Introduction
Theobroma cacao L. (hereafter cocoa) is considered one of the most important perennial crops, with an estimated world output of 4.65 million tons in 2017–2018 (ICCO, 2019). Most smallholders in tropical developing countries cultivate it using rain-fed agroforestry (Carr and Lockwood, Reference Carr and Lockwood2011). Yet, very often, their production is limited by water deficit as well as nutrients such as N (Balasimha et al., Reference Balasimha, Daniel and Prakash1991; dos Santos et al., Reference dos Santos, Almeida, Ahnert, Branco, Valle and Baligar2016). Amelioration in cocoa production should therefore consist of selecting clones able to cope with dry season (DS) without competing for water or nutrients.
Drought affects establishment and yield of cocoa (Bae et al., Reference Bae, Kim, Kim, Sicher, Lary, Strem, Natarajan and Bailey2008; dos Santos et al., Reference dos Santos, Almeida, Ahnert, Branco, Valle and Baligar2016), specially during the juvenile stage (Ayegboyin and Akinrinde, Reference Ayegboyin and Akinrinde2016; De Almeida et al., Reference De Almeida, Tezara and Herrera2016; García and Moreno, Reference García and Moreno2016). In adult trees, drought affected differently the water relations and photosynthetic responses of cocoa clones (Ávila-Lovera et al. Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; De Almeida et al., Reference De Almeida, Tezara and Herrera2016). Global climate change scenarios for South America predict lower rainfall averages during the DS (Gornall et al., Reference Gornall, Betts, Burke, Clark, Camp, Willet and Wiltshire2010). Cocoa plants submitted to drought showed different responses which depended on the plant organ evaluated, that is, reduction of leaf area, and increase of root system and stem biomass strategic mechanisms for the survival of cocoa progenies under drought (dos Santos et al., Reference dos Santos, Almeida, Ahnert, Branco, Valle and Baligar2016). Therefore, it is essential to know the physiological responses of different cocoa clones to prolonged dry periods.
Decreases in yield caused by low water availability may largely be associated with a decline in net photosynthetic rate (A), either by a reduction in stomatal conductance (g s) or by diminished metabolic processes (Lawlor and Tezara, Reference Lawlor and Tezara2009). In cocoa, low A was related to lower efficiency of Photosystem II (PSII) in the capture of excitation energy, as well as lower electron transport rate, J (Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016). The adjustment of physiological processes, especially photochemistry and g s feedbacks on A, and instantaneous water use efficiency (WUE = A/E, where E = transpiration rate) and yield in response to drought, which can provide functional details of growth and development in this tropical crop tree, are still not well understood (Daymond et al., Reference Daymond, Tricker and Hadley2011).
In C3 plants, the carbon isotopic ratio (δ 13C) of leaf biomass gives a better indication of intercellular CO2 concentration (C i) over a longer time interval than instantaneous gas exchange measurements (Farquhar et al., Reference Farquhar, Ehlenringer and Hubick1989). The value of C i in itself is a reflection of g s relative to A. As such, δ 13C provides an integrated or long-term WUE during the time when the carbon in leaf tissues was assimilated (Farquhar et al., Reference Farquhar, Ehlenringer and Hubick1989; Farquhar and Richards, Reference Farquhar and Richards1984). Long-term WUE is an important trait associated with yield and adaptation to water deficit (Monti et al., Reference Monti, Amaducci, Pritoni and Venturi2006; Rajabi et al., Reference Rajabi, Griffiths, Ober, Kromdijk and Pidgeon2008). The ability to recognize genotypes that combine high yield with high WUE is essential when breeding crops for drought-prone areas (Acosta-Rangel et al., Reference Acosta-Rangel, Ávila-Lovera, De Guzman, Torres, Haro, Arpaia, Focht and Santiago2018; Dias et al., Reference Dias, Araujo, Moraes, Barros and Damatta2007; dos Santos et al., Reference dos Santos, Almeida, Ahnert, Branco, Valle and Baligar2016). Although the importance of screening large numbers of crops under similar agricultural management has been recognized, many breeding programs fail due to inappropriate clone choice for a certain region as a consequence of inadequate knowledge of the potential of clones and their growth and productivity under different site conditions (Rajabi et al., Reference Rajabi, Griffiths, Ober, Kromdijk and Pidgeon2008). A target trait at the leaf level for the selection of drought-tolerant woody crops such as cocoa could be WUE.
Cocoa is commonly divided into three morpho-geographic groups (morphotypes) which differ in quality, vigor, and yield: Criollo, Forastero, and Trinitario. Motamayor et al. (Reference Motamayor, Lachenaud, da Silva e Mota, Loor, Kuhn, Brown and Schnell2008) suggested a new classification of cocoa grouped into 10 genetic groups, which would better reflect the genetic diversity available for breeders, among which the Criollo cultivars have a low genetic diversity (Motamayor et al., Reference Motamayor, Risterucci, Lopez, Ortiz, Moreno and Lanaud2002). On the whole, cocoa shows significant genetic variability regarding morphological and physiological traits (Daymond et al., Reference Daymond, Hadley, Machado and Ng2002a, b). However, studies of genotypic differences in photosynthetic performance of cocoa are few (Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Daymond et al., Reference Daymond, Tricker and Hadley2011; De Almeida et al., Reference De Almeida, Tezara and Herrera2016). Despite a wide genotypic variability (Motamayor et al., Reference Motamayor, Lachenaud, da Silva e Mota, Loor, Kuhn, Brown and Schnell2008), most of the studies on yield reductions under water deficit have been reported only for Trinitario and Forastero clones (Almeida and Valle, Reference Almeida and Valle2007; Daymond et al., Reference Daymond, Tricker and Hadley2011; De Almeida et al., Reference De Almeida, Tezara and Herrera2016; García and Moreno, Reference García and Moreno2016; Joly and Hahn, Reference Joly and Hahn1989), while little information on Criollo cocoa is available (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016).
In order to gain knowledge on the changes in physiological traits during seasonal drought, we evaluated water status, gas exchange, leaf δ 13C, photochemical activity, and leaf N and chlorophyll content in 31 cocoa clones grown in a common garden during the rainy season (RS) and DS. Our main question was Can the possible differences in physiological performance and responses of Venezuelan cocoa clones to drought suggest differences in drought tolerance? We hypothesized that during the DS, WUE would increase differently among clones due to differential physiological plasticity.
Materials and Methods
Plant material and field site
The study was carried out in the field germplasm bank at Instituto Nacional de Investigaciones Agrícolas (INIA)-Miranda Padrón Experimental Station, Tapipa, Venezuela (10º13′46′′ N, 66º18′39′′ W). The climate corresponds to a tropical rainforest with a mean annual precipitation of 2480 ± 120 mm and an air temperature of 25.8 ± 1.9 °C. The rainfall regime of the study area is characterized by an RS with 317 ± 7 mm mo–1 from June to December and a DS with 75 ± 24 mm mo–1 from January to April (a mean of 10 years). Measurements were made in September during RS and April at the end of the DS. Measurements were made on five-year-old 2-m-tall trees of 31 cocoa clones (17 Trinitarios, 6 Criollos, and 8 Modern Criollos, i.e. hybrids between Criollo and Forastero with many Criollo morphological traits; Motamayor et al., Reference Motamayor, Risterucci, Lopez, Ortiz, Moreno and Lanaud2002), which were selected from the germplasm bank on the basis of their high productivity, and were propagated on open-pollinated IMC 67 rootstock. Clones were collected from five different geographical regions of Venezuela (Table 1). From now on, clones will be referred to by the number assigned in Table 1, instead of their names. Trees were planted following the germplasm policy in a common garden divided into five 400 m2 square plots. Trees grew to 6–8 per clone per plot. Cocoa trees grew without irrigation under the shade of 15-m-tall trees of Erythrina poeppigiana (bucare) and lateral shade provided by 2.5-m-tall banana plants in every row. Banana trees were 1.5 m apart from each other and cocoa trees 3 m from each other. Trees were fertilized twice a year with commercial fertilizer triple 15-15-15 N:P:K. Soils in the study site were composed of 13.2% sand, 65.2 loam, and 21.6% clay, pH 6.4, available N, P, and K contents of 1.8, 4–8, and 535 mg kg–1, respectively, and a cation exchange capacity of 0.44 dS m–1. No tree died during the period of this study as a result of drought or, for that matter, due to any other damage, such as disease.
Cocoa clones used in this study. Clone code, name, morphotype, location of origin with coordinates, mean annual rainfall
* Trinitario are descendants of crosses between Criollo and Forastero.
† Pure Criollo, mostly with white seeds and high homozygocity percentage.
‡ Modern Criollo are hybrids with many Criollo morphological traits.
Microclimatic conditions
Photosynthetic photon flux density (PPFD) was measured above cocoa trees with an LI-185 quantum sensor meter (LI-COR Inc., Lincoln, NE, USA), air temperature with YSI 405 thermistors connected to a telethermometer (Yellow Springs, OH, USA), and relative humidity (RH) with a hair strand hygrometer (Abbeon model AB167B, Santa Barbara, CA, USA). Leaf water vapor pressure deficit (ΔW) was calculated as the difference in absolute humidity between the leaf intercellular spaces and the atmosphere, where water vapor pressure in the leaf was calculated with values of leaf temperature in the Infrared Gas Analyzer (IRGA) leaf chamber when gas exchange was measured (see below). Each microclimatic parameter was measured every hour for three consecutive days during each season when the physiological measurements were made.
Physiological measurements
Six trees of each clone were sampled to a total of 186 trees per season. All physiological measurements were made on the third mature, fully expanded leaf belonging to a mature branch.
Water relations
Morning leaf xylem water potential (ψ) was measured in one leaf of three trees (n = 3) per clone at 0600–0800 h with a pressure chamber (PMS Instruments Inc., Corvallis, OR, USA). Soil water content (SWC) at 30 cm was determined gravimetrically as SWC = water content/dry mass (%) in four replicates per plot. Soil samples were placed in metal containers, weighed, dried at 100 ºC for 72 h and re-weighed.
Leaf gas exchange
Measurements of A, g s, E, and WUE were made in one leaf per six individuals of each clone (n = 6) with a portable IRGA mod. LCA 4 used in conjunction with an assimilation chamber provided with a halogen lamp unit (Analytical Development Co., Hoddesdon, UK). Measurements were randomly made between 0900 and 1100 h in all clones during three consecutive days each season at ambient CO2 concentration 380 ± 10 µmol mol−1, PPFD 400 ± 10 µmol (photon) m−2 s−1, previously determined to be saturating for many clones (Baligar et al., Reference Baligar, Bunce, Machado and Elson2008; Bertolde et al., Reference Bertolde, Almeida, Pirovani, Gomes, Ahnert, Baligar and Valle2012; De Almeida et al., Reference De Almeida, Herrera and Tezara2018; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016), and leaf temperature 28 ± 1 ºC.
Specific leaf area, leaf N content, and δ 13C
Specific leaf area (SLA) was calculated as the ratio between leaf area and dry mass in six individuals per clone. Leaf area was allometricaly determined by weighing 1 cm2 of paper and paper replicas of the leaves. Leaves were oven-dried to constant mass at 70 ºC and weighed using a digital balance. Leaf samples of four individuals per clone (n = 4) were ground and then analyzed for δ 13C and leaf N content at the University of Illinois at Chicago using an element analyzer (Costech, Valencia, CA, USA) coupled to a Delta+XL isotope ratio mass spectrometer (Finnigan, Bremen, Germany) operated in continuous flow and run against standards from the National Institute of Standards and Technology and laboratory standards to a precision of 0.05% for C.
Biochemical determinations
Leaf chlorophyll and total soluble protein (TSP) contents were determined in discs (n = 6) taken from the same leaf in which gas exchange was measured; discs were enclosed in sealed polythene bags, transported to the laboratory, and maintained in the refrigerator until determinations were done. Total chlorophyll content (Clt) was determined in extracts obtained by grinding discs in chilled 80% acetone after Bruinsma (Reference Bruinsma1963); TSP was determined according to Bradford (Reference Bradford1976) in the macerate obtained by grinding discs over ice and following the procedure for the extraction of rubisco described by Tezara et al. (Reference Tezara, Mitchell, Driscoll and Lawlor2002).
Chlorophyll a fluorescence of PSII
Chlorophyll fluorescence was measured on light-adapted attached leaves under actinic light of 150 ± 10 μmol (photon) m–2 s–1 PPFD for 2 min; this low value of PPFD was used because trees growing under bucare trees received low PPFD. Measurements were made in six trees per cultivar with a Mini PAM fluorometer (Walz, Effeltrich, Germany). Relative quantum yield of PSII at steady-state photosynthesis was calculated as Φ PSII = (F´m−F s)/F´ m, where F s and F´m are fluorescence at steady-state photosynthesis and maximum fluorescence in the light, respectively. Non-photochemical quenching (NPQ) was calculated from measurements of fluorescence as NPQ = (F m−F´m)/F m. Electron transport rate of PSII was estimated as J = Φ PSII × PPFD × 0.84 × 0.5. Measurements were randomly made in all clones during three consecutive days each season.
Statistical analysis
Mean values of the measured parameters were tested for differences between seasons and among clones by two-way analysis of variance (ANOVA) using the statistical software R (v 2.15.0, R Foundation for Statistical Computing, 2012) and Agricolae version 1.2–3. When the ANOVA indicated a significant overall treatment effect, Tukey’s honest significance test (Tukey’s honestly significant difference (HSD), p < 0.05) was carried out to compare means between seasons within a given clone. Principal component analysis (PCA) was conducted using the PAST 3 (PAleontological Statistics) software.
Results
Microclimatic conditions and water relations
Values of PPFD between 0800 and 1600 h under full sunshine ranged from 200 to 1200 μmol (photon) m–2 s–1; maximum PPFD during the RS on bucare-shaded cocoa trees was approximately 30% that under full sunshine. Air temperature and RH varied diurnally from 26 ± 1 to 33 ± 2 ºC, and from 90 ± 5 to 60 ± 2%, respectively, showing no seasonal variation; maximum ΔW, though, was lower during the RS than the DS (1.8 ± 0.3 and 3.2 ± 0.1 KPa, respectively).
SWC decreased from 36 ± 3 in the RS to 16 ± 1% during the DS. All the clones had similar ψ during the RS; a significant difference between seasons was found, with ψ decreasing on average 40% during the DS in 81% of clones, except for Modern Criollo clones, which showed similar ψ (–0.50 MPa) during both seasons (p = 0.048, Figure 1). From RS to DS, values of ψ ranged from –0.37 to –1.18 MPa (Trinitario clones) and from –0.38 to –0.90 MPa (Criollos).
Seasonal changes in intact leaves of adult trees of cocoa clones grown in a common garden during the RS and the DS in (a) leaf water potential; (b) net photosynthetic rate; (c) stomatal conductance; and (d) WUE. Filled bars, RS; empty bars, DS. Values are mean ± SE (n = 6) of different clones (indicated at bottom of panel (d)). An asterisk indicates significant differences due to season for each clone (p < 0.05, Tukey’s HSD after a one-way ANOVA).
Leaf gas exchange
Significant differences in leaf gas exchange variables occurred among clones (Figure 1). On average, a decrease of 47 and 51% in A (p = 0.035) and of 63 and 50% in g s of Trinitario and Criollo clones (p = 0.025), respectively, were observed between the RS and the DS. In Modern Criollos, A did not change with drought. High variability in g s was observed in Trinitario clones (T6–T11) during both seasons and in Criollo clones (C1–C6) during the RS. The WUE was significantly reduced during the DS by 41% in Trinitario and 46% in Criollos (p = 0.007), while in Trinitarios T1–T5, with the lowest WUE (≈2 mmol mol–1), no change was observed between seasons. However, WUE in Modern Criollos was not only similar in both seasons but also significantly higher (≈5 mmol mol–1; p = 0.012) than in the other clones, the highest values occurring in clones MC1–MC5. The interaction clone × season for A and WUE was significant (p = 0.049 for A and p = 0.01 for WUE). During the DS, lower g s was observed in Trinitario and Criollo clones, while Modern Criollo and Trinitario clones T1–T5 were not affected by drought.
The relationship between A and g s was a straight line (Figure 2a, p = 0.005), changes in g s explaining 56% of the variation in A. Values of A ranged from 1.7 to 5.5 μmol m–2 s–1 associated to values of g s from 40 to 120 mmol m–2 s–1 during the RS and from 30 to 80 mmol m–2 s–1 during the DS. The relationship between WUE and ΔW was curvilinear in both seasons, the value of r 2 being slightly higher during the RS (p < 0.05, Figure 2b).
(a) The relationship between net photosynthetic rate and stomatal conductance of intact leaves of adult trees of 31 clones of T. cacao grown in a common garden during the RS (filled circle) and the DS (open circles). The linear regression including both seasons is represented by the solid line, A = 0.03 × g s +1.2; r 2 = 0.56. (b) The response of WUE to leaf water vapor deficit of adult trees of T. cacao grown in a common garden during both seasons. Values are mean ± SE (n = 6) of different clones. The non-linear regression fitted curves are included (solid for RS, broken for DS). The regression equations are WUE = 0.4 × ΔW2 −3.0 × ΔW +7.8; r 2 = 0.32 for RS and WUE = 0.5 × ΔW2 −2.9 × ΔW + 0.5; r 2 = 0.54 for DS.
Leaf carbon isotopic composition as indicator of long-term WUE
Values of δ 13C (Figure 3) showed significant differences among clones (p = 0.045) and between seasons (p = 0.039). The difference between seasons in leaf δ 13C was higher in Criollos (approximately 1.62%). Trinitarios showed less variability between seasons (<1%) but had higher variability among clones. Trinitario clones T2, T4, T9, T11, T15, T17, and T17 showed a large variation in δ 13C between seasons, while in the rest of them δ 13C had a small change. Increases in δ 13C were observed during the DS in most of clones independently of their site of origin (–28.7 ± 0.1%, compared with –30 ± 0.2% in the RS). The highest δ 13C during the RS was observed in clone T8 and during the DS in C4 and T12. Values of δ 13C of all clones under study bore no correlation with either instantaneous WUE (p = 0.08) or intrinsic WUE, A/g s (p = 0.07).
Seasonal changes in intact leaves of adult trees of cocoa clones grown in a common garden in (a) leaf carbon isotope ratio; (b) leaf nitrogen content; and (c) TSP. Filled bars, RS; empty bars, DS. Values are mean ± SE (n = 4 for δ 13C and leaf N, and 6 for TSP) of different clones (indicated at bottom of panel (c)). An asterisk indicates significant differences due to season for each clone (p < 0.05, Tukey’s HSD after a one-way ANOVA).
SLA and biochemical parameters
No significant seasonal changes in SLA were found in either most Trinitarios (T6–T17) or Criollos and Modern Criollos. However, variability in SLA was evident among clones from Barlovento (T1–T5) and between seasons, with the highest values in T2 and T3 (222 cm2 g−1) and the lowest in C2 (114 cm2 g−1). The interaction clone × season for SLA was significant (p = 0.000; data not shown).
A positive correlation between A and area-based N content was found (A = 1.07 × N + 0.11; r 2 = 0.45; p = 0.047; data not shown). No differences in TSP were found among clones but, during the DS, an increase was observed in seven Trinitarios, one Modern Criollo, and three Criollos (Figure 3). Leaf N and TSP contents were similar among clones, but leaf N content varied considerably between seasons in most clones (p = 0.007; Figure 3). On average, a reduction of 35% in leaf N content was observed during the DS. No significant interaction clone × season was found for leaf N or TSP.
Chlorophyll a fluorescence parameters and chlorophyll content
In most clones, J was lower than 90 μmol e− m–2 s–1, except for Trinitario clones T8 and T9; J was significantly higher during the RS in Trinitario from Eastern Venezuela and Margarita, with the exception of clone T8 (p = 0.049, Figure 4). No seasonally changes in J were found in Criollos (C2, C4, and C5) and Modern Criollos (MC2, MC3, MC4, and MC5), but in the rest of clones, differences were significant (p = 0.005) and there was interaction clone × season (p = 0.003). During the DS, NPQ increased in Trinitario and Modern Criollos clones, except for T4, T14, T17, and MC1, no changes being observed in Criollo clones (p = 0.92, Figure 4). A significant reduction of 29% in chlorophyll content was observed during the DS, with the exception of T2, T1, T3, T11, MC2, MC4, and MC5 (p = 0.048, Figure 4), and there was interaction clone × season (p = 0.000).
Seasonal changes in intact leaves of adult trees of cocoa clones grown in a common garden in (a) electron transport rate; (b) NPQ; and (c) total chlorophyll concentration. Filled bars, RS; empty bars, DS. Values are mean ± SE (n = 6) of different clones (indicated at bottom of panel c). An asterisk indicates significant differences due to season for each clone (p < 0.05, Tukey’s HSD after a one-way ANOVA).
Principal component analysis
We used PCA to examine the level of coordination among selected traits across seasons and clones (Figure 5). PCA identified two primary axes of variation (principal components) that explained 54.1% of total variation. The first axis (PC1), which explained 31.2% of total variation, was positively associated with g s, E, ΔW, PPFD, J, and NPQ and negatively associated with WUE and Φ PSII. The second axis (PC2), which explained 22.9% of total variation, was positive associated with ψ, A, WUE, and Clt and negatively associated with δ 13C, ΔW, Φ PSII, and TSP. The PCA fairly clearly showed separation of clones according to season.
Bi-plot of PCA of the trade-offs among physiological traits evaluated of 31 clones of T. cacao: Trinitarios from Barlovento (B, squares), Trinitarios from eastern (E, star), Trinitarios from Margarita (M, triangles), western Criollos (W, inverts triangles), and Modern Criollos from Central Coast (CC, circles), grown in a common garden during the RS (filled symbols) and the DS (open symbols). Traits used for PCA analysis were xylem water potential (ψ), net photosynthetic rate (A), stomatal conductance (g s), transpiration rate (E), instantaneous water use efficiency (WUE), SLA, total chlorophyll content (Clt), TSP, leaf nitrogen content (N), carbon isotopic ratio (δ 13C), photosynthetic photon flux density (PPFD), leaf water vapor pressure deficit (ΔW), relative quantum yield of PSII (ΦPSII), non-photochemical quenching coefficient (NPQ), and electron transport rate (J).
Discussion
Inter-clone variations: genotype or seasonal rainfall?
The results of our study show a great genotypic variation on physiological traits such as water status, photosynthesis, WUE, and N content among cocoa clones in response to drought, which suggests differences in cocoa tolerance to drought. Modern Criollo, Criollo and some Trinitario clones had high WUE during both seasons. The strategies of optimization of water loss and CO2 assimilation were different among clones; some Trinitarios, for example, had higher g s during the RS, while Modern Criollos and Trinitarios from Barlovento showed stomatal control under both seasons, causing differential patterns of 13CO2 enrichment in the leaf. Our hypothesis that during drought, WUE would increase differently among clones due to great physiological plasticity was validated, since instantaneous WUE and δ 13C were significantly different in all clones despite trees growing under similar climatic, soil, and agronomic conditions, with δ 13C being values higher during the DS than the RS, suggesting a higher long-term WUE.
The ψ was poorly correlated with changes in A and g s due to drought, whereas ΔW had a more marked influence on such changes, especially in Modern Criollo. Our results suggest that in cocoa, drought causes decreases in A due to stomatal closure as well as metabolic inhibition, given the observed reductions in N, chlorophyll content, and electron transport rate.
Water relations
The lower precipitation during the DS, which halved SWC, caused a reduction of ψ in most of the clones under study. Similarly, a reduction of ψ was observed in Criollo trees (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Rada et al., Reference Rada, Jaimez, García-Nuñez, Azócar and Ramírez2005) and seedlings of Trinitario clones subjected to water deficit in different experiments in the greenhouse (Bae et al., Reference Bae, Kim, Kim, Sicher, Lary, Strem, Natarajan and Bailey2008; Joly and Hahn, Reference Joly and Hahn1989). In our study, in Modern Criollo and some Trinitario clones, ψ remained unchanged with drought. Cocoa clones more tolerant to drought generally differ physiologically, with mechanisms such as maximization of water uptake by deep, dense root systems, and/or minimization of water loss by stomatal closure (Almeida and Valle, Reference Almeida and Valle2007; dos Santos et al., Reference dos Santos, Almeida, Ahnert, Branco, Valle and Baligar2016). Genetic variation in the ability to tolerate drought in cocoa clones has been documented in only few occasions (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Daymond et al., Reference Daymond, Tricker and Hadley2011).
Leaf gas exchange
The low values of A observed in all clones were associated with low g s. Similar ranges of A and g s have been reported for Trinitario and Forastero clones of different ages grown in different conditions and environments worldwide (Acheampong et al., Reference Acheampong, Hadley and Daymond2013; Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Bae et al., Reference Bae, Kim, Kim, Sicher, Lary, Strem, Natarajan and Bailey2008; Baligar et al., Reference Baligar, Bunce, Machado and Elson2008; Daymond et al., Reference Daymond, Tricker and Hadley2011; De Almeida et al., Reference De Almeida, Herrera and Tezara2018; Galyuon et al., Reference Galyuon, McDavid, Lopez and Spence1996; Joly and Hahn, Reference Joly and Hahn1989; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016).
We found variation in A and its response to drought in the clones studied, similarly to cocoa seedling in greenhouse conditions (Ayegboyin and Akinrinde, Reference Ayegboyin and Akinrinde2016; De Almeida et al., Reference De Almeida, Tezara and Herrera2016; García and Moreno, Reference García and Moreno2016) and cocoa trees in the field (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; De Almeida et al., Reference De Almeida, Tezara and Herrera2016). In most Trinitario and Criollo clones, A and g s decreased during the DS, while in Modern Criollos, these variables were unaffected by drought. Decreases with drought of A in cocoa have been reported previously (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Acheampong et al., Reference Acheampong, Hadley and Daymond2013; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Ayegboyin and Akinrinde, Reference Ayegboyin and Akinrinde2016; Balasimha et al., Reference Balasimha, Daniel and Prakash1991), indicating sensitivity to water deficit, as reviewed by Almeida and Valle (Reference Almeida and Valle2007) and Carr and Lockwood (Reference Carr and Lockwood2011).
There was a general trend toward stomatal closure with declining ψ during the DS, but this was not the cause for the decrease in WUE of Trinitario clones from Margarita and Eastern Venezuela and in Criollos because E was lower in these cases. Rather, diminished WUE was the result of lower A. Similar decreases in WUE with ψ have been reported in Forastero clones (Joly and Hahn, Reference Joly and Hahn1989). In contrast, in other Trinitarios, WUE increased with drought (Almeida and Valle, Reference Almeida and Valle2007; Baligar et al., Reference Baligar, Bunce, Machado and Elson2008). In some Criollo clones during the DS, WUE increased while ψ decreased (Rada et al., Reference Rada, Jaimez, García-Nuñez, Azócar and Ramírez2005) or even remained unchanged (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012), similar to our findings in Trinitario clones (T1–T5). Clones responded differently to ΔW, with most of them showing decreased WUE with increased ΔW during both seasons; in Forastero clones measured under controlled conditions, E increased significantly with ΔW (Baligar et al., Reference Baligar, Bunce, Machado and Elson2008).
The fact that g s and A did not change seasonally in Modern Criollos probably allows these clones to be more productive during the DS than the other clones studied. In both seasons, the highest WUE was observed in the Modern Criollo clones, due to their lower g s.
Carbon isotope ratio as indicator of long-term WUE
In both seasons, δ 13C was significantly different in all clones in spite of trees growing under similar conditions. Also, significant differences in δ 13C between seasons were observed, that is, values of δ 13C were higher during the DS than the RS, suggesting that a higher long-term WUE was mostly due to stomatal closure (low g s).The effects of physiological and biochemical factors on gas exchange are integrated by δ 13C in prolonged time scales and often related to WUE (Evans and von Caemmerer, Reference Evans and von Caemmerer1996); higher δ 13C can be a consequence of either lower g s or higher A, both leading to a higher long-term WUE. Values of δ 13C were lower in Modern and Criollo clones and Trinitarios from Barlovento during the RS, falling in the same range of values for a Trinitario and a Forastero clone under irrigation (Daymond et al., Reference Daymond, Tricker and Hadley2011).
A poor relationship between δ 13C and instantaneous WUE and between δ 13C and A/g s was obtained when data of all clones were pooled. A positive linear relationship between δ 13C and instantaneous WUE has been observed in fast-growing crop species such as wheat (Farquhar and Richards, Reference Farquhar and Richards1984), barley (Hubick and Farquhar, Reference Hubick and Farquhar1989), cotton (Hubick and Gibson, 1993), beans (Monti et al., Reference Monti, Amaducci, Pritoni and Venturi2006), and sugar beet (Rajabi et al., Reference Rajabi, Griffiths, Ober, Kromdijk and Pidgeon2008), and between A/g s and instantaneous WUE in avocado (Acosta-Rangel et al., Reference Acosta-Rangel, Ávila-Lovera, De Guzman, Torres, Haro, Arpaia, Focht and Santiago2018). In contrast, values of WUE in grapevines at different time scales did not follow the trend in values of δ 13C (Medrano et al., Reference Medrano, Tomás, Martorell, Flexas, Hernández, Rosselló, Pou, Escalona and Bota2015) and, in Poa pratensis (Ebdon et al., Reference Ebdon, Petrovic and Dawson1998), environmental factors such as air temperature and ΔW caused a negative relationship between δ 13C and WUE. These results indicate that WUE as evaluation tool among genotypes or treatments is not always advisable a priori.
Leaf characteristics and chlorophyll a fluorescence
The SLA across all clones was on average 143 ± 9 cm2 g–1, within the range reported in Forastero and Criollo clones in Eastern Venezuela (Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016) and lower than those reported in eight Forastero clones (170–240 cm2 g–1; Daymond et al., Reference Daymond, Tricker and Hadley2011). The only clones that showed seasonal variation in SLA were Trinitarios (T1–T5), which also had the highest values, indicative of a more shaded provenance habitat. As a structural trait, SLA was less responsive to seasonal water availability than physiological traits such as A or δ 13C.
Leaf N content was significantly higher in the RS than in the DS. Leaf N content ranged between 0.93 and 2.74 g m−2, similar to values reported in eight Forastero clones (Daymond et al., Reference Daymond, Tricker and Hadley2011) and lower than those reported in Criollo clones (Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016). A positive correlation between A and area-based N content was found, as previously found in eight cocoa clones (Daymond et al., Reference Daymond, Tricker and Hadley2011), indicating that A could have been affected during drought by both photochemical and biochemical limitations, since all processes of photosynthesis involve N.
Values of total chlorophyll content in the clones studied were consistent with those previously reported for Forastero and Criollo clones in the field (De Almeida et al., Reference De Almeida, Herrera and Tezara2018; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016) but higher than previously reported for two Trinitario clones (Bertolde et al., Reference Bertolde, Almeida, Pirovani, Gomes, Ahnert, Baligar and Valle2012) and half the chlorophyll content in Forastero clones (Daymond et al., Reference Daymond, Tricker and Hadley2011). Drought considerably reduced chlorophyll content in Trinitario clones from Eastern Venezuela and Margarita and in Criollos, suggesting a lower capacity of absorption of light and electron transport in these clones during the DS.
Chlorophyll a fluorescence
The low values of J (<90 μmol e− m–2 s−1) in the present research are similar to those reported before in Criollo clones (Araque et al., Reference Araque, Jaimez, Tezara, Coronel, Urich and Espinoza2012; Ávila-Lovera et al., Reference Ávila-Lovera, Coronel, Jaimez, Urich, Pereyra, Araque, Chacón and Tezara2016; De Almeida et al., Reference De Almeida, Herrera and Tezara2018; Tezara et al., Reference Tezara, Urich, Jaimez, Coronel, Araque, Azócar and Chacón2016) and lower than those found in other shade-tolerant crops such as tea (Mohotti et al., Reference Mohotti, Dennett and Lawlor2000) and coffee (Martins et al., Reference Martins, Galmés, Cavatte, Pereira, Ventrella and DaMatta2014), in which J also saturates at low PPFD, suggesting that low A might have been due to low J. Reduction of J during drought is probably related to the lower chlorophyll content in this season, as was previously reported in two xerophytic species (Tezara et al., Reference Tezara, Marín, Rengifo, Martínez and Herrera2005). Values of NPQ increased on some Trinitario clones and Modern Criollos, indicating a down-regulation of PSII activity.
Coordination of traits across clones and seasons
The two primary axes of variation in the PCA explained 54.2% of total variation. The PCA showed a clear separation in physiological performance during the RS and drought season in the cocoa clones studied, showing that some clones of Modern Criollos, Barlovento– and Eastern–Trinitarios were less affected by drought. The PC1 was positively associated with traits related to water use (g s, E, ΔW) and photochemical activity (PPFD, J, NPQ) and negatively associated with WUE and Φ PSII, whereas the PC2 was positively associated with water status (ψ) and carbon economy (A, WUE) and negatively associated with δ 13C.
Conclusions
The differences in physiological traits among the clones studied under the same microclimatic, soil, and agronomic conditions suggest high genotypic variation in Venezuelan cocoa. Variability in response to water deficiency over a range of variables suggests differences in tolerance to drought.
The ψ was poorly correlated with changes in A and g s due to seasonal drought, whereas ΔW had a more marked influence on such changes, especially in Modern Criollos. Our results suggest that in cocoa, drought causes decreases in A due to stomatal closure as well as metabolic inhibition, given the observed reductions in N content, chlorophyll content, and J.
All the Modern Criollos tested, Criollo C2 and C3, and Trinitario T8 and T10, with their better taste, flavor, and economic value, show potential for better performance during the DS. Complementary yield data are needed to evaluate which of these would be suitable as cultivars. The Trinitario clones from Barlovento, with lower WUE, seem to be less sensitive to adverse ambient factors (air temperature, humidity) than to intrinsic factors, since chlorophyll, N, and TSP content, along with g s did not change with drought, indicating a conservative use of water and smaller capacity for carbon acquisition (lower A). During drought, Modern Criollos showed no reductions in either ψ, g s or A. These clones showed the highest WUE in both seasons, suggesting that they would be more productive in environments with low water availability.
We conclude that clones MC1, MC2, C2, C3, T8, and T10 showed optimization of water use and potential tolerance to drought; therefore, these clones might be good alternatives for cocoa breeding programs in drier environments. However, in order to select clones tolerant to DS, data of pod production should be obtained to improve selection of better clones.
Acknowledgments
This study was financed by Sub- Proyecto N° 5 Proyecto Ruta del Chocolate No. 200500898, CDCH PG 03-00-6874-2007, CDCH PG 03-7981-2011, and FONACIT PEI N° 2012000649. We thank Valentina Villalobos, Armando González, and Oranys Marín for their help in field measurements, and Jessica Rucks for isotope analyses at the Stable Isotope Laboratory of the University of Illinois at Chicago. Thanks to the INIA Miranda Estación Experimental Padrón for the use of their facilities.
Financial support
None.
