Does the functional leaf anatomy of Justicia calycina (Acanthaceae) reflect variation across a canopy gradient in the Southern Brazilian Amazon?

Justicia calycina is a subshrub well distributed in northern South America. Despite its widespread occurrence, little is known about its anatomy and plasticity in different environments. Our study aimed to evaluate how leaf anatomical traits of J . calycina adjust in forest areas with variations in canopy openness in the Southern Amazon, Mato Grosso State, Brazil. Forest areas were: Open and Dense Ombrophilous Forest, and Seasonal Deciduous Forest (rocky outcrop). We performed anatomical measurements in leaves sampled on 10 individuals in each forest site, with the help of usual anatomy techniques. Foliar anatomy of J . calycina showed similar pattern among the studied forests, with significant difference (p<0.05) in four leaf traits among 22 studied: vascular bundle diameter, palisade parenchyma thickness, stomatal density and trichome density in adaxial epidermis. These leaf traits showed high plasticity index, with higher values in Dense Forest and in Deciduous Forest. Even though J . calycina showed overall little anatomical variation among forest types, the few attributes that differed are fundamental in the photosynthesis process. The adjustment in leaf anatomical traits under different luminosity conditions demonstrates a degree of phenotypic plasticity in the species, which contributes to its distribution in different forest phytophysiognomies in Brazilian Amazonia.


Introduction
Tropical plant communities show variable environmental conditions, especially in microclimatic and edaphic factors.The coexistence and development of plant species is directly related to the potential that they have to adjust phenotypic characteristics in response to environmental variations (Pireda et al. 2019).The ability of plants to alter their phenotype in different environments is defined as phenotypic plasticity (Sultan 2000).
Leaf traits are among the most plastic in plants in response to environmental variations (Sultan 2000;Gratani 2014).Leaves are directly involved in photosynthetic processes and ecosystem processes (Garnier et al. 2001;Diaz et al. 2004).Leaf morphological characteristics can significantly externalize the environmental stresses that plants are suffering or, conversely, show plasticity in response to anatomical adjustments (Garnier et al. 2001;Diaz et al. 2004;Rosado & Mattos 2007).In this sense, plants that have greater phenotypic plasticity are more apt to coexist in different environments, allowing for a wide geographic distribution (Gratani 2014).
Variations on thickness of tissues (e.g., epidermis and mesophyll parenchyma) and density of stomata and trichome are usually found related to luminosity, water availability and soil fertility conditions (Dickison 2000;Gratani 2014).These anatomical variations are associated with regulating light and CO 2 profiles within leaves and maximizing photosynthetic efficiencies (Dickison 2000;Crang et al. 2018).Sun leaves need to prevent high sunlight irradiance and/or water stress, and may have larger cells or more layers in palisade parenchyma making them more efficient in capturing and distributing of light through the mesophyll (e.g., Rabelo et al. 2013;Gratani 2014;Campbell et al. 2018).Sun leaves may also have more stomata per unit of leaf area, thicker epidermis and cuticle which favors greater light reflection, preventing leaf overheating (Rossatto et al. 2009;Fang & Xiong 2015).
The Acanthaceae family has high diversity in Brazil (471 native species; Flora e Funga do Brasil 2020 2022), but only few studies addressed leaf anatomical responses in different environmental conditions (e.g., Larcher & Boeger 2006;Cassola et al. 2019).For Justicia calycina (= Justicia acuminatissima), some studies pointed to great morphological variability (Wasshausen & Wood 2004;Wasshausen 2006;Verdam et al. 2012).This species is a perennial subshrub that occurs in several different phytophysiognomies and is well distributed in northern South America, including Brazil and a number of other countries (Tropicos 2022).The species also has potential medicinal properties (e.g., Corrêa & Alcântara 2012;Cordeiro et al. 2019).Thus, it is important to deepen studies on the leaf morphoanatomy of J. calycina to understand its wide geographic distribution.Cassola et al. (2019), for example, observed changes in leaf anatomy of three Justicia species under different growth conditions.Besides, anatomical studies associated with taxonomy, phylogeny and degree of plasticity are useful to evaluate environmental heterogeneity, contributing to the conservation and management plans of natural vegetation (e.g., Metcalfe & Chalk 1979;Gupta & Shukla 2012;Gratani 2014;Sokoloff et al. 2021).
In this context, our study aimed to evaluate how leaf anatomical traits of J. calycina vary among three forest areas with canopy openness variation in the Southern Brazilian Amazon.Considering that the forests in this region have different types of vegetation and environmental conditions, we expect that anatomical traits will differ among areas (structural adjustments), enabling acclimation and adaptation of J. calycina.We also expect that in forest areas with higher luminosity intensity, J. calycina will show typical variations of sun plants in leaf anatomical traits, such as thicker tissues and increased stomata and trichome density.These anatomical traits are among those that most vary in relation to environmental conditions, especially in sun-shade comparisons (Dickison 2000;Gratani 2014).

Studied areas
Our study with Justicia calycina (Nees) V.A.W.Graham (Fig. 1A) was developed in three sites with different forest types in Southern Amazon, Mato Grosso State, Brazil: Open and Dense Ombrophilous forest, and Seasonal Deciduous Forest (Fig. 1B).The species is abundant in the three areas.
The Dense Ombrophilous Forest (hereafter Dense Forest) was sampled at the Reserva Surucuá, a particular reserve with about 50 hectares in the urban area of Alta Floresta municipality.In the studied forest portion (9°52'42"S, 56°05'55"W, 290 m altitude; Fig. 1C), the degree of anthropogenic disturbance is relatively low, the forest canopy reaches 20-30 m in height and the understory is dense and rich in shrub species (E.Gressler, personal communication).
The Seasonal Deciduous Forest (hereafter Deciduous Forest) was sampled in the Reserva Particular do Patrimônio Natural Mirante da Serra reserve (1616.7 ha) located in the municipality of Novo Mundo, about 40 km away from the other two studied forests.We sampled Justicia calycina at two points in a rocky outcrop with transitional area of Does the functional leaf anatomy of Justicia calycina (Acanthaceae) reflect variation across a canopy gradient in the Southern Brazilian Amazon?
Rainy seasons in the region of Alta Floresta and Novo Mundo occur between October and April, and dry seasons between May and September (Caioni et al. 2014).Open and Dense Forest are under the same meteorological conditions.In these areas, in 2016, total annual rainfall was 1911 mm, mean annual temperature 27.5 °C and mean annual air humidity 75.0 %.For the same year, in the Deciduous Forest, total annual rainfall was 2566 mm, mean annual temperature 27.2 °C and mean annual air humidity 85.6 % (Fig. S1).Those data were collected in two meteorological stations, one installed in the Dense Forest and one installed about 3 km of distance from the Deciduous Forest (V Dubreuil, unpubl.data).There is no meteorological data available for 2017 in the Deciduous Forest, due to local equipment failure.
We measured luminosity conditions in each studied forest using hemispheric photographs taken with a digital professional camera (Nikon DSLR D600), equipped with a fisheye 8 mm lens (Sigma EX DG 1:3.5) and mounted on a tripod at 1.0 m aboveground.Photographs were taken on 27-28th September/2016, from 07:00 to 11:00 a.m., in points close to the sampled individuals in Dense Forest (9 points), Open Forest (3) and Deciduous Forest (4).We analyzed canopy openness (%) in the photographs using the program Gap Light Analyzer version 2.0 (Frazer et al. 1999).Hemispheric fisheye photographs are widely used in ecological studies to estimate light conditions in forest understory (e.g., Dahlgren et al. 2007).In the case of J. calycina, we took photographs when the equipment was available in our institution, which coincided with the beginning of the growing season (Lauton et al., unpubl. res.) and may represent the general luminosity conditions under which the leaves collected in April/2017 for anatomy developed.

Anatomical measurements
In April/2017 (end of the rainy season), we randomly sampled four adult leaves (fully expanded and developed) from the middle plant part (between the first, second and third knots) of 10 adult individuals of Justicia calycina in each studied forest, to minimize influence of developmental stage on the analysis.We also collected fertile material (Fig. 1A), which we deposited at the Southern Amazon Herbarium (HERBAM).Vouchers: M.B.Lauton 01 (26587 -25/IV/2017), M.B.Lauton 02 (26588 -27/IV/2017) and M.B.Lauton 03 (26589 -28/IV/2017).We fixed the leaves with FAA 50 (formaldehyde, glacial acetic acid and 50 % ethanol, with proportion of 5:5:90, respectively) for 48 hours, and later we preserved in 70 % ethanol according to the modified method of Johansen (1940).
From each individual we analyzed two leaves, totaling 60 samples, without visually signals of senescence or damage caused by herbivores or pathogens.We used the method of Roeser (1962) to make freehand cross sections (with razor blades), in the median region of each leaf, obtaining thin sections which were stained with Astra blue and Basic Fuchsin.The median region was selected because it is an area with more uniform tissue distribution in the leaf lamina (Appezzato-da-Glória & Carmello-Guerreiro 2012).For the analysis of the epidermis, we used the technique of printing both leaf surfaces with a drop of universal cyanoacrylate ester instant adhesive on histological slides, as described by Segatto et al. (2004).
We photographed each histological slide using a digital image capture device, coupled to a Leica ICC50 photomicroscope, with the support of software LAZ EZ 1.7.0.For the comparison among forest sites, we measured 22 leaf anatomical traits (anatomical variables) using the Anati Quanti software (Aguiar et al. 2007; Tab. 1, Fig. S2).

Statistical analysis and plasticity index
We compared luminosity conditions (canopy openness) among the three forest sites using Kruskal-Wallis test in STATISTICA 8.0 software.Forest types differed significantly in luminosity (H = 11.76,p = 0.0028) and Deciduous Forest showed higher values than the other two forest sites (Fig. 2).
All the statistical analyses with anatomical data were performed using R package version 4.0.3(R Development Core Team 2020).First, we tested the assumptions for normality and homoscedasticity using Shapiro and Levene tests.We tested the relationships among anatomical variables with Spearman's correlation (Tab.S1) and considered a strong correlation when the pairs of variables showed a value above 0.70 (Kendall 1938).In such cases, we eliminated one of the variables from the subsequent analysis using the potential adaptive explanation of the variable as an elimination criterion.Thus, the variables midrib size (MIDS), spongy parenchyma thickness (SPT) and stomatal index (STI) were excluded (Tab.S1).
We performed a PCA (Principal Components Analysis) with data of the three studied forests using the anatomical variables with the highest factorial load values for significant axes, selecting them for comparison among the forest types.To identify which anatomical variables contribute significantly to distinguish the studied forest sites, we performed the Kruskal-Wallis test.After that, we applied Dunn's multiple comparison test with Bonferroni's correction (PMCMR package; Pohlert 2014) on the variables with positive results in the previous tests.
For the variables which were significantly different among forest types we calculated the plasticity index (PI), which ranges from 0 (no plasticity) to 1 (maximum plasticity), with values higher than 0.6 interpreted as high plasticity (Valladares et al. 2000;2006).

Qualitative analysis
In the three studied forest sites, Justicia calycina showed hypostomatic leaves, midrib with a bicollateral vascular bundle in an open arc and 2-3 accessory bundles, one-layered adaxial and abaxial epidermis, dorsiventral mesophyll and both epidermal surfaces with trichome and cystoliths (Figs. 3,4).
We observed only slight qualitative differences in cross and paradermal sections among the forest sites.The epidermal cells showed sinuous contour, with visually less sinuosity and thicker walls in Deciduous Forest (Fig. 3).Both the adaxial and abaxial epidermal surfaces presented cystoliths and numerous trichomes (or their scars) mainly on the abaxial surface (Fig. 3).In the midrib we observed little visual variation among forest types in relation to the size of vascular bundles (larger in Dense Forest) and a slight difference in the collenchyma position, which in the Open Forest was more dispersed than the other two forests (Fig. 4A, C, E).
In the leaf lamina, anatomical traits were visually similar among forest sites (Fig. 4B, D, F).Palisade parenchyma showed a single layer and spongy parenchyma showed four to five layers.The cuticle was relatively thin (less than 25 % of the adjacent epidermal cell size) and stomata occurred in the same level of other epidermal cells.

Quantitative analysis
The PCA showed an overlap of the three studied forest sites regarding anatomical variables, indicating a common anatomical pattern among the areas, with a greater distribution of sample range in the Deciduous Forest (Fig. 5).The studied forests differed significantly in four anatomical variables among the 22 analyzed (p < 0.001; Tab. 1).
In the midrib, the vascular bundle diameter (VBD) showed higher values in Dense Forest.In the leaf lamina Does the functional leaf anatomy of Justicia calycina (Acanthaceae) reflect variation across a canopy gradient in the Southern Brazilian Amazon?
(cross-section), the thickness of palisade parenchyma (PPT) was significantly higher in Deciduous Forest.In the paradermal section, the variables which differ significantly among forest types were stomatal density (STD) and trichome density in adaxial epidermis (TRIAD), both with higher values in Open and Deciduous Forest (Tab.1).

Discussion
Our results showed that leaf anatomical traits of Justicia calycina were overall similar among the three forest sites -Dense, Open and Deciduous Forest -studied in the Southern Brazilian Amazon.We found only few significant differences among forest sites, which may be related to variations in luminosity conditions.The leaf traits which differed among forests showed high plasticity, representing a certain degree of acclimation/adaptation of J. calycina according to Valladares et al. (2006).
The thicker palisade parenchyma and higher stomatal and trichome density (PPT, STD and TRIAD, respectively) we verified in the leaves of Deciduous Forest may be interpreted as typical changes of sun plants (i.e., plants under higher luminosity conditions).This forest is located in a rocky outcrop and showed significant greater canopy openness due to the sparse distribution of woody plants (Silva et al. 2020).Along with this, at the time of leaf sampling for anatomical measurements (rainy-dry transition season), woody plants of the Deciduous Forest start to shedding leaves (Sasaki et al. 2010;E. Gressler, personal communication), which contributes to making this forest even more illuminated in comparison to the other two studied forests (both evergreen).Sun leaves usually contain an increased number of palisade layers and greater palisade cell elongation (Lange et al. 1981;Dickison 2000;Gratani 2014).The elongation of palisade cells in sun leaves has been related to a reduction in mesophyll resistance to CO 2 and optimization of light absorption allowing light to channel deeper into lower leaf cell layers, such as the spongy cells where the light is scattered (Vogelmann & Martin 1993;Théroux-Rancourt & Gilbert 2017).
Regarding stomata, the higher density (STD) we observed in the Open and Deciduous Forests may also be explained by higher luminosity in these forests.Stomata of a given species can largely vary in quantity, distribution, size and shape in response to environmental factors (Willmer Sun leaves usually present high stomatal frequency, being necessary to maximize CO 2 absorption rates, to water saving and to prevent excessive leaf heating under high irradiance (Givnish 1988;Bertolino et al. 2019;Lambers & Oliveira 2019).The lower stomatal density in Dense Forest can be an indicator of larger leaves in this forest, because stomata are more spread out in these leaves (Franks & Farquhar 2007).
The higher trichome density in adaxial epidermis (TRIAD) of J. calycina leaves from the Decidual Forest may be related to the protection of intense luminosity and desiccation, given the high light intensities of this forest.Highly pubescent surfaces can reflect sunlight and decrease transpiration rates, especially in the adaxial epidermis which is exposed directly to sun and therefore present higher temperature and transpiration than abaxial epidermis (Fahn    & Cutler 1992;Roy et al. 1999;Dickison 2000;Larcher 2003;Ichie et al. 2016).
The epidermal cells of the adaxial surface of J. calycina showed thicker walls and slightly less sinuous outline in Deciduous Forest compared to the other forest sites.In environments with high luminosity intensity, such as the Deciduous Forest, epidermal cells usually have thicker walls and less or no sinuosity (Evert 2013).According to Roth (1984), thicker walls increase the protective character of the epidermis.
The greater diameter of the central vascular bundle (VBD) in Dense Forest could indicate that in this forest, due to the significant lower luminosity, plants of J. calycina have larger leaves and therefore need to invest more in supporting tissues and major veins (e.g., Chazdon 1985;Niinemets & Sack 2006).Unfortunately, we did not measure the size and area of the collected leaves to confirm this hypothesis.
The few anatomical differences we detected among forests in J. calycina may be explained by the interaction of luminosity and other environmental variables such as temperature and humidity.Plants in rocky outcrops, as the studied Deciduous Forest, usually deal with stressful environmental conditions such as high insolation and temperature, desiccant winds, low humidity and poor and shallow soil (Prance 1996;Porembski & Barthlott 2000).According to Porembski (2007), rocky outcrops offer several microhabitats, such as crevices, depressions and soil islands, causing great microenvironmental variation.In the studied Deciduous Forest, the higher annual rainfall may contribute to maintaining a certain degree of humidity in the forest.Another explanation for the few anatomical differences among the studied forests is the intense herbivory observed in J. calycina leaves (MB Lauton et al., unpubl.res.), which may limit plant phenotypic plasticity (Valladares et al. 2007).
In the same region and similar forest types of our study, other studies have also found anatomical variations in understory species related to environmental conditions, though with overall low plasticity index (Ariano & Silva 2016;Dardengo et al. 2017;Rocha et al. 2019;Müller et al. 2020).On the other hand, in savannas and cerrado phytophysiognomies in the Cerrado-Amazon transition of Mato Grosso State, some studies revealed phenotypic plasticity with high occurrence of xeromorphic anatomical traits in most species (Simioni et al. 2017;Pessoa et al. 2019).It is well known that leaf anatomical traits vary among species, according to their ability to acclimation and intensity of light received (e.g., Rozendaal et al. 2006;Markesteijn et al. 2007;Liu et al. 2019).The plasticity in anatomical responses that we found in J. calycina individuals under different luminosity conditions probably contributes to its wide distribution in different forest phytophysiognomies of the Amazon region.
To the best of our knowledge, this is the first study of the functional leaf anatomy of J. calycina comparing different forest phytophysiognomies in Brazil.Contrary to our expectations, forest sites with different canopy openness (or light intensity) showed only few significant differences in leaf anatomy of J. calycina.The differences we found were related to stomatal/trichome density and thickness of palisade parenchyma/vascular tissues, which are fundamental leaf traits in photosynthetic processes and in plant water transport.Our results can help to deepen physiological studies, directing to integrated studies of plant anatomy and physiology, seeking to elucidate the photosynthetic mechanism of this species and its strategies to deal with different environmental factors.

Supplementary material
The following online material is available for this article: Table S1 -Lauton MB, Gressler E, Oliveira JA, Simioni PF, Ribeiro-Júnior, Yamashita OM, Silva IV.Does the functional leaf anatomy of Justicia calycina (Acanthaceae) reflect variation across a canopy gradient in the Southern Brazilian Amazon?Matrix of Spearman correlations for the 22 leaf anatomical variables of Justicia calycina evaluated in three forest sites, Southern Brazilian Amazon.High correlation (≥ 0.70).

Figure 2 .
Figure 2. Comparison of canopy openness (%) among the three forests where we sampled leaves of Justicia calycina and respective hemispheric photographs examples.Kruskal-Wallis test revealed significant differences among forest sites (H = 11.76,p = 0.0028).

Figure 3 .Figure 4 .
Figure 3. Paradermal sections of the leaf lamina of Justicia calycina in three forest types in Southern Brazilian Amazon: Dense Forest (A-B), Open Forest (C-D) and Deciduous Forest (E-F).Adaxial epidermis with cystoliths in detail (A, C, E).Abaxial epidermis with stomata positioning in detail (B, D, F). White arrows indicate trichome or scars and black arrows indicate cystoliths.Bars: 100 μm.

Figure 5 .
Figure 5. Principal Component Analysis (PCA) with 18 leaf anatomical variables of Justicia calycina in three forest sites in Southern Brazilian Amazon (A) and contribution of each variable to the distribution of samples (B).Forest sites: Dense Forest, Open Forest and Deciduous Forest.

Table 1 .
Means and standard deviations of 22 leaf anatomical traits measured in Justicia calycina in three forest sites, Southern Brazilian Amazon.VBS, SPT and STI were excluded from the analysis due to high correlation values with other leaf traits.*Significant differences among forest sites detected in Dunn's multiple comparison test (p < 0.001); different letters indicate statistical difference on the line.Ns: non-significant result.PI = plasticity index.

Figure S1 -
Meteorological data sampled in 2016 in the forest sites where we studied leaf anatomy of Justicia calycina, in the Southern Brazilian Amazon.Minimum temperature (Tmin), mean temperature (Tmean) and maximum temperature (Tmax) (A-B).Rainfall and air relative humidity (C-D).The rainy season (October to April) is represented by the boxes in the respective months.

Figure S2 -
Leaf anatomical traits measured in Justicia calycina in Southern Brazilian Amazon.