Influence of test tube sealing on the morpho-anatomy and ultrastructure of leaves of Aechmea bromeliifolia (Bromeliaceae) grown in vitro1

In vitro grown plants may have abnormal structural and physiological features. However, the type of the sealing material used in tissue culture may minimize such abnormalities. This study evaluates the influence of the type of sealing of test tubes on the anatomical and ultrastructural features of leaves of Aechemea bromeliifolia (Rudge) Baker (Bromeliaceae), an ornamental bromeliad native to Brazil, grown in vitro. Three types of sealing were used: rigid polypropylene cap (PC), polyvinyl chloride film (PVC), and PC coupled with a microporous membrane (PM). Seedlings germinated in a greenhouse were also studied for comparison. Plants grown in test tubes sealed with PM were more similar to those from the greenhouse, as far as the pattern of stomatal opening, the presence of starch grains, and the organization of the internal membrane system of the chloroplasts is concerned. Plants cultivated in test tubes sealed with PC had higher stomatal density and the chloroplasts had large areas without thylakoids in the stroma. Plants grown in test tubes sealed with PVC had few or no starch grains. These results suggest that microporous membrane used coupled with PC sealing provided natural ventilation, thus contributing to a better plant development.


Introduction
In vitro culture techniques are widely used for the propagation of many plant species, resulting in rapid multiplication and pathogen-free plants that are genetically identical to the original ones (Chen 2004). The in vitro environment is characterized by dim light, high relative humidity, low CO 2 concentration during the photoperiod, high ethylene concentration, and restricted air movement inside the culture vessel (Kozai & Smith 1995). Such features affect the anatomy and physiology of in vitro grown plants, resulting in absent or reduced cuticle, dysfunctional stomata, and poorly developed photosynthetic apparatus. As a consequence, in vitro grown plants may be vulnerable when exposed to the ex vitro environment (Hazarika 2006). Some materials used for sealing, such as screw caps, aluminum foils, transparent films (e.g., Parafilm ® ), and polypropylene caps, may restrict gas exchange between the culture vessel and the outer atmosphere (Zobayed et al. 2000;Zobayed 2008). The sealing material used for in vitro culture has significant influence on plant morphoanatomy, thus affecting its growth (Gonçalves et al. 2008;Ribeiro et al. 2009). Nonetheless, only a few studies Majada et al. (2002), Lucchesini et al. (2006) and Sáez et al. (2012a) have evaluated the effect of the different types of sealingby using ventilated and non-ventilated vessels, on the ultrastructure of plants grown in vitro.
It is a common practice for culture vessels be kept completely sealed to avoid contamination by microorganisms (Kozai & Kubota 2005;Kozai 2010). However, in order to enhance gas exchange and to promote normal plant growth, the vessels must be ventilated (Zobayed et al. 2000). Sealing materials that are more permeable can reduce the relative humidity inside the culture vessel, enabling an increased transpiration, and a higher uptake of water and nutrients by the plants (Xiao et al. 2011).
Given that the in vitro culture conditions are closely related to plant growth and that the sealing material may improve plant development therefore facilitating the acclimatization phase, herein we evaluate the effect of the type of test tube sealing on the morpho-anatomy and ultrastructure of leaves of Aechmea bromeliifolia (Rudge) Baker (Bromeliaceae) seedlings. A. bromeliifolia is a bromeliad native to Brazil with a huge ornamental potential, which is distributed in Central America, Northwest of South America and from North to South of Brazil. (Luiz-Santos & Wanderley 2012). The in vitro propagation of this species has currently been investigated by the Plant Tissue Culture Laboratory of the Federal University of Goiás (UFG) aiming at avoiding the exploitation of natural populations from the state of Goiás, Brazil. Therefore, in the present study, we compared the growth of plants obtained from seeds using three types of test tube seal with that of plants grown from seeds in a greenhouse. In vitro culture was initiated by placing the seeds in test tubes containing 20 mL half strength basal MS medium (Murashige & Skoog 1962) supplemented with 30 g L -1 sucrose, pH 5.7 ± 0.1, and solidified with 7.0 g L -1 agar. The total volume of the test tubes was 55 mL. The plants were kept in a growth room at 25 ± 1 o C with a photoperiod of 16 h, and photosynthetic photon flux density (PPFD) of 43.23 µmol m -2 s -1 provided by two cool-white fluorescent lamps. Plants grown in vitro were approximately 9 weeks old when collected for the analyses.

Plant material and cultivation conditions
For greenhouse cultivation, seeds (n = 40) were sown in polyethylene terephthalate (PET) bottles as described by Paula & Silva (2004). The substrate consisted of a mixture of sand and soil (2:1) which was covered with Plantmax ® (Aitkens, Glasgow, UK). The PPFD inside the greenhouse was of 350 µmol m -2 s -1 obtained from natural light and water was provided manually whenever needed. Plants grown ex vitro were approximately 10 weeks old when collected for the analyses and were about 1-2.5 cm high (aerial part).

Sealing of test tubes
The test tubes used for in vitro culture were sealed with three types of lids: rigid polypropylene cap (PC), single-layered polyvinyl chloride film (PVC), and PC with 5 mm-diameter holes covered with a membrane composed of two layers of microporous tape (Cremer ® , São Paulo, SP, Brazil) and a single layer of polytetrafluoroethylene film (Amanco ® , Grupo Mexichem, São Paulo, SP, Brazil) (PM) (Saldanha et al. 2012). The PPFD was measured inside the test tubes at the height of 6 cm, resulting in 42.34, 42.32, and 41.09 µmol m -2 s -1 for the tubes sealed with PC, PVC, and PM, respectively. We used one seed per tube and 40 test tubes per treatment.

Morpho-anatomical and ultrastructural characterization
For anatomical analyses, the plants were collected at 8 am and the leaves were fixed in FAA (formaldehyde, acetic acid, and 50% ethanol; 1:1:18, by volume) for 48 h (Johansen 1940) and stored in 70% ethanol. The middle regions of the most expanded leaves were used for all the analyses. Leaf samples were dehydrated in an ethanolic series (Johansen 1940) and embedded in Leica ® historesin (Leica Microsystems Nußloch GmbH, Heidelberg, Germany) following manufacturer's recommendations. Cross-sections (12-μm thick) of the samples were obtained using a rotary microtome and stained with toluidine blue (O'Brien et al. 1964), followed by a double stained with 0.1% basic fuchsin and 0.3% astra blue (1:3) (Kraus et al. 1998). Analysis of the surface view of the epidermis was performed using leaf samples that were cleared with a 10% sodium hydroxide solution for 30 min and then were immersed in 10% commercial sodium hypochlorite (2-2.5% active chlorine) until they were completely cleared and stained with 1% aqueous safranin solution (Johansen 1940). To confirm the presence of cuticle, Sudan III (Johansen 1940) was used in freehand cuts of fresh leaf samples.
The scanning electron microscope analysis and photography of fresh leaf samples were carried out using a JSM IT300LV scanning electron microscope (JEOL USA, Inc., Peabody, MA, United States) at the Centro Regional para o Desenvolvimento Tecnológico e Inovação, UFG.
For ultrastructural analysis, leaf fragments (1.0 mm 2 ) were fixed for 24 h in 2.5% glutaraldehyde diluted in 0.1 M sodium cacodylate buffer, pH 7.2, containing picric acid. Samples were washed in cacodylate buffer, post-fixed in 1% osmium tetroxide in the cacodylate buffer for 2 h. Subsequently, leaf fragments were dehydrated in acetone series and embedded in Epon resin EMbed-812 (Electron Microscopy Sciences, Industry Road, Hatfield, PA, United States) (Cotta-Pereira et al. 1976). Ultra-thin transverse sections were stained with 2% uranyl acetate solution (20 min), washed in distilled water, immersed in 2% lead citrate solution (6 min), and washed in double distilled water. Samples were analyzed using a JEM 2100 transmission electron microscope (JEOL USA, Inc., Peabody, MA, United States), equipped with an Energy Dispersive X-ray Spectrometer (EDS) (Thermo Fisher Scientific Inc., Waltham, MA, United States). The area of the chloroplasts was measured using the ImageJ software (National Institute of Health, Bethesda, MD, United States). For this analysis, the samples of all treatments were collected at the same time of the day (2 pm).
The anatomical analyses were carried out using five repetitions for each treatment. For cuticle and ultrastructural analysis, three repetitions of each treatment were used. The stomata index (SI) was calculated according to the formula of Salisbury (1927), SI (%) = [S/(S+E)] × 100, where S is the number of stomata per unit area and E the number of epidermal cells at the same unit area, these two parameters were counted in 10 field of 0.25-mm 2 per leaf.
For the height of the aerial part and number of leaves, 12 plants of each treatment were used.

Statistical analysis
The experimental design was completely randomized, and the data were submitted to analysis of variance (ANOVA) by using STATISTICA version 7.0 (StatSoft Inc., Tulsa, OK, United States), where significant main effects were obtained and compared by performing the Tukey's test (α = 0.05).

Results
The in vitro grown plants had longer aerial parts and a larger number of leaves compared to those cultivated in the greenhouse (Tab. 1). However, the leaves of plants cultivated in vitro were membranaceous, whereas the leaves of those grown in the greenhouse became thicker and rigid. Plants grown in tubes sealed with PC and PM showed similar heights (Tab. 1).
The same basic anatomical structure was found in all analyzed leaves: multicellular and uniseriate glandular trichomes (Fig. 1a) at the margin of the leaf (Fig. 1b), one-layered epidermis with thin cuticle; tetracytic stomata found on  the abaxial surface and on the margins of the adaxial surface (Fig. 1b) at the same level of other epidermal cells; heterogeneous mesophyll with aquifer parenchyma and chlorenchyma (Fig.  2a,b); collateral vascular bundles with fiber poles associated with the xylem and phloem (Fig. 2c-f). However, some differences were noticed on the aperture of stomata ostioles. Plants developed in the greenhouse had completely closed stomata (Fig.  2g), while plants grown in vitro had either open stomata in tubes sealed with PC and PVC (Fig.  2h,i), and completely closed in tubes sealed with PM sometimes with very reduced aperture (Fig. 2j).
The total thickness of the leaf, mesophyll, aquifer parenchyma, and chlorenchyma, as well as the height of the epidermal cells on the adaxial surface were significantly higher in plants developed in the greenhouse when compared to the in vitro grown plants, which did not differ statistically (Tab. 1). In addition, the epidermal cells on the abaxial surface were not significantly different in the leaves of plants developed in the greenhouse, and those grown in the tubes sealed with PM when compared to plants cultivated in the tubes sealed with PC and PVC (Tab. 1). Overall, the vascular bundles located at the leaves edges had more fibers associated with the xylem under all culture conditions (Fig. 2c-f). However, in the leaves of the plants grown in tubes sealed with PVC, the walls of these fibers were visually thinner (Fig. 2e).
The polar diameter of the stomata did not differ significantly under any culture conditions (Tab. 1). Nevertheless, the equatorial diameter was smaller in plants developed in tubes sealed with PM compared to those grown in the greenhouse. Consequently, the ratio between the polar diameter and the equatorial diameter was greater in the plants cultivated in tubes sealed with PM, whereas no significant differences were found for this parameter in the plants in vitro. The stomata of the plants grown under all treatments were elliptical (Tab. 1).
The stomatal density was significantly higher in leaves of plants grown in tubes sealed with PC compared to those obtained under all culture conditions. However, a stomatal index of leaves of plants grown in tubes sealed with PC was not significantly different to plants grown in greenhouse (Tab. 1).
The chlorenchyma cells of plants obtained in all treatments had typical organelles (Fig. 3a-f). We observed large central vacuoles, with other organelles adjacent to the cell walls (Fig. 3a). Mitochondria had well-developed crests (Fig.  3b-f), and were found associated to chloroplasts. The nuclei had conspicuous nucleoli and a similar shape in the cross-section of the leaf (Fig. 3b,c,e,f). Chloroplasts did not differ significantly in area among the treatments (Tab. 1). However, the shape of the chloroplasts, the organization of the internal membrane system, the number of starch grains, and the lipid bodies were different among the treatments.
The chloroplasts of plants grown in the greenhouse were oval, and had a less appressed arrangement of thylakoid membranes and a high amount of starch grains and plastoglobules, both large-sized (Fig. 4a). The leaves of plants developed in vitro had homogeneously small plastoglobules ( Fig. 4b-d) and a more appressed arrangement of thylakoid membranes when compared to the plants grown in the greenhouse (Fig. 4a-d). In plants grown in tubes sealed with PC, the chloroplasts were more elongated, and had large areas without thylakoids filled with the stroma (Fig. 4b). In plants developed in tubes sealed with PVC, the chloroplasts were semicircular, and the internal membrane system was irregularly arranged in some regions (Fig. 4c). Fewer chloroplasts contained starch grains in plants developed in tubes sealed with PVC and PC and, when present, they were smaller and less numerous (Figs. 3e;4c). In tubes sealed with PM, the chloroplasts ranged from semicircular to oval, had large starch grains (Fig. 4d) and, in some regions, a less appressed arrangement of thylakoid membranes.

Discussion
In vitro culture influenced the morphology of leaves of A. bromeliifolia, which were more elongated, slender and delicate when compared to those grown under greenhouse conditions. This result corroborates those previously reported (Johansson et al. 1992;Aoyama et al. 2012). In response to in vitro environment, plant physiology and morphology are significantly different from those of the plants grown in natural environments (Kozai 2010). The thicker leaves developed in the greenhouse indicate increased cell expansion, since no significant difference was observed in the number of cell layers in the leaves of all analyzed plants. This result is probably related to the intensity of light inside the greenhouse. In fact, several studies have reported thickening of leaves due to cell expansion under higher light intensity (Voltan et al. 1992;Hanba et al. 2002;Lima Jr. et al. 2006;Fernandes et al. 2014), with improved photosynthetic efficiency (Kim et al. 2005). According to Hazarika (2006), the delicate structure of plants developed in vitro makes them more vulnerable when exposed to ex vitro conditions. The leaves of A. bromeliifolia were anatomically similar under all culture conditions, and they are also similar to that reported for leaves of other species of Aechmea, grown under natural conditions (Proença & Sajo 2004;Souza et al. 2005;Silva & Scatena 2011). Nonetheless, we observed that some characteristics in adult plants of A. bromeliifolia are different from those of young plants. The leaves of the adult plants are hypostomatic, the epidermis has peltate trichomes, the mesophyll has aeration channels continuous a b c e f d to the subestomatic chambers with braciform cells, and groups of extravascular fibers dispersed throughout the chlorenchyma (Proença & Sajo 2004). In addition, Proença & Sajo (2004) do not report the occurrence of glandular trichomes. Despite the anatomical similarities, some characteristics differed between the leaves of A. bromeliifolia cultivated in vitro. The stomata of leaves of A. bromeliiflia grown both in a greenhouse and in tubes sealed with PM were closed by the time of collection (daytime period). Considering that A. bromeliifolia is a crassulacean acid metabolism (CAM) species (Griffiths & Smith 1983;Scarano et al. 2002), it is expected that stomata remain closed during the day under natural conditions. In contrast, the leaves of plants grown in tubes sealed with PC and PVC had opened stomata by the time of collection. The microporous membrane used in the PM sealing type allows greater amount of gas exchange per hour than other sealing types without membrane (Saldanha et al. 2012), and this probably favors the proper functioning of A. bromeliifolia stomata.
The leaves of A. bromeliifolia grown in vitro had elliptical stomata, similar to what was observed in leaves grown in a greenhouse. However, this result is not common, since the stomata of plants grown in vitro are usually rounded, which is often associated with low functionality (Khan et al. 2003;Afreen 2005), and incapacity to fully close when stimulated (Khan et al. 1999). The elliptical stomata found in A. bromeliifolia are believed to have good functionality since they are characteristically found in plants grown under photoautotrophic conditions (Khan et al. 2003). Plants developed in tubes sealed with PC had leaves with higher stomatal density. However, stomatal index of leaves from this condition was not significantly different from that observed in greenhouse leaves. This result suggests that the increase of stomatal density occurred due to a decrease on the expansion of ordinary epidermal cells. Greater stomatal density has been reported to occur due to high relative humidity in tubes with less ventilation (Zobayed et al. 2001;Mohamed & Alsadon 2010). Thus, the high number of stomata in plants grown in hermetically sealed tubes may be a strategy to enhance leaf transpiration, thus leading to higher uptake of water and nutrients.
The fibers associated with the vascular bundles in leaves of plants grown in tubes sealed with PVC were not very thick and this may provide less mechanical support for the plants during the process of acclimatization.
The in vitro environment affected the organization of chloroplasts of A. bromeliifolia, different from other organelles that were similar to those of plants gown in the greenhouse, as reported by Rodrigues et al. (2014). The most frequent changes in the ultrastructure of leaf cells of plants cultivated in vitro are observed in the chloroplasts, which vary in shape, size, organization of the internal membrane system and number of starch grains (Wetzstein & Sommer 1982;Lee et al. 1985;Majada et al. 2002;Sáez et al. 2012b;Kapchina-Toteva et al. 2014;Stefanova et al. 2015).
Little appressed thylakoid membranes are typical of chloroplasts exposed to high light intensity (Yano & Terashima 2001;Fernandes et al. 2014). The less appressed arrangement of thylakoid membranes in the chloroplasts of plants grown in the greenhouse may be associated with the formation of large starch grains which are due to high light exposure. Similarly, in the chloroplasts of Liquidambar styraciflua L. exposed to high light intensity, grana were poorly organized as a result of the presence of starch (Lee et al. 1985).
The chloroplasts of A. bromeliiflia grown in tubes sealed with PC have large areas without thylakoids and the chloroplast development in tubes sealed with PVC have irregularly arranged inner membranes. The abnormal development of chloroplasts has been reported in plants grown in non-ventilated vessels (Majada et al. 2002). Several authors have reported changes in the ultrastructure of chloroplasts in response to ethylene (Toyama 1980;Fukuda & Toyama 1982;Fan et al. 2013), which is a volatile plant hormone that accumulates due to lack of ventilation. In addition, chloroplasts of A. bromeliifolia grown in vitro have small and few plastoglobules. According to Sáez et al. (2012b) these features are associated to the lower oxidative stress in the in vitro environment when compared to the greenhouse conditions.
In conclusion, the analysis showed that the in vitro environment influences the texture and thickness of the leaves, the functionality of stomata, and the ultrastructure of chloroplasts. However, plants grown in tubes sealed with PM were similar to those developed in the greenhouse, especially considering stomata opening, occurrence of starch grains and organization of the internal membrane system of chloroplasts.