A direct one-to-one comparison was made between the transcriptomes of the samples S and LS at the same time of cold storage, given the notion that this analysis would outperform the general profile comparison to identify the candidates to be involved in tolerance/susceptibility to cold . Figure 2A shows how the number of differentially expressed genes at each time decreased with storage time , thus confirming PCA results . Functional enrichment analysis showed that by 1 week of cold storage, the transcripts with higher levels of expression in fruits CS1-LS were preferentially related to energy production, RNA translation and protein assembly, the antioxidant system, structure maintenance, and genes with unknown functions . As 1 week cold storage is critical timing i.e. when maximum differences were shown when later transferring fruits to shelf life ripening , these functions may play a prominent role in the tolerant/sensitive character of fruits . By 2 weeks of cold exposure, only the genes with unknown functions were over represented in the tolerant pool , whereas a significant enrichment was noted for the genes linked to amino acid metabolism, pyruvate, signal transduction and transport in the genes at higher levels in CS2S. Interestingly, most of the genes expressed at higher levels in S fruits by 2 weeks had already reached this state by 1-week of cold storage . As two weeks of cold exposure results in mealiness upon shelf life in both S and LS fruits , square plastic pot but with large differences in MI severity, high levels of these genes may correlate negatively with the tolerant character of fruits.

After 3 weeks in the cold, only the highly expressed genes in tolerant fruits showed signal transduction as an over represented class . In this case, the genes differed from those identified as being over represented at 1 and 2 weeks . At this time, both S and LS developed mealy fruits with MI 1.0 and MI 0.8, respectively , but S was probably much more severely affected or underwent other downstream processes. In order to analyze if the transcript program in the cold may have a direct effect on eventual mealiness development during shelf life, a Pearson correlation analysis was conducted between the gene expression values and the projected MI will be achieved when subjected to shelf life ripening after cold exposure . This ‘‘projected MI’’ correlation analysis resulted in 113 directly correlated genes and 159 inversely correlated genes according to their pattern of expression in the cold . The functional enrichment analysis indicated that genes directly correlated to projected MI were enriched in RNA transcription and RNA posttranscriptional regulation. A further inspection revealed genes related to RNA bio-genesis and processing, splicing, RNA transcription machinery and the transcription factors . In addition, genes correlated positively with the projected MI were also enriched in transport category , that includes transporters for auxin, anthocyanin, amino acid, peptides, sulfate, carbohydrates and metal-ions . No functional enrichment was observed for those genes which correlated negatively with projected MI . However, a detailed inspection indicated that this set of genes contained calcium-related genes, including a transcription factor of the CAMTA family, and genes related to antioxidant systems which could participate in the regulation of this transient tolerance mechanism.The possibility that, in addition to cold-inducible mechanisms, some sort of tolerance mechanism may already be partly preprogrammed in tolerant fruits was investigated. The direct comparison between S and LS fruits at mature stage resulted in 63 differentially expressed genes .

Out of them, 13 genes we high expressed in fruits T and some have to do with flavonoid metabolism , structure protection and that forms part of a cycle that generates asparagine for more energyeconomical nitrogen remobilization under darkness and stress conditions. Several cell wall modifying activities were also differentially expressed between fruits S and LS . As no differences at the maturity stage were between pools , it is likely that differences in the expression levels of these genes at harvest may protect fruits and/or contribute to develop the tolerance program at least in the early stages of the cold response. HCA of samples M, R and CS showed that genes differentially expressed between fruits S and LS at harvest qualified in fruits LS as ripening genes . Notwithstanding, it is most interesting to note these genes were characterized by continuing the ripening program during cold storage , which did not happen so clearly in fruits S . However and as expected this behavior of the differential M genes is the exception rather than the rule for ripening genes. As seen in Figure 3B, a similar analysis with a set of 862 ripening genes showed that although cold affect the expression many of ripening genes, is quite effective stopping the molecular ripening program in fruits LS. This result is in agreement with the findings from PC2 . The main expression differences between LS and S fruits involved changes occurring in the same direction in R and cold stored fruits. In fruits LS, the expression of several ripening genes during cold storage remained at the same or higher level that they were in the M stage, but achieved similar expression levels to fruits R in the sensitive backgrounds . Apart from the delayed or attenuated ripening program in the fruits LS during cold storage, these fruits also showed specific ripening processes that became activated during cold storage , which is in agreement with the findings for genes differentially expressed at harvest . A more detailed analysis of shelf life ripening conditions and mealiness will be addressed in a future manuscript .In this section we wanted to see if there were similarities between the adaptation mechanisms operating in peach fruits stored in cold and darkness and those well-characterized in the cold acclimation of Arabidopsis plants grown in day/night regimes.

We wanted to see if the patterns of gene expression for the peach homologues of Arabidopsis genes in cold/dehydration regulons were consistent with the differential cold responses in S and LS peaches. First we analyzed the overlap between the response of cold stored peach fruits and those to various stimuli, including abiotic/ biotic stresses and hormones . Gene-bygene comparisons revealed that the vast majority of the cold regulated genes in our peach cold storage experiment have Arabidopsis orthologs, which have been described as being regulated by cold , or by ABA . Similarly to Arabidopsis, approximately 30% of peach cold-regulated genes were found to be associated with drought and/or salinity treatments . More strikingly however, approximately 35% of the cold-responsive genes in peach were known pathogen responsive genes or have been postulated to play a role in pathogen resistance . Furthermore, the genes described as being regulated by darkness in Arabidopsis account for up to 3.7% of peach cold-regulated genes ,square pot indicating that, although its contribution to all cold-regulated genes was less than those also involved in other stresses, dark stress could contribute to the differences observed in the cold response between peach fruits and Arabidopsis plants . Second, a list of Arabidopsis genes reported in cold regulons and dehydration regulons was used to identify homologous peach genes that were present on Chillpeach microarray . In total, 163 Chillpeach unigenes corresponded to the genes found in at least in one of the previously defined cold and/or dehydration Arabidopsis regulons . The expression profiles of these genes in response to cold storage were compared to those described for Arabidopsis and scored as matching when they behave similarly. More than 60% of the genes associated to the regulons CBF, HOS9, ICE and DREB2 correlated well with both the known Arabidopsis WT cold response pattern and the Arabidopsis mutant expression . That is, the ortologs genes to those up-regulated in Arabidopsis in response to cold showed higher expression levels in LS peach fruits than in high sensitive ones, while the genes down-regulated in Arabidopsis had higher levels in high sensitive peach fruits than in low sensitive ones. In contrast, most of the genes in HOS15, ZAT12, ESK, AREB, MYB, ZF/HD-NAC presented low correlation levels . Therefore, these latter are less likely to contribute to the differences in response to cold between the S and LS pools of fruits. The individual participation of each regulon to the differential response to cold between fruits S and LS was assessed by studying their contribution to the traits/trends observed in the global dataset analysis. For this purpose, we performed both PCA and 2D-HCA using the gene expression values for all the genes in each regulon as input datasets and quantitatively evaluate the importance of each regulon to discriminate samples S from LS and to separate the samples that would eventually became mealy, or not, by assessing by the number of genes well correlated with Arabidopsis in the gene expression models . The importance to discriminate samples S from samples LS was calculated by multiplying the number of genes that correlated well by the variance explained by PC2.

The importance of an operon to separate the samples that would eventually become mealy, or not , was quantified by dividing the number of genes in that operon that correlated well by the weight of the nearest node to CS1-LS. Both PCA and 2D-HCA revealed that regulon ICE1 was the one most contributing to discriminate samples LS and S, as to separate samples CS1-LS from the rest of cold-stored fruits that developed mealiness when submitted to shelf life ripening . Furthermore, this analysis also indicated that the regulon CBF1 was the next major regulon in discriminating between samples LS and S , while emphasized the relevance of HOS9 to separate CS1-LS from the remaining samples . The rest of the cold operons produced no such separation between CS1 S and LS, or did so but to a lesser extent . The expression pattern of the subsets the genes appertaining to the regulons ICE1 , CBF and HOS9 across the different samples showed that although extended exposure to cold debilitated the response of ICE-CBF regulated genes, fruits LS were able to maintain a longer and greater response for many of the genes in the regulon in the cold. In the case of HOS9 regulon, many of its members were up-regulated or without change in LS fruits as compared to M fruits .The same bulked samples used in this microarray experiment were used to validate the results by using medium-throughput qRT-PCR over a set of genes selected because they 1) contributed to separate samples S from samples LS at 1 week of cold storage , 2) showed a differential expression in, both, the M stage and 1-week of cold storage , and 3) showed differences at harvest . In order to examine at the single sibling level the reliability of the differential gene expression patterns obtained from the pools, the analysis was performed also on 15 individual genotypes of the pop-DG population . The qRTPCR results obtained from the pools and from the individual lines making up this pools indicate that 72.5% of the genes had the same expression pattern in the microarray experiment as in the qRT-PCR experiment . However, the magnitude of expression varied slightly in many of the genes and samples tested . Furthermore qRT-PCR experiments conducted on individual pop-DG siblings revealed that 42 out of the 50 genes validated in the pools were consistent with the expected patterns for which they were selected . These results support the validity of our approach and indicate that the genes selected from the microarray analysis could be either involved in chilling tolerance and/or be associated with the differential response to chilling response, and for some of them could even prove to general enough to hold true in individual fruits/plants.Since cold induced mealiness is not observed until the cold stored fruit are allowed to ripen, the chilling sensitivity phenotype of each fruit in the cold was estimated from the protracted mealiness incidence observed for equivalent fruit samples after shelf life ripening . Although mealiness, probably, a downstream effect of cold stress in peach fruits , it is the best phenotyping tool to assess the effect of cold on peach fruit, and has be used successfully to identify CI QTLs in peach. For BSGA we use Chillpeach microarray, interrogating part of peach genome. This provides only an incomplete picture of the genes behind the process; that is partially compensated by Chillpeach microarray being enriched in fruit-specific and cold responsive genes.