Optimization of the real time PCR method
The real time PCR method for the quantification of genome copy numbers had been established for haloarchaea (Breuert et al., 2006), but, in the meantime, was also applied to methanogenic archaea and proteobacteria (Hildenbrand et al., 2011; Pecoraro et al., 2011). It has been validated against several independent methods, i.e. quantitative Southern blotting (Breuert et al., 2006), DNA isolation, and spectroscopic quantification (Hildenbrand et al., 2011), and the wealth of results published for E. coli, which was obtained by radioactive labeling and Fluorescence Activated Cell Sorter (FACS) analyses (Pecoraro et al., 2011). In each case, the results of the real time PCR method were in excellent agreement with the respective independent method.
To give a short overview, genomic DNA was used as a template in a conventional PCR reaction to amplify a fragment of about 1 kbp. A dilution series of this fragment was prepared and used for real time PCR analysis. A fragment of about 300 bp, internal to the standard fragment, was amplified. The results were used to generate a standard curve. To determine the genome copy number, cells were lysed and a dilution series of the resulting cell extract was analyzed using real time PCR in parallel to the standards. The results allowed calculating the number of genome copies in the cell extract and, in combination with the cell density of the culture, the ploidy level.
The following points have to be optimized for every new species under investigation and were optimized for the three species of cyanobacteria used in this study: (1) the cell density has to be quantified with a very low variance, (2) it has to be verified that culture growth is highly reproducible, (3) the method of cell disruption has to be about 100% effective yet leaving the genomic DNA intact, and (4) the real time PCR has to be truly exponential.
For cyanobacteria, the method for cell disruption turned out to be the most critical point. Several standard methods (sonification, enzymatic murein digestion, ‘normal shaking’ with glass beads) could not be used, either because the efficiency of cell lysis was too low or because damage of the genomic DNA was too high. Shaking the cells in a Speedmill with 0.1 mm glass beads led to satisfactory results, lysis efficiency was higher than 90%, and the genomic DNA was only slightly damaged (fragment sizes from 4 kbp to >20 kbp, data not shown). The amount of beads and shaking time were optimized for every species. To exemplify the results, Fig. S1 (Supporting Information) shows one typical example of a real time PCR analysis (Fig. S1a), a standard curve (Fig. S1b), a melting point analysis, and an analytical agarose gel of the analysis fragments (Fig. S1c, d). At least three independent cultures were analyzed (and each culture was analyzed at least in triplicates), and average values and standard deviations (SD) were calculated.
Synechococcus elongatusPCC 7942 and SynechococcusWH7803 are oligoploid
Synechococcus elongatus PCC 7942 grew with a doubling time of 24 h. An average growth curve of three cultures is shown in Fig. S2. The results of genome quantification of three independent cultures are summarized in Table 1. At an OD750 nm of 0.6, S. elongatus contained about four genome copies per cell and thus the species is oligoploid. This is termed ‘exponential phase’, although growth of the cultures was not truly exponential, but the OD750 nm of 0.6 was prior to the onset of the linear growth phase (compare Fig. S2). This value is in accordance with the previously published value of 3–5 genome copies that was obtained using FACS analysis (Mori et al., 1996). Stationary phase cells also contained 3–4 genomes per cell. Therefore, in S. elongatus, the ploidy level is not growth phase-regulated, in contrast to many other species.
Table 1. Genome copy numbers of Synechococcus elongatusPCC 7942 and Synechococcus sp. WH 7803
|Culture||OD750 nm||Exponential phase||Stationary phase|
|Genome copies||Average ± SD||Genome copies||Average ± SD|
|S. elongatus PCC 7942|
|1||0.6||4.2|| ||3.1|| |
|2||0.6||3.6||4.0 ± 0.3||3.5||4.0 ± 1.3|
|3||0.6||4.2|| ||5.5|| |
|Synechococcus sp. WH 7803|
|1||0.6||2.6|| || || |
|2||0.6||4.3||3.6 ± 0.9|| || |
|3||0.6||3.8|| || || |
The results of genome quantification for Synechococcus WH7803 are also summarized in Table 1. This species also contained between three and four genome copies at an OD750 nm of 0.6 and during stationary phase, and is thus oligoploid. Again, this is in accordance with an earlier study that applied FACS analysis for genome copy number determination and found 2–4 copies per cell (Binder & Chisholm, 1990). Taken together, the freshwater as well as the salt water species were found to be oligoploid, irrespective of the applied method for quantification (based either on one specific site of the genome (this study) or the average DNA content), growth in continuous light (this study) or growth in light–dark cycles (Mori et al., 1996), and the growth phase.
SynchocystisPCC 6803 is highly polyploid
First, the motile Synechocystic PCC 6803 wild-type strain was analyzed. An average growth curve of three independent cultures is shown in Fig. S3. The results of genome copy number determination are summarized in Table 2. The doubling time at the cell harvest in linear growth phase (OD750 nm = 0.6) was around 20 h. Synechocystis PCC 6803 turned out to be highly polyploid, and it contained nearly 60 genomes per cell, both in linear and in stationary growth phase. As this value is very high and in fact higher than any value published until now for any cyanobacterial species, the genome copy number in stationary phase cells was also determined using an independent method, namely spectroscopic determination of the DNA concentration. The average values of 57.9 (if the plasmid copy number would be low) and 53.3 (if the plasmid copy number would be high) genomes per cell were in excellent agreement with the real time PCR result, and thus underscored that Synechocystis PCC 6803 is highly polyploid.
Table 2. Genome copy numbers of two SynechocystisPCC 6803 strains
|No.||Exponential phase||Linear phase||Stationary phase|
|OD750 nm||Genome copies||Average ± SD||OD750 nm||Genome copies||Average ± SD||Genome copies||Average ± SD|
|Motile wild-type strain|
|1||0.1||210.3|| ||0.6||66.1|| ||52.9|| |
|2||0.1||218.3||218.0 ± 7.6||0.6||46.7||58.0 ± 10.1||61.0||57.6 ± 4.2|
|3||0.1||225.5|| ||0.6||61.2|| ||58.9|| |
|GT wild-type strain|
|1||0.1||144.5|| ||0.6||41.1|| ||42.1|| |
|2||0.1||136.1||142.2 ± 5.3||0.6||46.3||47.2 ± 6.6||45.2||43.3 ± 2.8|
|3||0.1||146.0|| ||0.6||54.1|| ||39.7|| |
An earlier study had also shown that this species is polyploid, but the reported value of 12 genome copies per cell for the ‘Kazusa’ wild-type of Synechocystis PCC 6803 (Labarre et al., 1989) is much lower than the value determined in this study. The reason for the discrepancy is not obvious, as in the previous study also the lysis efficiency was quantified, genome size was underestimated by only 32%, and the colorimetric assay for DNA quantification probably cannot be that wrong. The same medium was used, and a similar doubling time of 15–20 h was reported. Therefore, it might be that both reports are correct and that the ploidy level of various strains of the species Synechocystis PCC 6803 are different.
To test this hypothesis, another wild-type strain of Synechocystis PCC 6803 was used, i.e. the so-called GT wild-type. This strain was chosen because genotypes segregate much faster in the GT strain than in the motile strain in the course of chromosomal mutant construction (Annegret Wilde, personal communication). Average growth curves of three independent cultures are shown in Fig. S4, and again, cells in linear growth phase and in stationary phase were analyzed. The results are also shown in Table 2. The GT wild-type was also highly polyploid; however, the genome copy number was with 42 genome copies nearly 30% lower than that of the motile wild-type, verifying that different strains of PCC 6803 vary in their ploidy level.
Notably, the 12 genome copies reported for the ‘Kazusa’ strain (Labarre et al., 1989) are much lower compared with the 42 and 58 genome copies of the two other wild-type strains analyzed in this study. Three explanations appear possible: (1) the ‘Kazusa’ strain highly deviates from the other two strains, (2) the genome copy number changed during the last 20 years of cultivation in the laboratory and today the ploidy level of the ‘Kazusa’ strain is higher than in 1989, (3) strains cultivated for long times under identical names in different laboratories accumulated different mutations, including mutations that affect the ploidy level, and thus ‘identical’ strains have different ploidy levels in different laboratories.
The species Synechocystis PCC 6803 was isolated from freshwater in California more than 40 years ago (Stanier et al., 1971). Several mutations are known that occurred during its further ‘evolution in the laboratory’. The sequenced ‘Kazusa’ strain contains insertion elements at four places of the genome that were devoid of an insertion element in the original isolate (Okamoto et al., 1999). In addition, the sequenced ‘Kazusa’ strain contains a frameshift mutation in the gene encoding a protein kinase that is not present in other strains (Kamei et al., 2001). It will be interesting to unravel how different strains differ in their ploidy level. An in-depth analysis including several samples of each of the three wild-type strains obtained from different laboratories around the world will be needed to clarify the situation. In any case, all Synechocystis PCC 6803 strains analyzed until now are polyploid, and we could show that the ploidy levels of different strains vary. For experiments that are sensitive to the ploidy level, this should be taken into account and the ploidy level of the strain under investigation should be quantified.
Anonymous reviewers of the first version of this article pointed out that we only analyzed the linear and the stationary growth phase, and that an analysis of exponentially growing cells would also be desirable. Therefore, again three independent cultures of both strains were grown and were harvested during exponential growth at an OD750 nm of 0.1. The results are included in Table 3. Surprisingly, it turned out that the GT wild-type contained 142 genome copies per cells and the motile wild-type contained 218 genome copies per cell, much higher values than in linear and stationary growth phase. Several bacterial and archaeal species exhibit growth phase-dependent regulation of their ploidy levels and have higher genome copy numbers in exponential phase than in stationary phase (Hildenbrand et al., 2011; Pecoraro et al., 2011). Synechocystis PCC 6803 adds to this list, but is extraordinary in that the genome copy number is already down-regulated in linear growth phase.
Table 3. Overview of cyanobacterial species with experimentally determined ploidy levels and selected parameters
|Species||Growth temperature (°C)||Doubling time (h)||Genome size (Mbp)||Average genome copy No.||Ploidy||References|
|Anabaena cylindrica||30||18.5||–||25||Polyploid||Simon (1977)|
|Anabaena variabilis||30||–||7.1||5–8||Oligoploid||Simon (1980)|
|10 Microcystis strains||20||stat.ph.||–||1–10||Oligoploid||Kurmayer & Kutzenberger (2003)|
|Anabaena sp. PCC 7120||28||–||7.2||8.2||Oligoploid||Hu et al. (2007)|
|Prochlorococcus||–||–||1.7||–||Monoploid||Vaulot et al. (1995)|
|Synechococcus elongatus PCC 7942||28||24||2.8||3.9/3.3b||Oligoploid||This study|
|S. elongatus PCC 7942||30||11, LDC||2.8||3–5||Oligoploid||Mori et al. (1996)|
|Synechococcus sp. PCC 6301||38||5 to >50||2.7||2–6 to >1–2c||Oligoploid||Binder & Chisholm (1990)|
|Synechococcus sp. WH 7803||28||–||2.4||3.6||Oligoploid||This study|
|Synechococcus sp. WH 7803||25||–||2.4||2–4||Oligoploid||Binder & Chisholm (1995)|
|Synechococcus sp. WH 7805||25||15||2.6a||1||Monoploid||Binder & Chisholm (1995)|
|Synechococcus sp. WH 8101||25||17||3.2a||1||Monoploid||Armbrust et al. (1989)|
|Synechococcus sp. WH8103||25||22||2.7a||1–2||Monoploid||Binder & Chisholm (1995)|
|Synechocystis sp. PCC 6803 (motile)||28||20||3.6||218/58/58b||Polyploid||This study|
|Synechocystis sp. PCC 6803 (GT)||28||20||3.6||142/47/43b||Polyploid||This study|
|Synechocystis sp. PCC 6803 (‘Kazusa’)||30||15–20||3.6||12||Polyploid||Labarre et al. (1989)|
The genome copy numbers of 218 and 142 in exponentially growing cells of the two Synechocystis strains are considerably higher than the 120 genome copies per cell that have been reported for Buchnera, a symbiotic bacterium with a reduced genome size (Komaki & Ishikawa, 1999). To our knowledge, a higher value has been reported only for Epulopiscium sp. that contains tens of thousands of genome copies (Mendell et al., 2008). However, Epulopiscium sp. is a giant bacterium exhibiting cell lengths in excess of 600 μm. Therefore, Synechocystis PCC 6803 has the highest ploidy level of any ‘normal’ sized prokaryote. However, it is unclear whether such high ploidy levels also exist in natural habitats, or whether this is an artifact of decades of cultivation in the laboratory. Synechocystis PCC 6803 was isolated 40 years ago and has been cultivated in the laboratory since then (Stanier et al., 1971). Therefore, an in-depth analysis (see above) should also include fresh isolates of Synechocystis PCC6803 as well as samples from different culture collections that had been kept frozen since their submission.
Ploidy in cyanobacteria
A literature search was performed to identify (hopefully) all cyanobacterial species with experimentally determined ploidy levels. Table 3 summarizes the results together with selected features. Three species are polyploid and contain at least 10 genome copies. They belong to different genera and grow either as single cells or as filaments. More than ten species are oligoploid and contain between three and nine genome copies. Again, among them are unicellular and filamentous species of several genera. Four species are monoploid, and thus monoploidy is not the rule, but an exception in cyanobacteria. The ploidy level is highly variable in cyanobacteria similar to the proteobacteria (Pecoraro et al., 2011). One genus can harbor monoploid and oligoploid species (Synechococcus) or oligoploid and polyploid species (Anabaena). There is no obvious correlation between the number of genome copies and any of the listed features, i.e. genome size, growth temperature, and doubling time. Various evolutionary advantages of oligo- and polyploidy for prokaryotes exist. As has been extensively studied with D. radiodurans, one of the advantages is resistance against double strand breaks that can be induced by X-ray irradiation (an artificial situation) and desiccation (regularly occurring in natural habitats). In fact, it could be shown that the resistance of polyploid Synechocystis PCC 6803 against X-ray irradiation is much higher than that of the oligoploid Synechococcus PCC 7942 (Domain et al., 2004). Another advantage is that the cells can live without a stringent control of equal chromosome segregation, and for Synechocystis PCC 6803 and Nostoc sp. PCC 7120, it has indeed been shown that the amount of DNA in the two newborn daughter cells after cell division is not always identical, but can vary (Hu et al., 2007; Schneider et al., 2007). An additional advantage is gene redundancy, which opens the possibility that under unfavorable conditions, mutations are induced in some genome copies, whereas the wildtype information is retained in others. It has indeed been shown that heterozygous cells of S. elongatus PCC 7942 and of Synechocystis PCC 6803 can be selected, at least under laboratory conditions (Labarre et al., 1989; Spence et al., 2004; Takahama et al., 2004; Nodop et al., 2008). Heterozygous strains have also been selected of two halophilic and methanogenic archaea, Haloferax volcanii and Methanococcus maripaludis. In both cases, it was shown that in the absence of selection gene conversion leads to the equalization of genomes and reappearance of homozygous cells (Hildenbrand et al., 2011; Lange et al., 2011). By analogy, we predict that gene conversion also operates in oligo- and polyploid species of cyanobacteria. The higher efficiency of gene replacement with linear DNA compared with circular DNA in Synechocystis PCC 6803 indicates that this is really the case (Labarre et al., 1989).