- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
Flow cytometry provides a rapid, accurate, and simple means to determine nuclear DNA contents (C-value) within plant homogenates. This parameter is extremely useful in a number of applications in basic and applied plant biology; for example, it provides an important starting point for projects involving whole genome sequencing, it facilitates characterization of plant species within natural and agricultural settings, it allows facile identification of engineered plants that are euploid or that represent desired ploidy classes, it points toward studies concerning the role of C-value in plant growth and development and in response to the environment and in terms of evolutionary fitness, and, in uncovering new and unexpected phenomena (for example endoreduplication), it uncovers new avenues of scientific enquiry. Despite the ease of the method, C-values have been determined for only around 2% of the described angiosperm (flowering plant) species. Within this small subset, one of the most remarkable observations is the range of 2C values, which spans at least two orders of magnitude. In determining C-values for new species, technical issues are encountered which relate both to requirement for a method that can provide accurate measurements across this extended dynamic range, and that can accommodate the large amounts of debris which accompanies flow measurements of plant homogenates. In this study, the use of the Accuri C6 flow cytometer for the analysis of plant C-values is described. This work indicates that the unusually large dynamic range of the C6, a design feature, coupled to the linearity of fluorescence emission conferred by staining of nuclei using propidium iodide, allows simultaneous analysis of species whose C-values span that of almost the entire described angiosperms. © 2009 International Society for Advancement of Cytometry
Flow cytometry provides a means of unprecedented simplicity for determining the nuclear DNA contents of plant tissues and organs (1). The plant tissues are chopped, using a standard razor blade, in an appropriate buffer. The homogenate is filtered to remove large particulate materials that would otherwise clog the flow cytometer. The samples are then stained with DNA-specific fluorochromes and analyzed. Alternatively, staining and homogenization can be done concurrently.
In early work, the absolute values for the amounts of the plant nuclear DNA were determined by comparison to standards of known DNA content (such as chicken red blood cells (1) or plant standards (2)). Standardization is complicated by the types of fluorochrome employed. In general, it has been found that UV-excited fluorochromes, such as DAPI and Hoechst 33257, produce DNA histograms of the highest quality (operationally defined as the lowest half-CVs for the G0/G1 peak). However, these dyes display considerable base-pair bias (AT-specificity) (3, 4). Fluorochromes such as mithramycin and chromomycin A3 also display base-pair bias in the opposite direction (GC-specificity), but share with the UV fluorochromes an absence of overlap with endogenous plant fluorochromes. The intercalating fluorochromes, such as propidium iodide (PI), with staining in the presence of RNAase to eliminate binding to double-stranded RNA, display a slight GC bias (4), in addition, for most instruments they generally produce DNA histograms with G0/G1 peaks having elevated half-CVs as compared with the UV fluorochromes, and they suffer from the additional problem of simultaneous excitation and detection of endogenous plant fluorochromes (such as chlorophyll and the accessory photopigments). Questions of standardization of plant nuclear DNA content measurements have been thoroughly addressed by other groups (see for example, (3) and references therein), and further recommendations as to suitable standard species are available on-line (5).
In this technical note, an evaluation of the Accuri C6 flow cytometer for the analysis of plant genome sizes is described. The C6 produces DNA histograms from plant samples stained with PI for which the G0/G1 peaks display excellent CVs and which, uniquely, provide a simultaneous linear estimate of nuclear DNA content spanning at least two orders of magnitude (a range of 0.34–80.9 pg). This range encompasses most of the C-values encountered in the flowering plants.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
Table 1 provides binomial identifications for the plants that were employed in this study, and includes 2C DNA content values. These values were from two sources: the majority of the species are those recommended by Dolezel et al. (3, 5) as standards for DNA content estimation, for which the 2C DNA contents are provided in those references. The 2C nuclear DNA content values for three species, Arabidopsis thaliana, Medicago sativa, and Alstroemeria aurea, were obtained from the Kew C-value database (7).
Table 1. Names of plant species, source of seeds, 2C DNA contents, and attribution
|SPECIES||COMMON NAME||2C DNA CONTENT (PG)||DNA CONTENT ATTRIBUTION||SOURCE OF PLANTS|
|Arabidopsis thaliana ecotype Columbia||Thale cress||0.32||Kew C-value Database (7)||Lehle Seeds (14)|
|Raphanus sativus cv. Saxa||Radish||1.11||Olomouc website (3, 5)||J. Dolezel|
|Medicago sativa L. cv. Cimarron||Alfalfa||3.50||Kew C-value Database (7)||Lehle Seeds (14)|
|Pisum sativum L. cv. Ctirad||Garden pea||9.09||Olomouc website (3, 5)||J. Dolezel|
|Secale cereale L. cv. Dankovske||Rye||16.19||Olomouc website (3, 5)||J. Dolezel|
|Triticum aestivum L. line 812||Wheat||34.65||Kew C-value Database (7)||Lehle Seeds (14)|
|Alstroemeria aurea Grah.||Lily of the Incas||80.90||Kew C-value Database (7, 15)||Inter-American Products (16)|
Pisum sativum (garden pea) was chosen for the first set of analyses since its nuclear genome size is reasonably large, and it is known to produce uniparametric DNA distributions that are of high quality (3). The flow cytometric analyses would be expected to show a population of nuclei that would be reasonably well-separated from cellular and subcellular debris. Figure 1A illustrates a representative biparametric frequency distribution obtained for homogenized pea seedling leaves. Within this distribution (Fig. 1A) are two major clusters. One, comprising most of the detected objects, represents debris, and lacks correlation between the two fluorescence channels. The second, a very minor proportion (1.8%) of the objects, displays a strong correlation between the intensity values for the two channels, as would be expected for the fluorescence emission of the propidium iodide-DNA complex. Gating around this population (region P) provides a uniparametric DNA distribution having two well-defined peaks, respectively corresponding to the G0/G1 and G2 nuclei (Fig. 1B). The means of the peak positions were values of 533,898.3 and 1,065,516.0, or ratios of 1.000 to 1.996. The CV value for the G0/G1 peak was 2.1% (when determined using the CFlow software, based on manual positioning of windows at the points on either side of the peak, representing 50% of the peak value) and 2.55% (when determined using SigmaPlot, with fitting of a Gaussian to the extracted data). The general appearance of this histogram is typical of the nuclear DNA contents found within nonendoreduplicated tissues (1, 8).
Figure 1. Flow cytometric analysis of homogenates prepared from Pisum sativum (pea) seedling tissue. (A) Biparametric contour plot of FL2-A (585/40nm) versus FL3-A (>670nm) fluorescence emission. (B) Uniparametric histogram of F2-A fluorescence, gated on region P1 of panel A.
Download figure to PowerPoint
In the next series of experiments, we turned our attention to Arabidopsis thaliana (thale cress). This species was chosen for two reasons. First, Arabidopsis is the premier genomic model for plant biologists, and therefore is of considerable intrinsic interest. Second, it contains one of the smallest genomes of flowering plants, being ∼157 Mb (9), corresponding to a 2C value of 0.32 pg. For both reasons, the ability to readily distinguish nuclear DNA-specific PI fluorescence from endogenous autofluorescence and from other PI-stained components, would be highly desirable. Figure 2A illustrates a biparametric contour plot obtained from Arabidopsis leaf homogenates. This species displays somatic endoreduplication within most of its tissues and organs (8), and this is reflected in the form of multiple clusters within the distribution corresponding, in this case, to nuclei forming a 2C, 4C, 8C, 16C, etc., endoreduplicative series. These clusters again fall on a discrete diagonal which can be clearly observed following magnification of the appropriate region (Fig. 2B). Gating around this region (P1, corresponding to only 1.2% of the detected signals, provides uniparametric distributions with well-defined peaks (Fig. 2C). The mean fluorescence values of the peak positions were 24541.9, 47446.1, 91698.5, and 179634.1; these values fit a straight line almost perfectly (r2 = 0.9999). The CVs were 3.6%, 2.7%, 2.2%, and 2.4%. For Arabidopsis root homogenates, the clusters representing the nuclei in the biparametric contour plot were similar to those of the leaves (Figs. 2D and 2E). The nuclei comprise a slightly larger proportion of the signals detected by the flow cytometer (region P1), most likely due to the absence of chloroplasts from the homogenate. In addition, two additional discrete regions are observed having characteristics of PI staining, being parallel to the region occupied by the nuclei but lacking the discrete C-values imposed by the characteristics of the cell cycle. Again, gating on region P1 provides uniparametric histograms of excellent quality and linearity (Fig. 2F): the fluorescence values of the peaks were 22773.6, 44930.8, 87703.1, and 171246.2; the peak positions again fit almost perfectly to a straight line (r2 = 0.9999). The CVs were 3.5%, 2.8%, 3.1%, and 3.4%. The slight increase in the absolute fluorescence of leaf nuclei as compared with those of roots may reflect passive absorbance by the nuclei of endogenous orange/red fluorescent fluorochromes, such as those of the photosystems, following chopping.
Figure 2. Flow cytometric analysis of homogenates prepared from Arabidopsis thaliana (thale cress) leaf (A-C) and root (D-F) tissues. (A) Biparametric contour plot of FL2-A versus FL3-A fluorescence emission. (B) Enlargement of the square region-of-interest (R1) containing the nuclei. The nuclei are enclosed by polygonal region P1. (C) Uniparametric histogram of F2-A fluorescence, gated on region P1 of panel B. (D) As for panel A, except using roots. (E) As for panel B. (F) As for panel C. Abbreviations: 2C, 4C, 8C, 16C, 32C designate the appropriate C-values for the individual peaks.
Download figure to PowerPoint
To evaluate the day-to-day reproducibility of the C-value measurements, we repeated them seven times on separate days, using seedlings as the source of nuclei. The mean FL2-A value of the 2C peak positions was 21,633.6 ± 1588.5 (mean ± SD).
The next series of experiments explored the overall linearity of PI-based C-value measurements over the dynamic range of the vast majority of the reported values for angiosperms. We separately chopped plant tissues from the eight species indicated in Table 1, stained them with PI, and mixed them in various combinations. Figure 3 illustrates combined analysis of leaves of Arabidopsis thaliana, Pisum sativum, Triticum aestivum, and Alstroemeria aurea. Combining these different homogenates, containing many different sources of subcellular debris, does not obscure the characteristic diagonal region occupied by the PI-stained nuclei (Figs. 3A and 3B). On magnification of the region, it is clear that the Alstroemeria nuclei produce FL3-A signals that are off-scale, but these can be captured using a suitably placed polygonal window (P2), and the resultant FL2-A uniparametric distributions have well-defined peaks corresponding to the nuclei of the different species (Fig. 3C). Assignment of the six peaks to nuclei of individual species was done from the results of accumulation of histograms corresponding to analyses of the unmixed homogenates of individual species (data not shown). Thus in Figure 3C, the individual peaks and their positions and CVs are: Arabidopsis 2C, 20868.2, 4.2%; Arabidopsis 4C, 40333.2, 3.5%; Arabidopsis 8C, 77941.5, 2.8%; Pisum 2C, 453340.4, 2.1%; Triticum 2C, 1584795.7, 2.8%; Alstroemeria 2C, 3486157.8, 2.4%. The peak position values are strongly correlated with the reported nuclear DNA content values (Figs. 3D and 3E; r2 > 0.99). Results of identical quality were obtained on mixing homogenates of Arabidopsis, Raphanus, and Medicago (Fig. 4). In this case, the peak positions and CVs were: Arabidopsis 2C, 19650.6, 3.2%; Arabidopsis 4C, 39376.0, 2.7%; Raphanus 2C, 62541.0, 2.1%; Arabidopsis 8C, 77541.7, 2.9%; Raphanus 4C, 123377.9, 2.1%; Arabidopsis 8C, 152557.9, 3.2%; Medicago 2C, 199289.1, 2.6%.
Figure 3. Simultaneous analysis of mixtures of plant homogenates, illustrating the dynamic range and linearity of measurements that can be achieved. (A) Biparametric contour plot of FL2-A versus FL3-A fluorescence emission. (B) Enlargement of the square region-of-interest (R2) containing the nuclei. The nuclei are enclosed by polygonal region P2. (C) Uniparametric histogram of F2-A fluorescence, gated on region P2 of panel B. (D) Plot of DNA content versus the mean fluorescence values of the 2C peak positions for the four species. (E) As for panel D except employing a log scale, allowing identification of the six DNA content peaks. Abbreviations: As for Figure 2, and: At, Arabidopsis thaliana; Ps, Pisum sativum; Ta, Triticum aestivum; Aa, Alstromoeria aurea.
Download figure to PowerPoint
Figure 4. Simultaneous analysis of mixtures of plant homogenates, illustrating linearity of the measurements at low C values. Combined flow analysis was done using mixtures of homogenates of Arabidopsis, Raphanus, and Medicago. The mean fluorescence intensity (FL2-A) values of the nuclei are plotted as a function of DNA content. Abbreviations: As for Figure 3, and: Rs, Raphanus sativus; Ms, Medicago sativa.
Download figure to PowerPoint
In the final series of experiments, we investigated the impact on the measurement of C-values of combining instrument variation with experimental and biological variation. We separately chopped and analyzed Arabidopsis, Raphanus, Medicago, Pisum, Secale, and Triticum plants. We then determined the positions of the 2C peaks. The peak positions and CVs were: Arabidopsis, 20435.2, 3.6%; Raphanus 2C, 63006.1, 2.5%; Medicago 2C, 199289.1, 2.6%; Pisum 2C, 522700.8, 2.5%; Secale 2C, 918138.8, 2.4%; Triticum 2C, 2015272.8, 2.2%. These peak position values were almost perfectly correlated with DNA content (Fig. 5).
Figure 5. Sequential analysis of plant homogenates. Flow analysis was sequentially done using homogenates separately prepared and stained with PI from plants of Arabidopsis, Raphanus, Medicago, Pisum, Secale, and Triticum. Mean fluorescence values were determined for the 2C peaks, and plotted as a function of DNA content. (A) Linear scale. (B) Logarithmic scale.
Download figure to PowerPoint
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
The nuclear DNA contents (C-values) of higher plants span a remarkable range of values. The Kew C-value database (7) reports the smallest angiosperm 2C value to be that of Fragaria viridis Duch. (0.20 pg), estimated using flow cytometry combined with propidium iodide staining. The largest reported 2C value using the same method of estimation is that of Trillium apetalon Makino [190.00 pg; (10)], and the largest using any method of estimation (Feulgen staining) that of Fritillaria assyriaca Baker [254.80 pg; (11)]. Thus, angiosperm C-values span approximately three decades of DNA content. Nevertheless, of the 4,427 database entries, 4,365 (>98%) have 2C DNA contents lower than that of Alstroemeria aurea (80.9 pg), and no crop species has a 2C value higher than this.
Although the determination of plant nuclear DNA contents using flow cytometry is a well-established and robust technique (12, 13), a number of problems are typically encountered using standard instruments in these measurements, particularly if the C-value is unknown (and it should be pointed out that C-value estimations are available only for about 2% of the angiosperms). First, the chopping method (1) releases large quantities of cellular and subcellular debris. Thus, in comparison to situations typical in flow analysis of animal cell suspensions, in which the object of interest (the cell) represents most of the population, for analysis of C-values using plant homogenates, the objects of interest (the nuclei) at best represent a very minor population. Plant debris both scatters light and can exhibit autofluorescence. Since light scatter signals are generally used for triggering flow cytometric measurements, careful adjustment of thresholds is required to allow visualization of the nuclei; without appropriate thresholding, auto rescaling tends to obscure the presence of the nuclei on frequency distribution displays, and sample acquisition, based on total counts of scattering particles, arrests prior to accumulation of data for adequate numbers of nuclei. Autofluorescence, particularly that produced by the photosynthetic organelles of aerial tissues and organs, can also overlap the emission of DNA-specific fluorescent signals and, since chloroplasts greatly outnumber nuclei within the cells of green tissue, thereby obscure the presence of nuclei. The second problem relates to the choice of a standard against which to determine the 2C value for the unknown. This is ideally done in the form of an internal standard, using a plant of known 2C value, and is most accurate when the 2C values of the known and unknown are similar but not identical. Experience with previous instruments indicates data accumulation under conditions of linear signal amplification is preferred over logarithmic amplification, due to historical performance limitations of the log amplifiers. Given the extreme ranges of 2C values encountered in the plant kingdom, it can therefore be quite tricky to establish appropriate conditions and standards, for accurate flow analysis of nuclear DNA contents. Finally, for certain species, the process of homogenization can release secondary products that adversely affect the process of staining (3).
I was interested in evaluating the Accuri C6 flow cytometer for general plant C-value determinations. This instrument employs a 24-bit analog-to-digital converter (ADC) for signal processing, which provides an exceptional, six decade, dynamic range. Because this dynamic range is larger than that of the known 2C values for the described flowering plants, this suggested it might be possible to employ the C6 flow cytometer for accurate 2C value determinations for all angiosperm species, no matter what their genome size might be. To evaluate the instrument, a series of plant species having defined genome sizes were selected for analysis (Table 1). Use of propidium iodide/RNAase as the DNA-specific fluorochrome was a constraint of the laser excitation wavelengths available for the instrument, but this constraint also avoids significant base-pair bias inherent to other fluorochromes such as DAPI, the Hoechst dyes, and mithramycin/chromomycin (3, 4). Analysis of these species separately indicated it was very simple to define the region of interest within biparametric distributions that roughly split the fluorescence emission of PI-DNA into shorter and longer wavelength spectral components (Figs. 1–3). Correlation between the fluorescence emissions within these two spectral bands results in the appearance of an angled linear region within the biparametric contour plots containing discrete peaks of fluorescence, the major peak representing the predominance of individual 2C nuclei within G0/G1 cells (Figs. 1–3). Gating around this region then leads to the acquisition of very clean uniparametric histograms, displaying the expected distributions of nuclei occupying either a simple cell division cycle (Pisum; Fig. 1) or an endoreplicating cycle (Arabidopsis; Fig. 2). Debris from different tissue sources, for example leaves (Figs. 2A and 2B) and roots (Figs. 2C and 2D), can be readily excluded from consideration, and mixing homogenates from different plant species (Figs. 3A and 3B) does not impede nuclear analysis.
The linearity of measurement of plant genome sizes based on PI fluorescence extends over a dynamic range from 0.32 to 80.9 pg DNA (Figs. 3D and 3E). As far as the author is aware, this is the first time this relationship has been demonstrated over such a large range. The slight GC bias reported for PI staining and the positive correlation for angiosperms in GC content and genome size (4) do not represent impediments to this determination. This may be due to two factors: first, the observation of GC bias was based on animal species, not plants (4). Second, a study repeating the analysis, but using a much larger number of species than that used in reference (4) and spanning a much greater range of DNA content values, revealed no correlation between genome size and GC content across angiosperms as a whole (17).
The upper limit to C-value measurements using the Accuri C6 can be estimated from the observed mean position for the 2C peak of Alstroemeria (80.9 pg; channel 3486157.8) and the highest bin value that is available on the instrument (∼16 × 106), namely an upper 2C limit of about 370 pg DNA. This exceeds the largest record in the Kew C-value database by a factor of about 150%. Extending polygonal gate P2 (Fig. 3) around the PI-DNA region downward below the Arabidopsis 2C peak revealed no evidence of potentially confounding noise from debris for FL2-A values as low as 2,000 (data not shown). This value would correspond to a nuclear 2C DNA content of 0.032 pg, which is about six-fold smaller than the smallest record for the angiosperms (7).
Taken together, this suggests the C6 flow cytometer should be able to measure any angiosperm 2C value without modification to the instrument, and without the requirement for even simple adjustments such as insertion of neutral density filters. The accuracy of measurement of the nuclear DNA contents, reflected by the low CVs, was consistently high across the entire range of measurements. It is concluded that, for routine analysis of plant nuclear DNA contents, ploidy, or investigations of other issues requiring C-value determinations, the Accuri C6 flow cytometer provides an excellent measurement platform. The high degree of reproducibility in peak positions over different biological samples and preparation times further implies that the peak position values can be directly converted into DNA content values, which greatly facilitates these measurements in high-throughput situations. Our results also imply that an internal standard may not necessarily be needed during routine operation, since the combination of instrument, biological, and experimental (i.e. staining) errors did not significantly degrade the accuracy of the DNA content determinations. A number of empirical caveats should be noted: first, internal standards do permit recognition of species whose homogenization releases secondary products that can affect the staining process (3). The extent of this problem should become clear as the numbers increase of species sampled for C-value. Second, it is not recommended that any sample, standard or otherwise, be allowed to stand prior to analysis for longer than the times described in the experimental protocol. Finally, it is recommended to regularly repeat external standardization, using known plant homogenates or calibration microspheres, to ensure detection of any changes in overall instrument performance.
Taken together, setting a goal of determining the C-values for the remaining 98% of the angiosperms appears reasonable, at least in terms of the measurement platform. The homogenization and flow analysis pipeline can easily accommodate 12 samples per hour, which for ∼280,000 species would require about 24,000 h. Amortized over 20 instruments (a capital investment of ca. $600,000), this reduces to 1,200 h per instrument, which is clearly feasible. Caveats would include problems of sample collection and identification, and of samples recalcitrant to the chopping method (3). The small platform size of the Accuri C6, coupled to its minimal electrical and liquid input requirements, would allow this instrument to be placed in laboratories close to the sample sources, thereby avoiding the impact of regulations governing biodiversity export.