Assessment of Chilling Requirement and Threshold Temperature of a Low Chill Pear ( Pyrus communis L.) Germplasm in the Mediterranean Area

: In temperate climates, bud break and shoot and ﬂower emission of deciduous fruit tree species are regulated by precise chilling and heating requirements. To investigate this aspect, sixty-one accessions of European pear ( Pyrus communis L.) collected in Sicily were phenotyped for three consecutive years for harvest date, bud sprouting and blooming to determine both the chilling requirements and the threshold temperature using the Chill Days model. The whole germplasm collection was grown in two different experimental ﬁelds located at 10 and 850 m above sea level rep-resenting two Mediterranean-type climates in which pear is commonly cultivated. Results revealed a mean threshold temperature of 6.70 and 8.10 ◦ C for the two experimental ﬁelds, respectively, with a mean chilling requirement ranging from − 103 and − 120 days. Through this approach, novel insights were gained on the differences in chilling requirement for early ﬂowering cultivars to overcome dormancy. Furthermore, to better dissect differences in chilling requirement between accessions, the sprouting bud rate of six cultivars was assessed on excised twigs stored at 4 ± 0.1 ◦ C from 300 to 900 h followed by a period at 25 ± 0.1 ◦ C varying from seven to twenty-eight days. Results of both experiments highlighted that Sicilian pear germplasm is characterized by a low chilling requirement compared to other pear germplasm, making Sicilian local accessions valuable candidates to be used for selecting novel cultivars, coupling their low chilling requirements with other traits of agronomical interest.


Introduction
The Mediterranean area is characterized by a wide range of different microclimates, ranging from arid/semiarid to temperate and humid. In the last decades, several global and regional climate models indicated the Mediterranean scenario as a 'Hot-Spot' for future climate change prevision [1,2]. In this region, several areas are experiencing both a general increase in temperatures and a decrease in rainfall. This trend could influence the biological behavior of plants with direct repercussion on their distribution and alter the physiology and the phenology of many species of agronomical interest [3][4][5][6][7].
Temperature plays a fundamental role in promoting bud dormancy and bud break during winter and spring, respectively. In particular, bud dormancy is broken when the plant undergoes a period at low temperatures, while the vegetative and reproductive restart (bud break) is directly influenced by warm temperature. The cold exposure needed to bud break is called a chilling requirement (CR). This stage is followed by the break of quiescence when plants fulfill their heat requirements (HRs) [8][9][10][11]. When CR is not completely

Plant Materials, Site Description and Experimental Design
This research was conducted on three consecutive growing seasons from 1 October 2014 to 30 June 2017, in two experimental fields (EF) located in Catania district (Sicily, South Italy), the experimental farm of Catania University (EF1, 10 m a.s.l.) and the Germplasm Bank of 'Parco dell'Etna' located in the field of the Etna volcano (EF2, 850 m a.s.l.). Trees were grafted on the same rootstock and planted in the same year in both EFs. Accessions held in both EFs were clones of the same source tree. In Table 1, for each experimental field, the geographical characteristics, the elevation, the reached Chilling hours (CHs) (Weinberger, 1950) and Chill Units (Richardson et al., 1974), calculated from 1 October to 28 February from 2014 to 2017, are reported. Moreover, the growing degree hours (GDHs) according to Richardson et al. (1974), calculated as the number of hours accumulated between the end of dormancy and the end of fruit set from 1 March to 30 June from 2014 to 2017, are reported. Table 1. Geolocations and yearly and average data related to the Chill Hours, Chill Units (from 1 October to 28 February) and Growing Degree Hours (from 1 March to 30 June) (±standard deviation) calculated from data of the years 2014-2017 in the experimental field 1 (10 m above sea level) and experimental field 2 (850 m above sea level). Climatic data were provided by the Sicilian Water Observatory (www.osservatorioacque.it (accessed on April 2019). In Figure 1, for each EF, the daily maximum, mean and minimum air temperature and rainfall, registered over a period of 30 years (1984-2013) are reported. Both EFs were established in 2007, with plants grafted onto pear seedlings and subjected to standard agronomical practices. The germplasm consisted of 61 cultivars (each of those was planted in triplicates) as already described by Bennici and colleagues [34]; 18 of those were cultivated in both EFs (Supplementary Table S1).

130
The main phenological growth stage was monitored every four days according to the 131 Biologische Bundesanstalt, Bundessortenamt and Chemical industry-BBCH [36] from 132 the rest period till the end of blooming. Moreover, for each cultivar, the start of harvest 133 was registered. 134

135
The Chill Day model (CD) [28] was tested in both EFs. The threshold temperature 136 (TC) and the chilling requirement (CR) values were detected through an iterative process 137 aimed to identify the TC-CR combination that minimizes the root mean square error (RI) 138 between predicted and observed number of days from the end of the previous season 139 (harvest) to bud burst. Analysis was conducted using an in-house R script [37] available 140 upon request ( Figure 2). In particular, the CD model accuracy was tested with the root 141 mean square error (RMSE) between predicted and observed dates: where dpi is the predicted and doi is the observed bud-burst date for the ith season, and N 143 is the number of seasons.  Supplementary Table S2. Climatic data were provided by the Sicilian Water Observatory (www.osservatorioacque.it (accessed on April 2019)).

Phenological Monitoring in Open Field
The main phenological growth stage was monitored every four days according to the Biologische Bundesanstalt, Bundessortenamt and Chemical industry-BBCH [36] from the rest period till the end of blooming. Moreover, for each cultivar, the start of harvest was registered.

Chill Day Model Employment
The Chill Day model (CD) [28] was tested in both EFs. The threshold temperature (TC) and the chilling requirement (CR) values were detected through an iterative process aimed to identify the TC-CR combination that minimizes the root mean square error (RI) between predicted and observed number of days from the end of the previous season (harvest) to bud burst. Analysis was conducted using an in-house R script [37] available upon request ( Figure 2). In particular, the CD model accuracy was tested with the root mean square error (RMSE) between predicted and observed dates: where dpi is the predicted and doi is the observed bud-burst date for the ith season, and N is the number of seasons.

150
A total of six cultivars were selected for the assessment of the chilling requirement 151 after storage in a climatic chamber ('Bianchetto', 'Coscia', 'Gentile', 'Muscatello', 'Ucciar-152 done', 'Urzì'). For each cultivar, 40 twigs of about 30 cm were randomly collected in au-153 tumn. The number of nodes per twig varied according to the cultivar architecture from a 154 minimum of four to a maximum of 16. Residual leaves were hand-thinned, and the apical 155 bud was excised to facilitate the switch from para-dormancy to endo-dormancy [38]. The 156 axillary buds from a typical node with one central leaf bud and two flower buds were 157 conserved. Cuttings were bundled into four groups, each containing 10 twigs, and 158 wrapped in paper. Bundles were submerged in water and sterilized with benomyl [me-159 thyl-(butylcarbamoyl)-2-benzimidazolecarbamate], rinsed for 10 min, placed in sealed 160 plastic bags and stored at 4 ± 0.1 °C for 300, 500, 700, 900 h, respectively [8]. After each 161 chilling interval, the twigs were placed with their basal tip in water and forced in a growth 162 chamber at 25 ± 0.1 °C for 28 days with a photoperiod of 16 h of light. Flower bud-break 163 was assessed at 7 (T7), 14 (T14), 21 (T21) and 28 (T28) days. The bud-break was considered 164 reached when the inflorescence emergence stage (Principal growth stage 5: inflorescence 165 emergence. BBCH stage 53-Bud-burst: green leaf tips and visible flowers) was reached 166 [36]. 167 The end of the endo-dormancy was determined when the bud break exceeded 70% 168 and no further increase was observed [39]. 169

170
Data related either to chilling requirements, threshold temperature and twigs were 171 analyzed with IBM SPSS Statistics for Windows, version XXI (IBM Corp., Armonk, NY, 172 USA) for analysis of variance (One-way ANOVA) by testing the significance of each var-173 iable. Fisher's LSD pairwise comparison procedure at p ≤ 0.05 level was used to determine 174 significant differences among inflorescence emergence rates observed in excised twigs of 175 six accessions cultivated in the experimental field 2 (850 m a.s.l.) (Supplementary

178
In this research, 61 accessions from Sicilian pear germplasm were grown in two dif-179 ferent pedoclimatic areas representatives of different Mediterranean environments in 180 which P. communis L. is usually cultivated (Supplementary Table S1) [40,41]. In EF1, 181 monthly maximum temperatures range between 16 °C and 31 °C, while monthly mini-182 mum temperature spans from 11 °C to 25 °C in winter and summer, respectively. The total 183 annual rainfall, based on long-term observations, is 208 mm. Tree fruit crops represent the 184 main agricultural activity that, during the centuries, has been set up in the different envi-185 ronments of the Etna volcano. In fact, thanks to the wide varieties of microclimates that 186 Figure 2. Output of the R script employed for the definition of the best values for the threshold temperature (TC) (A) and the chilling requirement (CR) (B) found by the iteration that minimizes the root mean square error (RI) (C), between predicted and observed number of days from the end of the previous season (harvest) to bud-burst (figures referred to 'Coscia' cultivar).

Chill Requirements Analysis in Growth Chamber
A total of six cultivars were selected for the assessment of the chilling requirement after storage in a climatic chamber ('Bianchetto', 'Coscia', 'Gentile', 'Muscatello', 'Ucciardone', 'Urzì'). For each cultivar, 40 twigs of about 30 cm were randomly collected in autumn. The number of nodes per twig varied according to the cultivar architecture from a minimum of four to a maximum of 16. Residual leaves were hand-thinned, and the apical bud was excised to facilitate the switch from para-dormancy to endo-dormancy [38]. The axillary buds from a typical node with one central leaf bud and two flower buds were conserved. Cuttings were bundled into four groups, each containing 10 twigs, and wrapped in paper. Bundles were submerged in water and sterilized with benomyl [methyl-(butylcarbamoyl)-2-benzimidazolecarbamate], rinsed for 10 min, placed in sealed plastic bags and stored at 4 ± 0.1 • C for 300, 500, 700, 900 h, respectively [8]. After each chilling interval, the twigs were placed with their basal tip in water and forced in a growth chamber at 25 ± 0.1 • C for 28 days with a photoperiod of 16 h of light. Flower bud-break was assessed at 7 (T7), 14 (T14), 21 (T21) and 28 (T28) days. The bud-break was considered reached when the inflorescence emergence stage (Principal growth stage 5: inflorescence emergence. BBCH stage 53-Bud-burst: green leaf tips and visible flowers) was reached [36].
The end of the endo-dormancy was determined when the bud break exceeded 70% and no further increase was observed [39].

Statistical Analysis
Data related either to chilling requirements, threshold temperature and twigs were analyzed with IBM SPSS Statistics for Windows, version XXI (IBM Corp., Armonk, NY, USA) for analysis of variance (One-way ANOVA) by testing the significance of each variable. Fisher's LSD pairwise comparison procedure at p ≤ 0.05 level was used to determine significant differences among inflorescence emergence rates observed in excised twigs of six accessions cultivated in the experimental field 2 (850 m a.s.l.) (Supplementary Table S3).

Results and Discussion
In this research, 61 accessions from Sicilian pear germplasm were grown in two different pedoclimatic areas representatives of different Mediterranean environments in which P. communis L. is usually cultivated (Supplementary Table S1) [40,41]. In EF1, monthly maximum temperatures range between 16 • C and 31 • C, while monthly minimum temperature spans from 11 • C to 25 • C in winter and summer, respectively. The total annual rainfall, based on long-term observations, is 208 mm. Tree fruit crops represent the main agricultural activity that, during the centuries, has been set up in the different environments of the Etna volcano. In fact, thanks to the wide varieties of microclimates that are present at different heights and sides of the volcano, it is possible to cultivate many different species and with significant differences in the harvesting period as well. Wine grape, citrus and olive are the widespread species, but of peculiar importance are pistachio, nut and prickle pear apple, cherry, hazelnut and chestnut. Such cultivation strongly marks the landscape, the economy, and the culture. Although old genotypes are found on all slopes of the volcano, nowadays pears are largely cultivated in the west side of the volcano. In EF2 a particular climate is characterized by cold and wet winters and mild summers in which rain can occur. In the whole year, precipitation exceeds 900 mm. The mean lowest temperature in winter is 6 • C while in summer the monthly maximum temperature is 30 • C (Figure 1, Supplementary Table S2). This area is strictly related to citrus cultivation. In the last few years, some fruit tree crop species such as peach and apricot cultivars with a low chilling requirement, have been cultivated.
The analyzed accessions were previously characterized both morphologically [42,43] and genetically [34,35] and represent a valuable germplasm source for several traits of agronomical interest, such as ripening and harvesting periods, yield and adaptability with limiting environmental factors, and resistance to biotic stress both for fresh and transformed products, or for their antioxidant power [44,45] An assessment of their cold and warm needs to ensure optimal phenological development was carried out in both EF1 and EF2. The analysis was aimed to better dissect the genotype x environment interaction and to provide useful information to growers and breeders for a rational selection of the cultivar(s) for specific environments. Furthermore, the contemporary analysis of the germplasm in two environments is of great interest for a consistent and predictable investigation of the winter chilling accumulation. This aspect is of particular interest, especially in light of the great variability observed between the years of the observations [46]. The annual harvest dates of the two EFs were assessed and implemented in the CD model [28] to calculate the chilling days (Cd), the antichill days (Ca) and the corresponding chilling requirements and threshold temperature for each cultivar in each EF (Tables 2 and 3). These parameters allowed a precise estimation of the days needed to break the bud eco-dormancy and endodormancy, respectively. An accurate comparison of the performances and reliability of the most widely used models was carried out by Cesaraccio et al. [28]. Among the tested models, the CD model showed a much lower RMSE compared to the others due to a more accurate determination of the dormancy period.
Significant differences between the two EFs were registered in terms of mean Tc and CR (Table 4, Figures 3 and 4). In the EF1, a lower amount of CR (−103 days) and a higher mean TC (8.1 • C) compared to EF2 (Chilling requirements = −122 days; Threshold Temperature = 6.7 • C) were needed for bud break ( Table 4). The observed differences in TC between the two EFs can be due, to a certain extent, to differences in the genetic background between the two germplasm collections, but certainly environment played a significant role. In EF1, the mean RMSE was 2.82 and ranged between 1.94 and 3.17. In EF2 values ranged between 1.94 and 3.08.      of September or at the beginning of October while, for late-ripening cultivars, the CR was 251 reached later in February (Tables 2 and 3). Among the 18 cultivars grown in both EFs, 252 'Reale', 'Campana', 'Regina' and 'Coscia' showed an earlier flower bud sprouting in the 253 colder EF2 compared to EF1 (from 11 to three days). Anticipation of the flower bud sprout-254 ing in colder environments could be related to an easier fulfilment of the CR as already 255 reported for 'Coscia' (Figure 3) [28,48]. The remaining fourteen cultivars showed flower 256 bud sprouting anticipation in the warmer EF1 (six days on average), with 'Pasqualino' 257 and 'Virgolese' showing the highest precocity (11 days, Figure 3). As for TC, no major 258 differences were detected among the two experimental fields (Figure 4).  Figure 4 reports the registered TC for each season. As far as the EF1, several cultivars 266 showed a strong variability between the first (>10 °C) and the third season (7 °C). In the 267 EF2, for several cultivars the TCs registered in the first season were lower than those ob-268 served for EF1. 269 As reported in the literature, 'Coscia' is considered a low chilling requirement culti-270 var and, consequently, it is widely used in different warm winter countries. In our condi-271 tion, the bud break for this cultivar occurred between the second and the third week of 272 March, 14 days later to that observed in two different areas in Sardinia (Italy), in which 273 the bud break was registered between the end of February and the first week of March 274 [28], but earlier compared to Lleida (Spain) in which bud break was registered at the end 275 of March [49]. 'Spadona' and 'Coscia' are the main pear cultivars grown in the warm cli-276 mate of Israel, and several agronomical and genetic studies have been carried out to es-277 tablish the field performance and the genetic control of their bud-break time [50,51]. A 278 cross population obtained by crossing 'Spadona' (low CU-requiring cultivar = 300 CU) 279 and 'Harrow Sweet' (high CU-requiring cultivar = 800 CU) has been used for QTL fine-280 mapping of vegetative bud break time in European pear [46]. Our results highlighted that 281 more than 80% of the accessions in EF1 showed an evident anticipation of bud break com-282 pared to 'Coscia'. Most of these cultivars showed an anticipation ranging from 16 to six 283 days in bud break in EF1 and EF2 respectively, indicating the presence in this germplasm 284 of interesting variability sources for this specific trait. 285 Table 5 reports the rates of inflorescence emergence in excided twigs of one year after 286 an endodormancy period. After 14 days of observation only the cultivar 'Gentile' ex-287 ceeded the 70% of sprouting for buds stored for 900 h at 4 ± 0.1 °C. The threshold of 70% 288 of bud break is often considered the threshold to define the period of endo-dormancy 289 breaking [11]. After 21 day at warm temperature (25 °C) the highest values of bud sprout 290 In EF1, TC was higher compared to that reported in previous reports for pear and cherry (6-7.5 • C), while results were similar to those detected for kiwi (8 • ), olive and some woody species (≥10 • C) [28,47]. On the other side, the mean CR values were very close to those reported by Cesaraccio et al. [28] for pear.
As for the phenological traits, the length of the quiescence period was strongly influenced by the high variability in harvest date among the studied cultivars [48] (Tables 2 and 3). In fact, harvest date began in the last days of June in both EFs and ended at the end of October in the EF1 and in the first week of November in the EF2. Therefore, in both EFs, the fulfilment of the CR for several early-ripening cultivars was reached in the third week of September or at the beginning of October while, for late-ripening cultivars, the CR was reached later in February (Tables 2 and 3). Among the 18 cultivars grown in both EFs, 'Reale', 'Campana', 'Regina' and 'Coscia' showed an earlier flower bud sprouting in the colder EF2 compared to EF1 (from 11 to three days). Anticipation of the flower bud sprouting in colder environments could be related to an easier fulfilment of the CR as already reported for 'Coscia' (Figure 3) [28,48]. The remaining fourteen cultivars showed flower bud sprouting anticipation in the warmer EF1 (six days on average), with 'Pasqualino' and 'Virgolese' showing the highest precocity (11 days, Figure 3). As for TC, no major differences were detected among the two experimental fields (Figure 4). Figure 4 reports the registered TC for each season. As far as the EF1, several cultivars showed a strong variability between the first (>10 • C) and the third season (7 • C). In the EF2, for several cultivars the TCs registered in the first season were lower than those observed for EF1.
As reported in the literature, 'Coscia' is considered a low chilling requirement cultivar and, consequently, it is widely used in different warm winter countries. In our condition, the bud break for this cultivar occurred between the second and the third week of March, 14 days later to that observed in two different areas in Sardinia (Italy), in which the bud break was registered between the end of February and the first week of March [28], but earlier compared to Lleida (Spain) in which bud break was registered at the end of March [49]. 'Spadona' and 'Coscia' are the main pear cultivars grown in the warm climate of Israel, and several agronomical and genetic studies have been carried out to establish the field performance and the genetic control of their bud-break time [50,51]. A cross population obtained by crossing 'Spadona' (low CU-requiring cultivar = 300 CU) and 'Harrow Sweet' (high CU-requiring cultivar = 800 CU) has been used for QTL fine-mapping of vegetative bud break time in European pear [46]. Our results highlighted that more than 80% of the accessions in EF1 showed an evident anticipation of bud break compared to 'Coscia'. Most of these cultivars showed an anticipation ranging from 16 to six days in bud break in EF1 and EF2 respectively, indicating the presence in this germplasm of interesting variability sources for this specific trait. Table 5 reports the rates of inflorescence emergence in excided twigs of one year after an endodormancy period. After 14 days of observation only the cultivar 'Gentile' exceeded the 70% of sprouting for buds stored for 900 h at 4 ± 0.1 • C. The threshold of 70% of bud break is often considered the threshold to define the period of endo-dormancy breaking [11]. After 21 day at warm temperature (25 • C) the highest values of bud sprout were observed on twigs stored for 700 h, while the 900 h did not show the same effectiveness. Among the accessions in analysis, 'Bianchetto', 'Gentile' and 'Muscatello' exceeded 70% of bud sprout under the 700 h cold treatment. 'Gentile' exceeded the bud sprout threshold under 500 and 900 h of cold treatment as well, while bud sprout in 'Muscatello' occurred only at 900 h treatment. After 28 days at 25 • C, all accessions reached the 70% of bud sprout after 700 h. These results indicated that for the majority of the studied accessions, a cold exposition of at least 700 h, followed by a 21-day exposure to warmer temperature (25 • C in our conditions), are mandatory for break endodormancy and for sprouting. For all the accessions, the CU requirement was fulfilled after 700-900 h of cold exposition, except for 'Gentile' for which a lower chilling requirement between 500 and 700 h was evidenced. The low chilling requirements of 'Gentile' makes this cultivar an ideal candidate for cultivation in warm climates. The overall results evidenced a general low CU requirement of the analyzed accessions, compared to those reported for other varieties, indicating a CU requirement ranging from 1600 to 1800 hours [39]. In addition to the temperature effect, the two EFs were characterized by a different type of soil. EF2 in particular is characterized by a volcanic ash soil, thus, even though temperatures were lower, and rainfall was higher, plants were more stressed in EF2 compared to EF1. Table 5. Inflorescence emergence rates observed in excised twigs of six early genotypes cultivated in the experimental field 2 (850 m a.s.l.). Excised twigs (one year old) were stored in climatic chamber for 300, 500, 700 and 900 h at +4 ± 0.1 • C followed by 7 (T7), 14 (T14), 21 (T21) and 28 (T28) days at +25 ± 0.1 • C. Degrees of freedom for the error term (DFE), mean square error (MSE), Fisher's F-test (F), and significance level (Sig.) from one-way ANOVA within cultivars separately for days of storage in climatic chamber at +25 ± 0.1 • C are reported. Inflorescence emergence rates exceeding the 70% (end of endodormancy) are underlined.

Conclusions
The CD model adopted in this work showed its effectiveness in detecting site and genotype specific CR and TC. A precise determination of such parameters could allow a more precise identification of the cultivars that well adapt to specific environments, maximizing buds opening and ensuring an optimal development of the flower with positive influence in fruit yield and quality.
The studied accessions were selected by growers during the past centuries both for their good adaptability and fruit quality characteristics. Most of them showed a significant lower chilling requirement compared to that of the reference cultivar 'Coscia'. This trait is of particular interest for areas characterized by mild winters in which the adoption of low chilling varieties could allow the cultivation of temperate species, including pear, contributing to the enlargement of market availability thanks to a wide ripening calendar.
Supplementary Materials: The following are available online at https://www.mdpi.com/2311-7 524/7/3/45/s1, Table S1: List of the studied genotypes during the period 2014-2017 grown in the two s experimental field 1 (10 m a.s.l) and experimental field 2 (850 m a.s.l.). Accessions in bold were planted in both experimental fields. Table S2: Daily minimum, mean and maximum air temperatures and rainfall registered in the experimental field 1 (10 m a.s.l) (A) and experimental field 2 (850 m a.s.l.). Climatic data were provided by the Sicilian Water Observatory (www.osservatorioacque.it (accessed on 20 April 2020)). Table S3: Results of the LSD pairwise comparison procedure at the p prot Significant Difference (LSD). Excised twigs (one year old) were stored in climatic chamber for