New reports of nuclear DNA content for 407 vascular plant taxa from the United States

Abstract

Background and Aims

The amount of DNA in an unreplicated haploid nuclear genome (C-value) ranges over several orders of magnitude among plant species and represents a key metric for comparing plant genomes. To extend previously published datasets on plant nuclear content and to characterize the DNA content of many species present in one region of North America, flow cytometry was used to estimate C-values of woody and herbaceous species collected in Wisconsin and the Upper Peninsula of Michigan, USA.

Methods

A total of 674 samples and vouchers were collected from locations across Wisconsin and Michigan, USA. From these, C-value estimates were obtained for 514 species, subspecies and varieties of vascular plants. Nuclei were extracted from samples of these species in one of two buffers, stained with the fluorochrome propidium iodide, and an Accuri C-6 flow cytometer was used to measure fluorescence peaks relative to those of an internal standard. Replicate extractions, coefficients of variation and comparisons to published C-values in the same and related species were used to confirm the accuracy and reliability of our results.

Key Results and Conclusions

Prime C-values for 407 taxa are provided for which no published data exist, including 390 angiosperms, two gymnosperms, ten monilophytes and five lycophytes. Non-prime reports for 107 additional taxa are also provided. The prime values represent new reports for 129 genera and five families (of 303 genera and 97 families sampled). New family C-value maxima or minima are reported for Betulaceae, Ericaceae, Ranunculaceae and Sapindaceae. These data provide the basis for phylogenetic analyses of C-value variation and future analyses of how C-values covary with other functional traits.

INTRODUCTION

The size of an organism's genome reflects a fundamental aspect of its biology as well as a character of considerable practical use (Bennett and Leitch, 2005, 2011; Beaulieu et al., 2007a, b, 2008; Leitch et al., 2009, 2010; Chung et al., 2012; Janousek et al., 2012; Leitch and Leitch, 2012). Genome sizes vary by >2000-fold among the angiosperms, from fewer than 10⁷ base pairs (1C = 0·065 pg in Genlisea margaretae, Lentibulariaceae; Greilhuber et al., 2006) to more than 10¹¹ (1C = 152·23 pg in Paris japonica, Melanthiaceae; Pellicer et al., 2010). Such variation exceeds that of other groups. This could reflect the propensity of plants to accumulate repetitive DNA via transposable elements and chromosomal duplications and/or less efficient mechanisms for purging such redundant DNA via unequal crossing over and deletions. This variation presumably reflects a complex interplay of neutral and selective processes acting at several levels of organization. Changes in genome size among 87 species of Carex appeared to occur at similar rates across the phylogeny and did not covary with chromosomal rearrangements (Chung et al., 2012). More specifically, within the genus Gossypium, variation in the rates at which particular retrotransposons accumulated in different lineages account for most of its 3-fold increase in genome size over the last 5–10 Myr (Hawkins et al., 2006). Bennetzen et al. (2004) also found that retrotransposons had been rapidly removed from the smaller genomes of Oryza sativa and Arabidopsis leading them to hypothesize that variation in the efficiency of this process may account for much of the variation in plant genome sizes.

Genome sizes have been characterized in terms of their C-value, the number of base pairs or picograms of DNA present in an unreplicated haploid or gametic nucleus (Swift, 1950). Greilhuber et al. (2005) clarified the term ‘C-value’ by equating it with the holoploid genome size, i.e., the whole chromosome complement of an individual (versus the Cx-value, which is the DNA content of the unreplicated monoploid set of chromosomes in a polyploid series).

Since 1976, Bennett and colleagues have published nine compilations of C-value data (Bennett and Smith, 1976, 1991; Bennett et al., 1982, 2000; Bennett and Leitch, 1995, 1997, 2005, 2011; Zonneveld et al., 2005). Starting in 1997, these data have been shared via the Angiosperm DNA C-values Database (Bennett and Leitch, 1997). To date, researchers have estimated C-values for 7058 plant species as reported in the most recent Plant DNA C-values Database (Release 5·0) (http://data.kew.org/cvalues/) (Bennett and Leitch, 2010, 2011). The database includes summaries of statistical, taxonomic, cytological (i.e. ploidy level, chromosome number), technical (i.e. method used to estimate genome size) and bibliographic information, as well as search functions to facilitate sorting, filtering and analysing these data. This enhanced accessibility and utility has increased the use of C-value data across a variety of studies. The electronic databases have now been cited >230 times with >250 000 hits (Bennett and Leitch, 2011).

Early estimates of plant species C-values from the 1950s and 1960s relied on laborious chemical extraction methods. Subsequent techniques, like Feulgen microdensitometry and flow cytometry, made estimating genome sizes easier and considerably faster. Since 1983, flow cytometry has been the primary method used to estimate genome sizes in angiosperms (Galbraith et al., 1983; Galbraith, 2009). Analysis in the current database carried out by Bennett and Leitch (2011) shows that flow cytometry continues to be the dominant method for estimating C-values in recent years (84·5 %), while Feulgen microscopy methods now represent just 15·4 % of the estimates.

Researchers use DNA C-values to address questions in cellular, developmental, ecological, evolutionary and molecular biology as well as systematics, physiology and paleontology (Bennett et al., 2000; Bennett and Leitch, 2005, 2010; Leitch and Bennett, 2007). Recent work includes studies of the relationships between genome size and seed mass (Beaulieu et al., 2007b), photosynthetic rate (Beaulieu et al., 2007a), leaf cell size and stomata density (Beaulieu et al., 2008; Hodgson et al., 2010) and patterns of genome size evolution (Leitch et al., 2005, 2009; Beaulieu et al., 2010; Chung et al., 2012). Šmarda et al (2012) assessed the suitability of flow cytometry for estimating GC (cytosine + guanine) content by comparing it with DNA temperature melting analysis and found high correspondence. Flow cytometry continues to develop as a useful method, increasing the quantity and quality of genome-size estimates and thus our ability to access and analyse these data across a broad set of taxa.

Discussion continues regarding the occurrence, definition and significance of observed variation in estimates of angiosperm C-values. This reflects the multiple factors that can affect the stability of C-values, including the extraction buffer (with >28 different buffers in use), how raw flow cytometry data are analysed (forward scatter, side scatter and relative fluorescence intensity), which plant species are used as standards and the potentially confounding effects of cytosolic compounds (e.g. anthocyanin and tannic acid) (Doležel et al., 2007a, b; Bennett and Leitch, 2011; Loureiro et al., 2006, 2007; Greilhuber et al., 2007).

Here, we report estimated C-values for 514 taxa (i.e. species, subspecies and varieties) of wild plants collected in the states of Wisconsin and Michigan, USA. Most of these reports (79 %) are new to the Plant DNA C-values database. However, sufficient overlap exists to closely compare results from this study with previous work. We had two goals in collecting these data. First, we sought to augment the amount of genome size data available for research generally. More specifically, we sought to construct a database of C-values for Wisconsin plant species to explore phylogenetic patterns of C-value variation and relationships between genome size and other functional traits. This work represents an initial phase of a broader effort to explore how genetic, phylogenetic and functional diversity among Wisconsin plants affect their responses to drivers of ecological change (www.botany.wisc.edu/dob/).

MATERIALS AND METHODS

Plant material

First, we identified a list of vascular plant species for which we already had ecological data concerning their abundance and distribution over several hundred sites sampled in the 1950s and again in the 2000s (www.botany.wisc.edu/PEL/). We sought to sample fresh leaf material from these species from across the region (Table 1). Fresh leaf material was stored at 4 °C for no more than 3 d until it could be processed. For each sample, we collected a herbarium voucher to serve as a permanent record of the species identity and deposited these at WIS (the Wisconsin State Herbarium in Madison). Identities of all field-collected material were confirmed by taxonomists [W. S. Alverson and/or T. Cochrane (Wisconsin State Herbarium)] to ensure accurate identifications. Scanned images of all voucher specimens will become available by late 2012 at www.botany.wisc.edu/dob/. We also collected small back-up samples of leaf tissue in silica gel. These were used in the few cases where the fresh tissue did not provide a C-value. This sometimes worked, but neither sample produced a useable C-value in other cases.

Table 1.

General provenance of 674 samples collected for this study in Wisconsin and Michigan, USA

County of sample collection	No. of samples	Latitude N range	Longitude W range
Ashland County, WI	3	46°17′–46°18′	90°38′–90°39′
Bayfield County, WI	1	46°20′–46°21′	91°13′–91°14′
Dane County, WI	350	42°56′–43°17′	89°02′–89°49′
Door County, WI	27	44°52′–45°14′	86°59′–87°26′
Douglas County, WI	5	46°32′–46°37′	91°53′–92°08′
Goegebic County, MI	64	46°10′–46°14′	89°00′–89°12′
Grant County, WI	33	43°11′–43°12′	90°31′–90°32′
Green County, WI	23	42°31′–42°43′	89°21′–89°46′
Iowa County, WI	6	42°54′–43°06′	89°53′–90°01′
Iron County, WI	3	46°07′–47°08′	90°00′–90°01′
Jefferson County, WI	7	42°54′–42°55′	88°51′–88°52′
Langlade County, WI	6	Approx. 45°22′	Approx. 89°14′
Lincoln County, WI	1	45°11′–45°12′	89°42′–89°43′
Oneida County, WI	13	45°28′–45°51′	89°10′–89°41′
Sauk County, WI	93	43°21′–43°24′	89°45′–89°59′
Vilas County, WI	37	46°00′–46°10′	89°22′–89°41′
Walworth County, WI	2	42°47′–42°48′	88°39′–88°40′

Estimating DNA C-values

We estimated 2C-values of DNA content per leaf cell nucleus using flow cytometry based on the fluorochome propidium iodide (Sigma) as described by Doležel et al. (2007b). We relied primarily on the LB01 and Otto isolation buffers (Doležel et al., 1989; Otto, 1990; Doležel and Gohde, 1995) using the methods (reagent preparation, selection of standards, etc.) laid out initially by Otto (1990) and available on the flow cytometry methodology webpage (www.ibot.cas.cz/fcm/method.html). In cases where the LB01 buffer did not provide a reliable estimated C-value, we tried again using the Otto buffer to extract nuclei.

In brief, we placed approx. 1 cm² of leaf tissue from both the sample and a known plant standard into a 100-mm-diameter polystyrene Petri dish with 1·0–1·5 mL of ice-cold buffer. We then finely chopped the tissue using a new single-edge razor blade for 1 min. After mixing this solution using a disposable graduated transfer pipette, we filtered it through a section of 30-μm disposable nylon filter (Celltrics^®, Partec) into a 5-mL polystyrene round-bottom tube (Falcon^®). This isolated the somatic nuclei from much of the cellular debris in the isolation buffer. This suspension was kept on ice until tested. In contrast to the standard method, we did not centrifuge the filtered isolation buffer (finding better results without that step). Next we added 20 µL of a propidium iodide stock solution and 20 µL of an RNase stock solution to the nuclei suspension (both 1 mg mL⁻¹, giving final concentrations of 0·05 mg mL⁻¹) and vortex-mixed the suspension at 2000 rpm for 5–10 s. We then measured fluorescence to estimate C-values on a BD Accuri^® C6 flow cytometer. Usually we measured several samples, each with >5000 nuclei and evaluated the coefficient of variation (where CV = standard deviation/mean channel number, as per Ormerod, 2008). If the CV was >5 %, we obtained further estimates until the CV was <5 %. The 2C-value for each sample was estimated using the following formula:

(mean of sample peak/mean of standard peak) × 2C DNA amount (pg) of the standard.

When we observed other peaks corresponding to higher C-values (e.g. 4C and 8C), we only used these to infer the location of 2C samples.

Internal calibration standards

Three internal standards of different DNA content were used to calibrate our estimates: Raphanus sativus ‘Cherry Belle’, Pisum sativum ‘Dwarf Gray Sugar’ and Vicia faba ‘Bell Bean’. We determined the 2C-values of these in our laboratory (1·08, 8·77 and 26·53 pg, respectively) using Oriza sativa L. subsp. japonica var. nippobare (2C = 0·91 pg) as the reference. These values correspond closely to mean 2C-values of these taxa calculated from previous reports in the literature (1·10, 8·92 and 26·43 pg). We grew plants of our three standards in the greenhouse to ensure access to fresh young leaves. To choose the best standard for a given sample, we first searched the Plant DNA C-values Database for C-values of related species (ideally congeners). We then chose a species to use as the internal standard that was close to the published C-value but that differed by at least 1–2 pg from the expected C-value. This approach avoided the problem of having the material standard or fluorescence from cellular debris overlapping the sample peak. When the expected 1C-value of a sample was 0·1–3 pg, we used Pisum sativum as the standard. When the expected value was 3–6 pg, or >10 pg, we used Raphanus sativus or Vicia faba, respectively, as the standard. If we could find no C-value for a congener we defaulted to using the P. sativum standard, which in our pilot tests generated reliable, reproducible and discernible peaks with low coefficients of variation. When these procedures failed, we changed the standard and measured another sample. In the few cases where the sample could not be reliably estimated using this second standard (or a new buffer), we abandoned further efforts. In total, we used Pisum sativum, Raphanus sativus and Vicia faba as internal standards to estimate 569, 91 and 14 C-values, respectively.

RESULTS AND DISCUSSION

Accuracy and reliability of the estimated C-values

To assess the quality of the C-value data generated (see Supplementary Data Table S1), we made three comparisons. First we compared estimated C-values of samples independently collected within 125 taxa. Most (79, or 63 %) of these replicated samples varied by ≤5 % within taxa, which gave us confidence in those values. In the 46 other samples, estimates for particular taxa varied along an exponentially declining gradient from 6 % to 618 % with a conspicuous node at 100 % (Fig. 1). Thus, some of this variation is likely to reflect polyploidy levels within and among wild populations. Agrostis perennans showed the widest (618 %, likely a hexaploid) range in C-values.

Fig. 1.

Apparent variation in DNA content among 46 taxa for which we measured multiple independent samples that differed in estimated 2C-values by >5 %. The histogram shows the binned frequencies of the logarithm of percentage differences found among samples (highest 2C-value/lowest 2C-value). Raw values range from 6 % to 618 % (Agrostis perennans). Fitted exponential decay function parameter s = 1·409 with –2 log (likelihood) = 123·6.

We also compared our estimated C-values with 119 C-values published for the same taxa (Bennett and Leitch, 2010, 2011). The overall regression showed strong similarity (Fig. 2, R² = 0·88). For the subset of 72 values (61 %) that agreed within 30 %, the relationship was even tighter (R² = 0·99). In the remaining 47 cases, our estimates differed from previous C-value estimates by >30 %. The possible causes of these discrepancies are unclear but include ploidy variation, taxonomic errors and measurement errors. Polyploidy variation seems likely in 20 or more cases where our C-values were about 2 × , 4 × , 6× or 8× smaller or larger than previous estimates. For example, the outlying points in Fig. 2 include two independent samples of Dioscorea villosa both of which are half of the value reported in the Plant DNA C-values database. Likewise, one sample of Parthenocissus quinquefolia is one-eighth of the value in the database. Neither chromosome number counts nor ploidy-level values for these taxa are currently available in the database, though other species of Dioscorea are noted as tetra-, hexa- or octoploids. Our Leucanthemum vulgare sample also has a C-value half the size of the value in the database (where the sample is reported as a tetraploid). Another pair of outlying points, representing two independent samples of Rhamnus cathartica, are approximately one-sixth the value reported in the database (both 1C = 0·5 pg versus 1·33 pg). Here, it is noteworthy that the sample in the database was scored as a diploid – a likely mistake in light of our results. In another 16 taxa, our C-values were approx. 1·5× the previous estimates, suggesting aneuploid variation within taxa. Such aneuploid variation may be particularly tolerated in recent allopolyploids (Birchler, 2009). No chromosome counts were made to confirm polyploidy. We had confidence in our identifications of field-collected plants. We made no attempts to verify determinations of voucher material from previous studies of these taxa (where these exist). Such variation, however, does point up the importance of including proper vouchers in all serious studies of C-value variation.

Fig. 2.

Comparisons among 1C-DNA amounts of 119 taxa tested in this study versus values for the same species from the Plant DNA C-values database (Bennett and Leitch, 2010).

Over 90 % of our C-values were obtained by analysing fresh tissue with dried leaf tissue used for only a few recalcitrant taxa. Current best practices for flow cytometry stress analysing fresh tissue. Dried plant samples have proved to be suitable in some cases, though tissue dehydration has been observed to degrade the quality of flow-cytometry results (Kolář et al., 2012). Our estimates of C-values in fresh versus dried tissue were quite similar across the 37 taxa tested (R² = 0·98, P < 0·0001; Fig. 3). These encouraging results suggest that rapidly desiccating tissue over silica gel could be used to sample plants for genome sizes in situations where collecting fresh tissue is not feasible (Bainard et al., 2011). This could relax constraints on applying flow cytometry in field research in plant biosystematics, ecology and population biology.

Fig. 3.

Comparison showing 1C-values for dry and fresh tissue from 37 taxa included in this study.

Reported C-values

Reported here are the C-values for 674 samples representing 514 species, subspecies and varieties (see Supplementary Data Table S1). Of these, 407 (79 %) comprise reports for taxa with no previously reported C-value as of December 2010 (Bennett and Leitch, 2010). These new C-values represent a significant addition to the global pool of information on genome sizes, including several first values at the level of genus and family. For example, five of the 97 families and 129 of the 303 genera in Supplementary Data Table S1 represent new reports at that level for the Plant DNA C-value Database (Tables 2 and 3). Most of the C-values reported here are from angiosperms (482 taxa). However, we also report values for 17 monilophytes, nine gymnosperms and six lycophytes. New and total reports by taxonomic group, including values for the major clades within angiosperms, are provided in Table 2.

Table 2.

Number of C-value reports (new/total) in Supplementary Data Table S1 at the level of species, genus, and family, sorted by major clades (cf. APG III, 2009; Chase and Reveal, 2009; Stevens, 2012)

Clade	Species	Genera	Families
Lycophytes	5/6	2/4	0/1
Monilophytes	10/17	3/9	2/9
Gymnosperms	2/9	0/5	0/3
Basal angiosperms	1/1	0/1	0/1
Monocots	89/110	14/60	0/15
Eudicots	300/371	110/224	3/68
All angiosperms	390/482	124/285	3/84
All vascular plants	407/514	129/303	5/97

Table 3.

Families, and their representative species, first reported in this paper (i.e. not included in Bennett and Leitch, 2010: Plant DNA C-Values Database, December 2010)

Family*	Species	Clade	1C DNA (pg)
Cystopteridaceae	Gymnocarpium dryopteris Ching	Monilophyte	7·4
Elaeagnaceae	Shepherdia canadensis Nutt.	Eudicot	3·1
Molluginaceae	Mollugo verticillata L.	Eudicot	0·8
Onocleaceae	Onoclea sensibilis L.	Monilophyte	15·5
Phrymaceae	Phryma leptostachya L.	Eudicot	1·1, 1·2

* Families follow APG III (2009) and Stevens (2012).

The minimum, maximum, mean, median, mode and range of DNA amounts for the 674 samples listed in Supplementary Data Table S1 were compared with those reported for the 6287 species in the Plant DNA C-values Database (Table 4). The range of 1C-values reported here (0·3–76·3 pg; Fig. 4) represents a substantial subset of 1C-values reported in the Plant DNA C-values Database (0·0648–152·23 pg) with similar modes. Our mean (3·9 pg) and median (1·5 pg) were below those values in the Database. Our maximum–minimum range is also much narrower (254 ×, versus 2394× in the Plant C-value Database). These results indicate that the Plant C-values Database contains more species in the high tail of the C-value distribution than were found among Wisconsin plants (Leitch and Leitch, 2012).

Table 4.

Comparison of the minimum, maximum, mean, median, mode and range of C-values for 514 taxa in Supplementary Data Table S1 with data for 6287 species in the Plant DNA C-values database (Bennett and Leitch, 2010)

	Supplementary Data Table S1	Bennett and Leitch (2010)
Minimum 1C-value (pg)	0·3	0·0648
Maximum 1C-value (pg)	76·3	152·23
Mean 1C-value (pg)	3·9	5·48
Median 1C-value (pg)	1·5	2·35
Mode 1C-value (pg)	0·5	0·45
Fold range of variation (max/min)	254	2349

Fig. 4.

Log distribution of 1C-values of the 674 plant samples included in this study (from 0·3 to 76·3 pg).

The C-values here extend the ranges previous reported in a few families, most notably in the Ericaceae where a 1C-value of 29·9 pg for Monotropa uniflora greatly exceeds the value of 0·67 for Vaccinium corymbosum in the Plant DNA C-values Database. Another new maximum is Acer rubrum (Sapindaceae) with 1C = 1·7 pg (versus the previous 1·17 pg for Acer pseudoplatanus). Three species of Betula (Betulaceae) also exceed previous values with a peak value of 1C = 1·5 pg for B. papyrifera (versus the previous value of 0·90 pg for B. alba). A new minimum in Ranunculaceae was set by Coptis trifolia with 1C = 0·4 pg versus a previous low of 0·51 pg in Aquilegia vulgaris.

Research prospects

Genome size as reflected in these plant C-values represents a key diversity character as well as one that is often associated with other traits such as nuclear and cell size, seed mass, specific leaf area, growth rate and/or cell- and life-cycle length (Beaulieu et al., 2007 a, b, 2008, 2010; Hodgson et al., 2010; Leitch et al., 2005, 2009). To understand better the origins and functional significance of multiple components of plant diversity, we are also quantifying an array of other attributes for many of the 515 taxa reported here. These include sequence data which will allow us to trace phylogenetic relationships among these taxa. This will allow us to analyse the degree of phylogenetic inertia and lability of genome size within several plant groups (Soltis et al., 2003; Schneeweiss et al., 2006; Leitch et al., 2007, 2009, 2010; Lysak et al., 2009; Chung et al., 2012). Measurements of several other functional traits will also allow us to analyse how genome size covaries with these traits, correcting for phylogenetic relationships using comparative methods. We are also curious to know how genome size may be related to the local and regional shifts in plant abundance and colonization–extinction dynamics observed in many of these species (Wiegmann and Waller, 2006; Rogers et al., 2009). Characterizing the geographic distributions of these species may reveal how genome size covaries with geographic range. Finally, we aim to characterize patterns of population genetic structure and gene flow within at least 30 of these species in part so that we can examine whether genome sizes are associated with such population traits. Collectively, these and further studies will extend our ability to understand how plant traits, ecological conditions and phylogenetic history jointly influence genome size variation.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxfordjournals.org and consist of Table S1: C-values for 674 samples representing 514 species, subspecies and varieties.