1 Abstract
The predicted rise in global temperature is not only affecting plant functioning directly, but is also increasing air vapour pressure deficit (VPD). The yield of banana is heavily affected by water deficit but so far breeding programs have never addressed the issue. A reduction in transpiration at high VPD has been suggested as a key drought tolerance breeding trait to avoid excessive water loss, hydraulic failure and to increase water use efficiency. In this study, stomatal and transpiration responses under increasing VPD at the leaf and whole-plant level of 8 wild banana (sub)species were evaluated, displaying significant differences in stomatal reactivity. Three different groups were identified under increasing VPD. M. acuminata spp. errans (group I), M. acuminata spp. zebrina (group II) and M. balbisiana (group II) showed the highest transpiration rate limitations to increasing VPD. In contrast to group I, group II only showed strong reductions at high VPD levels, limiting the cost of reduced photosynthesis and strongly increasing their water use efficiency. Group II genotypes thus show favourable responses for high water use efficiency in regions with high VPDs. This provides a basis for the identification of potential parent material within their wild populations for drought tolerance breeding.
Highlight Wild banana species respond significantly different to water deficit caused by VPD increases and differ in the rate of stomatal reduction, revealing opportunities for drought tolerance breeding.
2 Introduction
Climate change projections predict that global temperatures will continue to increase this century (IPCC, 2021). This temperature rise is not only affecting plant functioning directly, but is also increasing air vapour pressure deficit (VPD) (Hatfield and Prueger, 2015; Ficklin and Novick, 2017; Grossiord et al., 2020). VPD represents the atmospheric water vapour demand and is defined as the difference between the saturation and actual vapour pressure in the atmosphere (Monteith and Unsworth, 2013). The saturation vapour pressure, the water vapour that air can hold, increases exponentially with temperature and has been increasing as global temperatures rise (Lawrence, 2005). The actual vapour pressure (i.e. absolute humidity in the air) on the other hand has not been rising at the same rate as the saturation vapour pressure, therefore increasing the worldwide VPD (Ficklin and Novick, 2017; Grossiord et al., 2020). The impact of this rising VPD is often underestimated compared to other climate change consequences, but periods of high VPD have recently been linked with large-scale tree mortality (Breshears et al., 2013; Williams et al., 2013) and strong yield reductions (Challinor and Wheeler, 2008; Lobell et al., 2013).
Plants respond to the vapour pressure deficit encountered at the leaf level, the leaf-to-air vapour pressure deficit (VPDleaf). The leaf temperature can after all deviate from that of the ambient air by transpirational cooling or heating through radiant energy. For a given stomatal opening, transpiration would increase linearly with VPDleaf, without any gain in carbon uptake. Stomatal conductance (gs) however decreases with increasing VPDleaf, avoiding excessive water loss, but restricting carbon uptake (Dai, Edwards and Ku, 1992; Monteith, 1995; Oren et al., 1999). In angiosperms the reduction of gs in response to an increase in VPDleaf is believed to be abscisic acid (ABA) mediated (Xie et al., 2006; Bauer et al., 2013; McAdam and Brodribb, 2015). Upon an increase in VPDleaf, gs is reduced by a rapid ABA biosynthesis (i.e. within 20 min) presumably located in the leaf phloem parenchyma cells and stomatal guard cells (Kuromori, Sugimoto and Shinozaki, 2014; McAdam, Sussmilch and Brodribb, 2016). The trigger for ABA interference under high VPDleaf is believed to be a drop in water status (McAdam and Brodribb, 2016; Sack, John and Buckley, 2018), which has been linked to a limited maximal hydraulic conductance at the leaf, stem and/or root level in comparison to the transpiration (Brodribb and Jordan, 2008; Zhang et al., 2013; Choudhary et al., 2014; Ocheltree, Nippert and Prasad, 2014; Schoppach et al., 2016). Essential gatekeepers for this hydraulic conductance are aquaporins. They are present all along the water transport pathway from root to stomata. Aquaporins were less abundant in soybean and pearl millet genotypes that showed a reduced transpiration rate at high VPDleaf (Sadok and Sinclair, 2010; Devi, Sinclair and Taliercio, 2015; Reddy et al., 2017).
Despite the reductions in gs, the transpiration rate usually increases with increasing VPDleaf. Only at high VPDleaf significant decreases in transpiration rates have been observed (Franks, Cowan and Farquhar, 1997; Fletcher, Sinclair and Allen, 2007; Gholipoor et al., 2010; Ryan et al., 2016). These transpiration responses are commonly described by a segmented pattern where the slope of transpiration rate versus VPDleaf is significantly reduced after a specified breakpoint. Significant differences in segmented transpiration responses to VPDleaf have been observed across- and within-species (Fletcher, Sinclair and Allen, 2007; Gholipoor et al., 2010; Ryan et al., 2016). While some species or genotypes already reduce transpiration rate significantly at low VPDleaf, others show only a reduction at higher VPDleaf or even maintain the increasing transpiration rate. Restricting transpiration rate at high VPD has been suggested as a key drought tolerance breeding trait as excessive water loss is avoided and might be saved for later in the growing season (Vadez, 2014; Sinclair et al., 2017). Limiting transpiration above a VPD threshold can increase the daily transpiration efficiency but the reduced water use may compromise the yield potential. Reduced transpiration limits carbon uptake, thereby hampering photosynthesis and yield (Richards, 2000; Lee et al., 2020; Eyland et al., 2021). Moreover, care must be taken that the so-called saved water is not merely lost by evaporation or transpiration by neighbouring plants.
The transpiration rate response to VPD was shown to be highly heritable in wheat (Schoppach et al., 2016). Models predict that in drought-prone environments limiting transpiration at high VPD would improve maize and soybean yields by maintaining more soil water available later in the season during flowering or grain filling (Sinclair et al., 2010; Messina et al., 2015). In these drought-prone regions, the negative effect of gs reduction on A during vegetative growth could be compensated later in the growing season (Sinclair et al., 2010; Messina et al., 2015). Improved maize hybrids which, amongst other traits, showed reduced transpiration at high VPDleaf indeed increased yields under water-limited conditions (Gaffney et al., 2015), while for durum wheat cultivars this was only the case under severe drought conditions (Medina et al., 2019).
The current set of edible bananas is complex and has resulted from different parental routes and several back crosses (De Langhe et al., 2010; Perrier et al., 2011; Martin, Baurens, et al., 2020; Cenci et al., 2021). The hybrid banana genomes are unbalanced with respect to the parental ones, and inter- and intra-genome translocation chromosomes are relatively common (Christelová et al., 2017; Němečková et al., 2018). Most, if not all, cultivars have genomes consisting of different proportions of A- and B-genome chromosomes and/or recombinant chromosomes originating from different parents. Similar to other tropical species, bananas are very sensitive to VPD, with reductions in transpiration when VPD exceeds 2 – 2.3 kPa (Aubert and Catsky, 1970; Carr, 2009; Eyland et al., 2022). Thomas et al. (1998) observed a diverse response in three banana cultivars with different genomic constitutions. Despite these efforts, the transpiration responses to VPD remain largely uncharacterized across diverse banana genotypes.
The main objective of this work was to evaluate 8 wild banana (sub)species for their stomatal and transpiration responses under increasing VPD at the leaf and whole-plant level. Transpiration rate limitations at high VPD have been indicated as a key breeding trait for high water use efficiency. This work could therefore provide the basis for systematically screening crop wild relatives of banana for their transpiration at high VPD, with the aim to identify potential parent material for drought tolerance breeding.
3 Materials & methods
3.1 Plant material & growing conditions
A diversity panel of 9 wild banana genotypes belonging to 8 (sub)species (Table 1) were phenotyped for their transpiration response to VPD. Plants were grown in 2.5 L pots filled with peat-based compost and maintained under well-watered conditions. Plants were grown in the greenhouse for 6 - 8 weeks before moving to the growth chamber (Bronson PGC-1400, the Netherlands). The growth chamber contained an air mixing fan and LED panels providing a light intensity of 250 µmol m-2 s-1 for a 12 h photoperiod and a light spectrum with blue:red:far-red ratio of 1 : 1.5 : 0.15. Plants were acclimated to the growth chamber for one day under a day/night temperature and relative humidity of 27/24.5 °C and 78 %, respectively. The next day the VPD step-changes were initiated by altering relative humidity, while temperature was maintained at 36 °C during this day. VPD was increased by decreasing relative humidity as temperature fluctuations would not only affect VPD but also aquaporin conductance and water viscosity in xylem and mesophyll cells (Matzner and Comstock, 2001; Yang et al., 2012). At light onset relative humidity was maintained for 90 min at 87 %, after which it was subsequently decreased to 78, 68, 62 and 56 %, each for 60 min. Average VPDs at each step were 0.77, 1.36, 1.93, 2.34 and 2.64 kPa. Plants were maintained under well-watered conditions by daily watering before light onset. Measurements were taken before 14:00 to avoid afternoon stomatal closure (van Wesemael et al., 2019; Eyland et al., 2021).
3.2 Leaf gas exchange measurements
Gas exchange responses to step increases in VPDleaf were measured every 60 s on the middle of the second youngest fully developed leaf using a LI-6800 infrared gas analyser (LI-COR, USA). Light intensity and CO2 concentration were maintained at 250 µmol m-2 s-1 and 400 µmol mol-1, respectively. Leaf temperature was maintained at 36 °C. Relative humidity went from 85 to 75, 65, 55, 45 and 35 %, reaching VPDleaf of 0.91, 1.50, 2.09, 2.69, 3.28 and 3.87 kPa. Note that measurements were stopped early if the drying capacity of the infra-red gas exchange system was saturated and unable to maintain reduced relative humidity. The intrinsic water use efficiency (iWUE) was calculated as iWUE = A/gs with A being the photosynthetic rate. At every VPDleaf level the steady-state gs, A, Erate (transpiration rate) and iWUE after 60 min was calculated. The maximum gs was calculated as the highest gs observed across all VPDleaf levels. Segmented regression was performed on the transpiration rate response to increasing VPDleaf for each genotype by using a nonlinear mixed effect model in which the intercept was assumed to vary at individual plant level (segmented R package, Muggeo, 2008). This analysis calculates the optimal breakpoint in the transpiration response with a different linear response before and after the breakpoint. To determine the effect of the reduction in stomatal opening on the transpiration, the transpiration reduction (ϕE) was determined according to Franks et al. (1999) and Ryan et al. (2016) (Fig. 1). For each individual, a linear regression was fitted through the transpiration rate at the first two VPDleaf levels (0.90 and 1.50 kPa). This linear regression was then extrapolated to predict the transpiration rate (Epred) at higher VPDleaf levels (2.69, 3.28 and 3.87 kPa) (Fig. 1). The percentage decrease of the actual measured transpiration rate (Emeas) compared to Epred (Fig. 1) was then quantified at each VPDleaf level:
The percentage of limitation of the photosynthetic rate (A) by gs reduction was calculated at every VPDleaf level by comparing the measured A (Ameas) with the overall maximally measured A (Amax): Stomatal reduction (ϕstom) with increasing VPD was defined as the absolute slope between stomatal conductance (gs) and loge(VPDleaf) as described by Oren et al. (1999): where a is the estimated gs at VPDleaf 1 kPa.
3.3 Whole-plant transpiration rate
Plants were placed on balances (0.01 g accuracy, Kern, Germany) to register their weight every 10 s. The soil was covered by plastic to avoid evaporation and ensure only water loss through transpiration. Transpiration during each VPD step was calculated by differentiating 5 min average total weight (mtot) at the start of the VPD level with the 5 min average total weight at the end of the VPD level: Transpiration was normalized by leaf area (LA) and the time (t) passed. LA was quantified by destructive leaf area imaging at the end of the experiment.
Segmented regression was performed on the transpiration rate response to increasing VPD for each genotype by using a nonlinear mixed effect model in which the intercept was assumed to vary at plant level. Transpiration reduction (ϕE) was determined according to Eq. 1 with linear regression between the two first VPD levels (0.77 and 1.36 kPa) and comparison between Epred and Emeas at the highest level (2.64 kPa).
3.4 Statistics
All data processing and statistical analysis were carried out in R (V3.6.2). Genotypic differences were tested by applying analysis of variance (ANOVA) with a post hoc Benjamini & Hochberg correction. Significance of the segmented response of transpiration rate to VPD compared to a linear response was determined by the Davies Test (segmented R package, Muggeo, 2008). K-means clustering of genotypes was performed on the average scaled output of the segmented regression, the transpiration reduction, the stomatal reduction and photosynthesis limitation, including measurements by leaf gas exchange and by whole-plant transpiration were included (Hartigan and Wong, 1979). Clusters were optimized across 10,000 random sets of cluster centres and plotted on the first two principal components.
4 Results
4.1 Diverse response to VPD: three phenotypic clusters
The transpiration response was measured at leaf and whole plant level while relative humidity was stepwise decreased and VPDs consequently increased. The response to increasing VPD at leaf and whole-plant level was described by the segmented regression of transpiration rate versus VPD, the transpiration reduction (Eq. 1), the photosynthetic limitation under increasing VPD (Eq. 2) and the stomatal reduction (Eq. 3). K-means clustering was performed on the output variables measured by both leaf gas exchange and whole-plant transpiration (Table 2). Three clusters were identified and plotted along the first two principal components (Fig. 2). The first principal component was mainly determined by the limitation of photosynthetic rate (A) at high VPDs and the transpiration reduction at leaf and whole-plant level (Table 2). Important variables in the second principal component were the slope before the breakpoint in transpiration rate with increasing VPD and the stomatal reduction (Table 2). Cluster I consisted of only one genotype: M. acuminata ssp. errans (Fig. 2). In group II M. acuminata ssp. zebrina and M. balbisiana clustered together (Fig. 2). Group III contained 6 genotypes: M. acuminata ssp. banksii, ssp. burmannica, ssp. burmannicoides, ssp. malaccensis and ssp. microcrocarpa (Fig. 2).
4.2 Leaf level responses of gs, transpiration rate and A to increasing VPDleaf
With increasing VPDleaf, gs decreased in all genotypes (Fig. 3A, Supplemental Table S1). The transpiration rate initially increased, but eventually reached steady-state or even declined (Fig. 3B). The transpiration rate and gs of M. acuminata ssp. errans were lowest and differed significantly from all other genotypes at VPDleaf exceeding 1.50 and 2.09 kPa, respectively (Fig. 3A-B, Supplemental Table S1). Under a VPDleaf ≤ 2.9 kPa, the highest transpiration rates and gs were observed for M. balbisiana and M. acuminata ssp. burmannica. However, when VPDleaf increased further, the gs of M. balbisiana decreased stronger than M. acuminata ssp. burmannica, translating only in M. balbisiana in a lower transpiration rate (Fig. 3A-B, Supplemental Table S1). As gs decreased with increasing VPDleaf, the CO2 uptake was limited and A decreased (Fig. 3C). The lowest A was observed for M. acuminata ssp. errans and ssp. burmannicoides, with significantly lower A compared to all other genotypes except M. acuminata ssp. zebrina (Fig. 3C, Supplemental Table S1). The intrinsic water use efficiency (iWUE) increased with increasing VPDleaf (Fig. 3D). iWUE was highest in M. acuminata ssp. errans and differed significantly from all other genotypes as VPDleaf exceeded 1.5 kPa (Fig. 3D, Supplemental Table S4.1). The lowest iWUE were observed for M. acuminata ssp. burmannica and ssp. burmannicoides (Fig. 3D).
In all genotypes there was a decrease in the slope of transpiration rate versus VPDleaf (Fig. 3B). This response was described by a segmented regression with a specified breakpoint after which the slope of the transpiration rate decreases. A significant breakpoint in transpiration rate in response to VPDleaf was identified in all genotypes (Fig. 4). Across genotypes the breakpoints ranged between 1.75 and 2.5 kPa with M. acuminata ssp. errans having a significant breakpoint at the lowest VPDleaf (Fig. 4, Fig. 5). Two M. acuminata ssp. banksii genotypes and ssp. microcarpa showed the highest breakpoint in transpiration rate (Fig. 4, Fig. 5). The groups defined by k-means clustering differed in their segmented transpiration response (Fig. 5). Group I consisted only of M. acuminata ssp. errans, the genotype with a breakpoint (a reduction in transpiration rate) at the lowest VPDleaf, as well as the lowest slope (the lowest Erate) before the breakpoint (Fig. 5). Group II, consisting of M. acuminata ssp. zebrina and M. balbisiana, had a breakpoint at a relatively low VPDleaf around 2 kPa and a negative slope after the breakpoint (Fig. 5). This negative slope indicates a net decrease in transpiration rate, which was not observed in the other genotypes. In group III all genotypes kept relatively high transpiration rates at relatively high VPDleaf. Musa acuminata ssp. burmannica, ssp. burmannicoides and ssp. malaccensis had a breakpoint at relatively low VPDleaf, but maintained a high slope of transpiration rate afterwards while the M. acuminata ssp. banksii genotypes and ssp. microcarpa showed only a significant breakpoint in transpiration rate at higher VPDleaf, (Fig. 5).
The transpiration reduction (ϕE) (Eq. 1, Fig. 1) representing the increase in stomatal resistance with increasing VPDleaf also differed significantly across genotypes (Fig. 6A, Supplemental Table S2). Reductions in transpiration ranged between 37 and 59 % at the highest VPDleaf of 3.87 kPa (Fig. 6A, Supplemental Table S2). The highest reductions in transpiration were observed for M. acuminata ssp. errans, ssp. zebrina and M. balbisiana (Fig. 6A). The transpiration reduction of group I and II was significantly higher compared to group III at all VPDleaf levels (Fig. 6A, Supplemental Table S2).
The decrease in stomatal opening with increasing VPDleaf limited the photosynthetic rate (A). In all genotypes there was a significant increase in the limitation of A with increasing VPDleaf (P < 0.01) and the limitation ranged from 7 to 17 % at the highest VPDleaf level (Fig. 6B, Supplemental Table S3). The limitation of A was highest in M. acuminata ssp. errans from VPDleaf 2.69 kPa onwards, followed by M. acuminata ssp. zebrina and M. balbisiana (Fig. 6B, Supplemental Table S3). The limitation of A was significantly higher in group I compared to group II and III from VPDleaf 2.69 kPa onwards (Supplemental Table S3). At VPDleaf of 3.28 and 3.87 kPa group II had a significantly higher A limitation compared to group III (Supplemental Table S3). Across genotypes the limitation of A at higher VPDleaf (≥ 2.69 kPa) was significantly correlated to the breakpoint in transpiration rate (R2 = 0.47-0.57; Supplemental Fig. S1). Similarly, the limitation of A and the transpiration reduction at higher VPDleaf (≥ 2.69 kPa) were significantly correlated (R2 = 0.53-0.58; Supplemental Fig. S1). These correlations indicate that strong reductions in transpiration at high VPDleaf result in higher A limitations.
The stomatal reduction (ϕstom), defined as the slope of gs versus loge(VPDleaf) (Eq. 3) differed significantly across genotypes (Supplemental Table S4). Highest stomatal reduction was observed in M. balbisiana, while M. acuminata ssp. errans showed lowest reduction (Fig. 7, Supplemental Table S4). The stomatal reduction was strongly correlated to the maximum observed gs (R2 = 0.88, Fig. 7, Supplemental Fig. S1). No significant differences across previously described groups was observed (Supplemental Table S4).
4.3 Whole-plant transpiration rate responses corroborate leaf measurements
The whole-plant transpiration rate increased between 98 and 197 % with increasing VPD (Fig. 8). The lowest transpiration rates were observed for M. acuminata ssp. errans with significant differences compared to all other genotypes from VPD 1.93 kPa and beyond (Fig. 8, Supplemental Table S5). Transpiration rates of all other genotypes were double compared to M. acuminata ssp. errans at the highest VPD level (Fig. 8, Supplemental Table S5).
A significant breakpoint in whole-plant transpiration rate response to VPD was identified in all genotypes (Fig. 9). The breakpoints ranged between 1.6 and 2.2 kPa, with M. acuminata ssp. errans and M. balbisiana having the lowest breakpoint (Fig. 9, Fig. 10). The slope after the breakpoint was strongly negative in M. acuminata ssp. errans and ssp. zebrina (Fig. 9, Fig. 10). Genotypes belonging to group I or II thus showed breakpoints in transpiration rate at lower VPD values and/or strongly negative second slopes (Fig. 10).
The whole-plant transpiration reduction (ϕE) (Eq. 1, Fig. 1) of M. acuminata ssp. errans was significantly higher compared to all other genotypes (Fig. 11, Supplemental Table S6). The second highest transpiration reduction was observed for M. acuminata ssp. zebrina and M. balbisiana (Fig. 11, Supplemental Table S6). Group I (M. acuminata ssp. errans) showed a significantly higher transpiration reduction compared to group II and III (Supplemental Table S6). Group II (M. acuminata ssp. zebrina and M. balbisiana) showed a significantly higher transpiration reduction compared to group III (Musa acuminata ssp. burmannica, ssp. burmannicoides, ssp. malaccensis, ssp. banksii and ssp. microcarpa) (Supplemental Table S6).
The whole-plant transpiration reduction was significantly correlated to the transpiration reduction measured at leaf level at similar VPD (R2 = 0.52, Fig. 11, Supplemental Fig. S1). Similarly, the whole-plant transpiration reduction was significantly correlated to the limitation of A measured at leaf level for VPDleaf exceeding 2.1 kPa (R2 = 0.50 - 0.73, Supplemental Fig. S1).
5 Discussion
Diversity in transpiration patterns with increasing VPD has been observed among different genotypes of many crops including chickpea, maize, peanut, pearl millet, sorghum and soybean (Fletcher, Sinclair and Allen, 2007; Gholipoor et al., 2010; Jyostna Devi, Sinclair and Vadez, 2010; Kholová et al., 2010; Yang et al., 2012; Ryan et al., 2016; Sivasakthi et al., 2017). We observed a significant change in the transpiration rate of 9 wild banana genotypes already at VPD levels between 1.6 and 2.5 kPa (Fig. 4, Fig. 9). These values are in line with the general transpiration rate reduction of banana at VPD 2 to 2.3 kPa reported by Carr (2009) and the modelled VPD responses of (Eyland et al., 2022). The breakpoints in transpiration rate were at similar VPDs compared to other crops (Gholipoor et al., 2010; Yang et al., 2012; Ryan et al., 2016). However, in other crops several genotypes were identified without a breakpoint as they maintained a linear increase in transpiration rate with increasing VPD (Fletcher, Sinclair and Allen, 2007; Gholipoor et al., 2010; Jyostna Devi, Sinclair and Vadez, 2010; Kholová et al., 2010; Yang et al., 2012; Ryan et al., 2016; Sivasakthi et al., 2017). Moreover, temperature and other environmental factors like radiation and soil water potential have been shown to interact with VPD in banana (Eyland et al., 2022). These complex interactions explain why a fixed VPD level per genotype, where a reduction in transpiration takes place, cannot be defined without taking the other environmental conditions in account.
The wild banana genotypes clustered in three groups based on their leaf gas exchange and whole-plant transpiration response to VPD (Fig. 2). Genotypes of group I and II, M. acuminata ssp. errans, M. acuminata ssp. zebrina and M. balbisiana, showed the highest transpiration rate limitations. This is in line with our previous observations under fluctuating conditions: M. balbisiana showed together with M. acuminata ssp. errans the most pronounced response by strongly decreasing their transpiration rate (Eyland et al., 2022). As reported by Oren et al. (1999), the stomatal reduction was significantly correlated to the maximum gs (Fig. 7, Supplemental Fig. 1). This indicates that genotypes with higher gs under low VPDleaf show higher stomatal closure at increasing VPDleaf. However, M. acuminata ssp. errans (group I) showed a very strong stomatal response, despite its low gs. As a consequence of this strong stomatal restriction, the iWUE of M. acuminata ssp. errans was significantly higher compared to all other genotypes (Fig. 3D). In contrast to the very conservative behaviour of M. acuminata ssp. errans, the genotypes of group II displayed high gs and A when VPDleaf was favourable in addition to early or strong transpiration rate reductions at high VPDleaf. This behaviour is assumed to be beneficial in drought-prone areas with periods of high VPD (Sadok and Sinclair, 2010; Vadez, 2014), as water is used efficiently and saved for later in the growing season. Some genotypes of group III also showed a breakpoint in transpiration at a relatively low VPDleaf, but a high transpiration rate was kept and a net transpiration increase continued with rising VPDleaf (Fig. 4, Fig. 5). Hence, these genotypes display a more risk taking behaviour, thereby risking hydraulic failure (Sade, Gebremedhin and Moshelion, 2012).
The transpiration reduction at leaf level was significantly correlated to the reduction at whole-plant level, suggesting similar responses to increasing VPD (Fig. 11). The conservative behaviour of genotypes of group I and group II was validated at the whole-plant level by breakpoints in transpiration rate at low VPDs and/or low increases in transpiration afterwards (Fig. 5, Fig. 9).
As demonstrated in other crops, identification of this conservative behaviour towards VPD, opens up possibilities to improve drought tolerance of cultivated banana hybrids. M. balbisiana is a parent to many edible bananas belonging to the AAB, ABB and AB genome groups and their subgroups. In line with the conservative behaviour of M. balbisiana in response to VPD (Fig 3-4, Fig 6, Fig 8), it has been indicated in many studies that edible bananas with a high portion of B genes are related to drought tolerance (Ekanayake, Ortiz and Vuylsteke, 1994; Thomas, Turner and Eamus, 1998; Turner and Thomas, 1998; Thomas and Turner, 2001; Vanhove et al., 2012; Kissel et al., 2015; Van Wesemael et al., 2018; van Wesemael et al., 2019; Eyland et al., 2021, 2022; Uwimana et al., 2021). Also M. acuminata spp. zebrina is a parent to several edible bananas (Carreel et al., 2002; Perrier et al., 2011; Němečková et al., 2018; Baurens et al., 2019; Martin, Baurens, et al., 2020; Martin, Cardi, et al., 2020; Jeensae et al., 2021), among others the East-African highland banana subgroup (i.e. Mutika/Lujugira). The East-African highland banana subgroup, endemic to the East-African highlands, is due to its risk taking behaviour sensitive to drought (Kissel et al., 2015; van Wesemael et al., 2019; Eyland et al., 2021; Uwimana et al., 2021). Hence, identification of drought tolerance traits in M. acuminata ssp. zebrina populations provides opportunities to mitigate climate change impacts in this and all other important subgroups. So far, not much is known about the contribution of M. acuminata ssp. errans to edible bananas. The accession screened in this study and representing M. acuminata ssp. errans, has been proved to be complex in genome with ancestries coming from ‘malaccensis’, ‘zebrina’ and ‘burmannica/siamea’ (Martin, Cardi, et al., 2020).
6 Conclusion
The reduction of transpiration response to high VPD is a key trait for water saving and diversity among wild banana relatives was observed. Reductions in transpiration ranging between 37 and 59 %, translated in an increased WUE of 54 to 166 %. M. acuminata spp. errans, on the one hand, responded most conservative, but was also characterized by low gs overall. M. acuminata ssp. zebrina and M. balbisiana, on the other hand, showed strong stomatal closure while maintaining relatively high carbon uptake under low VPD. These two genotypes thus show favourable responses for a specific sub-trait linked to high water use efficiency, providing a potential basis for identification of parent material for drought tolerance breeding.
7 Supplementary data
Table S1: Genotype-specific steady state response of gs, Erate, A and iWUE at increasing VPD
Table S2: Genotype-specific transpiration reduction at increasing VPD
Table S3: Genotype-specific photosynthetic rate limitation at increasing VPD
Table S4: Genotype-specific stomatal reduction
Table S5: Genotype-specific whole-plant transpiration rate at increasing VPD
Table S6: Genotype-specific whole-plant transpiration reduction
Fig. S1: Correlation matrix leaf and whole-plant traits
8 Acknowledgements
The authors would like to thank Edwige Andre for the plant propagation; Hendrik Siongers, Stan Blomme, Loïck Derette and Poi Verwilt for their technical assistance during plant growth and phenotyping.
9 Author contributions
SC and RS wrote the concepts for funding. DE performed the experiments and analyzed the data. SC supervised the experiments. SC, CG and DE wrote the manuscript. All authors reviewed and approved the final manuscript.
10 Conflict of interest
The authors declare no conflicts of interest.
11 Funding
This study was undertaken as part of the initiative ‘Adapting Agriculture to Climate Change: Collecting, Protecting and Preparing Crop Wild Relatives’ which is supported by the Government of Norway. The project is managed by the Global Crop Diversity Trust in partnership with national and international gene banks and plant breeding institutes around the world http://www.cwrdiversity.org/. DE was supported by a scholarship funded by the Global TRUST foundation project ‘Crop Wild Relatives Evaluation of drought tolerance in wild bananas from Papua New Guinea’ [Grant number: GS15024]. CG was supported by a PhD scholarship funded by the Belgian Development Cooperation project ‘‘More fruit for food security: developing climate-smart bananas for the African Great Lakes region”. The authors thank all donors who supported this work also through their contributions to the CGIAR Fund (http://www.cgiar.org/who-we-are/cgiar-fund/fund-donors-2/), and in particular to the CGIAR Research Program Roots, Tubers and Bananas (RTB-CRP) and to the ERA-Net transnational call European Research Projects LEAP Agri H2020 cofund project on food & nutrition security & sustainable agriculture, with funding from national funding agencies for the Project ‘PHENOTYPING THE BANANA BIODIVERSITY TO IDENTIFY CLIMATE SMART VARIETIES WITH OPTIMAL MARKET POTENTIAL IN AFRICA AND EUROPE’.
12 Data availability
All data supporting the findings of this study are available within the paper and within its supplementary materials published online
Footnotes
↵* first co-authorship
Email addresses clara.gambart{at}kuleuven.be
david.eyland1994{at}gmail.com
rony.swennen{at}kuleuven.be
Abbreviations
- A
- photosynthetic rate
- Amax
- maximally measured photosynthetic rate
- Ameas
- measured photosynthetic rate
- ABA
- abscisic acid
- Erate
- transpiration rate
- Emeas
- measured transpiration rate
- Epred
- predicted transpiration rate
- Eq
- equation
- gs
- stomatal conductance
- h
- hour
- ITC
- International Transit Centre
- kPa
- kilopascal
- L
- liter
- LA
- leaf area
- m
- meter
- min
- minutes
- mol
- moles
- mtot
- total weight
- PC
- principal component
- s
- seconds
- se
- standard error
- ssp
- subspecies
- R2
- R-squared
- t1
- timepoint 1
- t2
- timepoint 2
- VPD
- vapour pressure deficit
- VPDleaf
- leaf-to-air vapour pressure deficit
- iWUE
- intrinsic water use efficiency
- µmol
- micromoles
- ϕE
- transpiration reduction
- ϕstom
- stomatal reduction