted into sterile 1.5 ml eppendorf tubes and placed on ice. Then 50 µl of 1 mg/ml fucose dehydrogenase (Kikkoman Corporation, Walworth, WI, USA) were added to each tube followed by the addition of 50 µl of 200 µm nicotinamide adenine dinucleotide. The tubes were flicked to mix and incubated at 4 °C for 3 h. Then 50 µl of 1N NaOH were added to each tube. The tubes were flicked to mix and incubated for 10 min at 60 °C to stop the enzymatic reaction. The tubes were cooled to RT and neutralized with 50 µl of 1 m HCl. In total, 50 µl from each tube were removed and placed into a new sterile 1.5 ml eppendorf tube. Then 250 µl of the cycling reagent was added to each tube that was then flicked to mix. The cycling reagent consists of 200 mm Tris, pH 8.4, 50 mm ammonium acetate, 0.5 mm adenosine di-phosphate, 100 mm lactate, 5 mm alpha-ketoglutarate, 20 units/ml lactate dehydrogenase and 20 units/ml glutamate dehydrogenase. The tubes were incubated at RT for 1 h. The reaction was stopped by heating the tubes for 2 min in boiling water. The tubes were chilled on ice and 250 µl of the pyruvate reagent were added to each tube and flicked to mix. The pyruvate reagent consists of 800 mm imidazole buffer, pH 6.2, 0.45 mm NADH and 0.06 units/ml lactate dehydrogenase. The tubes were warmed to RT in a water bath for 2 min and then placed at 30 °C in an incubator for 20 min. The reaction was halted by addition of 200 µl of 1.5M HCl. The contents of each tube were transferred to 15 ml Falcon tubes. Then 2.5 ml of 6N NaOH were added to each tube and flicked to mix. The tubes were incubated at 60 °C for 10 min. The tubes were then cooled to RT and 4 ml of sterile water was added to each tube and inverted to mix. In total, 300 µl from each tube were placed in a white-bottomed 96-well microtiter plate (Thermo Labsystems, Thermo Scientific) for fluorescence measurements. Samples were assayed in triplicate on a fluorescence P-E plate reader (FLUOstar OPTIMA, BMG Labtech, Cary, NC, USA) set at the following parameters: excitation filter at 360 nm, emission filter at 465 nm, gain at optimal, lag time at 0 µs, integration time at 40 µsec, flashes at 10 µs and time between move and flash at 0 µs. The concentrations of fucose in the DNA samples were calculated on the basis of the linear regression data generated from the fucose standard curve.

DTAF labeling

This assay was used to detect lower MWpolysaccharides in plasmid DNA preparations. DTAF was purchased from Molecular Probes (Invitrogen) for use in labeling of polysaccharides in plasmid DNA. For each sample, 80 µg of DNA in 40 µl were transferred into 500 µl eppendorf tubes. One tube containing 40 µl of sterile water was included as a control that was treated identically to the DNA samples. The DNA was precipitated with 10 µl of 3 m NaOAc, pH 5.2, and 200 µl cold 100% EtOH. Each sample was vortexed for 5–10 s. The tubes were incubated for 30 min at −20 °C followed by centrifugation at 10 000 r.p.m. for 4 min at 4 °C. The supernatants were discarded, and 10 µl of 3 m NaOAc pH 5.2 and 200 µl cold 100% EtOH were added to the pellets. The tubes were vortexed and centrifuged as stated above. The DNA precipitation and wash process was repeated, except that only 200 µl cold 100% EtOH was added. The tubes were centrifuged and the supernatant discarded. The pellets were dissolved in 50 µl of 0.1 m Na2CO3, pH 10.5, at RT. A 60 mg/ml slurry of DTAF in 0.1 m Na2CO3, pH 10.5, was prepared in a 1.5 ml eppendorf tube that provided enough slurry to add a total of 15 µl to each tube. The DTAF slurry was kept cold on ice and in the dark in a covered ice bucket. 5 µl of the freshly vortexed DTAF slurry was added to each tube, and the DTAF slurry was vortexed every time before addition to each tube. Sample reactions were kept at RT and in the dark in a covered storage box during the course of the reaction. The addition of 5 µl DTAF to each tube was repeated two more times at 45 and 90 min after the initial addition. After a total of 2.5 h (3 × 45 min beyond each addition of DTAF slurry), the reactions were stopped by the addition of 10 µl of 3 M NaOAc, pH 5.2, and 325 µl cold 100% EtOH. All tubes were placed at −20 °C for 45 min. The tubes were centrifuged at 10 000 r.p.m. for 4 min at 4 °C. The pellets were washed with 25 µl of 3 m NaOAc, pH 5.2, and 500 µl cold 100% EtOH, vortexed and centrifuged as stated above. This process was repeated for a total of three washes. The final wash was performed with 500 µl of cold 100% EtOH, vortexed well and centrifuged as stated above. All of the supernatant was removed from each tube and the pellets were air dried for 20 min in the dark (a longer time would make the pellets difficult to dissolve). The pellets were dissolved in 40 µl of 1 × TAE buffer. For gel analysis a 1% agarose gel was prepared in 1 × TAE buffer without including EtBr. 2 µl of a clear gel loading solution (lacking bromophenol blue or any other dye) and 3 µl of 1 × TAE buffer were added to 5 µl of each DTAF-labeled sample. Bromophenol blue was not added to the DTAF-labeled lanes because it quenches the fluorescence. The DNAwas analyzed on the same gel by adding 5 µl of a 1:100 EtBr solution to 5 µg of DNA and 2 µl of gel-loading solution containing bromophenol blue. About 3 µg of a mix of the high MW DNA marker, lambda HindIII and low MW markers was included and was prepared identically to the other DNA samples containing EtBr. The samples were pipetted into the wells leaving an empty well between each DTAF-labeled sample. All lanes were heavily loaded with 10 µg of plasmid DNA per lane to show that polysaccharides migrate differently from plasmid DNA from identical samples. The 1% TAE-agarose gel was electrophoresed for approximately 3 h at 180 V in 1 × TAE running buffer that did not contain EtBr. The gel was checked periodically with a handheld UV lamp to monitor the progress. The bromophenol blue (MW = 670.0) in the DNA lanes was used to determine the progress of the separation by electrophoresis, and it has a higher MW than unbound DTAF (MW = 495.3). The data (Figure 1) show that all the DNA samples contain polysaccharides that are also found in LPS (lane 10) and in detoxified LPS (D-LPS, lane 8). Furthermore, D-LPS is LPS from which lipid A has been removed. Therefore, removal of lipid A and endotoxin, does not remove the majority of polysaccharides. The only polysaccharides that cannot be visualized on the gel are those that are extremely large MW or highly branched and cannot migrate into the gel pores when subjected to electrophoresis, including the long-chained, branched and high MW CA.

Preparation of CA

We produced CA as previously described13 from the growth medium of SC12078 bacteria K-12, a CA-overproducing strain.20 Briefly, SC12078 was inoculated into 2 l of Luria broth (LB) medium containing 0.4% glycerol and 10 µg/ml chloramphenicol. Use of chloramphenicol is required to maintain this CA-overproducing strain. The culture grew at 37 °C in a shaker incubator at 230 r.p.m. overnight (O/N) to an OD600 between 4.5–4.7. The cells were shaken manually in the flask briefly to obtain more CA into the culture medium. The culture was centrifuged at 6000 × g for 15 min at 4 °C. The supernatants were pooled and the pellet was discarded. The volume of the supernatant was reduced using an Amicon filter apparatus and the YM30 membrane according to manufacturer’s instructions (Millipore, Billerica, MA, USA). Three volumes of icecold EtOH were added to the retentate and placed on ice for 15 min to precipitate the CA. The precipitate was collected by centrifugation at 10 000 × g for 15 min at 0 °C to yield a clear supernatant. The pellet was dissolved in sterile water using the least amount of water required. The mixture was dialyzed in sterile water O/N at 4 °C with three changes. The empty tube that would store the pellet was weighed. The mixture was then placed into this tube and lyophilized to dryness. The tube containing the pellet was then weighed. The final weight of the pellet was determined by subtracting the weight of the empty tube plus cap from the weight of the tube and cap containing the pellet. A 2% solution of the pellet was prepared by dissolving in sterile water. Solid ammonium sulfate was added to saturate the sample to 90%. At this point in the process, both antigen O and CAwere precipitated. The precipitate was collected by centrifugation at 10 000 × g for 15 min at 0 °C to yield a clear supernatant. The pellet was dissolved in sterile water using the least amount of water required. The sample was dialyzed in sterile water O/N at 4 °C with three changes. The sample was lyophilized to dryness and dissolved in 150 ml of 0.1 M sodium phosphate, pH 7.2. Then 37.5 ml of hexadecyltrimethylammonium bromide (cetavlon; also named cetrimide) were added to precipitate the CA. Cetavlon was prepared by dissolving 2.5 g of cetrimide in 100 ml of 2% NaOH. This solution was heated briefly at 40 °C before use until a uniform solution appeared with no particulates. The precipitate was collected by centrifugation at 10 000 × g for 15 min at 0 °C to yield a clear supernatant. The pellet was dissolved in 100 ml of 1 m NaCl. Then three volumes of ice-cold EtOH were added, and the mixture was placed on ice for 15 min to precipitate the CA. The precipitate was collected by centrifugation at 10 000 × g for 15 min at 0 °C to yield a clear supernatant. The pellet was dissolved in sterile water using the least amount of water required and dialyzed in sterile water O/N at 4 °C with three changes. The sample was lyophilized to dryness and the weight of CA was determined as stated above. The pellet was dissolved in sterile water using the least amount of water required, and the percentage solution was recorded and listed on the storage vial labels. The vials containing CA were stored at −25 °C.

Bacteriophage production

We isolated a previously unidentified bacteriophage (phage), the NST1 phage, by its ability to form plaques on SC12078 bacteria. Briefly, SC12078 was inoculated into tubes with 5 ml of LB medium containing 0.4% glycerol and 10 µg/ml chloramphenicol and grown O/N. In total, 10-fold serial dilutions were prepared that contained five different NST1 phage particle numbers including 102, 101, 100, 10−1 and 10−2. Particle number was based on 1 µl of phage stock containing approximately 107 phage particles. 200 µl of the O/N growth were mixed with 1 µl of the phage stock to make the 107 concentration of phage. Dilutions containing 180 µl of O/N bacterial growth were mixed with 20 µl of the next higher concentration of phage. These dilutions were plated by quickly mixing each into 3 ml of LB+glycerol top agar, 0.7% agarose at 55 °C. The mixtures were quickly poured onto LB+10 µg/ml chloroamphenicol plates. The plates were incubated upside down at 37 °C for 5 h. After incubation, the plates containing phage plaques were stored at 4 °C. This process was used both to isolate the phage initially and was repeated once per month to maintain the NST1 phage. Growth of this phage was performed by inoculating phage plugs in the presence of SC12078 O/N cultures. Briefly, SC12078 was inoculated into LB medium containing 0.4% glycerol and 10 µg/ml chloramphenicol and allowed to grow at 37 °C in a shaker incubator O/N at 230 r.p.m. In total, 15 ml of the O/N culture were inoculated into each 1.5 l of LB medium containing 0.4% glycerol and 10 µg/ml chloramphenicol in three 4 l flasks. These cultures were allowed to grow at 37 °C in a shaker incubator at 230 r.p.m to an OD600 between 0.12–0.67, for approximately 2–4 h. Each 1.5 l SC12078 culture was inoculated with 30 NST1 phage plugs and grown at 37 °C with shaking at 230 r.p.m O/N. Growth was continued to an OD600 of 4.5–4.7. The cultures were centrifuged at 4200 r.p.m at 4 °C for 5 min. The supernatants containing the NST1 phage were collected for use in purifying the full-length CAE and were stored at −80 °C. The pellets containing bacterial cells and debris were discarded.

Purification of the CAE

Phenylmethanesulfonylfluoride (PMSF) was diluted to a final concentration of 0.1 mm and was added to 4 l of NST1 phage supernatant and placed at 4 °C. Using an Amicon filter apparatus and the YM30 membrane according to manufacturer’s instructions, the volume was reduced from 4 l to 4ml at 4 °C. The 4ml retentate was centrifuged at 40 000 × g for 60min at 4 °C. The supernatant was dialyzed with 10 mm Tris HCl, pH 7.5, 0.1 mm PMSF O/N at 4 °C with three changes. A Q Sepharose Fast Flow column (Pharmacia, GE Healthcare), 10 cm high and 1.5 cm diameter, was equilibrated with 10 mm Tris HCl, pH 7.5, and 0.1 mm PMSF until the pH of the solution eluting from the column was 7.5. The dialyzed supernatant was loaded onto the equilibrated Q Sepharose column and washed with two column volumes of 10mm Tris HCl, pH 7.5, 0.1 mm PMSF, approximately 30 ml. The column was eluted using a linear gradient from 10 mm Tris HCl, pH 7.5, 0.1 mm PMSF (150 ml) to 200 mm Tris-HCl, pH 6.5, 0.1 mm PMSF (150 ml) collecting 4 ml fractions (75 fractions in total) at a flow rate of 7 ml per h. The fractions were tested for CA-degrading activity using a viscometer postdigestion of CA (see protocol below), and the fractions containing the activity were pooled. The pooled fractions were concentrated on a disposable Amicon filter by centrifugation according to the manufacturer’s instructions. The protein concentration of the Amicon retentate was determined (Micro BCA; Pierce), and the sample was electrophoresed on a 4–12% gradient polyacrylamide gel and stained with Colloidal Blue (Novex, Invitrogen). A 120 cm column containing Toyopearl HW-50F (Catalog Number 07453; TosoHaas, Tosoh Bioscience LLC, King of Prussia, PA, USA) resin was equilibrated with phosphate-buffered saline pH 7.3–7.4, 0.1 mm PMSF. The retentate was fractionated on the equilibrated Toyopearl HW-50F column (TosoHaas, Tosoh Biosciences LLC), collecting 1 ml fractions. The fractions were tested for CAdegrading activity, and fractions containing the activity were pooled. The pooled fractions were concentrated on a disposable Amicon filter by centrifugation according to the manufacturer’s instructions. The protein concentration of the Amicon retentate was determined, and the sample was electrophoresed on a 4–12% gradient polyacrylamide gel and stained with Colloidal Blue. Approximately 206 µg of CAE at single-band purity was produced by this method.

Viscosity assay

A Wells-Brookfield Cone Plate viscometer (Catalog # LVDV-1+CP) with a CPE-40 cone was used to measure viscosity. This set-up provides the most sensitive measurement of changes in viscosity in the smallest volume, 500 µl. Viscometer accuracy was checked by measuring the viscosity of the standard, mineral oil. A solution of 1.2 mg/ml of CA was prepared in 0.05 m potassium phosphate buffer, pH 6.5. CA is highly viscous. All other samples were also prepared in this buffer. Dilutions of CAE (or other test samples) were added to the CA solution using identical volumes of this solution for each test or control sample. Approximately, 545 µg of CA were used in each sample, except for measurements of plasmid DNA alone in which no CA was used. CA alone or CAE, diluted in buffer alone was used as controls. Samples containing no protein were also added to CA for controls. The samples were incubated at 37 °C for 1 h. The samples were cooled to RT for 10 min. The viscometry readings for all samples were performed at RT for 30 s at 100 r.p.m. Viscosity values are measured in CentiPoise (cP). Calculations are based on the fact that sterile water has low viscosity and a cP of 1.00 at RT, and the dynamic viscosity of water at 20 °C is 1.002. Samples containing CAE and CA showed a decrease in viscosity, for example; whereas, control samples showed no change in viscosity.

Amino-acid analyses and database searching

The single protein band for the CAE, 84 354 MW, was prepared for Edman degradation and mass spectrometry analyses by the Baylor College of Medicine Protein Core Facility (Houston, TX, USA). Briefly, for Edman degradation, the CAE was electrophoresed on an SDS–PAGE gel poured at least 24 h before electrophoresis. The gel system included 0.1 mm thioglycolate as a scavenger in the upper-running buffer. The gel was electroblotted onto a polyvinylidene fluoride membrane using a Trisglycine buffer at RT for 2 h at 300 mA. The buffer consisted of 25 mm Tris, 192 mm glycine and 10% MeOH, pH 8.3. The gel was rinsed with sterile water for 5 min and stained with 0.05% Coomassie Blue in 1% acetic acid and 50% methanol. The gel was then destained in 50% methanol for approximately 15 min until the background turned pale blue. The gel was rinsed for 10 min in sterile water. The band was cut from the gel using a clean, fresh scalpel blade, placed into a 1.5 ml eppendorf tube, and submitted to the Core Facility. For mass spectrometry analyses, the CAE was electrophoresed on an SDS–PAGE gel using standard conditions. The gel was stained with 0.05% Coomassie Blue in 5% acetic acid and 10% methanol for 30 min to visualize the protein band. The gel was destained with 5% acetic acid and 10% methanol to clearly visualize the protein band. The gel was rinsed in sterile water for 15 min. The protein band was cut from the gel using a clean, fresh scalpel blade, placed into a 1.5 ml eppendorf tube, and submitted to the Core Facility. In a separate tube, an equal amount of gel containing no protein was submitted as a control.

The Edman degradation provided amino acid sequences for the following peptide fragment: ANSYNAYVANGSQTA

The mass spectrometry data provided amino acid sequences for the following eight-peptide fragments:

  • LLEQGTGEALTDGVLR
  • VPNSEVSLNALPNVQR
  • LADYEFTSAPSNSK
  • YSDLSTLN
  • QLLFDTAPLA
  • APYQVDDNL
  • FGAYLPDD
  • LGTLGG

In each of these peptide fragment sequences the amino acid leucine (L) may actually be either leucine (L) or isoleucine (I), the amino acid aspartic acid (D) may actually be aspartic acid (D) or asparagine (N), the amino acid glutamine (Q) may actually be glutamine (Q) or lysine (K) and the amino acid phenylalanine (F) may actually be phenylalanine (F) or oxidized methionine.

Peptide fragments were searched against the Non-Redundant Protein Database, excluding the eucaryota organisms. Peptides were also searched using TBLASTN version 2.2.1044 against the database from the NCBI including the GenBank, EMBL, DDBJ and PDB sequences excluding the EST, STS, GSS, environmental samples and HTGS sequences. The search parameters used the BLOSUM62 matrix with gap existence penalty of 11 and gap extension penalty of 1. Peptides were also searched using BLASTP version. 2.2.1044 against the NCBI, Non-Redundant Protein Sequence Database that included the Genbank CDS translations, PDB, SwissProt, PIR and PRF, excluding environmental samples. There were 2 316 421 sequences with 787 539 419 total letters. The search parameters used the PAM30 matrix with Gap open penalty of 9 and extension penalty of 1. The peptide sequences were also searched against the phage sequences using a local installation of BLASTP. Three independent searches were performed by Kim Worley (Human Genome Sequencing Center, the Baylor College of Medicine), the Baylor College of Medicine Protein Core Facility and by Wei Zhang and Florante Quiocho (Baylor College of Medicine).

DNA sequencing of the CAE ORF

Genomic DNA was prepared from the NST1 phage. On the basis of the amino-acid sequences for peptide fragments determined by mass spectrometry and Edman degradation listed above, the following 12 degenerate oligonucleotides (oligos) were designed for sequencing the CAE ORF from the NST1 genomic DNA (sequences listed 5′–3′):

  • TAYAAYGCNTAYGTNGCNAAYGG
  • CCRTTNGCNACRTANGCRTTRTA
  • GARCARGGNCANGGNGARGC
  • GCYTCNCCNTGNCCYTGYTC
  • TNCCNAAYGTNMARMG
  • CKYTKNACRTTNGGNA
  • GTNCCNAAYWSNGARGT
  • ACYTCNSWRTTNGGNAC
  • TNGCNGAYTAYGARWT
  • AWYTCRTARTCNGCNA
  • GCNCCNWSNAAYWSNAA
  • TTNSWRTTNSWNGGNGC

These degenerate oligos were designed by Kim Worley and prepared by Operon Technologies (Qiagen) and used for initial sequencing. Sequencing of both strands of the CAE ORF was performed by Lark Sequencing (Houston, TX, USA).

Chymotrypsin digestion and other proteases

Proteolytic digestion of the CAE with chymotrypsin was performed at 4 °C O/N in 20 mm Tris-HCl, pH 7.0, 100 mm NaCl and 5% glycerol. The following proteases were also tested including elastase, endoproteinase GluC, thermolysin, trypsin and proteinase K; however, these proteases did not cleave the CAE.

Cloning of CAE

PCR amplification of the NST1 phage genomic DNA using Deep Vent DNA Polymerase (New England BioLabs, Ipswich, MA, USA) for the CAE ORF extending from amino acids 107–790 with inclusion of an Nterminal methionine followed by glycine and then 6 histidines in the forward primer was performed. An NcoI digestion site was also included in the forward primer and an XhoI digestion site was included in the reverse primer. The forward are reverse primer sequences were the following (listed 5′–3′):

  • Forward: TAGACCATGGGCCATCATCATCATCATCATCTCGCGCCACTTGACCGCGCC
  • Reverse: TAGACTCGAGTTAGATGTAAGCCACCGGGTAGGTAGC

The amplification product was digested with NcoI and XhoI and ligated in proper orientation into the cloning site of the similarly digested pET-28a-c(+) vector (Novagen) using the Clonables Ligation/Transformation Kit (Novagen). The ligation mixture was transformed into Novablue competent cells (Novagen) and plated onto LB+kanamycin (80 µg/ml) plates. CAE-positive clones were identified by restriction enzyme digestion analyses.

Large-scale preparation of CAE

The CAE-recombinant clone was transformed in the E. coli host, BL21(DE3) (Invitrogen). A small-scale culture was first grown O/N in LB medium containing 50 mg/l of kanamycin at 37 °C with shaking at 225 r.p.m. In total, 2 l of hyper broth medium containing 50 mg/l of kanamycin was inoculated with 24 ml of the O/N culture and grown at 37 °C with shaking at 225 r.p.m until the OD600 reached 0.5, approximately for 2–4 h. Then the temperature was changed to 16 °C, and growth was continued for 30 min. Isopropyl β-D-1-thiogalactopyranoside was then added to make a final concentration of 0.015 mm, and the culture was grown for an additional 20 h. The growth medium was centrifuged at 4800 × g for 15 min, and the pellet was stored at −80 °C.

CAE was purified on a Ni-NTA column (Invitrogen) that selectively bound the 6 His-tag at the N-terminus using the following procedure. A 20 ml bed volume Ni-NTA column was equilibrated with 200 ml of buffer A consisting of 20 mm Tris-HCl, pH 8.0, 0.25M NaCl, 10% glycerol, 10 mm imidazole, 0.1 ml/l of β-mercaptoethanol and 1 mm PMSF. The cell pellet from the large culture was thawed and suspended in 150 ml of buffer A. The cell paste was disrupted using the Microfluidizer Processor, Model M-110Y, according to manufacture’s instructions (Microfluidics Corporation, Newton, MA, USA). The mixture was centrifuged at 40 000 × g to obtain a clear supernatant. The supernatant was loaded onto the equilibrated Ni-NTA column, washed with 600 ml of buffer A, and then washed with 80 ml of buffer B consisting of 20 mm Tris-HCl, pH 8.0, 0.25 m NaCl, 10% glycerol, 125 mm imidazole, 0.1 ml/l β-mercaptoethanol and 0.1 mm PMSF. The CAE was eluted from the column with 80 ml of buffer C consisting of 20 mm Tris-HCl, pH 8.0, 0.25MNaCl, 10% glycerol, 500 mm imidazole, 0.1 ml/l β-mercaptoethanol and 0.1 mm PMSF. The eluate was added to an Amicon Ultra-15 centrifugal filter unit with a 30 k cutoff (Millipore) and centrifuged at 2800 × g at 4 °C to concentrate the eluate. Buffer D, the storage buffer, consisting of 20 mm Tris-HCl, pH 8.0, 0.25 m NaCl and 10% glycerol was added to the eluate to provide a final volume of 13 ml. This concentration process using the Amicon Ultra-15 Centrifugal Filter listed above was repeated three times. The CAE protein was stored in buffer D at 4 °C at a 15 mg/ml concentration.

Plasmid DNA preparation and digestion with CAE

All plasmid DNAs made for Super Clean DNA processing were initially purified from 2.5 l LB cultures using Qiagen EndoFree Plasmid Giga Kits according to manufacture’s instructions. In addition, it was noted that bacterial growths reach an OD600 of 4.0 at 11–12 h post-inoculation of the starter culture (Qiagen Plasmid Purification Handbook 07/99, page 60). Furthermore, the maximal amount of plasmid DNA is produced from growth to an OD600 of 4.0. However, bacteria continue to grow to an OD600 of 5.0 if left in the incubator longer, with no further production of plasmid DNA. Therefore, we inoculated the ‘O/N’ growth late in the evening and started monitoring the OD600 in the morning, about 10 h post-inoculation. All bacterial growth was ended at an OD600 of 4.0 measured accurately using a turbidity cell holder in a DU 600 spectrophotometer (Beckman). The plasmid DNA was dissolved in 4 ml of 0.05 m potassium phosphate buffer, pH 6.5. CAE was added to the Qiagen EndoFree purified plasmid DNA in a weight ratio of 1:10 CAE:plasmid DNA. This mixture was incubated at 37 °C for 3 h. Then the temperature was increased to 50 °C and the mixture was incubated for 21 h.

Purification of CAE-digested plasmid DNA

The CAE-digested plasmid DNA was prepared for boronic acid column chromatography. The mixture was phenol-extracted twice and centrifuged at 13 000 r.p.m. for 10 min at RT. The aqueous phase was transferred to a sterile eppendorf tube and extracted with chloroform. The mixture was centrifuged at 13 000 r.p.m. for 10 min at RT, and the aqueous phase was transferred to a sterile eppendorf tube. The plasmid DNAwas precipitated with two volumes of 100% cold EtOH and 1/10 volume of 3 m sodium acetate, pH 5.2. The sample was placed at −20 °C for 1 h. The sample was centrifuged at 13 000 r.p.m. for 15 min at 4 °C. The plasmid DNA pellet was washed twice with 1 ml of 70% EtOH at RT and centrifuged at 13 000 r.p.m for 5 min at 4 °C. The plasmid DNA pellet was air-dried and resuspended in 1 ml of 0.2M ammonium acetate, pH 8.8, at approximately 3 µg/µl concentration.

Poly-prep chromatography columns (0.8 × 4 cm; BioRad, Life Science Research, Hercules, CA, USA) were equilibrated by pipetting 3.7 ml of 50% aqueous slurry of immobilized boronic acid (Pierce). Then 8 ml of 0.2 m ammonium acetate, pH 8.8, were added to each column, pipetted to disperse the resin and allowed to flow through until a small amount of buffer remained at the top. This step was repeated again. The plasmid DNA was loaded onto the equilibrated boronic acid columns without exceeding 3 ml volumes of DNA and not greater than 9 mg of DNA for each column. 5 ml fractions were collected in 1.5 ml eppendorf tubes using 0.2M ammonium acetate, pH 8.8, as the running buffer. The OD260 was measured for each fraction. The fractions containing significant amounts of plasmid DNA were pooled and loaded onto a second boronic acid column equilibrated with 0.2 m ammonium acetate, pH 8.8. In total, 5 ml fractions were collected in 1.5 ml eppendorf tubes using 0.2 m ammonium acetate, pH 8.8, as the running buffer. The OD260 was measured for each fraction. The fractions containing significant amounts of plasmid DNA were pooled. In total, 1/10 volume of 3 m sodium acetate and two volumes of −25 °C stored 100% EtOH were added to the chilled plasmid DNA and incubated on ice for 15 min. The mixture was placed at −25 °C for 1 h. The tubes containing plasmid DNA were centrifuged at 13 000 r.p.m for 30 min at 4 °C. The EtOH was aspirated from each tube. The DNA pellets were washed with 70% EtOH at RT and centrifuged at 13 000 r.p.m for 10 min at 4 °C. The EtOH was aspirated from each tube. This 70% EtOH wash-step was repeated again. The DNA pellets were air-dried. The boronic acid columns were also regenerated for reuse in purifying identical plasmids by first eluting the bound material, polysaccharides, with 0.1 m formic acid and then washing with five column volumes of 0.1 m formic acid. The column was then rinsed and stored in 0.02% sodium azide according to the manufacturer’s instructions.

The plasmid DNA was then prepared for Macrosep clean up. The plasmid DNA pellets were resuspended in 14 ml of 10 mm Tris-HCl, pH 8.0, containing 0.1% zwittergent 3–14 detergent (catalog No. 693017; Calbiochem, EMD4 Biosciences). The plasmid DNA tubes were incubated at 37 °C for 15 min. This mixture was placed into a Macrosep 100 Centrifugal Concentrator unit with 300 k cutoff and processed according to manufacture’s instructions. The unit containing the mixture was centrifuged at 4500 × g for 1 h at RT to concentrate the plasmid DNA. The retentate was brought to a 14 ml final volume by adding 10 mm Tris-HCl, pH 8.0, containing 0.1% zwittergent as stated above. The mixture was centrifuged at 4500 × g for 1 h at RT. This process was repeated three more times without inclusion of the zwittergent in the final two concentration steps. The plasmid DNA was precipitated with two volumes of 100% cold EtOH and 1/10 volume of 3 M sodium acetate, pH 5.2. The tubes containing plasmid DNA were placed at −20 °C for 1 h and then centrifuged at 13 000 r.p.m for 15 min at 4 °C. The DNA pellets were washed twice with 1 ml of 70% EtOH at RT and centrifuged at 13 000 r.p.m for 5 min at 4 °C. The EtOH was aspirated from each tube. The DNA pellets were airdried, resuspended in sterile water at a final concentration of 5 mg/ml and stored in nunc vials at −85 °C.

BCA assay

This BCA assay was performed to detect any residual reducing ends from polysaccharides remaining in the Super Clean processed plasmid DNA. However, it can also be used to determine the activity of CAE. A modified version of the published, miniaturized highly sensitive BCA assay was used.26,27 Briefly, this assay used an optical 96-well reaction plate (Applied Biosystems, Life Technologies, Carlsbad, CA, USA), with each well containing a mixture of 2 µg of CAE (test sample) or bovine serum albumin (control sample) and 100 µg CA or plasmid DNA in a total volume of 110 µl of 0.05 m potassium phosphate buffer, pH 6.5. All wells were sealed with strip caps, and the plates were centrifuged at 1500 r.p.m at RT for 1 min. The caps were removed from the wells, and the solution mixed in each well by gently pipetting up and down three times using a multi-channel pipettor. The wells were resealed using strip caps and incubated at 37 °C for 3 h and then at 50 °C for 21 h. The plates were centrifuged at 1500 r.p.m at RT for 1 min. Then 100 µl of each reaction mix was transferred to a new 96-well plate using a multi-channel pipettor and mixing gently by pipetting three times up and down. Then 100 µl of freshly prepared BCA reagent (MicroBCA Protein Assay Kit; Pierce) were added to each well. Lids were placed over the plates and incubated at 37 °C for 2 h, cooled to RT for 15 min, and read on a multi-plate reader at 550 nm. Each test or control sample was assayed in triplicate. The results were determined by subtraction of any values obtained for the control sample from the test sample.

Liposome and complex preparation

The BIV liposomes and complexes were prepared as previously described except that synthetic cholesterol (Sigma) was used in place of extracted cholesterol at a DOTAP:cholesterol ratio of 50:45. The complexes were filtered through a 1.0 µm pore size, polysulfone filter of 13 mm diameter (catalog No. 6780-1310; Whatman, GE Healthcare) before the administration into mice.

In vivo toxicity assay

Mice were IV injected into the tail vein slowly and evenly over one minute using a 30-gauge syringe needle. Male Balb/c mice were purchased from Harlan Laboratories (Houston, TX, USA). The SCID mouse strain used was HsdIcr:Ha(ICR)-Prkdcscid bred in the Barrier Animal Facility, Baylor College of Medicine. Human pancreatic PANC-1 tumor-bearing mice were established in these SCID mice. The PANC-1 cells were purchased from the (Manassas, VA, USA) ATCC. The PANC-1 cells were harvested and resuspended in 1 × phosphate-buffered saline. A 200 µl cell suspension containing 5 × 105 PANC-1 cells was intraperitoneally injected into each 8–10 week-old male SCID mouse. Tumors were allowed to grow for 7 weeks before IV injections. All animal procedures were performed in accordance with the Baylor College of Medicine institutional guidelines using an approved animal protocol.

Table 4
Toxicity of Qiagen-endofree purified DNA versus SC-DNA–BIV complexes (in SCID mice)

Footnotes

Conflict of interest

The authors declare no conflict of interest.

References

1. Tousignant JD, Gates AL, Ingram LA, Johnson CL, Nietupski JB, Cheng SH, et al. Comprehensive analysis of the acute toxicities induced by systemic administration of cationic lipid:plasmid DNA complexes in mice. Hum Gene Ther. 2000;11:2493–2513. [PubMed]
2. McLachlan G, Stevenson BJ, Davidson DJ, Porteous DJ. Bacterial DNA is implicated in the inflammatory response to delivery of DNA/DOTAP to mouse lungs. Gene Ther. 2000;7:384–392. [PubMed]
3. Templeton NS, Lasic DD, Frederik PM, Strey HH, Roberts DD, Pavlakis GN. Improved DNA: liposome complexes for increased systemic delivery and gene expression. Nat Biotechnol. 1997;15:647–652. [PubMed]
4. Lu C, Stewart DJ, Ji L, Ramesh R, Jayachandran G, Erasmus J, et al. A Phase I trial of intravenous therapy with tumor suppressor FUS1-nanoparticles for recurrent/metastatic lung cancer. ASCO Meeting. 2009 abstract e19065.
5. Lu C, Sepulveda CA, Ji L, Ramesh R, O’Connor S, Jayachandran G, et al. Annual Meeting of the American Association of Cancer Research. Los Angeles, CA: 2007. Systemic nanoparticle FUS1 tumor suppressor gene therapy for stage IV lung cancer.
6. Edelman M. Purification of DNA by affinity chromatography removal of polysaccharide contaminants. Anal Biochem. 1975;65:293–297. [PubMed]
7. Do N, Adams RP. A simple technique for removing plant polysaccharide contaminants from DNA. Biotechniques. 1991;10:162–164. 166. [PubMed]
8. Chan JW, Goodwin PH. Extraction of genomic DNA from extracellular polysaccharide-synthesizing gram-negative bacteria. Biotechniques. 1995;18:418–422. [PubMed]
9. Ingalls RR, Monks BG, Savedra R, Jr, Christ WJ, Delude RL, Medvedev AE, et al. CD11/CD18 and CD14 share a common lipid A signaling pathway. J Immunol. 1998;161:5413–5420. [PubMed]
10. Neudecker F, Grimm S. High-throughput method for isolating plasmid DNA with reduced lipopolysaccharide content. Biotechniques. 2000;28:107–109. [PubMed]
11. Jann B, Jann K. Structure and biosynthesis of the capsular antigens of Escherichia coli. Curr Top Microbiol Immunol. 1990;150:19–42. [PubMed]
12. Wicken AJ. Bacterial cell walls and surfaces. In: Savage DC, Fletcher M, editors. Bacterial Adhesion: mechanisms and physiological significance. New York, NY: Plenum Publishing Corporation; 1985. pp. 45–70.
13. Goebel WF. Colanic acid. Proc Natl Acad Sci USA. 1963;49:464–471. [PubMed]
14. Shioda M, Murakami-Murofushi K. Selective inhibition of DNA polymerase alpha by a polysaccharide purified from slime of Physarum polycephalum. Biochem Biophys Res Commun. 1987;146:61–66. [PubMed]
15. Aoki Y, Koshihara H. Inhibitory effects of acid polysaccharides from sea urchin embryos on RNA polymerase activity. Biochim Biophys Acta. 1972;272:33–43. [PubMed]
16. Aoki Y, Koshihara H. Inhibitory effects of acid polysaccharides from sea urchin embryos in RNA synthesis in vitro. Exp Cell Res. 1972;70:431–436. [PubMed]
17. Kuhn HM, Meier-Dieter U, Mayer H. ECA, the enterobacterial common antigen. FEMS Microbiol Rev. 1988;4:195–222. [PubMed]
18. Fields PA, Kowalczyk DW, Arruda VR, Armstrong E, McCleland ML, Hagstrom JN, et al. Role of vector in activation of T cell subsets in immune responses against the secreted transgene product factor IX. Mol Ther. 2000;1:225–235. [PubMed]
19. Yew NS, Zhao H, Przybylska M, Wu IH, Tousignant JD, Scheule RK, et al. CpG-depleted plasmid DNA vectors with enhanced safety and long-term gene expression in vivo. Mol Ther. 2002;5:731–738. [PubMed]
20. Trisler P, Gottesman S. Ion transcriptional regulation of genes necessary for capsular polysaccharide synthesis in Escherichia coli K-12. J Bacteriol. 1984;160:184–191. [PMC free article] [PubMed]
21. Knutson CA, Jeanes A. A new modification of the carbazole analysis: application to heteropolysaccharides. Anal Biochem. 1968;24:470–481. [PubMed]
22. Morris JB. Enzymatic assay for subnanomole amounts of L-fucose. Anal Biochem. 1982;121:129–134. [PubMed]
23. Hughes KA, Sutherland IW, Clark J, Jones MV. Bacteriophage and associated polysaccharide depolymerases–novel tools for study of bacterial biofilms. J Appl Microbiol. 1998;85:583–590. [PubMed]
24. Hughes KA, Sutherland IW, Jones MV. Biofilm susceptibility to bacteriophage attack: the role of phage-borne polysaccharide depolymerase. Microbiology. 1998;144(Part 11):3039–3047. [PubMed]
25. Sutherland IW. Phage-induced fucosidases hydrolysing the exopolysaccharide of Klebsiella arogenes type 54 (A3(S1)) Biochem J. 1967;104:278–285. [PubMed]
26. Verhoef R, Beldman G, Schols HA, Siika-aho M, Ratto M, Buchert J, et al. Characterisation of a 1,4-beta-fucoside hydrolase degrading colanic acid. Carbohydr Res. 2005;340:1780–1788. [PubMed]
27. Meeuwsen PJ, Vincken JP, Beldman G, Voragen AG. A universal assay for screening expression libraries for carbohydrases. J Biosci Bioeng. 2000;89:107–109. [PubMed]
28. Schott H, Rudloff E, Schmidt P, Roychoudhury R, Kossel H. A dihydroxyboryl-substituted methacrylic polymer for the column chromatographic separation of mononucleotides, oligonucleotides, and transfer ribonucleic acid. Biochemistry. 1973;12:932–938. [PubMed]
29. Singh N, Willson RC. Boronate affinity adsorption of RNA: possible role of conformational changes. J Chromatogr A. 1999;840:205–213. [PubMed]
30. Jerkovic B, Kung HC, Bolton PH. Purification of thymine glycol DNA and nucleosides by use of boronate chromatography. Anal Biochem. 1998;255:90–94. [PubMed]
31. Pruess-Schwartz D, Sebti SM, Gilham PT, Baird WM. Analysis of benzo(a)pyrene:DNA adducts formed in cells in culture by immobilized boronate chromatography. Cancer Res. 1984;44:4104–4110. [PubMed]
32. Larsson AM, Andersson R, Stahlberg J, Kenne L, Jones TA. Dextranase from Penicillium minioluteum: reaction course, crystal structure, and product complex. Structure. 2003;11:1111–1121. [PubMed]
33. Yoder MD, Lietzke SE, Jurnak F. Unusual structural features in the parallel beta-helix in pectate lyases. Structure. 1993;1:241–251. [PubMed]
34. Steinbacher S, Baxa U, Miller S, Weintraub A, Seckler R, Huber R. Crystal structure of phage P22 tailspike protein complexed with Salmonella sp. O-antigen receptors. Proc Natl Acad Sci USA. 1996;93:10584–10588. [PubMed]
35. Ishii KJ, Akira S. Innate immune recognition of, regulation by, DNA. Trends Immunol. 2006;27:525–532. [PubMed]
36. Templeton NS. Cationic liposome-mediated gene delivery in vivo. Biosci Rep. 2002;22:283–295. [PubMed]
37. Johnson GB, Brunn GJ, Kodaira Y, Platt JL. Receptor-mediated monitoring of tissue well-being via detection of soluble heparan sulfate by Toll-like receptor 4. J Immunol. 2002;168:5233–5239. [PubMed]
38. Arancibia SA, Beltran CJ, Aguirre IM, Silva P, Peralta AL, Malinarich F, et al. Toll-like receptors are key participants in innate immune responses. Biol Res. 2007;40:97–112. [PubMed]
39. Altenbach D, Robatzek S. Pattern recognition receptors: from the cell surface to intracellular dynamics. Mol Plant Microbe Interact. 2007;20:1031–1039. [PubMed]
40. Krishnan J, Selvarajoo K, Tsuchiya M, Lee G, Choi S. Toll-like receptor signal transduction. Exp Mol Med. 2007;39:421–438. [PubMed]
41. Dam TK, Brewer CF. Lectins as pattern recognition molecules: the effects of epitope density in innate immunity. Glycobiology. 2010;20:270–279. 2009 Nov 24. [PubMed]
42. Lavelle EC, Murphy C, O’Neill LA, Creagh EM. The role of TLRs, NLRs, and RLRs in mucosal innate immunity and homeostasis. Mucosal Immunol. 2010;3:17–28. [PMC free article] [PubMed]
43. McGreal EP. Structural basis of pattern recognition by innate immune molecules. Adv Exp Med Biol. 2009;653:139–161. [PubMed]
44. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–3402. [PMC free article] [PubMed]