The action of the Cu2+, Ag+ and time inducing the in vitro anther culture-derived barley regenerants

Plant regeneration via anther cultures is a world-wide approach as it allows for the regeneration of uniform and homozygous double haploids. Recent studies have shown that in vitro cultures are the origin of the so-called tissue culture-induced variation (TCIV) that may lead to off-type regenerants. Moreover, the regeneration of green plants may be limited by the presence of albinos. It was demonstrated that the presence of Cu 2+ and Ag + ions in the regeneration medium might increase the number of green plants. DArTseqMet markers were evaluated based on regenerants and donor plants derived via in vitro anther cultures of barley. The regenerants were obtained under varying Cu 2+ and Ag+ ion concentration in the regeneration medium during distinct time conditions of the tissue cultures. The DArTseqMet markers were quantified using a semi-quantitative MSAP approach delivering data on CG and CHG sequence contexts de novo methylation and demethylation. Under each tissue culture conditions, the number of regenerated green plants per 100 anthers was evaluated. Conditional moderation analysis was applied to test for the role of Cu 2+ and Ag + ions in the medium. Moreover, the importance of the time of in vitro anther cultures were analyzed. affecting moderated ions the medium


3
Our data are congruent with the putative function of these ions in the ethylene and DNA methylation pathways.

Background
An in vitro plant regeneration via anther cultures requires cell reprogramming [1], including DNA demethylation and de novo methylation of cytosine residues [2] accruing in symmetric (CG and CHG) and asymmetric (CHH) methylation contexts [3,4]. The methylation of symmetric CG sequence contexts is performed during DNA replication cycle [5,6], whereas CHG sequence methylation changes are controlled by genetic and epigenetic mechanisms [7][8][9][10]. The asymmetric methylation change affecting CHH sites is regulated by epigenetic mechanisms [11] related to stressful conditions influencing plants [12].
Plant regeneration via anther cultures requires precise tuning of the in vitro tissue culture conditions, including the concentration of ingredients (i.e., the balanced concentration of ions) that may influence cellular processes promoting plant regeneration. Such ingredients may encourage induction of cell reprogramming [13][14][15][16] and potentially change the balance of biochemical pathways. For example, the addition of Cu 2+ ions may affect mitochondrial Complex IV [17] belonging to the electron transport chain and is involved in the copper-delivery pathway and creates functional ethylene receptors [18]. It may also form complexes with ethylene. Its improper functioning may result in ATP deficiency [19,20], or burst of reactive oxygen species (ROS) [21]. On the other hand, silver ion regulates the polyamine pool in a plant [22], ethylene-and calcium-mediated pathways [23], and plays a role in the physiological process including morphogenesis and prone the uptake of Ca 2+ into a cell [24]. The presence of Ca 2+ may be a stress signal for the nucleus [25]. If the uptake of Ca 2+ ions due to the presence of Ag + in the medium precedes stressful conditions, then burst of Ca 2+ and ROS from mitochondria [26,27] and other organelles may be mitigated winning the cell time for reprogramming [28] changing a haploid cell fate for green plant regeneration [29]. The process affects nuclear DNA [1] at the methylation level [2].
To study the DNA methylation changes and their relation with the number of green plants, several

Results
In vitro anther tissue cultures performed under nine distinct conditions (trials M1-M9) varying in the Cu 2+, Ag + ion concentrations, and time allowed the regeneration of 35 plants. As indicated earlier [43], no morphological differences among regenerants were observed, and all of them were in the type of an anther tissue donor plant. DNA isolation from fresh leaves of donor and regenerated DH plants using commercial kits resulted in integral samples without impurities. The new generation sequencing approach exploiting HpaII and MspI endonucleases that differ in sensitivity towards site DNA methylation [48] was used to evaluate DArTseqMet DNA markers. The markers were classified to those related to de novo methylation and demethylation within CG and CXG contexts, following the procedure described earlier [38]. The DNA methylation characteristics were evaluated, as indicated in Table S1 (Additional file 1). A minimum population size required to achieve actual power of statistics more that 0.31 was estimated for 35 (F(2,31)=2.9113, f 2 =0.111).
Conditional moderation analyses performed for eight analyses (Additional file 1: Table S2 (Table 1). Our results demonstrated that de novo methylation and DNA demethylation affecting CG and CXG sequence contexts were moderated by Cu 2+ and Ag + ions present in the medium conditional on the time of in vitro anther cultures of barley and that such conditions of tissue cultures influenced the number of green plants regenerated per 100 anthers (Table 1).  Table S13).
Using combined delta = DNM -DM variable moderated by unified moderator Cu + Ag conditional on the time of the in vitro anther tissue cultures, the analysis was significant (Fig. 3, Additional file 1:  Table S14).

Discussion
In vitro anther tissue cultures is an essential approach for the evaluation of uniform plant materials useful for breeding programs [50][51][52][53]. The approach is being world-wide used, but a growing number of data clearly shows that plant regeneration via anther culture is affected by an in vitro TCIV [54][55][56][57] that could be transmitted to a progeny [58]. Among others, the TCIV is due to changes in the DNA methylation patterns affecting distinct methylation contexts (symmetric and asymmetric) that are under either genetic or epigenetic control [44]. As plant regeneration requires cell reprogramming [1] involving DNA methylation pattern changes, it is of value to understand whether ingredients present 9 in the in vitro medium and other factors such as the time of in vitro cultures may affect methylation changes and the number of the regenerated green plants that may be limited by the presence of albinos [59]. Among many components used in the in vitro anther culture, Cu 2+ and Ag + are being considered as promising [60,61]. There are shreds of evidence that they may increase the number of green regenerants in cereals [62,63]. However, the putative role of the ions is not explicit.
Our data are in agreement with prior studies suggesting that for the regeneration of green plants, cell reprogramming [65] is needed. We have demonstrated that delta equal to de novo methylation less demethylation (all symmetric contexts taken together) may be used as a predictor in moderation conditional on time. As delta was calculated without focusing on any of the methylation contexts ( Fig.   1 and 2, Additional file 1: Table S2, Analyses I and J; Table S12-S13), and any conditional moderation analyses (Additional file 1: Interestingly, plant regeneration takes place when delta = DNM-DM is negative; demethylation events are either higher than or precede de novo methylation, and such a sequence of changes is vital for plant regeneration in anther cultures. The regeneration of GPs is also promoted when delta is positive under the highest (Ag + Cu) concentration conditional on a long time of in vitro tissue cultures (Fig. 3, Additional file 1: Table S2, Analysis K; Table S14). Moderation analysis shows that there is a window of time when GPs regeneration is hardly possible, but even under positive delta (when de novo methylation exceeds DNA demethylation), regeneration may still be significant in barley anther 11 cultures. The presented data may be interpreted in terms of full cell reprogramming to regenerate under a long time of in vitro tissue culture conditions that require increased (Ag + Cu) concentration.
The molecular basis of Ag + and Cu 2+ ions in tissue cultures are lacking; however, they evidently "cooperate" with each other affecting the number of GPs conditional on the time of in vitro tissue cultures. Our analysis has not only shown that the delta equal to de novo methylation less demethylation could be used as a predictor in conditional moderation analysis, but the sum of Cu 2+ and Ag + is also indicative here. We have noticed that the two ions play in concert and moderate DNA methylation in a similar manner resulting in GPs conditional on the time of tissue cultures (Fig. 3, Additional file 1: Table S2, Analysis K, Table S14). The cooperative effect of silver and copper ions could be explained by silver ions substitution for copper ions [68], interfering with ethylene pathways contributing to the regulatory pathways of DNA methylation [69]. DArTSeqMet markers were converted into quantitative methylation characteristics following the MSAP approach described earlier [38].

Power of the statistics
The minimum population size was calculated in G-Power software [70]. Squared multiple correlation p 2 was set to 0.1 to calculate effect size f 2 at a=0.05 with three predictors and power (1 -b err prob) set to 0.31.
Model quality was tested using second-order Akaike's Information Criterion for small sample size (n/k < 40), where n is a sample size, k states for model parameters and log-likelihood is a measure of model fit using the formula: AIC = -2(log-likelihood) + 2k + 2k*(k+1)/(n-k-1). AIC scores were reported as ΔAIC (the relative difference between the best model which has a DAIC of zero) using the following formula: ΔAIC = AICi -min AIC, where AICi is the AIC of the i model and min AIC is the score for the best model. If DAIC was less than 2, the model was assumed substantial. Akaike weights were used in model averaging. They represent the relative likelihood of a model. For each model, the relative likelihood (RL) of the model, which is exp(-0.5 * ∆AIC score for that model), was calculated.
The Akaike weight (AW) for a model is the exp(-0.5 * ∆AIC score for that model) value divided by the sum of all such values across all models [49] for the given type of models.

Supplementary Information
Additional file 1:

Table S1
The arrangement of the MSAP DNA methylation characteristics for regenerants obtained via in vitro anther culture Table S2 The arrangement of statistics for conditional moderation analyses based on model 3. RLrelative likelihood, AW -Akaike's weight, LLCI -ULCI is a 95% confidence interval   This is a list of supplementary files associated with this preprint. Click to download.