Condensation patterns of prophase/prometaphase chromosome are correlated with H4K5 histone acetylation and genomic DNA contents in plants

Mitotic prophase chromosome condensation plays an essential role in nuclear division being therefore regulated by highly conserved mechanisms. However, degrees of chromatin condensation in prophase-prometaphase cells may vary along the chromosomes resulting in specific condensation patterns. We examined different condensation patterns (CPs) of prophase and prometaphase chromosomes and investigated their relationship with genome size and distribution of histone H4 acetylated at lysine 5 (H4K5ac) in 17 plant species. Our results showed that most species with small genomes (2C < 5 pg) (Arachis pusilla, Bixa orellana, Costus spiralis, Eleutherine bulbosa, Indigofera campestris, Phaseolus lunatus, P. vulgaris, Poncirus trifoliata, and Solanum lycopersicum) displayed prophase chromosomes with late condensing terminal regions that were highly enriched in H4K5ac, and early condensing regions with apparently non-acetylated proximal chromatin. The species with large genomes (Allium cepa, Callisia repens, Araucaria angustifolia and Nothoscordum pulchellum) displayed uniformly condensed and acetylated prophase/prometaphase chromosomes. Three species with small genomes (Eleocharis geniculata, Rhynchospora pubera, and R. tenuis) displayed CP and H4K5ac labeling patterns similar to species with large genomes, whereas a forth species (Emilia sonchifolia) exhibited a gradual chromosome labeling, being more acetylated in the terminal regions and less acetylated in the proximal ones. The nucleolus organizer chromatin was the only chromosomal region that in prometaphase or metaphase could be hyperacetylated, hypoacetylated or non-acetylated, depending on the species. Our data indicate that the CP of a plant chromosome complement is influenced but not exclusively determined by nuclear and chromosomal DNA contents, whereas the CP of individual chromosomes is clearly correlated with H4K5ac distribution.


Introduction
general, Arabidopsis-and Solanum-type CPs are found in species with low nuclear DNA content, whereas Hordeum-and Allium-types are usually found in species with higher DNA content [16,17]. Similarly, a correlation between genome size and chromosomal distribution of some posttranslationally modified histones has been reported for plant species. For example, the chromosomal distribution of the histone H3 dimethylated at lysine 4 (H3K4me2) in Arabidopsis thaliana (130 Mpb), H. vulgare (5440 Mpb), and V. faba (12810 Mpb), was distinct to each species, but it was almost identical among species with similar chromosome size and nuclear DNA content [18]. In V. faba and H. vulgare, both with large chromosomes, the labeling pattern observed with H3K4me2 was very similar to that obtained with the histone H4 acetylated at lysine 5 (H4K5ac) [19,20]. Likewise, in Costus spiralis and Eleutherine bulbosa, both with small chromosomes and Solanum-type CP, the distribution of H3K4me2 and H4K5ac signals were identical [21], suggesting that a general correlation between the distribution of these histones and CPs may exist. Actually, chromosome regions enriched in H3K4me2 and H4K5ac were found to be early replicating [18,22], therefore, they are expected to be late condensing.
H4K5ac is a universal mark for euchromatin and is directly involved in the control of chromatin replication and indirectly associated to chromatin condensation [23][24][25]. In general, early condensing chromatin (such as heterochromatin, G bands, and epignetically silenced chromatin, as the inactive X chromosome of female mammals) is hypoacetylated in H4K5 [25][26][27]. In the few plant species where the chromosomal distribution of H4K5ac was investigated by immunodetection, different patterns of acetylation were observed [21,22,28], but systematic efforts to correlate acetylation and CP have not been undertaken. It is interesting to note that the acetylation level of H4K5 in some chromosome sites may change during the cell cycle [20,[23][24][29][30]. Further, specific chromatin types, as the nucleolus organizer region (NOR), may be hyperacetylated in some species [19,31] and hypoacetylated or non-acetylated in others [21], indicating that not all epigenetic marks are conserved in plants.
In this work, we analyzed the CPs, the nuclear and chromosomal DNA contents and the distribution of H4K5ac in 17 plant species with different chromosome sizes. To identify the acetylation levels of eu-or heterochromatic regions, the chromosomes were sequentially stained with the fluorochromes chromomycin A3 (CMA) and 4'-6-diamidino-2-phenylindole (DAPI) (which differentially stains most of the C-band heterochromatin) after immunodetection of H3K5ac. The distribution of heterochromatic bands was previously determined by C or CMA/DAPI banding for most of the species examined here (e.g., [12,[32][33][34][35][36][37]). We have formerly reported that the hyperacetylated chromatin fraction observed in prophase chromosomes remains roughly the same in metaphase chromosomes [21]. Since the distribution of H4K5ac signals is more clearly defined when chromosomes are more condensed and separated, during prometaphase to metaphase, the characterization of the distinct acetylation patterns in the present sample was mainly performed in cells in these stages.

Plant material
Seeds of Araucaria angustifolia, Phaseolus vulgaris, and P. lunatus and bulbs of A. cepa were obtained from commercial sources. Seeds of Solanum lycopersicum cv IPA-5 and Arachis pusilla were kindly supplied by Instituto Agronômico de Pernambuco (IPA) and Prof. Reginaldo de Carvalho (Universidade Federal Rural de Pernambuco), respectively. The remaining species were weeds or cultivated plants growing on the campus of the Federal University of Pernambuco, Recife, Brazil. A list of all investigated species is presented in Table 1.
For Giemsa and CMA/DAPI staining, root tips were pretreated in 0.002 M 8-hydroxyquinoline (for species with small to medium-sized chromosomes) or 0.2% colchicine (for large chromosomes) for 24 hours at 10˚C, fixed in a 3:1 ethanol-acetic acid solution, and subsequently stored at −20˚C. For immunostaining, the roots tips were pretreated as above, fixed for 40 min in 4% paraformaldehyde, and subsequently rinsed in PBS.

Giemsa and CMA/DAPI staining and H4K5ac immunodetection
The protocols used for Giemsa and fluorochrome staining were identical to those previously used with most of the species examined here (see e.g., [12]). Briefly, for conventional staining with Giemsa, young root tips were washed in distilled water, hydrolyzed in 5N HCl, and squashed in 45% acetic acid. The coverslips were removed in liquid nitrogen and the slides stained with 2% Giemsa. For CMA/DAPI double staining, the root tips were digested in a 2% cellulase-20% pectinase solution and squashed in 45% acetic acid. The coverslips were removed in liquid nitrogen and the slides were left to age for three days. They were then stained with CMA (0.1 mg/ml) for 60 min and subsequently counterstained with DAPI (2 μg/ ml). The slides were then mounted in glycerol/McIlvaine (1:1) and kept in the dark for three days before analysis.
The protocol for immunostaining was identical to that used with Costus and Eleutherine by Feitoza and Guerra [21]. The root tips were digested in a cellulase-pectinase solution and squashed in PBS buffer. The coverslips were removed in liquid nitrogen and the slides incubated in a blocking solution with 3% BSA (w/v) containing 0.1% Triton X-100 in PBS. The primary anti-H4K5ac antibody used (rabbit polyclonal IgG, Upstate Biotechnology, USA) was Nuclear DNA contents DNA content estimations were performed using a CyFlow SL (Partec) flow cytometer and Flomax software (Partec), following the protocol described by Doležel et al. [38]. Young leaves of the specimens examined were fragmented (using a razor blade) together with an internal standard, in 0.5 ml of cold Otto I buffer (0.1 M citric acid + 0.5% Tween 20), and filtered through a nylon membrane (45 μm pores). After the addition of 0.5 ml Otto II buffer (0.4 M Na 2 H-PO 4 .12H 2 O) supplemented with RNAse (50 μg/ml) and propidium iodide (PI) (50 μg/ml), the samples were immediately analyzed. Exceptionally, for Callisia repens, fragmentation was performed in WPB (woody plant buffer), followed by filtering, and supplemented with RNAse and PI, following Loureiro et al. [39].  [38]. Seeds of all of the internal standards were furnished by Dr. J. Doležel (Institute of Experimental Botany, Olomouc, Czech Republic). An internal standard was selected for each plant analyzed so that its genome size was close to, but not overlapping, that of the analyzed sample. Nuclear DNA contents (2C) were calculated by the equation: (Sample peak mean/Standard peak mean) × 2C DNA content of standard (pg). Each species was evaluated three different times on three different days. We also estimated the genome sizes of E. bulbosa and C. spiralis from the same cultivated samples previously analyzed for their H4K5ac distributions [21].

Estimates of nuclear DNA content
The DNA contents of Callisia repens, Eleocharis geniculata, Emilia sonchifolia, Indigofera campestris, Rhynchospora pubera, R. tenuis, C. spiralis, and E. bulbosa were estimated here for the first time and are presented in Table 1, together with the internal standard used for each. The 2C values previously estimated for the remaining species (A. cepa, A. angustifolia, Arachis pusilla, Bixa orellana, Nothoscordum pulchellum, Phaseolus vulgaris, P. lunatus, and Poncirus trifoliata) and their respective average chromosomal DNA contents (2C/2n values) are also indicated in Table 1. The value obtained for B. orellana (2C = 0.66 ± 0.03) was~50% greater than previously reported [40]. In order to check for technical errors, we analyzed a second sample collected in a private garden using a different standard species (Raphanus sativus cv. Saxa, 2C = 1.11, [38]) as well as a different buffer (WPB), but the 2C value obtained remained nearly the same (0.65 ± 0.02 pg). The smallest 2C and 2C/2n values observed here were for B. orellana and P. trifoliata (0.66 and 0.04 pg respectively) while the highest value found was with N. pulchellum (2C = 49.94 pg; 2C/2n = 4.99 pg).

Correlation between CP and DNA content
Chromosomes of most species with low DNA content (2C < 5 pg) displayed typically well condensed regions (deeply stained with Giemsa), and less condensed regions (weakly stained) (Figs  1 and 2). In species with the Solanum-type CP, the amount of early condensing chromatin was quite variable, depending on the species investigated. A typical Arabidopsis-type CP was not found. The group of species with Solanum-type CP included B. orellana (2C = 0.66 pg), P. trifoliata (2C = 0.78 pg), P. vulgaris (2C = 1.20 pg), P. lunatus (2C = 1.40 pg), I. campestris (2C = 1.85 pg), S. lycopersicum (2C = 2.13 pg), E. bulbosa (2C = 2.63 pg), C. spiralis (2C = 3.32 pg), and A. pusilla (2C = 4.18 pg) ( Table 1). Bixa orellana stood out for having a bimodal karyotype, with six small chromosome pairs and a larger one. During prometaphase, one of the arms of the largest chromosome pair was more deeply stained than the other (Fig 2C1 and 2C4). This chromosome pair had a proximal secondary constriction that was usually highly distended, leaving the chromosome arms far from each other (arrows in Fig 2C4 point to the heterochromatic arms whereas the other two arms were associated with each other by the CMA + NORs). Likewise, E. bulbosa had a bimodal karyotype with a single pair of large chromosomes (~8 μm, corresponding to 42.5% of the total chromosome complement length or 0.56 pg) that were almost entirely condensed during prometaphase, except for a very small terminal region, whereas the five pairs of small chromosomes displayed overall low condensation [41].
The genome of E. sonchifolia (2C = 1.44 pg) was slightly larger than that of the other species with low nuclear DNA content (Table 1). Its chromosomes were almost uniformly condensed during prophase and prometaphase, with the short arms being more condensed than the long arms ( Fig 3A1) due to their pericentromeric heterochromatin [42], whereas the interstitial to terminal regions of the long arms were slightly less condensed (Hordeum-type CP).

Heterochromatin differentiation with CMA/DAPI
Staining with the fluorochromes CMA and DAPI identified CMA + bands in one or more chromosome pairs in all of the species investigated. These bands corresponded to all of (or most of) the known C-bands for those species [12,34,36,39,[43][44]. No DAPI + band was clearly observed. Noteworthy, the heterochromatic bands in all species with Solanum-type CP constituted only a small fraction of the condensed chromatin.
The CMA + bands of I. campestris, P. lunatus, P. vulgaris, and A. pusilla were located in the proximal regions of all of the chromosomes (Figs 1 and 2). In B. orellana and A. angustifolia there was a single pair of CMA + /DAPI ˗ band, which was located in the proximal region and corresponded to the NOR (Figs 2C7 and 4D4). Additional pericentromeric heterochromatin on the smaller chromosomes of B. orellana and interstitial to terminal bands observed on its largest chromosome pair were identified only by C-banding [36] but not by CMA/DAPI staining. In both Phaseolus species, besides the proximal CMA + bands, there were additional terminal CMA + bands corresponding to the NORs (Fig 1B5 and 1C4). NORs were identified as secondary constrictions distended at metaphase or the only CMA + bands in the karyotype, as in B. orellana (Fig 2C7), E. sonchifolia (Fig 3A5), N. pulchellum (Fig 4C4), and A. angustifolia (Fig 4D4). In P. lunatus, N. pulchellum, P. trifoliata, E. sonchifolia, and S. lycopersicum the NORs had also been previously identified by FISH with rDNA probe [12,33,35,39,43]. In S. lycopersicum, besides the NORs there were some very small bands on other chromosomes ( Fig  1D4). Small CMA + bands were observed in the terminal regions of all of the chromosomes of A. cepa, including the NORs localized terminally in the short arms of a submetacentric pair (Fig 4A4). In P. trifoliata there were 16 CMA + bands, six of which correspond to rDNA sites [33] (Fig 2A3).
the terminal region to the proximal one); (c) localized (signals restricted to the low condensed regions). The localized pattern was found in all species with the Solanum-type CP (Table 1; Figs 1 and 2). In P. trifoliata, the chromosomes with a terminal CMA + band displayed often a subterminal labeling (Fig 2A4, upper row). The small chromosomes of B. orellana differed from the largest pair in that the former were hyperacetylated in the both interstitial and terminal regions, while the large chromosome pair was weakly and irregularly labeled in these regions (Fig 2C5-7).
Uniformly acetylated prophase/prometaphase chromosomes were found in all species with large genomes (A. cepa, C. repens, N. pulchellum, and A. angustifolia) or with holokinetic chromosomes (R. tenuis, R. pubera, and E. geniculata) (compare Fig 3B-3D with Fig 4). Labeling had often a granular appearance, sometimes displaying small gaps, quite different from the compact and more brilliant blocks observed in species with differentially condensed chromatin. A gradual labeling pattern was observed only in E. sonchifolia, which displayed chromosomes more acetylated in the terminal region of the long arms and less acetylated in the short arms and pericentromeric regions (Fig 3A3-4). Remarkably, NORs were the only chromosomal regions that could be unequivocally identified as non-acetylated (Figs 1B4, 2C5-7 and 3A3-5), hypoacetylated (Fig 4A3, 4C3 and 4D3) or hyperacetylated (Fig 1C3 and 1D3), depending on the species.
In prophase to metaphase cells of A. pusilla and most other species of the section Arachis, there is a small, typically under-condensed chromosome pair, called A chromosome [45]. Even at metaphase this kind of chromosome displayed only a very small amount of completely condensed chromatin and was entirely hyperacetylated (Fig 2B2-5). Observe that the A chromosomes were very weakly stained with both DAPI (Fig 2B2) and CMA (Fig 2B5).

Chromosome condensation pattern and DNA content
All of the species with the Solanum-type CP that were investigated here, including E. bulbosa and C. spiralis, had low nuclear DNA content (2C < 5 pg) and low mean DNA content per chromosome (2C/2n < 0.5 pg). The species with the Allium-type CP had either large genomes (2C > 5 pg) and large amount of DNA per chromosome (> 0.5 pg), or small genomes and holokinetic chromosomes. Holokinetic chromosomes are characterized by multiple centromeric units intercalated with small patches of euchromatin [46]. This regular and punctate distribution of euchromatin could explain the uniform acetylation pattern found in these species, independent of genome size. A small genome size was also found in Emilia sonchifolia, the only species in the present sample displaying the Hordeum-type CP. Therefore, the amount of DNA per genome (or per chromosome) is associated with the different CPs but it is not enough to determine them. A similar correlation between nuclear/chromosomal DNA content and CP had been previously observed in 17 species of Rutaceae [17]. Likewise, species with areticulate and semi-reticulate interphase nuclei (associated with Arabidopsis/Solanum-type CP) had low DNA content (< 5 pg) whereas those with reticulate and eureticulate nuclei (Hordeum/Allium-type) had higher DNA content [16]. In both these studies there were some exceptions to the general trend, suggesting that other factors may also contribute to the establishment of a specific CP.

Chromosome condensation pattern and H4K5 acetylation
The H4K5 acetylation patterns observed here are in agreement with those reported for other plant species with known CPs, such as A. thaliana, Silene latifolia, V. faba, and H. vulgare [9,20,22]. Our data further confirmed the acetylation pattern previously described for A. cepa [22] and Phaseolus vulgaris [28], reinforcing the reliability of the methodology used.
In the present sample we found three different patterns of H4K5ac distribution: an uniform labeling, in species with the Allium-type CP; a gradual labeling, associated with the Hordeumtype CP; and a localized labeling, in species with the Arabidopsis/Solanum-type CP. These patterns were quite similar to those observed for H3K4me2 [18,47], suggesting that both histone modifications contribute to determine the chromatin condensation timing and CPs.
Notably, all of the chromosomes of a karyotype display the same CP and similar distribution of H4K5ac [19,22,28], H3K4me2 [18,47], and some other epigenetic marks [21,30,47]. However, the proportion of high/low condensed chromatin may vary between prophase/prometaphase chromosomes of a karyotype exhibiting the Arabidopsis/Solanum-type PC. For example, during the prometaphase of E. bulbosa only the chromosome pair I was almost entirely condensed and unlabeled with anti-H4K5ac [21], whereas in A. pusilla only the A chromosome pair was almost completely undercondensed. The differential proportion of high and low condensed chromatin observed in these chromosome pairs was directly correlated with the amount of repetitive DNA sequences accumulated in each one. In E. bulbosa, the main satellite DNA sequences and most retroelements are largely accumulated in the chromosome pair I [48], while in Arachis species the A chromosome pair exhibit the lowest concentration of repetitive sequences [49]. Likewise, the more condensed chromatin fraction of prophase/prometaphase chromosomes of S. lycopersicum is highly enriched in retrotransposons and has a gene density 10-100 times lower than the less condensed euchromatin [50,51]. Therefore, the CP should be primarily determined not by chromosome size or chromosomal DNA content but rather by the amount of repetitive sequences and by the way they are distributed along the chromosomes.

H4K5 acetylation levels in euchromatin and heterochromatin
In prophase/prometaphase chromosomes of species displaying the Arabidopsis/Solanum-type CP, the less condensed euchromatin fraction differed strongly from the more condensed one in their acetylation levels, indicating that hyperacetylation is not a universal mark of euchromatin (at least when we consider the euchromatin concept based on C-banding). For example, the interstitial to proximal condensed region of most prophase/prometaphase chromosomes of S. lycopersicum includes only a small fraction of heterochromatin [12] but it is uniformly non-acetylated.
Heterochromatic bands, on the other hand, are known to be cytologically lacking of H4K5ac signals [18,21] and indeed most of them were non-acetylated. However, the pericentromeric C-bands present in all of the chromosomes of E. sonchifolia [42] and the terminal heterochromatic bands of A. cepa, as well as the pericentromeric C-bands of A. cepa reported by some authors [32,52], were weakly acetylated (see also Vyskot et al. [22]). The largest chromosome pair of B. orellana had one chromosome arm completely heteropycnotic, mainly composed by C-bands [34], which was partially labeled with anti-H4K5ac. These data indicate that some heterochromatic regions may be considerably enriched in H4K5ac, as previously observed in other organisms [24,29,[53][54].
A particular kind of heterochromatin positively stained for C-banding and CMA is the NOR [42]. In the present sample, NORs could be hyperacetylated (Phaseolus vulgaris, P. lunatus, and S. lycopersicum), weekly acetylated (A. cepa, A. angustifolia, and N. pulchellum), or non-acetylated (B. orellana, E. bulbosa, P. trifoliata, and E. sonchifolia). The acetylation level in these cases was independent of the condensation status of the NORs in prophase and metaphase cells. This is an unusual condition where a CMA + band (NOR) can be hyperacetylated [19][20][21]23] and the acetylation pattern does not match perfectly the heteropycnotic pattern.

Conclusions
Our data indicate that chromosome condensation is a karyotype feature that affects all chromosomes of a complement in a similar way. This is most evident in chromosomes with Allium-type CP, but it is equally true for all other CPs. The different patterns are strongly associated with the chromosomal distribution of H4K5ac and only partially correlated with nuclear and mean chromosomal DNA contents. The repetitive DNA sequences of species with the Arabidopsis/Solanum-type CP are largely concentrated in the more condensed and non-acetylated proximal chromosome regions, whereas in Allium/Hordeum-type CP they are apparently finely spread throughout of the chromosomes. The change from one distribution pattern to the other is in some way related to changes in the average DNA amount per chromosome, but other karyological features, as for example massive expansion of repetitive sequences or the presence of holocentromeres, may play a dominant role in the establishment of a new CP in all chromosomes of the complement.