“Ponchi Re” and “Tarzan” dwarf tomato plants were grown from seeds bought from Prudac (Enkhuizen, The Netherlands) in a greenhouse compartment of Wageningen University and Research in Wageningen, The Netherlands. Seeds were sown on September 11th, 2017 in a greenhouse with a photoperiod of 16 h per day with temperature settings at 21°C (day and night), and a relative humidity of 75%. When daylight was insufficient, additional artificial lighting (Philips 400 watt Son-T natrium lamps, Eindhoven, The Netherlands) was employed. Plants were watered manually and supported with a plastic peg for stability during fruiting. Fertilizer was applied according to commercial growth practices.

When plants were fully flowering, all flowers were removed, and the temperature treatments were started. The plants were divided randomly and placed in three climate cabinets (HPS1500S, Weiss Technik, Germany) set to either 16, 22, and 28°C. For each temperature treatment, a corresponding relative humidity was set as to ensure a constant vapor pressure difference of 0.8 kPa for all treatments (Table 1). Every week, the position of the plants within a chamber was randomized. MG tomatoes were harvested twice, 1 week apart. Forty five MG tomatoes per cultivar were selected in each harvest and transported within 10 min to the Horticulture and product physiology laboratory, Wageningen University for further storage. Per temperature/cultivar combination, six extra tomatoes were harvested for immediate destructive measurements. Tomatoes of the first harvest were stored first at 4°C for 15 d and then placed at 20°C for 20 days (shelf life), whereas tomatoes from the second harvest were stored immediately at shelf life condition (20°C).

The SSC was measured using a digital refractometer (Atago refractometer PR-32α, Fukaya-Shi, Saitama, Japan). Six tomatoes were measured for each treatment. The tomatoes were cut in half. Using one half per tomato, the juice was squeezed out and used to determine the SSC (°Brix) using the refractometer. The other half of the fruit was used for dry matter content measurement. The refractometer and the knife used for cutting were cleaned with distilled water before each individual measurement. Results were expressed in ° Brix.

The remaining halves of tomatoes from SSC measurement were weighed after cutting (fresh weight), and oven-dried at 70°C until constant weight (24–48 h). Fresh and dry weight (in g) were measured using a balance (Mettler PM 480, Mettler-Toledo, Leicester, UK) and subsequently expressed as the percentage fresh to dry weight (Kaiser et al., 2019).

The fruit size was determined by measuring the diameter (in mm) of the tomato along the equator, using a digital caliper.

Color was assessed non-destructively by a hand-held photodiode array spectrophotometer (Pigment Analyzer PA1101, CP, Germany). Remittance was assessed at 570 (R570) and 780 (R780) nm by calculating the normalized different vegetative index (NDVI, Equation 1) and normalized anthocyanin index (NAI, Equation 2) which are normalized value between −1 and 1 (Affandi et al., 2020).

The measurement head of the pigment analyser was adapted by adding a small plastic cup on top of the measurement head to collect light for color measurements of dwarf tomatoes.

Firmness was measured using a Zwick Z2.5/TS1S materials testing machine (Ulm, Germany). A probe (diameter 0.8 cm) was placed on the skin of the tomato and compressed the tomato 0.5 mm with the maximum force recorded and regarded as tomato firmness (Schouten et al., 2007). Each tomato was measured twice, and the average was taken. Measurement were taken on the equator of the fruit. After the first measurement the tomato was turned 45° for the second measurement. Results were expressed in Newton.

Each tomato was judged for the degree of CI using the chilling injury index (CII) (Vega-García et al., 2010). Three CI symptoms were measured; uneven ripening and color development (U), pitting (P), and decay (D). The severity of each symptom is based on the percentage of affected tissue (0 = No injury 1 = <10%, 2 = 11–25%, 3 = 26–40%, 4 = >40%) with the CII score calculated as CII = (U+P+D)/3. Tomato weight loss over time was expressed as the percentage weight loss compared to the initial weight (Affandi et al., 2020).

One-way ANOVA was employed to analyse data measured at harvest whereas data obtained over the course of shelf life were subjected to mixed ANOVA, applying SPSS ver.21 (SPSS, Chicago, USA) at P < 0.05. In the mixed ANOVA, growth temperature treatment and storage temperature serve as between subject factor whereas shelf life days as within subject factor. The Shapiro-Wilk was applied to test the normality of the variables. Mauchly's test of sphericity was carried out to test whether variances of the differences between all possible pairs of within-subject conditions were equal. Greenhouse-Geisser's correction was applied to calculate the degrees of freedom in the case that the sphericity assumption was not fulfilled. When mixed ANOVA resulted in a significant interaction, a pairwise comparison was carried out for each shelf life day with LSD (Least Significant Difference) values estimated.

Tomato plants started flowering after 7 weeks after sowing. At this moment, they were transferred from the greenhouse compartment to the climate cabinet. Plant morphology was affected by the growth temperature: higher growth temperature during cultivation resulted in larger and darker leaves compared to plants grown at 22°C. Whereas plants grown at 16°C were smaller than that of 22°C (Supplementary Figure 1). The fruiting period was also affected; tomatoes were ready to be harvested at mature green (MG) at 12, 13, and 14 weeks after sowing for tomatoes cultivated at 28, 22, and 16°C, respectively. A higher growth temperature also correlated with increased fruit production (data not shown).

A growth temperature of 16 and 28°C resulted in higher SSC (°Brix) values at harvest for both “Ponchi Re” and “Tarzan” tomatoes, compared with tomatoes cultivated at 22°C (Figure 1A). No differences with respect to growth temperature were found for fruit fresh weight and fruit dry matter content (data not shown). Diameter of the MG tomatoes was not affected by growth temperature for “Ponchi Re” tomatoes. “Tarzan” tomatoes grown at 22°C, had a larger diameter than those grown at both 28 and 16°C (Figure 1B).

After harvest all green harvested tomatoes matured to red ripe, values were measured non-destructively by the pigment analyser (Figure 2). NAI and NDVI values are representative for lycopene and chlorophyll levels in the tomato pericarp, respectively (Schouten et al., 2014). The point in time where the NAI and NDVI values cross (cross point, CP), is an indication of the synchronization of the chlorophyll decay and lycopene formation. In “Ponchi Re,” the CP was ~14, 12, and 3 days after the start of shelf life without cold storage for tomatoes cultivated at 16, 22 and 28°C, respectively. For prior cold stored “Ponchi Re,” the CP was 17, 9, and 7 days (Figures 2A–F). For non-cold stored “Tarzan” tomatoes the delay in color formation due to growth temperature was less pronounced (Figures 2G–L): 10, 8, and 5 days without storage, and 12, 7, and 9 days with prior cold storage for cultivation at 16, 22, and 28°C, respectively. “Ponchi Re” tomatoes showed a difference in the CP of ~11 days between the lowest and highest growth temperature without cold storage (Figures 2A–C). This difference is approximately only 5 days for “Tarzan” tomatoes (Table 2). In fruit cultivated at 16 and 28°C, prior cold storage resulted in a delay in the start of color development. This delay is about 2–4 days for tomatoes of both cultivars (Figure 2). Interestingly the CP was shorter for prior cold stored compared to non-cold stored tomatoes for especially “Ponchi Re” tomatoes (3 days) when cultivated at 22°C (Figures 2B,E).

Firmness at harvest was lower for tomatoes cultivated at 22°C (Figures 3A,B). Cold storage resulted in less softening for tomatoes cultivated at 16°C (Figures 3C,D). Softening during shelf life without prior cold storage was faster for “Ponchi Re” tomatoes cultivated at 28°C than tomatoes cultivated at 16 and 22°C (Figure 3A). Softening for non-cold stored “Tarzan” tomatoes was not affected by growth temperature (Figure 3B). Cold storage had no effect on the speed of softening in Ponchi Re. However, the reduced softening of Tarzan fruit grown at 16°C during cold storage was compensated by an increased softening during shelf life, to end at the same firmness as the other fruit, after 20 days of shelf life.

Weight loss during storage and shelf life was generally correlated with cultivation temperature. Weight loss was lowest at 16°C, and increased for tomatoes of both cultivars cultivated at 22 and 28°C (Figure 4). Weight loss during shelf life of prior cold stored tomatoes was similar as for non-cold stored tomatoes, with the exception of “Ponchi Re” fruit cultivated at 22°C (Figure 4C). Reduced weight loss for “Ponchi Re” grown at 16 and 28°C compared with 22°C, could be considered a positive effect of cultivation stress on quality.

Chilling injury was judged by observing three symptoms; uneven ripening and color development, pitting, and decay (Affandi et al., 2020). For “Ponchi Re” tomatoes, increasing growth temperature resulted in lower CI incidence (P < 0.0001) while the opposite occurred for “Tarzan” tomatoes (Figure 5). “Tarzan” tomatoes, cultivated at 28°C, experienced higher CI index values due to higher decay and pitting. In contrast, CI index of “Ponchi Re” cultivated at 16°C was mainly due to higher uneven coloring and decay (data not shown).

A lower (16°C) than optimum growth temperature increased SSC of “Ponchi Re” tomatoes (Figure 1). This is in line with the findings of Klopotek and Kläring (2014), who found increased accumulation of soluble sugars when tomato plants are exposed to growth temperature between 14–16°C. Increased levels of soluble sugars, especially glucose, induce the ascorbate–glutathione cycle that scavenges H₂O₂ (Shao et al., 2013). Soluble sugars are also thought to act as osmoprotectant that confers freezing tolerance in plants (Ruelland et al., 2009). In addition, higher soluble sugars are associated with better cold tolerance such as in tomato (Liu et al., 2012; Zhang et al., 2019), loquat (Shao et al., 2013) and nectarines (Zhao et al., 2019). Higher SSC at harvest for 16°C cultivated “Ponchi Re” tomatoes was, however, associated with higher CI incidence (Figure 5). This indicates that SSC is here likely not associated with chilling tolerance. In addition, higher SCC values for 16°C cultivated “Ponchi Re” tomatoes softened at the same rate, indicating that this higher SSC was not related to difference in developmental stage. On the contrary, firmness at harvest was generally higher for 16°C cultivated “Ponchi Re” tomatoes (Figure 3). It is possible that a higher SSC may have an osmotic effect resulting in increased water containment causing higher firmness (Shackel et al., 1991; Farneti et al., 2013; Lahaye et al., 2013). “Ponchi Re” fruits cultivated at higher than optimum (28°C) also had higher SSC, and in this case, a lower CI index was observed. Plants cultivated at the lowest growth temperature showed smaller plants and less number of fruit and leaves (Supplementary Figure 1), it is likely that the plants grown at 16°C were able to allocate more assimilates to fruit which caused higher SSC (Heuvelink, 1997; Beckles, 2012).

The crossing point (CP) indicates the level of synchronization of pigment degradation and synthesis. Cold storage was reported to delay and reduce lycopene accumulation in MG tomatoes despite uninterrupted chlorophyll degradation during shelf life (Affandi et al., 2020). The CP could therefore be regarded as indicator of chilling sensitivity. “Ponchi Re” tomatoes showed a larger difference in CP between the lowest and highest growth temperature with or without cold storage than “Tarzan” tomatoes (Figure 2). This indicates that color development is delayed by lower growth temperatures for especially “Ponchi Re” tomatoes. NAI and NDVI behavior is highly synchronized by growth temperature, with and without cold storage. This might indicate that a system that synchronizes chlorophyll decay and lycopene synthesis is dependent on growth temperature. Such a system might be under the regulation SGR (STAY-GREEN) protein. SGR genes induces chlorophyll breakdown and simultaneously blocks lycopene synthesis until SGR levels decrease to a threshold that allows synthesis of lycopene to start (Hu et al., 2011; Luo et al., 2013; Sánchez-González et al., 2016). Possibly, synthesis and breakdown of SGR in “Ponchi Re” tomatoes is more temperature dependent than in “Tarzan” tomatoes.

Prior cold stored showed a similar delay with the non-cold stored tomatoes in the start of the color development for both cultivars when cultivated at 16 or 28°C (Figure 2). The opposite was shown for tomatoes cultivated at 22°C. This means that cold storage benefitted the start of color development for tomatoes when cultivated at 22°C. Likely, cold storage induced the accumulation of lycopene precursors, but only when cultivated at 22°C. It is currently unclear why this happens. But it is clear that growth temperature affects color development. For instance, Hernandez et al. (2015) showed that a high (32°C) growth temperature decreased and later increased the lycopene content when the number of days of high temperature exposure was increased.

A lower growth temperature resulted in less softening during cold storage for both cultivars (Figures 3C,D). In addition, tomatoes cultivated at 16°C showed less weight loss in general (Figure 4). Firmness retention during chilling and shelf life is associated with higher structural cell wall integrity and reduced decay (Mirdehghan et al., 2007; Rodoni et al., 2010; Gang et al., 2015). Lower weight loss is thought to be related to higher membrane integrity leading to less decay (Cohen et al., 1994; Ali et al., 2019; Ahmed et al., 2021). Finally, a higher SSC could also lead to increased osmolarity, retaining more water resulting in higher firmness (Shackel et al., 1991; Farneti et al., 2013; Lahaye et al., 2013).

Higher firmness retention and lower weight loss in low temperature cultivated “Tarzan” tomatoes than the control and higher temperature was associated with lower CI incidence (Figure 5). This might indicate that the cold tolerance of “Tarzan” from low temperature cultivation is, at least partly, based on increased membrane integrity. Increased cold tolerance in terms of ability to restore development after prior exposure to low temperature (4°C) was also found in tomato plants (Barrero-Gil et al., 2016), watermelon (Lu et al., 2020), and sweet pepper (Ferguson et al., 1999) when cultivated at lower temperatures. There could also be a genetic component. It is possible that “Tarzan” may have a better regulation of the components of the chilling tolerance inducing CBF pathway (Singh et al., 2011) or produce small heat shock proteins (Luengwilai et al., 2012) when cultivated at lower cultivation temperatures.

The delay in lycopene formation in “Ponchi Re” tomatoes cultivated at the lowest growth temperatures (Figure 2) is probably associated with the increased CI incidence at lowest growth temperature (Figure 5). Reduced lycopene formation might result in less ROS scavenging capacity as lycopene and lycopene precursors, such as phytoene and phytofluene, scavenge singlet oxygen and peroxyl radicals generated by chilling stress (Engelmann et al., 2012; Lado et al., 2016; Rey et al., 2020). This delay might trigger a cascade of oxidative processes that lead to chilling damage (Hodges et al., 2004; Martínez et al., 2014). Cold stored “Tarzan” tomatoes cultivated at 16°C, also experienced a delay in color development, although to a lesser extent (Figure 2). However, this delay did not result in more CI symptoms (Figure 5). This implies that the development of CI symptoms depends not only on the ROS scavenging capacity but also on other factors, undoubtedly dependent on genotypical differences between the two cultivars used in this study, such as membrane integrity, cell wall integrity, and transpiration rate of the fruit (Cohen et al., 1994; Rodoni et al., 2010; Gang et al., 2015).