Drought Tolerance Assessment of Okra (Abelmoschus esculentus [L.] Moench) Accessions Based on Leaf Gas Exchange and Chlorophyll Fluorescence

Physiological and complementary phenotypic traits are essential in the selection of drought-adapted crop genotypes. Understanding the physiological response of diverse okra genotypes under drought stress conditions is critical to the selection of drought-tolerant accessions for production or breeding. The objective of this study was to assess the levels of drought tolerance in preliminarily selected okra accessions based on leaf gas exchange and chlorophyll fluorescence to determine best-performing genotypes for drought-tolerance breeding. Twenty-six genetically diverse okra accessions were screened under non-stressed (NS) and drought-stressed (DS) conditions under a controlled glasshouse environment using a 13 × 2 alpha lattice design in three replicates, in two growing seasons. Data were subjected to statistical analyses using various procedures. A significant genotype × water condition interaction effect was recorded for transpiration rate (T), net CO2 assimilation (A), intrinsic water use efficiency (WUEi), instantaneous water use efficiency (WUEins), minimum fluorescence (Fo′), maximum fluorescence (Fm′), maximum quantum efficiency of photosystem II photochemistry (Fv′/Fm′), the effective quantum efficiency of PSII photochemistry (ɸPSII), photochemical quenching (qP), nonphotochemical quenching (qN) and relative measure of electron transport to oxygen molecules (ETR/A). The results suggested variable drought tolerance of the studied okra accessions for selection. Seven principal components (PCs) contributing to 82% of the total variation for assessed physiological traits were identified under DS conditions. Leaf gas exchange parameters, T, A and WUEi, and chlorophyll fluorescence parameters such as the ɸPSII, Fv′/Fm′, qP, qN, ETR and ETR/A had high loading scores and correlated with WUEi, the ɸPSII, qP and ETR under DS conditions. The study found that optimal gas exchange and photoprotection enhance drought adaptation in the assessed okra genotypes and tested water regimes. Using the physiological variables, the study identified drought-tolerant accessions, namely LS05, LS06, LS07 and LS08 based on high A, T, Fm′, Fv′/Fm′ and ETR, and LS10, LS11, LS18 and LS23 based on high AES, Ci, Ci/Ca, WUEi, WUEins, ɸPSII and AES. The selected genotypes are high-yielding (≥5 g/plant) under drought stress conditions and will complement phenotypic data and guide breeding for water-limited agro-ecologies.


Introduction
Okra (Abelmoschus esculentus [L.] Moench), belonging to the Malvaceae family, is an important crop mainly cultivated as fruits, vegetables and seed oil. It is extensively grown in tropic and subtropic regions [1] and arid and semi-arid regions with limited and erratic Life 2023, 13,682 rainfall conditions [2]. The tender and immature pods of okra are consumed as cooked vegetables [3]. The pods are rich in protein content (25 %) and amino acids, notably lysine and tryptophan [4], fat, fibre, vitamins (A, C and K), vital mineral elements such as calcium, potassium, sodium, magnesium, iron, zinc and manganese [5], and soluble sugars such as sucrose (110.4 g/100 g FW), fructose (34.8 g/100 g FW and glucose (30.9 g/100 g FW [6]. In addition, minor quantities of organic acids, including citric, oxalic and malic acid, are present in the succulent pods [6]. The mature and dry seeds are a vital source of edible oils. The seed oil content ranges from 20-40%, consisting of the following major fatty acids: linoleic, palmitic, oleic, diacylglycerols and triacylglycerols acids [7]. Continental Asia accounts for a total annual okra production of 6 million tons from 592,375 million hectares of cultivated land, whereas Africa is the second major producer, with 3 million tons per annum from approximately 1.9 million ha of cultivated land [8]. Commercial and small-scale farmers produce okra. In sub-Saharan Africa (SSA), the crop is mainly grown in marginal conditions characterised by low and erratic rainfall, with minimal agricultural inputs and production technologies. In SSA, okra is mainly cultivated under rainfed conditions, and these agro-ecologies face moderate to severe droughts during the growing season [9]. Drought stress significantly reduces growth, biomass and yield [10]. Drought alone accounts for yield losses ranging between 30 and 100% in okra, primarily when the stress occurs during the flowering and pod-filling stages [3]. Breeding okra cultivars with drought adaptation is the major objective in improvement programs. Physiological and complementary phenotypic traits are critical in the selection of drought-adapted crop genotypes.
Phenotyping of plants using gas exchange and chlorophyll fluorescence traits has been reported as a preferred approach for selecting drought-tolerant okra accessions [11]. Some gaseous exchange traits used to assess drought tolerance include photosynthesis rate, stomatal conductance, chlorophyll content and transpiration rate. Further, chlorophyll fluorescence parameters (e.g., minimum fluorescence, maximum fluorescence, effective quantum efficiency of PSII photochemistry, photochemical quenching and non-photochemical quenching) have been used in phenotyping for drought tolerance [11][12][13]. Drought stress affects okra growth and productivity, disrupting physiological functions and the photosynthetic rate, resulting in yield losses [11,13,14]. Mkhabela et al. [3] reported that okra yield loss under drought stress could be significantly minimised by breeding drought-tolerant ideotypes with intrinsic water use efficiency. Hence, understanding the physiological response of diverse okra genotypes under drought stress conditions is essential for the selection of drought-tolerant accessions for production or breeding.
There has been limited progress in the breeding of okra for drought tolerance. This could be due to limited accessions identified as good drought and heat tolerance sources and insect pests and disease resistance [15]. Some unique accessions, including Sabz Pari [16], NHAe 47-4 [17], Pusa Sawari, Iraq P, Hala [1] and Xianzhi [18], were identified as useful sources of genes for enhancing drought tolerance under water-limited conditions. Compared to the highest genetic diversity reported in the cultivated okra [19], the identified accessions with tolerance to drought are relatively few. Therefore, there is a need for concerted research and development in okra to develop market-led and improved varieties for water-limited conditions.
In South Africa, okra is an important but under-researched and under-utilised crop. It is grown under rainfed conditions using local and unimproved accessions with poor adaptation and low yield potential. Genetically unique okra accessions could be sourced from different geographical regions to enhance okra pre-breeding programs [3]. Morphological traits associated with drought tolerance in okra include the number of pods per plant, fresh pod length, number of seeds per pod, hundred seed weight, number of branches per plant, plant height and total pod production [1,19]. Reportedly, a higher number of branches, pod length and number of pods per plant, plant height between 150 and 170 cm and pod weight have a direct influence on pod yield [19]. Drought tolerance assessment of okra accessions using the combination of morphological and physiological traits could increase the efficiency of identifying and selecting drought-tolerant accessions for cultivar development under dry environments. Therefore, this study aimed to assess the levels of drought tolerance in preliminarily selected okra accessions based on leaf gas exchange and chlorophyll fluorescence to determine best-performing genotypes for drought-tolerance breeding.

Plant Materials and Study Site
Twenty-five genetically distinct okra accessions were used for the study. The accessions were sourced from the Agricultural Research Council, Vegetable, Industrial and Medicinal Plants (ARC-VIMP) gene bank, and one local variety was included. The accessions were previously studied for their morphological responses to drought stress under field and glasshouse environments [3]. Detailed information on their geographical origin and drought resistance index are presented in Table 1. The experiment was conducted under glasshouse conditions at the Controlled Environment Facility (CEF) of the University of KwaZulu-Natal during the 2020/2021 growing seasons. The first experiment was conducted from September 2020 to December 2020, and the second from February 2021 to May 2021. The accessions were evaluated under non-stressed (NS) and drought-stressed (DS) conditions in the glasshouse environment. Drought tolerance index was calculated as DTI = (Ys/Yn)/(Ms/Mn), where Ys and Yn are the genotype yields under stress and non-stress, and Ms and Mn are the mean yields of the accessions under stressed and non-stressed conditions, respectively [20].

Experimental Design and Crop Establishment
Five seeds were initially planted in 5 L capacity plastic pots filled with composted pine bark growing media. Later, two plants were established per pot for each genotype. The day and night temperatures in the greenhouse (GH) were 30 • C and 20 • C, respectively, and the relative humidity ranged between 45 and 55% during the study. Inorganic fertilizers consisting of nitrogen (N), phosphorus (P) and potassium (K) were applied at a rate of 120, 30 and 30 kg ha −1 , based on soil fertility recommendations using urea (46-0-0), phosphorus pentoxide (P 2 O 5 ) and potassium oxide (P 2 O), respectively.
The trials were established using a 13 × 2 alpha lattice design under drought-stressed and non-stressed conditions with three replications. Drought stress was imposed at 50% flowering until physiological maturity by withholding irrigation until the soil water content reached 30% field capacity for plants under DS. The duration of stress was seven days before sampling. Plants under NS conditions were irrigated regularly to maintain soil moisture content at field capacity until physiological maturity. To determine pod yield, plants reached maturity, and pods were harvested sequentially at the soft, most digestible and immature stage. Tensiometers, moisture monitors (Spectrum Technologies, Inc, Chicago, IL, USA), were used to detect soil moisture levels at the root zone. Agronomic performance of the test genotypes was reported in Mkhabela et al. [19].

Data Collection
Gas exchange and chlorophyll fluorescence parameters were measured using an LI-6400 XT Portable Photosynthesis system (Licor Bioscience, Inc. Lincoln, NE, USA) integrated with an infrared gas analyser (IGRA) attached to a leaf chamber fluorometer (LCF) (640040B, 2 cm 2 leaf area, Licor Bioscience, Inc, Lincoln, NE, USA). External leaf CO 2 concentration (C a ) and artificial saturating photosynthetic active radiation (PAR) were set at 400 µmol mol −1 and 1000 µmol m −2 s −1 , respectively. Water flow rate and relative humidity were maintained at 500 µmol and 43%, respectively. The leaf-to-air vapour pressure deficit in the cuvette was maintained at 1.7 kPa to avoid stomatal closure due to low air humidity. Gas exchange and chlorophyll fluorescence measurements were taken on the third half fully formed leaf inside the sensor head. Under both NS and DS conditions, measurements were taken from five plants of each accession.
The following gas exchange parameters were determined: stomatal conductance (gs), net CO 2 assimilation rate (A), transpiration rate (T), intercellular CO 2 concentration (C i ) and the ratio of intercellular and ambient CO 2 (C i /C a ) concentrations. The ratio of net CO 2 assimilation rate to intercellular CO 2 concentration (A/C i ) was computed according to Kitao et al. [21]. The ratio of A and gs was used to compute intrinsic water use efficiency [22] and the ratio of A and T was used to calculate instantaneous water use efficiency) [23].
To estimate chlorophyll fluorescence variables, a saturation flash intensity of 1300 µmol m −2 s −1 was applied. The following parameters were recorded. The minimum (Fo ) and maximum fluorescence (F m ) of light-adapted leaves under natural glasshouse conditions. The steady-state fluorescence (F s ) was also determined in light-adapted photosynthesis. Equation (1) was used to de-termine the variable fluorescence in light-adapted leaves, while Equation (2) calculated fluorescence changes [24].
Additional chlorophyll fluorescence parameters were estimated according to Evans [25], Fv /Fm , the maximum quantum efficiency of photosystem II photochemistry, the effective quantum efficiency of photosystem II photochemistry (φPSII), photochemical quenching (qP), non-photochemical quenching (qN) and electron transport rate (ETR). The ratio of ETR and A was used to calculate a relative measure of electron transport to oxygen molecules. The alternative electron sink (AES) was calculated as the ratio of photosystem II effective quantum efficiency to net CO 2 assimilation (A) [26]. Chlorophyll fluorescence was measured using a pulse-amplitude modulated (PAM) fluorometer, which applies a short pulse of light to the sample and measures the resulting fluorescence emitted by the chlorophyll. This measurement provided information on the photosynthetic efficiency and health of the crop. Gas exchange and chlorophyll fluorescence parameters were measured on fully expanded leaves. At the end of the second experiment, yield per plant (YPP) was determined by harvesting fresh pods when 50% of the pods were 3-5 cm long by hand every third day.

Statistical Analysis
Data were subjected to analysis of variance using Genstat 20th edition (VSN International, Hempstead, UK). The mean data for the two seasons were combined for analysis. Means were separated using Fisher's protected least significant difference (LSD) test at the 5% significance level. Pearson's correlation coefficients were calculated using IBM SPSS Statistics 25.0 (SPSS Inc., Chicago, IL, USA) to determine the magnitude of the relationship among physiological traits. Principal component analysis (PCA) based on a correlation matrix was used to identify influential traits under NS and DS conditions using R Studio version 4.0, ggplot2 (R Core Team, 2018). Biplots were built using XLSTAT to determine relationships among the accessions and response variables (physiological traits). Principal component biplot diagrams were used to identify drought-tolerant and drought-susceptible okra accessions using XLSTAT. ClustVis (https://biit.cs.ut.ee/clustvis_large (accessed on 23 November 2022)) was used to visualise the heatmap analysis of physiological traits.
Non-significant (p > 0.05) differences were observed among accessions under NS and DS conditions for Ci. Okra genotypes LS02 and LS21 exhibited high A/C i values of 0.23 and 0.28 µmol. mol −1 , respectively, under DS conditions compared to other accessions. Significant (p < 0.05) differences were observed in C i/ C a values among accessions under both NS and DS conditions. Intrinsic water use efficiency and instantaneous water use efficiency were increased by drought stress (Table 4). Accessions LS13 and LS20 had the highest WUE i under drought-stress conditions, with 1438.80 and 1256.10 µmol CO 2 m −2 , respectively. The highest WUE ins values under drought stress were recorded for accessions LS04 (2164.70 µmol·mol −1 ) and LS22 (2161.00 µmolmol −1 ).
The effect of drought stress on chlorophyll fluorescence parameters among the tested okra accessions are highlighted in Table 2. Chlorophyll fluorescence parameters indicated significant differences for genotype, water regime and genotype x water regime interaction, showing that the evaluated genotypes responded differently under non-stress and droughtstress conditions. Non-significant differences were observed for Fo under non-stress, while significant (p < 0.001) differences were recorded under drought-stress conditions (Tables 3 and 4). Genotypic variability (p < 0.001) with respect to F m was observed under non-stress and droughtstress conditions. Drought stress decreased Fv /Fm , from 0.51 under non-stressed to 0.35 under drought-stressed conditions. The φPSII varied significantly among the tested genotypes under non-stress and drought-stress conditions. LS07, LS12 and LS19 revealed considerably higher values for φPSII ≥ 0.40 compared to other genotypes under non-stress conditions. Photochemical quenching was significantly reduced from 0.32 to 0.13 by drought stress among the evaluated genotypes, of which LS04, LS12 and LS13 had the highest values of qP > 0.40. A variable genotypic response was observed with respect to qN under non-stress and drought-stress conditions. The mean for qN was higher under drought-stress (1.96) than non-stress conditions (1.39). The qN values ranged from 0.68 to 2.80 under non-stress (Table 3) and from 0.66 to 3.75 under drought-stress conditions (Table 4). LS02, LS03 and LS11 revealed qN values ≥ 2 under non-stress conditions. Genotypes LS01, LS02, LS10, LS11 and LS18 showed qN values ≥ 3 under drought-stress conditions. Non-significant differences were observed for ETR under non-stress conditions, while genotypic variation was observed for ETR under drought-stress conditions. LS08, LS09 and LS17 revealed the highest ETR value of ≥34,541 µmol e −1 m −1 s −1 , whereas LS16, LS22 and LS26 showed the lowest ETR ≤ 8071 under DS conditions. Drought stress significantly increased ETR/A ( Table 4). The highest ETR/A (≥1542 µmol e µmol-1 CO 2 ) was recorded from LS03, LS08, LS09 and LS17 under drought-stress conditions. Drought stress significantly increased AES (154.72) compared to NS (26.98). AES ranged from 12.77 to 61.12 under non-stress and from 64.90 to 562.80 under drought-stress conditions. Yield per plant was significantly reduced, from 7.20 g/plant to 4.31 g/plant, by drought stress among the evaluated genotypes. Accessions LS11, LS19, LS21, LS22 and LS24 had the highest yield (>9 g/plant) under NS conditions, whereas LS05, LS06, LS07, LS08, LS10, LS11, LS18 and LS23 exhibited the highest yield (>5 g/plant) under DS conditions.

Principal Component Analysis (PCA) for Leaf Gas Exchange and Chlorophyll Fluorescence Traits
Values of PCA, eigenvalues, percent, and cumulative explained variances are summarised in Table 5. Under NS conditions, seven principal components exhibited eigenvalues > 1 and accounted for 81% of total phenotypic variation. Net CO 2 assimilation, C i , C i /C a , qP, ETR, ETR /A, AES and YPP were positively correlated with PC1, which accounted for 22% of the total variation. PC2 was positively correlated with gs, Fo and φPSII, whereas WUE i and WU Eins were negatively correlated with PC2, which accounted for 17% of the total variation. Transpiration rate was negatively correlated with PC3, whereas WUE ins , qP and YPP were positively correlated with PC3, which contributed 11.42% of total variation. A/Ci positively correlated with PC4 accounted for 10.67% of total variation. PC5 was positively correlated with Fm and F v /F m , contributing 8% of total variation, whereas PC6 was positively correlated with Fm , contributing 7% of total variation.
Similarly, under DS conditions, seven PCs with eigenvalues > 1 were detected, which contributed 80% of the total phenotypic variability. Yield per plant was negatively correlated with PC1, whereas φPSII, ETR and ETR/A were positively correlated with PC1, which accounted for 20% of total variation. Transpiration rate, net CO 2 assimilation, C i , WUE i , F v /F m , qN and YPP were positively associated with PC2, accounting for 18% of the total variation. Stomatal conductance and A/Ci were positively correlated with PC3, whereas Ci and WUE ins negatively associated with PC3 contributed 14% of the total variation. Net CO 2 assimilation and qN were positively correlated with PC4, whereas ETR/A was negatively correlated with PC4, accounting for 10% of total variation. Instantaneous water use efficiency was positively correlated with PC5, which accounted for 7% of total variation, whereas stomatal conductance and photochemical quenching were positively correlated with PC6, which contributed 6% of total variation.
Principal component biplots based on PCA analysis were used to indicate the relationships among okra accessions for leaf gas exchange and chlorophyll fluorescence parameters under NS ( Figure 1A) and DS ( Figure 2B) conditions. Traits presented by parallel vectors or those close to each other revealed a strong positive association, and those located nearly opposite (at 180 • ) showed a highly negative association, while the vectors toward sides expressed a weak relationship. Under NS conditions, accessions LS06, LS11, LS22, LS05 and LS20 were grouped based on high qN. Accessions LS19, LS17 and LS18 were grouped together based on high gs, T and A/Ci. LS02, LS10 and LS24 were grouped based on high φPSII, Fo , F v /F m , A, ETR, ETR/A, AES and qP. Accessions LS25, LS01, LS23 and LS16 were grouped together based on high C i /C a , WUE i and WUE ins . Under DS conditions, accessions LS10, LS24, LS25, LS05 and LS06 were clustered together based on high Fm , AES, T, C i and YPP. LS13, LS15, LS17 and LS09 were grouped together based on high C i /Ca,

Heatmap Analysis for Leaf Gas Exchange and Chlorophyll Fluorescence Traits
A heatmap based on leaf gas exchange and chlorophyll fluorescence traits under NS and DS conditions was constructed using a hierarchical clustering method to discern the relationship of 26 okra accessions based on Jaccard's coefficient (Figure 2). Under NS (Figure 2A) conditions, physiological traits were grouped into four main clusters. The first cluster consists of two subclusters, dominated by eight accessions, including LS19, LS12, LS06, LS18, LS13, LS07 and LS02, which were grouped based on high negative correlations with WUEins, qN and YPP. The second subcluster consisted of accessions LS22, LS11, LS1, LS08, LS20 and LS14, which were negatively correlated with A/Ci and T. LS25, LS01, LS21, LS16, LS24 and LS23 dominated the fourth subcluster under NS conditions and positively

Heatmap Analysis for Leaf Gas Exchange and Chlorophyll Fluorescence Traits
A heatmap based on leaf gas exchange and chlorophyll fluorescence traits under NS and DS conditions was constructed using a hierarchical clustering method to discern the relationship of 26 okra accessions based on Jaccard's coefficient (Figure 2). Under NS (Figure 2A) conditions, physiological traits were grouped into four main clusters. The first cluster consists of two subclusters, dominated by eight accessions, including LS19, LS12, LS06, LS18, LS13, LS07 and LS02, which were grouped based on high negative correlations with WUEins, qN and YPP. The second subcluster consisted of accessions LS22, LS11, LS1, LS08, LS20 and LS14, which were negatively correlated with A/Ci and T. LS25, LS01, LS21, LS16, LS24 and LS23 dominated the fourth subcluster under NS conditions and positively

Heatmap Analysis for Leaf Gas Exchange and Chlorophyll Fluorescence Traits
A heatmap based on leaf gas exchange and chlorophyll fluorescence traits under NS and DS conditions was constructed using a hierarchical clustering method to discern the relationship of 26 okra accessions based on Jaccard's coefficient (Figure 2). Under NS (Figure 2A) conditions, physiological traits were grouped into four main clusters. The first cluster consists of two subclusters, dominated by eight accessions, including LS19, LS12, LS06, LS18, LS13, LS07 and LS02, which were grouped based on high negative correlations with WUEins, qN and YPP. The second subcluster consisted of accessions LS22, LS11, LS1, LS08, LS20 and LS14, which were negatively correlated with A/Ci and T. LS25, LS01, LS21, LS16, LS24 and LS23 dominated the fourth subcluster under NS conditions and positively correlated with qP. Under DS, physiological traits were grouped into three main clusters and six subclusters. The first cluster is dominated by accessions LS19, LS09, LS03, LS15, LS14 and LS17, based on their positive correlations with ETR and ETR/A. LS26, LS22, LS04, LS20, LS13 and LS11 dominated the second cluster under DS conditions, with positive correlations with WUEi, WUEins, qN and YPP. AES was positively correlated with LS25 and LS05 in the third cluster under DS conditions.

Discussion
Okra is one of the most important commercial vegetable crops grown for its fresh fruits and dry seeds. Drought is the major impediment to okra production in dry regions. To adapt to drought stress, plants have undergone many biochemical, molecular, and physiological changes. These changes increase the plants' tolerance to drought stress. Drought stress influences plant performance by reducing gas exchange and altering chlorophyll fluorescence formation. Gas exchange and chlorophyll fluorescence confer drought tolerance in okra [11,27]. Plants alter gene expression, disrupting the production of photosynthetic pigments and regulating stomatal function to adapt to and tolerate stress conditions [27]. Developing new strategies for maintaining high yield under drought-stress conditions is one of the major challenges in the current crop production system.
In this study, various physiological drought responses were assessed in okra accessions. Reductions in okra's stomatal conductance and transpiration rates have been associated with water conservation that allows plants to tolerate drought stress and the loss of physiological functions [9]. Stomatal closure leads to a reduction in CO 2 assimilation and minimises the rate of water loss through transpiration. This role of drought-induced stomatal closure limits CO 2 uptake by the leaves and possibly leads to increased susceptibility to photodamage [11]. Similar findings were reported for okra accessions under water shortages [11,27]. These physiological changes increase the plants' resistance to drought stress, enabling the crop to survive in environments with limited water availability.
Drought tolerance should be considered as a comprehensive evaluation of carbon assimilation during global climate change challenges [28]. In the current study, okra accessions exhibited a reduction in net CO 2 assimilation under drought-stressed conditions ( Table 4). The decrease in net CO 2 assimilation during water-stressed conditions might be reversible initially. However, drought in the pod-filling stage might cause irreversible damage to the photosynthetic pathway, thereby affecting carbon assimilation [29]. Further, utilisation of assimilates is relevant in addition to the photosynthetic performance of leaves. The evaluated okra accessions revealed high water use efficiency under drought-stressed conditions (Table 4). Enhancing water use efficiency to sustain okra production under water-limited conditions remains the most important task for water management. Hence, specific responses for enhancing water use efficiency could be achieved with more precise data on crop stress detection [11]. Drought-tolerant accessions exhibited high WUE i and WUE ins compared to drought-susceptible accessions ( Table 2). This indicates that the evaluated accessions use water efficiently, attributed to drought escape mechanisms such as the transpiration rate. Drought-tolerant accessions use water efficiently, maintain tissue water status, reduce water loss and produce stable yield during water shortages [30].
Chlorophyll fluorescence is a non-invasive measurement detecting the authenticity of photosystem II [31]. Chlorophyll fluorescence parameters, including photosystem II photochemistry, minimum fluorescence, maximum fluorescence, photochemical quenching and electron transport rate are useful for detecting drought-stress severity, genetic variation and determining damage to PSII [32]. F v /F m is considered the most important parameter of chlorophyll fluorescence, widely used to evaluate drought-stress response. In this study, a reduced F v /F m value was recorded under drought-stress conditions, corroborating with results reported by Ahmed and El-Sayed, [27]. According to Paknejad et al. [33], reduced F v /F m under drought-stress conditions indicates the presence of a protective mechanism of light absorption in response to water shortages. Hence, the F v /F m parameter can be applied to determine the potential efficiency of PSII.
In the present study, drought-tolerant okra accessions showed an efficient photosynthetic affinity compared to sensitive accessions. Photosystem II is highly drought tolerant. However, under drought-stress conditions, photosynthetic electron transport through PSII is inhibited [24]. The decrease in PSII might be due to the photo-protective increase in thermal energy dissipation induced by the excess of absorbed light [34]. However, there are contradictory reports on the direct effect of PSII functionality under drought-stress conditions. A study reported that, under mild water stress, PSII is not affected [35], while another study reported that, under drought-stress conditions, damage occurs to both photosystem I and photosystem II [36]. The current study found that PSII was significantly affected by drought stress. Under drought-stress conditions, the PSII thermal energy dissipation was strongly limited due to damage to PSII structure and functionality. A decrease in photochemical quenching was observed in the studied okra accessions under droughtstress conditions. Similar results were reported by Ashraf et al. [37] in the study of gas exchange characteristics and water relations in some elite okra cultivars under water-deficit conditions. The decrease in qP is attributable to either a decrease in the rate of consumption of reductants and ATP produced from non-cyclic electron transport relative to the rate of excitation of open PSII reaction centres or damage to PSII reaction centres [24].
Positive correlations were observed between non-photochemical quenching and intrinsic water use efficiency under drought-stress conditions, indicating a protective mechanism by the plants against reactive oxygen species that harm antenna pigments and closing reactions in the photosystem. Drought stress also affects the electron transport rate (ETR) and alternative electron sink (AES) [38]. An increase in alternative electron sink was observed among the studied okra accessions under drought-stress conditions. Drought-tolerant accessions indicated higher AES values. An increase in AES was reported as an indicator of drought stress [39]. Alternative electron sink is the second most important mechanism after photosynthesis used to remove electrons, which occurs at high rates in the leaves under drought stress conditions [40].

Conclusions
Drought is one of the most important factors affecting physiological traits and yield in crop plants, including okra. In the present study, it was observed that drought stress affected physiological processes such as reduced stomatal conductance, transpiration rate, net carbon dioxide assimilation, maximum quantum efficiency, effective quantum efficiency of PSII photochemistry, photochemical quenching and electron transport rate among the studied okra accessions. These physiological traits could be useful for drought-tolerance breeding in okra. Principal component analysis-based biplots allowed the identification of drought-tolerant accessions such as LS05, LS06, LS07 and LS08 based on high A, T, Fm , Fv /Fm and ETR, and LS10, LS11, LS18 and LS23 based on high AES, Ci, Ci/Ca,WUEi, WUEins, φPSII and AES. The selected genotypes are high yielding (≥5 g/plant) under drought-stress conditions. These accessions are suitable candidates for parental genotypes for drought-tolerance breeding in okra to enhance water use efficiency under waterlimited conditions.