Synchrotron micro-computed tomography unveils the three-dimensional structure and origin of staminodes in the Plains Prickly Pear Cactus Opuntia polyacantha Haw. (Cactaceae)

Abstract

Floral appendages display an array of shapes and sizes. Among these organs, staminodes are morphologically diverse structures that have lost the ability to produce pollen, but in some instances, they produce fertile pollen grains. In the family Cactaceae staminodes are uncommon and range from simple linear to flat to spatulate structures, but studies describing their structural attributes are scanty. This study highlights the advantages of synchrotron radiation for sample preparation and as a research tool for plant biology. It describes the internal morphology of floral parts, particularly stamen, tepal, and staminode in the Plains Prickly Pear Cactus, Opuntia polyacantha, using synchrotron radiation micro-computed tomography (SR-μCT). It also shows the different anatomical features in reconstructed three-dimensional imaging of reproductive parts and discuss the advantages of the segmentation method to detect and characterize the configuration and intricate patterns of vascular networks and associated structures of tepal and androecial parts applying SR-μCT. This powerful technology led to substantial improvements in terms of resolution allowing a more comprehensive understanding of the anatomical organization underlying the vasculature of floral parts and inception of staminodes in O. polyacantha. Tepal and androecial parts have uniseriate epidermis enclosing loose mesophyll with mucilage secretory ducts, lumen, and scattered vascular bundles. Cryptic underlying structural attributes provide evidence of a vascularized pseudo-anther conjoint with tepals. The undefined contours of staminodial appendages (pseudo-anther) amalgamated to the tepals’ blurred boundaries suggest that staminodes originate from tepals, a developmental pattern supporting the fading border model of floral organ identity for angiosperms.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00709-023-01846-6.

Introduction

The advent of diverse imaging techniques for the visualization of plant structures in two and three dimensions using high-resolution X-ray computed tomography (Heeraman et al. 1997; Dhondt et al. 2010; Mooney et al. 2012), neutron radiography (Oswald et al. 2008; Leitner et al. 2014), and magnetic resonance imaging (Pohlmeier et al. 2008; Borisjuk et al. 2012) has led to significant advances in understanding the intricate organization of plant parts and organs.

The unique properties of synchrotron beamlines, such as high brightness, strong polarization, wide and tunable energy spectrum, and time-structured emissions combined with advances in optical technologies has seen significant findings, e.g., drug structure (Hubbard 2008), cancer diagnosis and treatment (Heidari and Gobato 2019), COVID-19 structure (Baker 2020), and other medicine discoveries (Kaushik and Raj 2020). Synchrotron-based technology has also been progressively employed for different purposes in plant science revealing novel leaf venation networks (Blonder et al. 2012), calcium oxalate crystals (Matsushima et al. 2012), root growth (Leitner et al. 2014), cell walls (Hiraki et al. 2018), sulfur accumulation in N₂-fixing legume nodules (Rivard et al. 2019), host-pathogen interactions (Brar et al. 2019), and mapping and quantification of cellular organelles (Lombi and Susini 2009; Karunakaran et al. 2015), among other inquiries.

Spectral imaging employing synchrotron radiation is one of the main current trends to investigate cultural and natural heritage materials because it can be applied to preserved specimens in natural history collections. Synchrotron radiation-based micro-computed tomography (SR-μCT) is suitable for the morphological study of non-destructive extant and fossil samples (Smith et al. 2009) making it an ideal tool for the inspection of the external and internal structure of biological samples in a vast diversity of fields, including biology, medicine, palaeontology, and many engineering disciplines as it can provide insight into anatomy, function, and detection of disease(s). Within this context, the versatility of the Biomedical Imaging and Therapy–Bending Magnet (BMIT-BM 05B1-1) facility at the Canadian Light Source, Inc. (https://www.lightsource.ca) makes it suitable for numerous plant specimens because of the advanced imaging modalities allowing visualization in ways not possible using conventional systems. For instance, it performs ultrafast CT scans that can be completed 100× faster than with laboratory scanners, it achieves very high spatial resolution (sample features with dimensions less than one micron can be resolved), and it differentiates between soft tissues which do not normally exhibit contrast when imaged with conventional laboratory and clinical X-ray scanners. Overall, this imaging station has the capability to accelerate research in the areas of material sciences, environment, museological collections, and several programs in the life sciences because it eases imaging of small and large specimens with high spatio-temporal resolution. Furthermore, the spectral range of this equipment is capable of penetrating through large and thick specimens and avoids some steps of the often-lengthy manual preparation of plant tissue, e.g., dehydration, sectioning, and the use of chemical fixatives, e.g., formalin, that can be harmful to human health.

At present, few studies (nutraceutical and physiological) using synchrotron technology exist for cacti. These include the investigation of selenium accumulation, distribution, and speciation in the prickly pear fruit (Opuntia ficus-indica) using microfused X-ray fluorescence and chemical mapping (Bañuelos et al. 2011, 2012), nanocellulose characterization employing Fourier transform infrared (Orrabalis et al. 2019), and aspects dealing with hydraulic conductivities in O. microdasys (Kim et al. 2018), but to our knowledge, floral parts from cacti have not been inspected with this technology. Because of the BMIT advantages, this study uses synchrotron X-ray micro-imaging techniques to visualize and compare the morpho-anatomical fingerprint in floral parts of the Plains Prickly Pear Cactus, Opuntia polyacantha Haw. (Cactaceae), specifically regular stamens and developing staminodes, which are distinctive floral inventions occurring in ca. one-third of angiosperm species (Walker-Larsen and Harder 2000).

The presence of staminodes in Opuntia polyacantha flowers was reported earlier (Cota-Sánchez et al. 2013), but a recent scanning electron microscopy investigation of regular and staminodial flowers indicate that the staminodes of this prickly pear cactus display different shapes and derive from androecial parts, specifically from fertile stamens via a series of transformations involving gradual widening (Rice et al. 2022). The purpose of this research is to further explore androecial parts in staminodial flowers of this cactus. The specific objectives were (1) to apply synchrotron technology to investigate the morphology and vascularization of regular stamens and developing staminodes of O. polyacantha, and (2) to pinpoint the putative origin of staminodes in this species. This study leads the formal micromorphological inspection of these appendages rendering a better perception about the inception and evolution of floral parts in the Cactaceae. It also provides wider understanding of the intricate vascularization network in regular stamens and staminodes and highlights the benefits of synchrotron radiation in the investigation of soft, delicate floral tissues.

Material and methods

Source of plant material

This study used flowers of O. polyacantha collected ca. 20 km south of Saskatoon, Saskatchewan, Canada, from a population near the coordinates of 51° 55′57″ N and 106° 44′11″ W. The species is native to prairie habitats and thrives on south/south-east exposed rolling hills and banks of coulees with dry, sandy soils (Cota-Sánchez 2002). The blooming season begins in mid-June and lasts for around 4–5 weeks. The flower buds emerge from the areoles, solitary or in small groups, on the marginal upper part of the cladode along the apical stem margins (Fig. 1A, B), open around mid-morning and close in the evening.

Fig. 1.

Plants of Opuntia polyacantha in the wild bearing regular (A) and staminodial (B) flowers. Note the larger flower size of the staminodial flower. C Staminodial flower showing staminode whorls (o = outer; mo = mid-outer; mid-inner; i = inner whorl(s)). D Representative staminode series in flowers of O. polyacantha. Left to right progressively arranged floral parts from the innermost androecial whorls showing normal stamens gradually changing to tepal-like staminode, and the normal tepal at the end. bf, stamens with broad filament; cs, conjoined stamens; fs, furcated staminodes (bi- and trifurcate); s, stamen; sf, separated filament; t, tepal. Scale bar in A and B = 2 cm. Scale bar in C and D = 1 cm. Photos C and D with permission of Rice et al. (2022. Braz J Bot 45:665-678)

The species bears actinomorphic hermaphroditic regular and staminodial flowers. This study utilized the staminodial type because of the unusual presence of staminode-like structures in addition to regular stamens. A representative voucher specimen (Acc. No. SASK 180957) supporting this study was deposited in the permanent collection of the W.P. Fraser Herbarium (SASK) of the University of Saskatchewan.

Preparation of stamens and staminodes for micromorphological characterization

For synchrotron scanning purposes regular stamens, staminodes and other floral organs were dissected and fixed in 70% EtOH until the color completely faded, and then the stamens and staminodes were dissected from the floral cup, working from the center (innermost) whorl outwards and placed in separate 1.5-ml Eppendorf tubes. The samples were then dehydrated according to Rice et al. (2022). Afterwards, the samples were critical point dried with liquid CO₂ (Polaron Instruments E3000) and affixed on aluminum stubs following Almeida et al. (2012).

Once the samples were prepared, the diagnostic approach included synchrotron-based phase contrast X-ray imaging or SR-μCT of stamens and tepal-like staminodes to investigate their micromorphology. Images were collected at the BMIT-BM, 05B1-1. The 05B1-1 bending magnet beamline generates X-ray spectrum between 12.6 and 40 keV (0.8–0.3 Å) with a capable beam size of 240 mm (H) × 7 mm (V). A 0.200 μm aluminum filtered white beam is used to give a peak flux around 20 KeV. A PCO Edge 5.5 (2560 × 2048 pixel) CCD detector with a 6.5-μm pixel size coupled with an Optique Peter (Lentilly) 1.8 × objective lens and YAG (yttrium aluminum garnet): Ce 200-μm thick scintillator gives an effective pixel size of 3.61 μm with a field of view of 9.24 mm² (Chen et al. 2021). A total of 20 flat and dark images were captured at the start of the scan followed by 6000 projection images over 360° while sample was offset from the centre axis of rotation by approximately 4.5 mm to roughly double the horizontal field of mode. This is known as half acquisition mode collection method.

Image processing and three-dimensional reconstructions of selected stamens and staminodes were completed with the UFO-KIT software and EZ-UFO, EZ_Stitch, and EZ MView GUIs (https://ufo.kit.edu/dis/index.php/software/) following Vogelgesan et al. (2012). Half-acquisition mode scans need to be first horizontally stitched to produce 3000 stitched projection images and 5 flat and dark images. Images are then reconstructed with phase retrieval (Willick et al. 2020). Centre of rotation was determined by dividing the stitched images pixel width by 2. Removal of ring artefacts (concentric rings in reconstructed images) was supressed using the ring removal filter (“ring removal 3”). The histogram was utilized on a test reconstruction to determine the maximum and minimum grey values in 32-bit to use for conversion to 8-bit. The following parameters were used for phase retrieved images: Beam energy = 20.0 keV, pixel size = 3.61 μm, sample to detector distance = 8 cm, δ/β = 75. The resulting 1101 images representing a separate stamen and tepaloid staminode picture consisted of multiple vertical scans stitched to gather using ImageJ. The 3D volume renderings of the reconstructed images were performed using Avizo 2020.2 (Thermo Scientific™ Amira-Avizo™ Software, Thermo Fisher Scientific, Waltham, MA, USA) to further clean the data of reconstruction artifacts and visualize the structural components of the staminode in several angles, including cross sections at different levels and a panoramic 360° view. Final image editing was performed with Adobe Photoshop CC2019, version 20.0.0.

Results

General floral morphology and staminode types

Plants of Opuntia polyacantha bear bisexual regular and staminodial flowers (Fig. 1A, B) with a perianth made of 3–5(6) series of spatulated tepals (Figs. 1A–C, 2(A)). The androecium contains numerous stamens and is organized in concentric series. The stamens develop from a meristematic ring located in the floral cup (Fig. 2(A)) and have basifixed anthers (Fig. 2(B)). In staminodial flowers, and towards the periphery, the anther filaments gradually widen and flatten becoming tepal-like structures (Fig. 1C, D; see also staminode series, upper panel in Fig. 2) bearing incipient or primordial anthers with pollen grains. The shape of staminodes is diverse and varies from linear, elongated, mucronated, bifurcated, truncated, and spatulated (Fig. 1D). These appendages may be numerous and those nearer to the periphery of the perianth are larger and wider and tend to mimic the appearance of tepals. The unilocular ovary has several ovules (Fig. 2(A)) and a stigma with several papillose lobes (Fig. 2(C)) and is typically located above the androecium, making the flowers herkogamous.

Fig. 2.

Floral attributes in Opuntia polyacantha. (A) Staminodial flower of O. polyacantha with staminode series (upper panel). (B) Micro-CT volume rendering of the upper region of stamen. (C) Micro-CT volume rendering of the stigma showing papillate lobes with pollen grains attached. pe: pericarpel; std: staminode; stm: stamen; t: tepal; ov: ovary. Scale bar in A = 1 cm. Scale bar in B and C = 0.5 mm

Reconstruction and structure of regular stamen

The 3D reconstruction of regular stamens in O. polyacantha showed that the filaments bear small external papillae dispersed all through the entire strand, the epidermis of which bears elongated cells of various densities. The surface of anther’s epidermis is irregular and papillose and beneath it is subtended by a layer of endothecium. The anther is dithecal, each theca shows two empty pollen sacs (Fig. 3(A, B)) because of pollen release. The sterile (connective) tissue between the locules of the anthers has vascular strands and acts as an extension of the filament. The distal part of the filament is constricted at the point of attachment with the anther (Figs. 2(B), and 3(A, B)). Cross-sections of the filament reveal its rounded nature lengthwise with several anatomical features consistent throughout the strand. These include a compact one-layer epidermis, a few layers of parenchyma with loose cells, large mucilage secretory cavities, and a central vascular bundle (Fig. 3(A–D)), possibly of the amphicribral type (phloem surrounding the xylem).

Fig. 3.

Micro-CT volume rendering of a stamen of Opuntia polyacantha at four different cross sections along the XY plane (A–D) showing volume rendering of the filament and anther. Labels: an: anther; co: connective tissue; en: endothecium; an: anther; ep: epidermis; f: filament; me: mesophyll; sc: secretory cells; vb: vascular bundles. Color scale indicates cell density. Scale bar for stamen and for volume rendering of cross sections = 1 mm

Reconstruction and structure of tepaloid staminode and pseudo-anther

The reconstruction of the tepaloid staminode reveals some structures not easily seen with the naked eye or optical microscopy (Fig. 4). Longitudinal view of a tepal reveals that this ordinary leaf-like structure has many superficial longitudinal stretch marks. A nascent anther, i.e., pseudo-anther (Fig. 4(A)) containing pollen grains (Fig. 5), is also observed in the apex. The internal tissues and structures of the slender tepal are elaborate and organized. There is an outermost compact uniseriate epidermis followed inwards by mesophyll composed of loose cells, several mucilage secretory ducts and lumen, and scattered vascular tissue represented by small vascular bundles (Fig. 4). These anatomical features are consistent throughout; however, the structures are relatively larger and more evident at the base of the tepal because of its thicker nature at the bottom. Overall, there is an apparent concentration of vascular tissue, possible collateral vascular bundles, that contributes to the conformation of staminodes and their parts (filaments and pseudo-anthers), among the different parts and intricate organization of tepal.

Fig. 4.

Micro-CT volume rendering of a tepal of Opuntia polyacantha with a developing staminode (dotted lines) showing forerunner structures at four different cross sections along the XY plane (A–D). Abbreviations: aw: anther wall; ep: epidermis; fp: filament precursor; me: mesophyll; pg: pollen grain; psa: pseudo-anther; sc: secretory cavity; vb: vascular bundle. Color scale indicates cell density. Scale bar for stamen and for volume rendering of cross sections = 1 mm

Fig. 5.

Micro-CT volume rendering of a pseudo-anther (psa structure in Fig. 4) originating at the margin of a tepal of Opuntia polyacantha. aw: anther wall; co: connective region; pg: pollen grains. Scale bar = 300 μm

Another obvious feature is the presence of elongated branching conformation originating from the base of the tepal and running upwards in direction to the pseudo-anther (dotted line in Fig. 4(A–D)). The pseudo-anther is subtended by a vascularized area (see cross-section scan in Fig. 4(A)). The same pattern of an elongated bundle strand without anther-like structure is evident on the opposite side forming a bifurcation near the base of the tepal (Fig. 4(B–D)). These extended elements contain relatively larger vascular bundles. It is feasible that the more robust and vascularized packets are associated with precursor areas of origin for new staminode filaments. The outermost portion of the tepal reveal the formation of staminodes evidenced by the precursor. Two additional appendages are present on the margin and seem to have continuity with a vasculature apparently leading to the origin of a new filament. The pseudo-anther is round with numerous pollen grains and subtended by a short peduncle-like (stalk) structure (Fig. 5). The pseudo-anther is a small undifferentiated protuberance bearing an anther-like structure differing in morphology and size from a regular anther. Both the pseudo-anther and filament-like structure are minute in size due to incipient development and originate from the tepal rather than the receptacle.

Finally, in the reconstruction of the stamen (Fig. 3) and the tepaloid staminode (Fig. 4), a mucilage conductive system is visible. In the former, the mucilage cavities are arranged in a circular fashion, located closer to the epidermis in the upper part of the filament (Fig. 3(A–B)) and below the 2–4 parenchyma layer at the base of the filament (Fig. 3(C–D)). In the tepaloid staminode, the mucilage cavities are intertwined with vascular bundles. However, in the region suggested as the “precursor” of a filament (Fig. 4(D)) and the sporangium (Fig. 4(A–B)), we identified cellular regions with different densities resulting from the morphological organization of the staminode that is clearly connected to the pseudo-anther bearing connective tissues and pollen grains (Fig. 5). In addition, the number of cellular strata, including secretory cells, is higher in the specified region, indicating that cell density ranges from medium to high and the increasing layering number in a circular fashion may be associated with the developing filament initiation or precursor structures.

Discussion

This investigation demonstrates one of the greatest advantages of synchrotron radiation to unveil structures not perceived with conventional microscopic techniques. It also offers substantial advances in terms of precision regarding the anatomical organization and different cellular components in stamens and staminodes and revealing cryptic floral features.

Vasculature in stamens and tepaloid staminode

Spatial-resolution of synchrotron X-ray-based tomography images was high enough to accurately investigate the complex anatomical structure of tissues in stamens and tepaloid staminodes. These include presence of epidermis, vascular bundle elements, parenchyma/mesophyll, and secretory cells. There seems to be a uniform anatomical pattern in the organization of the vascular system in the androecial structures of the Cactaceae. Similar anatomical features as those described in this study, i.e., tepals with several small collateral bundles and stamens with more amount of vascular tissue made of xylem as phloem, have also been observed in terrestrial and epiphytic lineages, of the Cactaceae, specifically in species of Cereus (Silva 2020), Opuntia (Fuentes-Peres et al. 2009), Epiphyllum phyllanthus (Almeida et al. 2010), Hatiora, and Rhipsalis (Almeida et al. unpub.). Likewise, the tepals in Opuntia display several vascular bundles derived from a central group that differs in size from the remaining bundles (Rosas-Reinhold et al. 2021); thereby, the central bundle is larger than the others and comparable to vascular bundles in leaves. Thus, it makes sense that the filament’s vascular bundle is more complex than the single strand inside the tepals. Similar organization of vasculature has been observed in tepals of Opuntia (Fuentes-Perez et al. 2009) and E. phyllanthus (Almeida et al. 2010). Finally, the digital reconstructions throughout the entire stamen and tepal in cross-section reveal small vascular bundles, possibly of the collateral type, i.e., strand of phloem present external to the strand of xylem on the same radius side by side. These are distributed across the tepal and towards the pseudo-anther. See also Supplemental videos 1, 2, and 3 showing the slow-motion examination of the stamen and tepal with pseudo-anther.

On the other hand, secretory cells (idioblasts) of mucilaginous material are common in the Cactaceae but little is known about the arrangement of this conductive network in the family. The characterization of the complex vascular networks and histo-localization of mucilage channels in floral parts (stamens and tepals) of O. polyacantha is important because these elements lack stomata, but the presence of mucilage ducts suggests that this conductive system plays a role in secretion of mucopolysaccharides in and around the flowers. In fact, in most cacti, mucilage is abundant, especially is stem, with lesser extent in floral appendages, i.e., perianth parts (Mauseth 1983). As suggested by Fahn (2000), external flux attracts floral visitors and promotes pollination, because it is in the idioblast cells where mucilage is commonly produced and then released into the apoplast (Trachtenberg and Fahn 1981, reviewed in Nobel et al. 1992). Another important role of mucilage is its complex hydrophilic nature made of proteins and mucopolysaccharides responsible to hydrate tissues and protecting the organs from drought (Terrazas and Mauseth 2002) and facilitate water transport (Nobel et al. 1992), two important aspects to maintain flower turgor during anthesis. We suspect that the same method of apoplastic mucilage discharge by the secretory cells in stamen filaments and tepals may also function in water transport and maintenance of turgor in all floral whorls of this species.

Staminode origin in Opuntia polyacantha

Developmental studies in angiosperms indicate that staminodes originate from different floral parts (Walker-Larsen and Harder 2000). These floral appendages have been researched in several species of the Ranunculaceae. For instance, in Aquilegia, the foundation of staminodes is accompanied by physical interaction along the lateral margin of neighboring organs facilitating their amalgamation (Meaders et al. 2020). In turn, Li et al. (2021) classified the stamens of Clematis macropetala into four forms based on shape and anther size, specifically tepaloid staminode (St₁), spatulate staminode (St₂), linear-spatulate fertile stamen (St₃), and linear fertile stamen (St₄), all of which share similar early development but gradually differentiate in the later, final stage. Among these, the first two forms can be included within the context of this discussion because St₁ has tardy development and lacks anther differentiation (as the O. polyacantha tepaloid staminode) whereas St₂ develops abnormally in early stages of anther differentiation and bears a primordium of an anther, which also evokes the incipient pseudo-anther reported by Rice et al. (2022) and further portrayed here (Figs. 4 and 5A). The morphological analogies of Li et al. (2021) St₁ and St₁ and those of the tepaloid staminodes in O. polyacantha suggest similar ontogenetic progression in the inception of these structures.

Several studies, e.g., Ronse De Craene and Smets (2001), Hufford (2003), Endress and Mathews (2006), Botnaru and Schenk (2019), Meaders et al. (2020), and Li et al. (2021), are useful in the understanding of staminode structure, origin, and development in flowering plants. However, in-depth descriptions of staminodes are limited in the cactus family, except for that of Rice et al. (2022) and several general reports of these appendages in unisexual flowers of several cacti. In the Cactaceae, the stamens are initiated in centrifugal progression from a primary ring in the floral cup (Ross 1982) as demonstrated in the conspicuous androecia of Disocactus (Buxbaum 1953) and Schlumbergera truncata (Apolônio 2022). Recent developmental studies in the family report non-vascularized staminodes originating in the upper portion of the nectar chamber of Denmoza rhodacantha (González et al. 2021), a major difference from the rather vascularized staminodes and tepals seen in O. polyacantha in this study.

In members of the Hylocereeae and Rhipsalideae, specifically in Selenicereus setaceus and Schlumbergera truncata, the vascularization system is also made of bicolateral vascular bundles distributed throughout the floral nectary and the androecium, areas that lack staminodes (Apolônio 2022; Sampaio and Almeida, unpub. data). Thus, the conductive tissue extends to the subnectariferous parenchyma vascularizing every stamen filament in the androecium as well as the adjacent tissues (nectary, floral tube, and androecium), merging in a common vascular network. Conversely, the vasculature in the floral nectary of D. rhodacantha exhibits traces of branching xylem and phloem ducts infiltrating the nectariferous parenchyma (González et al. 2021). These contradictory data suggest that additional investigations in other groups of the Cactaceae are required to fully understand the organization of the vascular system and the presence of a common or different points of origin for these structures.

The morphological foundations for the existence of diverse shapes of tepals and staminodes in O. polyacantha may be due to the intricate aggregation of several organs (Rice et al. 2022), which is also the case in several angiosperms (Endress and Matthews 2006). Because of the possible involvement of more than one floral series in the origin of staminodes of O. polyacantha, it is reasonable to consider the ABC model of floral development in angiosperms. This model postulates that B-class genes control the formation of petals and stamens (Coen and Meyerowitz 1991). In the Cactaceae, the perianth is multiseriate and derived from bracts and sepal structures (Ronse De Craene 2013). In Opuntia, the flowers have a succulent and thick pericarpel (receptacular tissue) and the foliar organs become tepaloid towards the apex (Rosas-Reinhold et al. 2021). Thus, the perianth, which displays a progression from bracts to petal-like appendages, lacks the typical demarcation of a green calix and corolla (Rice et al. 2022); therefore, sepals and petals cannot be distinguished based on morphology and pigmentation (Hofmann 1994). Similarly, the staminodes in O. polyacantha have blurred boundaries between tepals and stamens; consequently, they are deemed transitional organs deriving from fertile stamens along the androecial margins via a series of transformations involving gradual widening. Furthermore, according to Rice et al. (2022), the fusion of floral whorls gives rise to the amalgamation of androecial and perianth structures, suggesting that the O. polyacantha staminodes are transitional because show a mixture of characteristics, supporting homeosis rather than heterotopy, and idea also in agreement with Ronse De Craene (2003).

Considering the blurred boundaries of tepals and the undefined nature of staminodial appendages conjoined to tepals unveiled by synchrotron images, we propose that staminodes originate from tepals because this development scheme supports the fading border model of floral organ identity for angiosperms. This archetypal is characterized by a gradual transition from bracts to tepals, from outer to inner tepals, and from tepals to stamens (Soltis et al. 2007). Nonetheless, in view of the lack of developmental and genetic studies, we can only speculate that staminode origin in O. polyacantha is controlled by cascading transcription factors of the ABC genes in the fading border model. The evo-devo approach seems quite suitable to unveil the inherent properties dealing with the origin of staminodes in this species in future investigations.

Concluding remarks

Synchrotron micro-computed tomography goes beyond the useful external description of staminodes in O. polyacantha by providing a comprehensive look at their micromorphology through tomographic imaging methods. The beamline allowed the visualization and imaging of novel cryptic plant parts useful to pinpoint the putative homeotic process for the origin and initiation of staminodes. The rather simple sample manipulation and characterization of the putative floral innovations further demonstrate the value and extraordinary resolving capability of this innovative non-destructive synchrotron technology to quickly investigate plant structures without the use of toxic or invasive agents in specimen preparation.

Author contribution

JHCS conceived the initial concept and design. JAS set up beamline and led the synchrotron image and video capturing and reconstruction along with DJF. JAS and RVV assisted with data collection and analyses. OJGA, RSR, and JHCS designed final figures. All authors participated in sample preparation, contributed ideas, writing, discussions, and editing the final manuscript.

Funding

This study received financial support from the National Geographic Society (Grant No. 7382-02) and the University of Saskatchewan Tri-Council Bridge (Grant No. 411051) to JHCS and from the São Paulo Research Foundation (FAPESP - Grant No. 2018/19634-3) and the Brazilian National Council for Scientific and Technological Development (CNPq-Grant No. 423273/2018-3) to OJGA.

Declarations

Conflict of interest

The authors declare no competing interests.