1 Introduction

Heterochromatin is a characteristic component of the eukaryotic nucleus which, as opposed to euchromatin, is highly compacted, non-coding and contains highly repetitive DNA sequences (^{Swanson et al., 1981}; ^{Brutlag, 1980}). The heterochromatin in the interphase nucleus can be visualized as easily discernible heteropycnotic bodies, called chromocenters, which can vary in number and size. In the brine shrimp Artemia, a conspicuous inhabitant of hypersaline lakes and lagoons, the chromocenter number may have a diagnostic value as an indicator of species or populations (^{Badaracco et al., 1987}; ^{Abreu-Grobois and Beardmore, 1989}; ^{Colihueque and Gajardo, 1996}; ^{Papeschi et al., 2000}; ^{Gajardo et al., 2001}; ^{Torrentera and Abreu-Grobois, 2002}; ^{Papeschi et al., 2008}). Previous studies indicate that the heterochromatin in Artemia varies in quantity and quality both within and among species (^{Gajardo et al., 2002}). For instance, the New World Artemia species, A. franciscana and A. persimilis, have a significantly different chromocenter number (^{Badaracco et al., 1987}; ^{Abreu-Grobois and Beardmore, 1989}; ^{Colihueque and Gajardo, 1996}; ^{Papeschi et al., 2000}; ^{Gajardo et al., 2001}; ^{Torrentera and Abreu-Grobois, 2002}; ^{Lipko et al., 2004}). While the former exhibits a high number of these structures (5-18 chromocenters) the latter has lower numbers (<5 chromocenters). Such variation in the amount of heterochromatin is also associated with the presence of repetitive AluI sequences in the genome, with around a tenth less of these sequences being observed in A. persimilis in comparison with A. franciscana (^{Barigozzi et al., 1984}; ^{Badaracco et al., 1987}). In addition, chromosome studies of these species have revealed an interspecific variation in the diploid chromosome number, consisting of 42 in A. franciscana and 44 in the case of A. persimilis.

The evidence available on plants and animals indicates that variation in the DNA content per genome, usually positively associated with the increase in the amount of heterochromatin (^{Rayburn et al., 1985}; ^{Kao et al., 2001}; ^{Bosco et al., 2007}), may produce changes at the cellular (termed nucleotypic effect), or organismal level (^{Gregory and Hebert, 1999}). In other words, the phenotype expression would not just depend on the interaction between genotype and environment, but also on the expression of the DNA quantity, irrespective of its informational content (^{Swanson et al., 1981}; ^{Hartl, 2000}). For instance, at the cellular level, this modification may affect either the cell or nucleus size, or the duration of the cell cycle. These changes in DNA quantity have been related to variations in the biomass content, breeding season and even the physiological responses of organisms (^{Swanson et al., 1981}; ^{Hartl, 2000}).

Artemia is a crustacean that can deploy numerous physiological adaptations; these enable it to tolerate abrupt abiotic changes in the brines inhabited by different species or populations that involve mainly changes in temperature, salinity, ionic concentration and dissolved oxygen. Such changes are driven by the high evaporation rate or the rain regime (^{Gajardo and Beardmore, 2012}). Among the adaptations developed by Artemia to tolerate abiotic changes are a high osmoregulatory capacity, the efficient utilization of dissolved oxygen and the conditional switch in offspring quality between cyst (oviparous reproduction) and nauplii (viviparous reproduction) depending on unfavorable and favorable environmental conditions, respectively. Although most of these adaptive traits have a relatively well-known physiological and molecular basis (^{Gajardo and Beardmore, 2012}), changes in heterochromatin could be another factor in the Artemia repertory adopted to withstand extreme conditions. This has been mentioned in other animal and plant studies, where some of their ecological features are associated with variation in the heterochromatin content (^{Walker et al., 1991}; Ceccarelli et al., ¹⁹⁹², ²⁰⁰²).

In this study the interphase heterochromatin content in different American Artemia populations belonging to the A. franciscana and A. persimilis species were determined through the analysis of interphase nuclei from nauplii cells. This parameter was related to variations in nucleus size in order to explore the existence of nucleotypic changes. We also investigated the relationship of these heterochromatic changes with the ionic composition of the brines inhabited by these populations as a proxy of their adaptive nature.

2 Material and Methods

3 Results

The interphase nuclei from nauplii cells, which display the chromocenters observed in the 12 populations studied, are shown in Figure 1. Quantification of these heterochromatic areas per nucleus indicated that the mean N-CHR values varied widely and significantly among populations (ANOVA, F_{[11, 440]}=31.08, p<0.001), from 0.81 ± 1.17 to 12.58 ± 3.78 (as shown in Table 3), with a variation index of 15.5 fold. Thus, there were populations whose means were low (SCC), medium (CHX, LLA, LVI, PCH) or high (RIN, CIS, RGB, MAC, SFB, GSL, TPA). In 30 out of 66 pairwise comparisons, differences in means were statistically significant (Tukey’s test, p<0.05). The analysis of this parameter at species level also indicated significant variation among populations for A. franciscana (ANOVA, F_{[8, 304]}=10.94, p<0.001) and A. persimilis (ANOVA, F_[2,136]=152.52, p<0.001). With regard to the reference populations, the mean N-CHR in the SFB population (A. franciscana) was significantly higher than in the SCC population (A. persimilis) (10.54 ± 3.55 vs. 0.81 ± 1.17, Student’s t-test, p<0.05). The mean N-CHR obtained in previous studies (as shown in Table 1) revealed less chromocenters (two to seven) in some populations, in contrast to the result found in this study, such as the LLA and LVI populations.

Figure 1 The interphase nuclei of nauplii cells, displaying the chromocenters stained by Hoechst 33258 fluorescent dye (bright bodies) observed in the 12 studied population of Artemia. Populations: (a) Salina la Colorada Chica (SCC), (b) Laguna Amarga-Torres del Paine (TPA), (c) Laguna de Los Cisnes (CIS), (d) San Francisco Bay-1258 (SFB), (e) Great Salt Lake (GSL), (f) Salar de Llamara (LLA), (g) Chaxas (CHX), (h) La Rinconada (RIN), (i) Palo Colorado-Los Vilos (LVI), (j) Salinas de Pichilemu (PCH), (k) Río Grande (RGB) and (l) and Macao (MAC). Bar represent 5 µm. 

The mean R-CHR values also varied significantly among populations (ANOVA, F_[11,440]=115.05, p<0.001), from 0.19 ± 0.34% to 11.78 ± 3.71%, but with a higher level of variation (62 fold) than the N-CHR parameter. Within this distribution, populations presented mean values that were grouped into low (SCC), medium (CHX, LLA, LVI, RGB, PCH), or high (RIN, MAC, SFB, CIS, GSL, TPA) categories (as shown in Table 3). The pairwise comparison of means for the R-CHR parameter indicated that 50 out of 66 had significant differences (Tukey’s test, p<0.001). The result of this parameter at species level also indicated significant variation among populations for A. franciscana (ANOVA, F_{[8, 304]}= 49.74, p<0.001) and A. persimilis (ANOVA, F_{[2, 136]}=297.07, p<0.001). In the case of the reference populations, mean R-CHR in the SFB population was significantly higher than in the SCC population (8.34 ± 3.32% vs. 0.19 ± 0.34%, Student’s t-test, p<0.05). The mean N-CHR among A. franciscana and A. persimilis did not differ significantly (9.19±4.03 vs. 8.95±5.55, Student’s t-test, p>0.05). However, the mean R-CHR was significantly lower in A. franciscana than in A. persimilis (4.59±3.21 vs. 8.31±5.76, Student’s t-test, p<0.05).

The mean S-NUC values ranged from 155.43 ± 88.32 µm² to 372.32 ± 160.03 µm². The mean differences among populations were statistically significant (ANOVA, F_[11,440]=19.67, p<0.001), but only 23 out of 66 pairwise comparisons were significant (Tukey’s test, p<0.05). At species level, this parameter also indicated significant variation among populations for A. franciscana (ANOVA, F_{[8, 304]}=21.26, p<0.001) and A. persimilis (ANOVA, F_{[2, 136]}=5.53, p<0.01). In addition, the S-NUC parameter, showed much less variation than the R-CHR parameter according to the variation index (2.39 vs. 62 fold). The mean S-NUC was significantly higher in A. franciscana than in A. persimilis (253.94±125.10 µm² vs. 205.76±90.48 µm², Student’s t-test, p<0.05).

According to the regression analysis (as shown in Table 4), there was a positive and significant relationship between N-CHR and R-CHR in eight out of 12 populations (r= 0.297-0.792, Student’s t-test, p<0.05,). All populations pooled exhibited a significant correlation between both parameters (r= 0.641, Student’s t-test, p<0.05). This pattern was also observed for populations of A. franciscana (r=0.525, Student’s t-test, p<0.05) and A. persimilis (r=0.807, Student’s t-test, p<0.05). The coefficient of determination (r²=0.41) in the relationship of pooled data revealed that only 41% of R-CHR was determined by N-CHR, reflecting a low level of determination between both parameters. Likewise, there was a significant association between R-CHR and S-NUC in five out of 12 populations (Student’s t-test, p<0.05), either negative, in four populations (CHX, r= -0.643; RIN, r= -0.464; RGB, r= -0.443; PCH, r= -0.540), or positive, in one population (CIS, r=0.367). All populations pooled showed a negative and not statistically significant association (r=-0.032, Student’s t-test, p>0.05) between both parameters. A similar pattern was observed for populations of A. franciscana (r=-0.009, Student’s t-test, p>0.05) and A. persimilis (r=-0.005, Student’s t-test, p>0.05).

Piper’s diagram (see Figure 2) grouped populations studied into two classes according to the predominant ions in each location: 1) waters containing sodium and/or chloride such as SCC, LLA, RIN, LVI, PCH, SFB and GSL populations; and 2) water with irregular ionic composition, represented by TPA and CIS populations, where the maximum percentage of SO₄^2- was relatively high (23-67%). With regard to the correlation between the ion concentration of each location and R-CHR (as shown in Table 5), the pooled data indicated a positive and significant association (Student’s t-test, p<0.05) for the Mg²⁺ ion and negative and significant association (Student’s t-test, p<0.05) for the Cl^-, Na⁺ and Ca²⁺ ions. At species level, this analysis indicated that there were different associations for some ions, particularly, the SO₄^2- and Ca²⁺ ions displayed negative and significant correlations in A. franciscana, in contrast to A. persimilis.

Figure 2 Piper’s diagram displaying ionic concentration (%) of brine waters of nine American Artemia populations studied in this work. The populations were the following: Salina la Colorada Chica (SCC), Laguna Amarga Torres del Paine (TPA), Laguna de Los Cisnes (CIS), San Francisco Bay (SFB), Great Salt Lake (GSL), Salar de Llamara (LLA), Palo Colorado Los Vilos (LVI), La Rinconada (RIN) and Salinas de Pichilemu (PCH). Antofagasta (ANT) sample was included in the analysis as reference of sea water. 

4 Discussion

Analysis of the interphase heterochromatin in Artemia has been based mostly on counting the number of chromocenters using visual methods. Despite the taxonomic value attributed to this cytogenetic trait (^{Gajardo et al., 2001}), some authors have stated that these studies have probably been subject to considerable experimental error (^{Papeschi et al., 2008}). This situation may occur due to following sources of error: 1) the chromocenters tend to merge in the interphase nucleus, and thus the actual chromocenter number counted in different nuclei might differ; 2) the chromocenter counts are often carried out regardless of size, thus small chromocenters may be compared to large ones, despite the fact that the former may have less heterochomatin than the latter; this situation leads to a misinterpretation of the amount of heterochromatin present in the nucleus; and 3) some researchers may exclude the tiny chromocenters on the final count, based on a subjective decision. To address this problem, we included the determination of the relative amount of heterochromatin per nucleus in this study (i.e. R-CHR parameter), to ensure more reliable heterochromatin quantification in the nucleus. Indeed, large interpopulation differences were observed in the mean percentage of heterochromatin per nucleus, compared to the traditional method based on counting the number of chromocenters (62 fold vs. 15.53 fold), which confirms the robustness of this methodology. In other species, such as Arabidopsis (^{Soppe et al., 2002}), a similar strategy has been adopted to improve the quality of this analysis.

The relationship between the percentage of interphase heterochromatin per nucleus and nuclear size showed no significant associations at species level. However, our results indicate that, in some cases, significant associations occur at a population level. For example, CHX, RIN, PCH and RGB populations exhibited a significant reduction in nuclear size associated with an increase in the amount of interphase heterochromatin, while the opposite was observed for the CIS population. This finding suggests the existence of a nucleotypic effect in Artemia from the Americas, mediated by variation in the amount of heterochromatin present in the nucleus. However, the effect would be specific to certain populations. It is important to note that the natural habitats of these population may vary widely throughout the year, for example, in temperature (18-29.8 ºC for RIN) (^{Gomez-Silva et al., 1990}) and salinity (30-120 ppt for PCH) (^{Gajardo et al., 1998}). Therefore, the associations observed between both parameters may reflect their adaptations to particular ecological conditions. On the other hand, the results showing reduction in nucleus size to be negatively correlated with heterochromatin content appears intriguing, although this effect has recently been demonstrated (^{Wang et al., 2013}) in another organism, specifically in the Arabidopsis mutant Crowded Nucleus (CRWN). In the case of increasing nuclear size, the evidence available in animals and plants has revealed a positive correlation, although mainly with nuclear DNA content (^{Swanson et al., 1981}; ^{Jovtchev et al., 2006}).

The strong and positive association between the content of interphase heterochromatin and magnesium concentration in the brine sites is a significant result. In other words, heterochromatin variation would induce changes that extend from the cell to the organism level, expressed as the differential ability to survive in brines with different ionic composition. Although this effect must be substantiated in further studies, especially relating the actual ion composition of the brine to the moment when samples of Artemia are obtained, this is a plausible working hypothesis worthy of future study. The two populations (TPA and CIS) considered in this study, located in Chilean Patagonia (below latitude 50º S) are atypical for Artemia standards (^{Clegg and Gajardo, 2009}). According to data previously collected (^{Campos et al., 1996}; ^{Zuñiga et al., 1999}; ^{De Los Ríos and Soto, 2009}), both hypersaline sites would differ from the traditional description of environments classified as Athassalohaline (inland waters) and Thassalohaline (marine waters). For instance, magnesium content (9.85-46.35%) is considerably higher (up to 2.5 times) than that of sea water. Despite the fact that the adaptive role of heterochromatin might still be considered controversial, evidence now available indicates that its role in the cellular function cannot be ignored, for example, this element is involved in the stabilization of the chromosome structure, chromosome segregation and gene silencing (^{Grewal and Jia, 2007}). Furthermore, variation in heterochromatin content has been associated with the particular distribution of the species in response to divergent environments (^{Walker et al., 1991}; Ceccarelli et al., ¹⁹⁹², ²⁰⁰²), or with the biomass level of the organisms (^{Edelman and Lin, 1996}). Previous studies showing a latitudinal variation in chromocenter numbers in A. franciscana would be an indirect indication of such heterochromatin abilities (^{Gajardo et al., 2001}). In fact, the Artemia used in this study, collected from Torres del Paine, present a larger relative heterochromatin content in the nucleus than that of the other populations analysed. Clarification of this paradox requires further analysis, but a preliminary hypothesis would be that particular environmental and water conditions may have effectively selected particular genomic and phenotypic features for this population. For instance, cyst size of this population, larger than that of other Chilean populations, appears to be associated with the particular ionic composition of the water in this site (^{Castro et al., 2006}). In addition, analysis of the association between amount of interphase heterochromatin and the ionic concentration of the brines revealed a species-specific negative association with the sulphate and calcium ions, only in the case of A. franciscana. Given that the brine habitats of populations of this species are mainly Thassalohaline waters, whose ionic composition is characterized as being rich in chloride and sodium and poor in sulphate and calcium ions (a result that was confirmed by the Piper’s diagram), this association could be explained by the scarce presence of these ions in their natural habitats.

The result indicating interspecific or interpopulation differences in the relative amount of interphase heterochromatin found in Artemia is significant. The origin of this variation may have other causes, such as difference in genome size, or the particular evolutionary process of the karyotype. In the case of differences in genome size, greater amount of heterochromatin is expected in species with a large genome size than in those with a smaller genome size, a pattern that has been demonstrated in evolutionary closely related organisms (^{Rayburn et al., 1985}; ^{Kao et al., 2001}; ^{Bosco et al., 2007}). Although this hypothesis is interesting, to date it is not possible to contrast Artemia species, since data is only available for the genome size of A. franciscana, reaching a value of 0.97 pg per haploid genome (^{De Vos et al., 2013}). Differences in the evolutionary process of the karyotype could also be taken into account given that A. franciscana appears to be at more advanced stage of evolution than A. persimilis (^{Parraguez et al., 2009}). Thus, it is expected that the former would accumulate more heterochromatin in the chromosomes than the latter. With the exception of the TPA and CIS populations of A. persimilis, which presented large amounts of interphase heterochromatin, most populations of both species studied followed this pattern. Hence in our case, this hypothesis seems to be consistent.

According to the evidence available in Arabidopsis, nuclear changes induced by heterochromatin cannot be ruled out, given the identification of various genes that specifically control such processes (^{Fransz et al., 2003}; ^{Wang et al., 2013}). For example, in the ddm1 mutant, it is observed that the heterochromatin content is significantly reduced in the nucleus (-30%) in comparison to the wild type. This is reflected at the cytological level in lower chromocenter numbers with small sizes. Likewise, in the CRWN4 mutant the chromocenters are notoriously disorganized or dispersed when compared to those observed in a normal nucleus. Based on these cytological patterns, it is possible to classify the organization of the nucleus in Arabidopsis according to the appearance of chromocenters, either as adispersed phenotype (CRWN4 mutant) or as acompact phenotype (wild type). It is important to note that the arrangement of chromocenters in the nucleus of some Artemia populations analysed in this study may match these phenotypes. For example, LLA, CHX and RGB populations would represent a dispersed chromocenter phenotype, while LVI population would correspond to a compact chromocenter phenotype (see Figure 1). Thus, we cannot discard the possibility that these phenotypes may reflect the existence of a particular organizational process of the heterochromatin in the nucleus, across different Artemia populations, whose control may depend on the action of specific genes. In addition, chemical and physical factors may also be involved in the packing status of the chromocenters in Artemia populations, given that natural populations of this organism may be subject to many abiotic stressors, such as changes in temperature, ionic composition and salinity (^{Gajardo and Beardmore, 2012}). Although the participation of these factors has not generally been demonstrated in Artemia, experimental data collected in other organisms reveals their existence. For instance, in NIH 3T3 mouse cells treated with valproic acid produce decondensation in the chromatin structure of the nucleus, including the heterochromatin areas (^{Felisbino et al., 2014}); while in Malpighian tubule cells of the blood-sucking insect, Triatoma infestans, heterochromatin decondensation occurs either after treatments with heavy metals (copper and mercury) or heat shocks (^{Mello et al., 1995}, ²⁰⁰¹). Moreover, heterochromatin decondensation after heat shock treatements in Malpighian tubule cells of a vector of Chagas' disease in Brazil, Panstrongylus megistus, infected and non-infected by Trypanosoma cruzi, has been also reported (^{Garcia et al., 2011}). In our case we believe that such factors would not have affected the results obtained since the samples analysed are from laboratory cultures that were kept under standardised conditions of temperature and salinity. However, the chemical or physical factors that might be affecting the condensation of heterochromatin in the interphase nucleus in natural populations of Artemia emerge as an interesting topic for future studies.

Further studies on heterochromatin in Artemia aimed at providing more in depth information about its organization and structure may help to clarify the biological meaning of variation across populations. This type of analysis may contribute additional insight into the source of variation in the interphase heterochromatin, both at intraspecific and interspecific levels in Artemia.

5 Conclusion

The amount of interphase heterochomatin per nucleus that was estimated based on chromocenter number and relative area of chromocenter, in A. franciscana (n=9) and A. persimilis (n=3) populations from different locations of the Americas, varied significantly both within and between species. The relationship between relative area of chromocenter and nuclear size revealed a significant association, either negative or positive, in five out of twelve populations analysed. All populations pooled or categorised by species did not show a statistically significant association between both parameters. There was a significant association between relative chromocenter area and ionic concentrations of natural brines in the populations studied, with a positive correlation for magnesium and negative correlations for chloride, sodium and calcium. When populations were categorised by species, a negative and significant correlation with sulphate and calcium ions was found for A. franciscana. These findings suggest the existence of a nucleotypic effect on nuclear size in some populations of Artemia from the Americas, mediated by the variation in the amount of interphase heterochromatin of the nucleus. Moreover, the association of this parameter with the ionic composition of natural brines suggests that it could be involved in the ability of this organism to survive in these environments, which are habitually subject to strong physicochemical changes.