DNA Content Variation and Its Significance in the Evolution of the Genus Micrasterias (Desmidiales, Streptophyta)

It is now clear that whole genome duplications have occurred in all eukaryotic evolutionary lineages, and that the vast majority of flowering plants have experienced polyploidisation in their evolutionary history. However, study of genome size variation in microalgae lags behind that of higher plants and seaweeds. In this study, we have addressed the question whether microalgal phylogeny is associated with DNA content variation in order to evaluate the evolutionary significance of polyploidy in the model genus Micrasterias. We applied flow-cytometric techniques of DNA quantification to microalgae and mapped the estimated DNA content along the phylogenetic tree. Correlations between DNA content and cell morphometric parameters were also tested using geometric morphometrics. In total, DNA content was successfully determined for 34 strains of the genus Micrasterias. The estimated absolute 2C nuclear DNA amount ranged from 2.1 to 64.7 pg; intraspecific variation being 17.4–30.7 pg in M. truncata and 32.0–64.7 pg in M. rotata. There were significant differences between DNA contents of related species. We found strong correlation between the absolute nuclear DNA content and chromosome numbers and significant positive correlation between the DNA content and both cell size and number of terminal lobes. Moreover, the results showed the importance of cell/life cycle studies for interpretation of DNA content measurements in microalgae.


Introduction
Streptophycean green algae are a sister group to land plants [1][2][3]. Desmids are the unicellular representatives of Zygnematophyceae (Streptophyta), and characterized by thousands of morphotypes. They are a logical target group for investigating nuclear DNA content variation [4]. They possess both asexual and sexual reproduction within their life cycle, however sexual reproduction is rare and its study lags behind that of other groups of organisms. Particularly the nuclei behaviour during the reproduction has not been studied in the majority of genera. Thus we still presuppose that desmids with other members of the Zygnematophyceae are haploid in the vegetative stage, only the zygospore being diploid [5]. Although some cases of higher ploidy level in vegetative stage has already been published in the species Closterium ehrenbergii [6], the life cycle of this species seems to be extraordinary. In contrast to C. ehrenbergii producing 2 zygotes per pairing cells, other genera (Micrasterias, Cosmarium, Euastrum) are producing 1 zygote per pair of sexualized cells [7], [8].
Cytogenetic research on Desmids is intriguing due to the presence of holocentric (holokinetic) chromosomes [9] which have a kinetochore located along the whole chromosome. In general, this type of chromosome is uncommon though described in some higher plant families (Cyperaceae and Juncaceae) and in the genera Drosera, Myristica, Cuscuta and Chionographis [10]. It is also described in some algae, arthropods and nematodes, including the model Caenorhabditis elegans. By their nature, holocentric chromosomes have high potential for rapid chromosome rearrangements as chromosome fragments can undergo both normal meiosis and mitosis. For this reason, holocentric organisms are characterised by enormous diversity in both number and size of chromosomes. Chromosome number in desmids ranges from 14 to 592 [11] and numbers characteristic of an aneuploid series have also been found [12], [13]. These phenomena suggest that symploidy or agmatoploidy prevails over polyploidy in these taxonomic groups.
Polyploidization, i.e. whole genome duplication (WGD), is considered one of a major evolutionary process in higher plants [14], [15]. Since early reviews of polyploidy in the late 1930's, the number of known polyploid events in the evolution of higher plants has been continuously growing along with the development of new research methods [15]. Recently, with an advent of new sequence techniques it has become evident that by far the vast majority of flowering plants have experienced (ancient) polyploidisation in their evolutionary history [16][17][18][19]. Polyploid series have their origin in reproductive isolation at different ploidy levels due to unbalanced meiosis. These are often treated as different taxa in higher plants at ranks of varieties to species. Several processes at the genome level, such as the reciprocal loss of duplicated genes (fractionation), gene silencing, chromosome rearrangements and others, might lead to cytotype differentiation at different levels, i.e. morphologically, ecologically. Desmidiales has been characterized by extensive polyploidy, with both interand intraspecific variation in chromosome complements reported [20]. Although there is practically no information about natural populations, changes in the chromosome complements of desmid cells have been artificially induced in cultures. Stable polyploid forms (triploid, tetraploid) of numerous Micrasterias species have been produced by Waris and Kallio [21]. Diploid cells were always larger than haploid in all Micrasterias species in which these conditions were induced [5].
In higher plants, detection of polyploid taxa within polyploid series became a routine procedure with the development of DNA flow-cytometry (FC) using several different protocols [22][23][24][25][26]. However, the microscopic size and highly variable cell wall composition of microalgae has limited its use for microalgae [27][28][29]. Thus, study of genome size variation in microalgae lags behind that of higher plants and seaweeds [4], [30]. The multi-step protocol proposed by Mazalová et al. [31] has been found useful for quantification of DNA content in Streptophyta, particularly desmids and a microalgal standard for FC measurement has been suggested.
Recent taxonomic research on microalgae suggests that traditional species/genera boundaries based largely on cell morphology have underestimated the real species diversity [32], [33]. In addition, numerous traditional genera and higher taxa proved to be polyphyletic so that they have to be revised using molecular phylogenetic methods [34][35][36]. Micrasterias, a model genus of desmids (Desmidiales, Streptophyta), includes about 60 morphospecies and its cryptic diversity, phylogeny and biogeography have recently been reported [37][38][39][40][41]. The results of the recent phylogenetic studies indicate monophyletic origin of the genus followed by substantial morphological transformation of individual infrageneric lineages [36], [37], [40], [42], [43]. According to a multigenic phylogeny of 41 Micrasterias taxa, the genus comprises at least eight lineages [40]. Mapping morphological diversification of the genus, on the phylogenetic tree has revealed profound differences in the phylogenetic signal of selected phenotypic features. Whereas the branching pattern of the cells clearly correlates with the phylogeny, the morphological complexity possibly reflects their adaptive morphological response to environmental conditions [40].
Kasprik [44] recognized four groups within the Micrasterias species based on chromosome morphology. The first group have small chromosomes with a tendency to aggregation and includes mostly representatives of clade A (, Figure 1), with the exception of M. americana from clade H. The second group possessing wellseparated chromosomes, includes representatives of clade G (, Figure 1), with the exception of M. rotata from clade C. The third group with short, thick, relatively compact chromosomes belongs mostly to clades C and D except for M. muricata from clade H (, Figure 1). The fourth group, characterized by long, compact chromosomes which appear to be joined together, includes M. thomasiana.
In this study, we asked whether the phylogeny of the genus Micrasterias is associated with DNA content variation. To answer this question, we focussed on: 1) assessment of overall DNA content variation; 2) recognition of the significance of DNA content in the evolution at generic and species levels, and 3) the correlations between DNA content and selected cell morphometric parameters.

Materials and Methods
No specific permits were required for the described field studies. No specific permission was required for any locations and activity. The locations are not privately owned or protected in any way. No activity during field study involved any endangered species or protected species.

Origin and cultivation of strains
The strains we used were obtained from five public culture collections: Sammlung von Conjugaten-Kulturen, University of Hamburg (SVCK); Culture Collection of Algae, Charles University in Prague (CAUP); Culture Collection of Algae, University of Vienna (ASW), currently deposited in the Culture Collection of Algae at the University of Cologne (CCAC); Culture Collection of Algae, University of Göttingen (SAG); and Microbial Culture collection, National Institute for Environmental Studies, Tsukuba (NIES). Some strains come from the personal collection of Jiří Neustupa [39] (Table 1). They were grown in 50 mm plastic Petri dishes in a liquid oligotrophic medium used in the CAUP culture collection (OGM; [45]). Storage cultures were kept at a temperature of 16uC, under an illumination of 20 ìmol. m 22 . s 21 with 12:12 light:dark cycle (cooling box Helkama C5G). Subsequently, two weeks before planned flow cytometric measurements, a rich inoculum of each strain (ca 1 ml) was transferred to fresh medium in a 100 mm Petri dishes and kept at a higher irradiation (40 ìmol. m 22 . s 21 ) with 16:8 light:dark cycle. All cultures were regularly examined under inverted microscope (Zeiss Axiovert), to check their fitness, cell density and estimation of cell division intensity. Well grown cultures were used for cytometric measurements in exponential phase of growth (14-18 days after subculturing), few slow growing cultures were maintained longer to obtain sufficient cell density. Although the cell density was priority, intensively dividing cultures were preferred because their cell walls were more sensitive to enzymatic disintegration [31]. For three independent repetitions of DNA measurements of each algal strain, we used three Petri dishes of material.

Flow cytometry
Protoplasts were isolated using an enzymatic mixture of 2% Cellulase Onozuka R-10 (Duchefa Biochemie, Netherlands), 0.5% Macerozyme R-10 (Duchefa Biochemie, Netherlands) dissolved in modified rinsing solution PGly [31]. The absolute quantity of nuclear DNA of the algae was estimated by FC using a ML CyFlow instrument (Partec GmbH, Münster, Germany). Raphanus sativus cv. Saxa (2C = 1.11 pg, [46]), Lycopersicon esculentum cv. Stupicke (2C = 1.96 pg, [46]), Zea mays, CE-777 (2C = 5.47 pg, [22]) or Pisum sativum cv. Ctirad (2C = 8.76 pg, [47]) were used as standards. Algal suspension was measured alone at first (without plant standard) to find out the best standard. Young leaves of standard plants were grown separately from algae. They were not affected by enzyme mixture, but chopped with a razor blade in lysis buffer. Then the algal suspension was added to the chopped standard, filtered, incubated and stained. As a fluorescent dye for DNA staining for FC measurements, propidium iodide was used. The whole multi-step protocol has been published in detail elsewhere [31].
All desmid strains were measured at least twice, usually three times, except of strain CAUP K604 with unsufficient amount of desmid cells (extremely slow growth). Altogether 3000 particles were measured (algae + standard together). Each of the repetitive measurements of the same strain was independent (desmid suspension originated from different culture vessels, flow-cytometric measurements were done other day after new calibration of cytometer). Thus the possibilities of errors in DNA measurements were minimized. As the majority of cultures represented proliferating cultures with sufficient amount of dividing cells (synchronized by photoperiod), the histogram of relative nuclear DNA content usually shows 2 peaks -first peak represents vegetative haploid cells (1C), second peak (2C) represent cells in G2 mitosis phase ( [48], [49]; Figure 2). The presumption, that the representatives of the genus Micrasterias are haploid in vegetative stage is based on the fact, that all published and observed Micrasterias species are producing 1 zygote per two pairing cells [7], [8], [50][51][52], no records of two zygotes as in the case of Closterium ehrenbergii [6] has been found. All clonal cultures used for measurement were strictly vegetative, neither sexual reproduction nor zygotes were observed during regular microscopic inspection, thus none of the peak represent generative polyploidy caused by hybridization. In few cases the histogram shows 1 peak only. In these cases we used microscopic inspection for decision, whether this peak represents extremely synchronized cells in G2 mitosis phase, or extremely slowly growing strains with dominating vegetative (haploid) cells.

Phylogenetic analysis, character evolution
The phylogenetic tree used to map the character of the evolution was created using Bayesian inference (BI), MrBayes, version 3.1 [53]. The analysis was conducted on the concatenated SSU rDNA, psaA, and coxIII alignment published by Š kaloud et al. [40], reduced to 26 taxa for which the cytometric data were obtained. The genetic identity of M. 'pusilla' (NIES 783) and M. radians var. evoluta (SVCK 519) strains with those molecularly characterized by Š kaloud et al. [40] has been described by Neustupa et al. [37] and Nemjová et al. [39]. Subsequently, eight M. rotata and M. truncata strains having identical SSU rDNA sequences with those included in the alignment ( [38]; [39]) were added to the dataset. The final alignment included 34 taxa.
The existence of a phylogenetic signal in DNA content variation was tested by calculating Pagel's lambda [54]. The maximum likelihood optimization of Lambda value was performed using the 'phylosig' function of the phytools package [55]. The ancestral states of DNA content data were calculated using the Ape and Geiger packages [56], [57]. The ancestral states were reconstructed by the function ''ace'', using the maximum likelihood optimization. All calculations were done in the program R, ver. 3.0.2. released on 2013-09-25 (website The R Project for Statistic Computing. Available: http://www.r-project.org/. Accessed 2013 Dec 19). The output from R was mapped onto the Bayesian phylogenetic tree with TreeExtender v1.03 [58], using a simple list parser (option -p list). Finally, the evolution of DNA content was  traced on a Bayesian phylogenetic tree as colours along a gradient with TreeGradients v1.03 [58].

Correlation analyses, geometric morphometrics
Correlation analyses were conducted separately on 1) a dataset of all taxa for which the cytometric data were obtained, and 2) a subset of five M. rotata strains. The relationship between DNA content and selected cell morphometric parameters (cell length, cell complexity and number of terminal lobes) was evaluated by linear correlation analyses in PAST, ver. 2.01 [59]. The morphometric parameters tested were obtained from Š kaloud et al. [40]. To analyze the correlation between the DNA content and cell size in the M. rotata subset precisely, we determined the centroid size (CS) of the cells per each strain by a landmark-based shape analysis.
Centroid size, a widely used dimension-free size measure is linearly correlated with traditional univariate cell size measures of desmids, such as cell length or width [60]. It is defined as the square root of the sum of squared distances from all the analyzed landmarks to their centroid [61].
For each M. rotata strain, 46-51 randomly chosen cells were photographed. In total, we defined 40 landmarks spanning the outline shape of the cells (see Neustupa et al. [38] and [60]

Overall DNA content variation
In total, DNA content was successfully determined for 34 strains of Micrasterias species, including the phylogenetically nested species Triploceras gracile (Table 2). To analyse both interspecific and intraspecific DNA content variation, several different strains belonging to the species M. rotata and M. truncata were analysed, as well. The estimated absolute 2C nuclear DNA quantity varied from 2.1 to 64.7 pg. The smallest genomes belonged to morphologically distinct species M. dickiei, M. swainei, M. pinnatifida, M. decemdentata and M. conferta. In contrast, the biggest genomes were detected in M. ceratofera, one strain of Triploceras gracile and several strains of M. rotata (Table 2). Generally, the majority of investigated strains possessed rather small DNA content. In fact, the estimated 2C DNA content did not exceed 10 pg in almost half of the strains. Interestingly, high variability in DNA content was also detected among different strains belonging to a single species. In M. truncata, the estimated absolute 2C nuclear DNA quantity varied from 17.4 to 30.7 pg; in M. rotata strains from 32.0 to 64.7 pg.
To test the correlation between the estimated DNA content values and real number of chromosomes, we compared our measured values with the chromosome data published by Kasprik [44]. For the purpose of this test, we used six Micrasterias SVCK strains used in both studies (Table 2 and 3). As can be seen from the regression analysis (Figure 3), a linear regression fits the data very well (R 2 = 0.974, P-value,0.001), suggesting a strong correlation between absolute nuclear DNA content and chromosome number.

Evolution of DNA content
To better understand the evolutionary history of DNA content changeover during the diversification of the genus Micrasterias, we mapped the estimated DNA content values along the phylogenetic tree. A phylogram was constructed by the Bayesian inference method, based on the concatenated SSU rDNA, psaA, and coxIII alignment published by Š kaloud et al. [40]. Since the protoplast isolation and subsequent DNA content determination was not successful in all strains [31], our phylogram contains only 26 of 36 Micrasterias species with known sequence data. The topology of the resulting phylogenetic tree was highly congruent with the phylogram constructed by Š kaloud et al. [40], with strains inferred as members of 6 main clades (A, C, D, E, G, and H).
The existence of a phylogenetic signal in nuclear DNA content was tested using Pagel's lambda calculations. The related species were significantly different in their DNA content (ë = 0.413, pvalue = 0.34). However, the existence of a phylogenetic signal was revealed by the maximum likelihood reconstruction of the ancestral states ( Figure 1). Whereas all species of clade G showed a clear tendency to rather low DNA content, all six M. rotata strains forming the clade C tended to have much higher amounts of DNA. On the other hand, clear differences were found in some closely related strains, indicating potential rapid evolutionary dynamics in the genome size. For example, two Triploceras gracile strains differed six times in their 2C nuclear DNA content (9.4 and 56.4 pg, respectively). Similarly, M. crux-melitensis (14.3 pg) and M. radians var. bogoriensis (6.6 pg) had 2C DNA content lower than the closely related M. ceratofera (39.2 pg).

Correlations between DNA content and cell morphometric parameters
In addition to analysis of the evolutionary significance of the DNA content variation, correlations between the genome size and selected morphometric characteristics were determined. Statistical relationships, described by correlation coefficients between the estimated DNA content and cell length, cell complexity and number of terminal lobes, are presented in Table 4 and illustrated in Figure 4A-C. A significant correlation was found between the DNA content and both average cell length ( Figure 4A) and number of terminal lobes ( Figure 4B). The correlation with the    average cell length was the most significant. In fact, 42.3% of the overall variability in cell lengths could be explained by the genome size data. As mentioned above, significant differences in DNA content were also found between different strains of a single species. To test whether these differences could also affect the cell size, we analyzed the correlation between DNA content and centroid size in five M. rotata strains having identical SSU rDNA sequences. Strong positive correlation was found (Table 3; Figure 4D) and altogether 84.5% of cell size variability detected could be explained by the different nuclear DNA contents of analyzed strains.

Discussion
Overall DNA content variation and correlations with chromosome numbers Genome size measurement in algae is presently at the beginning and, frequently methodologically more demanding than in embryophytes. This is also reflected in the amount of the Kew Plant DNA C-values database algae entries of only 253, in contrast to the presently over 7500 angiosperm C-value entries [62]. The results reported in this paper constitute one of few flow cytometric analysis in microalgae [31], [63][64][65] and first such analysis in desmids.
Measurements of DNA content in 34 strains of the genus Micrasterias, ranging from 2C = 2.1 to 64.7 pg, are in congruence with previously published data, based mostly on microspectrophotometric methods. For Streptophyta, 2C nuclear DNA content range from 0.2 to 6.4 pg, excluding the highly polyploid Charales and Desmidiales, which have genome sizes of up to 14.8 and 46.8 pg, respectively. In general, nuclear DNA content is smaller in Zygnematales than in Desmidiales [30]. Algae are a highly diverse organism group and phylogenetically at the basis of land plants. Therefore, a broader knowledge of genome size would be enormously valuable for the evaluation of the role of nuclear DNA amount in evolution. Nuclear DNA content data for Streptophyta superimposed on a contemporary molecular phylogeny indicated that early diverging lineages, including some members of Chlorokybales, Coleochaetales and Klebsormidiales, have genomes as small as 2C = 0.1-2.7 pg [30], [31]. It has been proposed that the Streptophyte ancestral nuclear genome, common to both the charophyte and the embryophyte lineages, can be characterized as 1C = 0.2 pg and 1n = 6. Moreover, the DNA contents of the freshwater charophyceans and zygnemataleans are significantly larger than that of Rhodophyta (2C = 0.2-2.8 pg) and Phaeophyta (2C = 0.2-1.8 pg) [4]. Although greater values for DNA content exist in polyploid bryophytes, more than 80% of the nuclear DNA 1C-values in mosses have been reported to occur in a narrow peak between 0.25 and 0.6 pg [66]. The size of algal genomes is best appreciated when compared with the minimum amount of DNA estimated in angiosperms. The smallest angiosperm genome is known in the   [68]. Many of the Zygnemataceae appear to be characterized by polyploid ''species complexes'' [20] and the large cell sizes reported for many Desmidiales, suggest that polyploidy in these uninucleate, unicellular organisms has produced some of the largest nuclear genome sizes known in plants. These suggestions are in congruence with our measurements. Broad variation has been found in different clones of the same species: Micrasterias rotata, M. truncata, and Triploceras gracile (Tables 2 and 3). Angiosperms with holocentric chromosomes have in general, large genome size variation. Individual species of the genus Carex have almost 8-fold variation in their 4C nuclear DNA content [69], genus Luzula up to 6-fold variation in their 2C nuclear DNA content [70], [71]. Likewise, in the genus Schoenus up to 14.8-fold variation in 2C DNA content [72] and in Eleocharis even up to 22.1-fold (0.25 pg in E. acicularis and 5.53 pg in E. palustris; [73]) were reported.
Although some of the desmid strains were maintained as long as 40 years and changes in chromosome numbers should be expected in cultures, our results support the earlier findings within Desmidiales [4] that chromosome complements and nuclear DNA contents are highly correlated, providing circumstantial evidence for the pervasive role of polyploidy in the evolution of this group of algae. In contrast, in the marine species of Ulvophyceae there is low correlation, consistent with a high occurrence of aneuploidy, i.e. chromosomal fusion and/or fission events. On the larger scale, genome variation in holocentric genera does not always correlate with chromosome numbers. Although the chromosome number in Carex varies greatly, between 6 and 62, genome size remains nearly constant [10], [74]. Similar dissonance between the chromosome number and DNA content is found in another holocentric organisms, such as Schoenus [72] and Juncus [75].
The dynamic nature of holocentric chromosomes can be demonstrated by the nuclear DNA variation within a single variety of sedge species Carex scoparia var. scoparia, in which the chromosome counts varies from 2n = 62 to 2n = 68 and the 1C DNA content varies from 0.342 pg to 0.361 pg [76].
Because of overall lack of experience with flow cytometric measurements of the DNA content in microalgae, we were solving a lot of methodological problems, during our study. Beside the problems with nuclei isolation due to the complexity of cell wall discussed previously [31] we noticed the peak pattern variation in histograms obtained from individual measurements.
Histograms of well growing cultures in exponential phase of growth showed two peaks, first representing vegetative cells (G1), second representing dividing cells (G2). We believe, that for desmid genus Micrasterias first peak represent haploid cells, i.e. 1C DNA content. The height differences of both peaks (G1.G2 or G1,G2) depend on stage of cell cycle [65]. Although Chlamydomonas represent different microalgal group and type of cell/life cycle, authors found that G1 peak is always observable and never represents less than 29%, because in each stage exist some cells, which are not dividing. The percentage of both peaks depend on degree of synchrony and division rate which both depend on complex of factors including culture conditions and cell size [65]. Thus flow cytometric measurements should be accompanied by careful inspection of measured material, which serve as a base for interpretation of such histograms, which show 1 peak only (see Methods).
The exact knowledge of cell cycle and degree of synchrony of each species/sample should lead to exact interpretation, but represent another complication which makes flow cytometry in microalgae extremely time consuming. Moreover, the lack of information on reproduction of the majority of microalgal species, complicate our interpretation of ploidy level. Vegetative cells (G1) can be haploid (some desmids) or diploid (e.g. diatoms), depending on their life histories.

Significance of DNA content in evolution
Genome size is an important species-specific characteristic in organisms with centromeric chromosomes and was extensively studied in past decade in order to describe its changes during the evolution of angiosperms [77]. The correlation between the Cvalue and other traits was extensively investigated, not only from the taxonomic point-of-view, but also from a broader biological perspective. In general, the DNA amount serves as a reliable taxonomic indicator which can help to understand the taxonomic problems in angiosperms [78]. Unfortunately limited data are available for microalgae.
Some studies on holocentric angiosperms [72] suggest that some clades demonstrate a narrow range of genome size variability, whereas others exhibit great variation. Although, this phenomenon can be found in taxa with monocentric chromosomes too it seems to be more typical/frequent in holocentric genera. Indeed, similar results were obtained in Micrasterias. For example, clade G consists of strains with a narrow range of genome size variability. However, within other clades, clear differences were found in some closely related strains, indicating the potentially rapid evolutionary dynamics in the genome size. Even more, great variability in genome size was detected in different strains of a single species (Micrasterias rotata, M. truncata and Triploceras gracile).

Influence of DNA content on cell size and morphology
Small genome size in angiosperms appears to be correlated with phenotypic characteristics such as rapid seedling establishment, short minimum generation times, reduced cost of reproduction and increased reproductive rate [79][80][81][82]. It has been recognized that although nuclear genome size is highly correlated with many cellular and ecological parameters, ''correlation'' and ''causation'' are far from interchangeable [83][84][85].
Ploidy level in conjugating green algae may be of taxonomic significance as cell dimensions are considered to be diagnostic [20] and highly correlated with genome size [86]. Diploid cells in Micrasterias were found to be usually larger than that of haploids. Waris and Kallio [21] and Brandham [87] observed that larger cells of Closterium, Cosmarium and Staurastrum were polyploid.
Kasprik [44] reported an interesting case of an aneuploid series correlated with morphological differences in Micrasterias thomasiana. The basic chromosome number of this desmid is n = 39, but morphologically irregular variants contained n = 40, n = 46 or even n = 70 and 75. Some of the cells of this latter clone were inclined to develop more or less typical morphologies, except that they were significantly larger [44]. Correlations between the DNA content and cell length, cell complexity and number of terminal lobes were tested in this study. The significant positive correlations were found between the DNA content and both average cell length and number of terminal lobes. Changes in the degree of radiation in some desmids, result from an increase in the level of ploidy, as concluded by Starr [88] in his study of a heterothallic strain of Cosmarium turpinii. Starr considered the production of large forms to be a response to increase in nuclear quantity; whereas the change in shape (bi-, tri-, quadriradiate cells) he explained as a response to the increase in cell volume.
We can support partially this hypothesis by our data. In contrast to the number of terminal lobes variation, cell complexity was not significantly correlated with DNA content.

Conclusions
This is the first flow cytometric analysis of both interspecific and intraspecific DNA content variation within one microalgal genus. There was strong correlation between nuclear DNA content and chromosome number in strains of the genus Micrasterias and between DNA content and cell size and morphology in the species Micrasterias rotata.
Moreover, the results showed the importance of cell/life cycle studies for interpretation of DNA content measurements in microalgae.