Date Published: January 13, 2017
Publisher: Public Library of Science
Author(s): Magdalena Plotka, Mateusz Wozniak, Tadeusz Kaczorowski, Hideyuki Doi.
Bacteria can be considered as biological nanofactories that manufacture a cornucopia of bioproducts most notably recombinant proteins. As such, they must perfectly match with appropriate plasmid vectors to ensure successful overexpression of target genes. Among many parameters that correlate positively with protein productivity plasmid copy number plays pivotal role. Therefore, development of new and more accurate methods to assess this critical parameter will result in optimization of expression of plasmid-encoded genes. In this study, we present a simple and highly accurate method for quantifying plasmid copy number utilizing an EvaGreen single colour, droplet digital PCR. We demonstrate the effectiveness of this method by examining the copy number of the pBR322 vector within Escherichia coli DH5α cells. The obtained results were successfully validated by real-time PCR. However, we observed a strong dependency of the plasmid copy number on the method chosen for isolation of the total DNA. We found that application of silica-membrane-based columns for DNA purification or DNA isolation with use of bead-beating, a mechanical cell disruption lead to determination of an average of 20.5 or 7.3 plasmid copies per chromosome, respectively. We found that recovery of the chromosomal DNA from purification columns was less efficient than plasmid DNA (46.5 ± 1.9% and 87.4 ± 5.5%, respectively) which may lead to observed differences in plasmid copy number. Besides, the plasmid copy number variations dependent on DNA template isolation method, we found that droplet digital PCR is a very convenient method for measuring bacterial plasmid content. Careful determination of plasmid copy number is essential for better understanding and optimization of recombinant proteins production process. Droplet digital PCR is a very precise method that allows performing thousands of individual PCR reactions in a single tube. The ddPCR does not depend on running standard curves and is a straightforward and reliable method to quantify the plasmid copy number. Therefore we believe that the ddPCR designed in this study will be widely used for any plasmid copy number calculation in the future.
Plasmids play an important role in molecular biology and biotechnology, primarily as vectors for molecular cloning to facilitate the overproduction of recombinant proteins , but also as sophisticated nanotools for specialized applications in the genome engineering . In a rapidly growing field of gene therapy and genetic vaccination, naked or lipid-coated plasmid DNA is also successfully applied to administer therapeutic genes  and is considered to be much safer and easier to use than genetically modified viruses [4, 5]. Moreover, plasmid-oriented studies provide insights to improve understanding of DNA replication, maintenance and transfer strategies which are essential to all microorganisms [6, 7]. In this respect, among many features that characterize these mobile genetic elements, the one which defines the number of plasmid units that are contained inside one bacterial cell is especially important, both from a practical and a biological point of view.
To produce recombinant proteins or non-proteinous recombinant products appropriate expression system needs to be chosen. Among different parameters describing expression vectors, such as structural and segregational plasmid stability, plasmid copy number is an essential feature with strong impact on system productivity. In the present study, we developed a method for plasmid copy determination based on droplet digital PCR and EvaGreen, a next-generation DNA binding dye. This method is compatible with an approach that calculates the PCN parameter as a number of plasmid copies per chromosome . However, there is also an alternative approach that estimates the PCN as a number of plasmid copies per cell . It is important to stress that the results obtained with both approaches can differ. In the case of pBR322 present in fast growing bacteria (log phase), the number of plasmids per cell was calculated to be 39–55. However, in a parallel experiment the plasmid copy number per chromosome in the same phase of growth was estimated as 15–32 . From those experiments it can be concluded that the average number of plasmid copies per cell is always higher than that calculated per chromosome. This phenomenon can be explained by multiple openings of the replication forks during bacterial exponential phase of growth that lead to the decrease of the plasmid per chromosome ratio and in consequence lower the plasmid copy number [18, 45]. Therefore, in the literature, depending on the methodology used, two different ranges of the pBR322 copy number exist: 15–20  and 30–70 , and our results correspond to the lower PCN ranges. Apart from the different PCN calculation strategies (per chromosome or per cell), another source of variations in the PCN determination is the method used for DNA purification . Real-time PCR requires only a small amount of the template DNA, but different efficiencies in the total DNA isolations can lead to the PCN miscalculation. Indeed, the most commonly used method for qPCR template purification is DNA isolation with the use of commercial kits or multi-step procedures involving cell disruption, often with the use of lysozyme, enzymatic protein digestion, DNA extraction with phenol-chloroform, precipitation and rehydratation [13, 20, 43]. However, it is already known that each step added to the DNA purification procedure increases the probability of the sample loss . Therefore, for efficient DNA isolation, we replaced the multi-step DNA isolation procedure by a simple, mechanical disruption of bacterial cells by the bead-beating method. In our hands, the bead-beating seemed to be a simple and fast strategy to isolate total DNA that was ready to be used in the qPCR experiments. In recent years, the bead-beating method was successfully used for efficient DNA isolation for various applications [47–49]. For example, among five different mechanical cell disruption methods, including sonication, nebulization, homogenization, microfluidization, and bead milling (bead-beating; BB in this paper), the bead milling was found to be the most efficient for intact plasmid extraction from bacterial cells, with the recovery yield reaching over 90% . Moreover, it was shown that only bead-beating was effective for isolating DNA from such difficult samples as Bacillus globigii (B. subtilis subsp. niger) endospores or Fusarium moniliforme conidia . In our hands, two different methods used for DNA isolation gave two different ranges of the pBR322 copy number (Tables 2 and 3 and S1 Table). For the E. coli bacteria in the log-phase, the pBR322 PCN was in a range of 6–7 for the bead-beating method, and 20–23 for the QIAamp DNA Mini kit used for DNA isolation. It is a common knowledge that there is a difference in efficiency of isolation of plasmid or chromosomal DNA, especially when the DNA binding columns are used. We have shown that in defined experimental settings, when mixture of genomic and plasmid DNA was loaded onto a purification column, on average only 46.54% of the initial amount of E. coli chromosomal DNA was present in the eluate. The same applied to as much as 87.38% of the plasmid DNA (Fig 4). This difference may distort the ratio of the plasmid to chromosomal DNA and lead to over-estimation of the plasmid copy number. Problems with the multi-step DNA isolation procedures were also noticed by other scientists [18, 37, 43]. In many laboratories, the researchers had started to prepare DNA for the PCN determination by heating the cell samples at 95°C , and different heating protocols were tested for an optimal template preparation [18, 37]. In our project, we performed the direct comparison between a multi-step procedure and a simple, mechanical cells disruption method for DNA extraction. Similar analyses had not been performed by others so far.
To our knowledge, for the first time, we applied a single colour ddPCR (with the use of EvaGreen) for determination of the plasmid copy number. Because, the β-lactamase is a common selection marker present in many expression systems, and E. coli is a popular bacterial host, we believe that ddPCR method developed in this study can be easily adopted by other researches to evaluate plasmid copy number. Moreover, with a little effort the designed method can be optimized to be used for other selection markers (such as genes encoding resistance to chloramphenicol or tetracycline) and different bacterial hosts.