Plant material.

A collection of 37 clonally propagated blueberry genotypes, including southern highbush blueberry (SHB; V. corymbosum L. interspecific hybrids), northern highbush blueberry (NHB; V. corymbosum L.), and rabbiteye blueberry (RBE; V. ashei Reade), were used to investigate stomata morphological traits. Genotypes sampled included publicly released cultivars, proprietary cultivars, and breeding selections growing in seven different locations in North America, including Auburn, AL, USA (lat. 32°35′49.7″N, long. 85°29′17.3″W), Tallassee, FL, USA (lat. 32°49′54.0″N, long. 85°53′30.4″W), Gainesville, FL, USA (lat. 29°38′15.696″N, long. 82°21′51.026″W), Citra, FL, USA, (lat. 29°24′54.652″N, long. 82°8′42.986″W), Venus, FL, USA (lat. 27°10.1″N, long. 81°32′22.6″W), Mt. Horeb, WI, USA (lat. 42°57′23.0″N, long. 89°39′41.5″W), and Tangancicuaro, Michoacan, Mexico (lat. 19°51′55.1″N, long. 102°11′44.7″W). Plants were grown either in fields or in greenhouses and managed according to commercial practices (Table 1).

Table 1. List of genotypes, type, and sampling location of plants studied during this experiment.

View Fullscreen

Stomata density and area.

Stomata morphological traits of five plants per genotype were evaluated. Two fully developed leaves from the upper canopy and two fully developed leaves from the center of the canopy were collected from each plant. The S_d and S_a were determined using the imprint method (Rogiers et al. 2011). Stomata imprints were made using a modified version of the protocol described by Das and Santakumari (1977). Briefly, a thin coat of clear nail polish (Color 103; Sally Hansen, NY, USA) was applied on both the abaxial and adaxial sides of each leaf. After drying, the nail polish film, which had an imprint of the leaf epidermis, was removed and mounted on a microscope slide with clear adhesive tape. All imprints were made in the center of the leaf lamina while avoiding the leaf margin and midvein. Imprints were digitalized using light microscopes [Micromaster II (Fisher Scientific, Newington, NH, USA) or B490B (Amscope, Irvine, CA, USA)] and a mobile phone camera (iPhone 7, 28-mm 12-megapixel camera with optical image stabilization; Apple, Cupertino, CA, USA). The S_d was determined by counting the number of stomata present in 1 mm² of the imprint at 40× magnification using the multipoint counter of ImageJ (version 1.53r) (Rueden et al. 2017). Stomata length and width were manually measured in 30 imprints per genotype at 100× magnification. A stage micrometer was used to calibrate images. The S_a (µm²) was modeled as an ellipsoid and calculated using measurements from three stomata per slide using the following equation: S_a = π*A*B, where A is length of major radius and B is length of minor radius. Leaf porosity (P_l) was calculated as P_l = S_a*S_d. This variable represents the allometric relationship between the stomata area and density in 1 mm² of leaf lamina.

Stomata speed.

Plants of SHB genotypes ‘Arcadia’, ‘Colossus’, ‘Meadowlark’, and ‘Optimus’ (n = 6 per genotype) were planted into 11.35-L pots filled with a substrate mixture of 3:2:1 (v/v) of peat, pine bark, and perlite, respectively. These genotypes were selected for further testing based on the results from the stomata morphology survey. Plants were watered as needed for the length of the experiment and fertilized twice per week. The fertilizer solution was prepared with 100 mg/L of a soluble fertilizer (21N–3.05P–5.81K; Peters Professional Acid Special; ICL Specialty Fertilizers, Summerville, SC, USA) with 11.5% and 9.5% of the N provided as ammonium and urea, respectively. Plants were pruned and allowed to grow in a temperature-controlled greenhouse located in Gainesville, FL, USA, for a duration of 10 to 12 weeks. Stomata morphology of the same leaves used for gas exchange was determined as aforementioned. Leaf gas exchange was recorded with an infrared gas analyzer (model CIRAS-4; PP-systems, Amesbury, MA, USA) equipped with a leaf cuvette and an integrated light-emitting diode light unit. Air flow inside the cuvette during measurements was 300 μmol·min⁻¹. The cuvette CO₂ concentration was 400 μmol·mol⁻¹. Leaf temperature was maintained at a constant 25 °C, while the relative humidity level was controlled at 60%. Plants were subjected to dark acclimation during the night. Gas exchange measurements commenced at sunrise (between 6:30 AM and 6:50 AM in Nov and Dec 2023) following a custom-built script, with leaves kept in the dark for 15 min before being exposed to an instantaneous increase in irradiance (800 μmol·m⁻²·s⁻¹) that lasted 150 min. Temporal responses of g_s of one young fully expanded leaf per plant were measured. Data were collected every 20 s after stabilization of conditions inside the chamber. Gas exchange measurements were collected inside the greenhouse.

Response curves were fitted using the asymmetric sigmoidal model described by Vialet-Chabrand et al. (2013) as implemented by Gerardin et al. (2018). The asymmetric model was formulated according to a standard Gompertz function with the following equation:$$g_s\hspace{0.17em}=g_{\mathit{min}}\hspace{0.17em}+\hspace{0.17em}\left(g_{\mathit{max}}-g_{\mathit{min}}\right)\text{* }{e}^{-e^{\left(\text{l}-\frac{t}{\text{t}}\right)}}$$ [1] where g_s is the fitted stomatal conductance, g_min is the minimal starting value of stomatal conductance, g_max is the maximum stomatal conductance, λ is the lag time of the stomatal response (the time needed to reach the inflection point of the curve from the moment of the irradiance change from 0 to 800 μmol·m⁻²·s⁻¹), and τ is the response time from the first response to change in irradiance until g_max is reached. The total response time (T) was calculated by summing τ and λ. The stomatal speed response (SL_max) was calculated as follows:$$S{L}_{\mathit{max}}\hspace{0.17em}=\hspace{0.17em}\left(\frac{1}{\text{t}}\right)\text{*}\frac{(g_{\mathit{max}}-g_{\mathit{min}})}{e}$$ [2] where (g_max – g_min) represents the amplitude of the stomatal response and e represents the Euler constant. This model was fitted in R (version RStudio 2021.09.2) using the NLIMB function with a large uniform grid of possible random starting values for τ and λ. The range of values selected was between 1 and 100 for both parameters, varying incrementally by 1. They were optimized through a grid search until stability and convergence were achieved.

Experimental design and data analysis.

Experiments were performed as a completely randomized design, with each plant considered a replication. For S_d and S_a, leaves from the same plant were considered pseudo-replications. Data were analyzed via an analysis of variance using the agricolae package (Mendiburu 2021) in R (version RStudio 2021.09.2). Mean separation was performed using Fisher’s protected least significant differences. Unpaired t tests were used to compare stomata density in leaves from the upper and central parts of the canopy. The Pearson correlation was calculated between stomata traits and model parameters in R. P values were considered significant at α ≤ 0.05.

The importance of genotype, leaf position, and species type and the estimate of their effect on leaf density and area were analyzed by fitting a linear mixed model. All model components were set as random terms except the intercept. The model was fitted using the Asreml R package (Butler et al. 2023) as follows:$$y\hspace{0.17em}=\hspace{0.17em}\mu \mathrm{\hspace{0.17em}+}{Z}_{1}g\hspace{0.17em}+Z_{2}p\hspace{0.17em}+Z_{3}t\hspace{0.17em}\hspace{0.17em}+\hspace{0.17em}e$$where y is the vector of the response variable, μ is the overall mean, g is the vector of the genotype effect, p is the vector of the leaf position effect, t is the vector of the species type effect, and e is the random error. The incidence matrices are denoted as $Z_1$, $Z_2$, and $Z_3$. All random effect terms are assumed to follow a multivariate normal distribution. Variance components were estimated using the restricted maximum likelihood. The significance of random terms was assessed using the likelihood ratio test, which compares the full model to a reduced model excluding the term being tested. Data visualization and graphing were performed using the ggplot2 package (Wickham 2011) in R.

Stomata density and area.

Our survey confirmed that Vaccinium spp. are hypostomatous plants because stomata were not observed on the adaxial side of leaves of all genotypes evaluated (1110 imprints). Blueberry stomata were composed by two symmetric dumbbell-shaped guard cells. Subsidiary cells were not discernible using light microscopy.

The S_d and S_a varied considerably among different blueberry genotypes (Fig. 1), and the genotype effect accounted for the highest variance explained compared with blueberry type and position of the leaves (Table 2). There was an approximately three-fold difference in S_a between the genotype with the smallest (‘Optimus’) and largest (‘Northland’) stomata. Overall, NHB genotypes exhibited more stomata than SHB and RBE genotypes studied in this experiment (Fig. 1A). There was an approximately four-fold difference between the genotype with lowest (FL16-194) and highest (‘Brightwell’) S_d. Additionally, variations in stomata traits were observed among genotypes of the same blueberry type (Supplemental Figure 1). In other species, increases in S_d co-occur with smaller stomata (Carlson et al. 2016; Franks and Farquhar 2001; Hetherington and Woodward, 2003). This allometric relationship was not observed in highbush blueberries during our survey (Fig. 1C). Correlations between S_a and S_d were not significant among the genotypes tested (P = 0.18).

View Figure

• Download as Powerpoint
• Download Figure

Table 2. Estimate of variance component for genotype, leaf position, and species type for leaf stomata density and leaf area in blueberry.

View Fullscreen

The S_a has been used to predict ploidy levels of many plant species, including blueberry (Chavez and Lyrene 2009), wheat (Wang et al. 1989), alfalfa (Bingham 1968), coffee (Mishra 1997), Crataegus spp. (McGoey et al. 2014), and Bromus inermis (Teare et al. 1971). Blueberries can be found at different ploidy levels (Hummer et al. 2015; Lyrene 2006; Sakhanokho et al. 2018). We showed that S_a is not an ideal tool for assessing blueberry ploidy because the presumed ploidy level did not correlate with the stomata area (r = 0.008; P = 0.87).

Light intensity and quality can influence stomata traits. Kim et al. (2011) reported fewer and larger stomata in leaves of NHB ‘Bluecrop’ plants as shade levels increased. We considered leaf position in the canopy as a proxy for sun exposure during leaf development. In the present study, this factor affected S_d only in a subset of the genotypes tested. ‘Sentinel’, ‘Sweetcrisp’, ‘Meadowlark’, ‘Victoria’, ‘Elliott’, ‘Duke’, ‘Patriot’, ‘Bueray’, ‘Northland’, ‘Bluegold’, ‘Climax’, ‘FL18-188’ and ‘Brightwell’ had significantly higher (P < 0.05) S_d on leaves collected from the top of the canopy (full sun exposure) compared with that on leaves collected from the center of the canopy (Fig. 2). These differences are likely the result of differences in light exposure because leaves in the upper canopy receive higher light intensities, which are known to influence stomatal development (Idris et al. 2018). All genotypes sampled in Wisconsin exhibited differences in S_d on leaves collected from the top of plant canopy compared with that on leaves from the middle to lower canopy. These results could be related to pruning practices or environmental conditions in the northern United States. Pruning varies by region based on climate, blueberry genotype, and farming practices (Fang et al. 2020; Kovaleski et al. 2015; Retamales and Hancock 2018).

View Figure

• Download as Powerpoint
• Download Figure

Stomata responses to irradiance.

The SHB genotypes grown under common environmental conditions were used to confirm stomata survey results and study the effects of stomatal traits on stomatal responses (Fig. 3). Genotypes ‘Colossus’ and ‘Optimus’ exhibited lower stomata area (P < 0.05) and higher stomatal density (P < 0.01) than those of ‘Meadowlark’ and ‘Arcadia’. In this subset of genotypes, cultivars ‘Arcadia’ and ‘Colossus’ maintained their stomata characteristics between the survey and follow-up experiment, but ‘Optimus’ and ‘Meadowlark’ did not. Dynamic models with acceptable fit metrics (g_min, g_max, λ, $\text{t}$, and SL_max) were fit for most of the studied leaves (n = 23). Model parameters from each genotype were compared (Supplemental Table 1). The stepwise increase in irradiance led to stomatal responses in all genotypes (Fig. 4). ‘Colossus’ exhibited the fastest stomata response to changing irradiance, which was significantly different (P = 0.001) from that of ‘Meadowlark’. The amplitude of the stomatal response ranged from 19.48 μmol·m⁻²·s⁻¹ to 164.47 μmol·m⁻²·s⁻¹ across genotypes. Notably, λ and τ exhibited different relations with stomatal morphology. Additionally, λ differed significantly (P < 0.01) among the genotypes, such that the inflection point of the curve from the moment of irradiance changed rapidly for all genotypes except ‘Optimus’. Stalfelt (1927) introduced the term “Spannumgsphase” to describe a phenomenon characterized by anticipatory reactions in the stomata that are not accompanied by movement. While it has been suggested that the speed of signal transduction is unlikely to cause significant delays in stomatal opening or closing relative to the mechanisms of movement (Franks and Farquhar 2007), our results indicate that λ can significantly affect the total time from the moment of irradiance change until g_max is achieved (P < 0.005). The larger λ of ‘Optimus’ suggests a prolonged Spannugsphase, which may delay the osmotic adjustments required for guard cell turgor changes. This delay could be caused by limitations in the rate of osmolarity generation, such as K⁺ influx and the production of organic solutes (Talbott and Zeiger 1996). However, our data suggest that neither S_a nor S_d affects λ.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

No correlation was observed between S_a and SL_max (r = 0.02; P = 0.91), suggesting that the speed of the stomatal response is not directly influenced by S_a in SHB (Table 3). These results are in agreement with those of previous research of a large spectrum of species, including ferns, cycads, conifers, and angiosperms (Elliott-Kingston et al. 2016). However, they differ from those of previous reports that suggest that a smaller S_a leads to a faster stomatal response caused by a higher surface-to-volume ratio and, consequently, the lesser solute transport needed to drive stomatal movements (Drake et al. 2013; Lawson and Blatt 2014). It is possible that parameters other than stomatal morphology, such as variations in ion and water transport within guard cells (Kübarsepp et al. 2020; Lawson and Blatt 2014), impacted the speed of the stomatal response in the surveyed SHB genotypes.

Table 3. Correlation table of model parameters and the stomata morphological traits. Correlations and P values are provided in parentheses between stomata density (S_d), stomata area (S_a), porosity (P_l), minimal stomatal conductance (g_min), lag time (λ), response time to changes in irradiance (t), maximum stomatal conductance (g_max), SL_max, total response time to changes in irradiance (T), and amplitude of the stomatal response (SA). Data are from four southern highbush blueberry genotypes ‘Arcadia’, ‘Colossus’, ‘Meadowlark’, and ‘Optimus’ grown under the same conditions in a greenhouse.

View Fullscreen

A significant positive correlation was found between S_d and SL_max (P = 0.04), associating more stomata with faster responses. Similar results were found in A. thaliana (Vialet-Chabrand et al. 2016). Morphological traits related to the regulation of water movement within and out of the leaf such as higher vein density (Westbrook and McAdam 2021) and a more favorable boundary layer (Defraeye et al. 2013) have been related to this response. Morphological phenes that create a more efficient hydraulic network in the leaf–atmosphere continuum might allow speedy turgor pressure adjustments in the guard cells, facilitating stomata opening.

In this study, g_max was not different among the tested genotypes (P = 0.18), despite their differences in S_d. While in other species an increase in S_d can lead to higher g_s (Maruyama and Tajima 1990; Ohsumi et al. 2007; Schlüter et al. 2003), our results indicate no correlation between these variables (P = 0.06). This is consistent with the findings in the literature (Jones 1977; Kawamitsu et al. 1996). Additionally, we observed no correlation between g_s and S_a (P = 0.41). Therefore, our hypothesis was not supported by the data.

In this study, g_min differed significantly among the tested genotypes (P < 0.001). The g_min, which is often referred as to g_cuticular or g_residual, reflects the conductance measured at maximal stomatal closure and is not directly regulated by guard cells (Caird et al. 2007). Previous studies have reported g_min estimates ranging from 0.004 to 0.020 μmol·m⁻²·s⁻¹ for genotypes of wheat, grape, helianthus, and European trees and shrubs (Boyer et al. 1997; Burghardt and Riederer 2003; Howard and Donovan 2007; Kerstiens 1995; Rawson and Clarke 1988). Blueberry g_min ranged from 14 to 52 µmol·m⁻² ·s⁻¹ at predawn. Genotypes ‘Arcadia’ and ‘Meadowlark’ had almost two-fold higher g_min values compared with those of ‘Optimus’ and ‘Colossus’ (Fig. 3). This could be a consequence of incomplete stomatal closure at night. However, it also may be influenced by factors such as the presence of dust or other materials preventing complete stomatal closure (Caird et al. 2007). Alternatively, differences in cuticle thickness, composition, and development among leaves could have caused the high g_min (Bi et al. 2017; Buschhaus and Jetter 2011; Pollard et al. 2008; Qiao et al. 2020). Finally, like other plants, SHB might exhibit predawn stomatal opening (Caird et al. 2007; Dodd et al. 2005; Howard and Donovan 2007; Schwabe 1952).

In conclusion, our study provides insights into the stomatal traits and responses of various blueberry (Vaccinium spp.) genotypes. We confirmed that blueberries are hypostomatous plants and documented significant diversity in stomatal morphology. This diversity extended to stomatal responses to irradiance, where S_d, but not S_a, affected the speed of stomatal response. Additionally, we documented differences in g_min among SHB genotypes. These results provide foundational information for the study of stomatal biology of a globally important food crop such as blueberry.