Pineapple SWEET10 is a glucose transporter

Abstract SWEET transporters are a unique class of sugar transporters that play vital roles in various developmental and physiological processes in plants. While the functions of SWEETs have been well established in model plants such as Arabidopsis, their functions in economically important fruit crops like pineapple have not been well studied. Here we aimed to investigate the substrate specificity of pineapple SWEETs by comparing the protein sequences of known glucose and sucrose transporters in Arabidopsis with those in pineapple. Our genome-wide approach and 3D structure comparison showed that the Arabidopsis SWEET8 homolog in pineapple, AcSWEET10, shares similar sequences and protein properties responsible for glucose transport. To determine the functional conservation of AcSWEET10, we tested its ability to complement glucose transport mutants in yeast and analyzed its expression in stamens and impact on the microspore phenotype and seed set in transgenic Arabidopsis. The results showed that AcSWEET10 is functionally equivalent to AtSWEET8 and plays a critical role in regulating microspore formation through the regulation of the Callose synthase5 (CalS5), which highlights the importance of SWEET transporters in pineapple. This information could have important implications for improving fruit crop yield and quality by manipulating SWEET transporter activity.


Introduction
Sugar is critical for the lives of all living organisms as it provides energy and carbon atoms for metabolic processes.Plants produce their sugar through photosynthesis, but the distribution of sugar from source to sink, sink to cell, and cell to cell requires specific sugar transporters [1].Multiple sugar transporters control the mobility and availability of sugar content in plants during their growth and development.To date, several families of sugar transporters have been discovered in plants, including sucrose uptake transporters (SUTs/SUCs), sugar transport proteins (STP) and SWEETs (sugars will eventually be exported transporters) [2,3].SWEETs are the most recently discovered and functionally diversified sugar transporters.SWEETs, being evolutionarily conserved, have been identified in a wide range of plant genomes, including model species like Arabidopsis, as well as economically important crops such as rice, maize, and wheat [4,5].Notably, these transporters act as uniporters and exhibit pH-independent activity under in vivo conditions [6,7].Interestingly, their subcellular localizations have been observed in the plasma membrane [7][8][9], vacuolar membrane (tonoplast), and Golgi membranes [8,10,11].Such diversification of SWEET transporter localization in eukaryotes or plants differs from prokaryotes.As prokaryotes lack membrane-bound organelles, the diversification of SWEET localization provides a prime example of eukaryotic evolution.
Additionally, the evolution of prokaryotic SWEETs (semiSWEETs) into eukaryotic SWEETs is thought to have arisen through gene duplication [12].
To date, significant progress has been made in understanding the transport route and mechanism of SWEET transporters through the use of crystal structures of semiSWEETs, which served as a valuable foundation for the structural studies of SWEETs in eukaryotes [13][14][15].Structurally, eukaryotic SWEETs are composed of two triple helix bundle (THB) domains (TM1-TM2-TM3 and TM5-TM6-TM7), each of which is derived from semiSWEETs and separated by a linker helix (TM4).The Nterminal and C-terminal THBs share a similar sequence and are arranged in a parallel orientation [7,15].The distinct conformations of the structure allow them to be uniporters to access substrates (sugar molecules) from either the extracellular or cytosolic environments [7,16,17].The substrate-binding pocket is located above the center of the transmembrane region of the transporter protein.Computational analysis has indicated that disaccharides, such as sucrose, are able to easily pass through larger pockets.At the same time, monosaccharides, such as glucose, due to their smaller size, can escape even if the pocket is partially open.However, extracellular residues of the transporter play a significant role in substrate recognition [18,19].
Phylogenetic tree of SWEET proteins from pineapple (Ac), Arabidopsis (At), rice (Os), grape (Vv), and maize (Zm).The distances in the tree were calculated from multiple sequence alignment (CLC software) using the maximum likelihood method.Bootstrap values are displayed in percentages.Arabidopsis SWEET8 protein is highlighted with red color and pineapple SWEET6/8/10 are represented with blue color.
The evolution of SWEET transporters provides valuable insights into their substrate preferences.Within the plant genome, ∼20 paralogs of the SWEET gene family have been identified, and they are commonly categorized into four distinct clades: clades I, II, III, and IV.This classification scheme helps to elucidate the evolutionary relationships and functional diversity among SWEET transporters.Interestingly, the preference for sucrose and hexoses of SWEETs appears to be correlated with phylogeny; for example, SWEETs from clade III mediate sucrose transport.Functional characterization of SWEET transporters in Arabidopsis has revealed their crucial roles in various processes, such as seed filling [8], nectar production [20], and pollen nutrition [21,22].
While significant progress has been made in understanding the function, localization, and substrate specificity of SWEET transporters in Arabidopsis, our knowledge of the vastly diverse SWEET transporters in the plant kingdom remains limited.In this study, we aim to expand the knowledge base of SWEET proteins by exploring the available information on SWEETs in pineapple, an economically important plant for which there is currently limited information.Instead of relying solely on linear protein sequences, we employed transmembrane topology and AlphaFold to decipher the structural and physiological roles of SWEETs in pineapple.This study combines 3D structure-based transporter screening with an evolutionary perspective to gain insights into the substrate specificity of SWEET transporters.

Three-dimensional structure-based similarity among Arabidopsis and pineapple SWEET transporters
We first studied the close relatedness of SWEET transporters by constructing a phylogenetic tree using existing information from rice, grape, maize, Arabidopsis, and pineapple.Most pineapple SWEETs were found to be closely associated with their corresponding SWEETs from Arabidopsis, forming clades I, II, III, and IV (Fig. 1).The average sequence identity of pineapple SWEETs was found to be 50%.Notably, the C-termini of the pineapple SWEET proteins displayed less conservation and exhibited varying sequence lengths (Supplementary Data Fig.S1).To determine their potential target sugar for transport, the pineapple  The alignment includes the protein sequences of AtSWEET8, a glucose transporter, and AtSWEET13, a sucrose transporter, as references.The numbers displayed at the top of the table correspond to the residues of AtSWEET8, while the ones below represent the residues of AtSWEET13.The four conserved residues among di-and monosaccharide-specific SWEETs are highlighted in grey and magenta, respectively.b 3D structures of SWEET proteins.In the figure, the numbers i, ii, iii, and iv represent AtSWEET8, AcSWEET6, AcSWEET8, and AcSWEET10 proteins, respectively.The numbers on transmembrane helices (TMs) are represented as TM1 to TM7.TM4 (linker helix) and the C-terminal tails are represented in blue and red colors, respectively.
We then selected three pineapple SWEET proteins (AcSWEET6, 8, and 10) from clade II for comprehensive structural and functional characterization.Our choice was based on their similarity to the well-characterized AtSWEET8 protein.Moreover, the selection of these three pineapple SWEET proteins was also due to their phylogenetic relationship, as well as their similarity in physiochemical properties and predicted subcellular localization (Supplementary Data Table S2) in relation to AtSWEET8.Given the comparable nature of pineapple SWEET6, 8, and 10 to AtSWEET8, we proposed that they might facilitate glucose transport in a manner similar to the SWEET8 protein in Arabidopsis.This hypothesis is supported by the conserved functions observed in SWEET proteins across various plant species.

Screening of glucose transporter among AcSWEETs via homology-based search
The observed similarities between AtSWEET8 and AcSWEET10, as well as the differences from AcSWEET6 and 8, indicate that AcSWEET10 has closer resemblance to AtSWEET8 and may exhibit a greater affinity for glucose transport compared with AcSWEET6 and 8 (Supplementary Data Fig.S2).An analysis of the 3D structures of selected pineapple SWEET transporters in reference to the AtSWEET8 transporter could enable the prediction of the substrates transported by them and provide insights into their role in facilitating sugar transport within the plant.Therefore, to determine whether AcSWEET6, 8, and 10 would transport glucose, the 3D structures of AtSWEET8, AcSWEET6, 8, and 10 were predicted using the AlphaFold server.The structures indicated that the proteins have a common structural arrangement with seven transmembrane segments organized into two units, connected by an inversion linker, TM4.The structural arrangements of AcSWEET6, 8, and 10 were in line with previous findings in rice and Arabidopsis (Fig. 2b).Despite the similarities, significant variations were observed in the C-terminal tails of AcSWEET6 and 8.For instance, AcSWEET8 had two additional β-sheets compared with AtSWEET8, while AcSWEET6 had an α-helix instead of β-sheets.In contrast, AcSWEET10 was structurally similar to AtSWEET8 (Fig. 2b).These variations in the 3D structures of AcSWEET6 and AcSWEET8 could play a role in determining substrate specificity and transport activity and may affect the glucose transport activity of AcSWEET6 and AcSWEET8.Altogether, the results of substrate specificity and comparative structural analysis indicate that only AcSWEET10 shares a close structural similarity with AtSWEET8, suggesting that it has the highest affinity for glucose transport among AcSWEET6, 8, and 10.

AcSWEET6, AcSWEET8, and AcSWEET10 localize to the plasma membrane
The substrate specificity and comparative structural analysis indicated that AcSWEET6, 8, and 10 could transport glucose, so we checked their subcellular localization to confirm the putative functional location.The subcellular localizations of AcSWEET6, 8, and 10 were identified using AcSWEET6, 8, and 10 GFP fusion proteins in tobacco epidermal cells (Fig. 3).The GFP signals were compared with the earlier reported plasma membrane-localized AtSWEET8-GFP [7,22] and the GFP control vector (35S::GFP), which was distributed throughout the cells.The f luorescence signals for AcSWEET6, 8, and 10 were observed in the plasma membrane (Fig. 3).The localization results also indicated that the SWEETs show conservation for subcellular localization.

Expression of AcSWEET genes during pineapple pollen development
The plasma membrane-specific cellular localization of AcSWEET6, 8, and 10 resembled that of the well-characterized AtSWEET8.Subsequently, we examined the organ-specific expression of AtSWEET8, which was observed in the stamen (Supplementary Data Fig.S3).Nevertheless, the localization of AcSWEETs during developmental stages of stamens remained unknown.To address this gap, we analyzed the expression of AcSWEETs using RNAseq data from six developmental stages (St1-St6) of the stamen (Fig. 4a and b).The results showed that 66% of the AcSWEET genes were upregulated during all the developmental stages of stamens.Among them, AcSWEET6, AcSWEET8, and AcSWEET10 expressions were detected during pineapple stamen development.The expression patterns of AcSWEET6, 8, and 10 were further validated using qRT-PCR in the developing stamen (Fig. 4c).Altogether, these results suggest that pineapple SWEET transporters are similar to AtSWEET8 not only in terms of 3D structure and cellular localization but also of tissue-specific expression.

Pineapple SWEET10 transports glucose
Although the homology-based identical pineapple SWEET transporters have comparable cellular localization and tissue-specific expression, the question remained whether pineapple SWEET transporters are capable of glucose transport.Due to technical difficulties with pineapple genetic modification, it is almost impossible to knock out each glucose transporter in the plant.As an alternative and the most reliable approach, we took advantage of the hexose transport-deficient yeast mutant EBY.VW4000 and performed the glucose transport assay in vivo.To achieve this, expression vectors harboring the coding sequences of AcSWEET6, AcSWEET8, and AcSWEET10 were transformed into EBY.VW4000.Transformants were allowed to grow in synthetic deficient (SD/−Trp) medium supplemented with maltose (as growth control) and glucose.The yeast complementation assay showed that AcSWEET10 complements the mutant strain and grows well on glucose-supplemented medium similarly to AtSWEET8, while AcSWEET6 grows weakly.However, the yeast carrying AcSWEET8 and the vector (without insert) did not show detectable growth on medium rich in glucose (Fig. 5).These heterologous transport assays highlight that AcSWEET10 has a maximum affinity to transport glucose compared with AcSWEET6 and 8.
To investigate the glucose transport ability of AcSWEET10 in more detail, we conducted site-directed mutagenesis experiments targeting conserved amino acids.Specifically, we focused on the predicted significance of Y60, Y183, and N196 residues in AcSWEET10, as well as Y60, Y185, and N199 residues in AtSWEET8 (Supplementary Data Fig.S5).These amino acids were substituted with alanine residues, following the conservation pattern observed in OsSWEET2b, a well-known transporter with established glucose transport capabilities [18].To evaluate the impact of these mutations, we expressed the mutated versions of both AtSWEET8 and AcSWEET10 in the EBY.VW4000 yeast strain.As anticipated, yeast cells expressing the mutated proteins of AtSWEET8 and AcSWEET10 exhibited normal growth on maltosesupplemented media.However, when cultivated on glucosesupplemented plates, the yeast cells expressing the mutated proteins displayed significantly reduced growth, indicating a defect in glucose transport (Fig. 6).These findings highlight the critical role of the substituted amino acids, Y60, Y183, and N196, in facilitating glucose transport.Moreover, they suggest that these specific amino acids are likely involved in recognizing and transporting glucose across the pineapple cell membrane.

AcSWEET10 functions as a glucose transporter in Arabidopsis
To investigate whether the AcSWEET10 of pineapple has a similar role in anther development, ectopic expression of AcSWEET6, 8, and 10 in the Atsweet8 mutant was studied to determine their complementation ability.In agreement with our substrate specificity and comparative structural analysis results, AcSWEET6 and 8 failed to complement the mutant phenotype.This was evident from pollen viability analysis using Alexander staining, which showed that most of the pollen in these lines was non-viable (Fig. 7a and b).The percentages of pollen abortion and viable seed numbers in these lines were similar to those in the mutant (Fig. 7c and d In contrast, the ectopic expression of AcSWEET10 was able to restore the defective pollen phenotype of the Atsweet8 mutant.AcSWEET10 expression fully rescued the low-fertility phenotype of the mutant plants.The complemented lines displayed a significant improvement in silique development and seed set (Fig. 8a and b), with the viable seed number nearly identical to that of the wild type (Fig. 8c) and a significant reduction in the pollen abortion rate in the complemented lines (Fig. 8d).The anthers showed viable pollen in complemented lines, similar to the wild type (Fig. 8e).Alexander staining showed that most of the pollen in the AcSWEET10 lines was viable, resulting in a restoration of the mutant phenotype (Fig. 8f).
In a previous study, mutation of Atsweet8 significantly reduced the expression of CalS5 (key enzyme for callose biosynthesis), resulting in the thinning of the callose wall of the Atsweet8 microspore [22,24].To investigate whether fertility restoration by AcSWEET10 is associated with callose deposition, we checked the deposition of callose and CalS5 expression in the wild-type, mutant, and complemented lines.DIC microscopy showed an improved callose wall around developing microspores at the tetrad stage in complemented lines compared with thinner, degenerated walls around mutant microspores (Fig. 9a).
Callose f luorescence was comparable in wild-type and complemented lines in contrast to Atsweet8 (Fig. 9b and d).The percentage of defective tetrads was significantly reduced to up to 20% in complemented lines (Fig. 9c).The relative expression of CalS5 in complemented lines was similar to that of wild type, in contrast to the mutant (Fig. 9e).Taken together, these results suggest that the abnormal callose deposition related to the reduced fertility in the Atsweet8 mutant is complemented by AcSWEET10, indicating that AcSWEET10 has a similar function to SWEET8 in pineapple.

Discussion
The SWEET gene family is recognized for its pivotal role in the facilitation of sugar transport across cellular membranes, a function critical for various physiological and developmental processes in plants [5,21,[25][26][27][28][29][30].Extensive investigations into the SWEET family in model plants such as Arabidopsis and rice have yielded valuable insights into their structural and functional characteristics.In Arabidopsis, the 17-member SWEET family is categorized into four clades, each associated with distinct sugar transport functions, contributing to our understanding of sugar allocation and homeostasis within plants [23].In general, clade III SWEET proteins tend to transport sucrose, whereas clades I, II, and IV exhibit a preference for hexose transport [31].Drawing inspiration from this knowledge base, our study sought to extrapolate these insights to the pineapple (Ananas comosus) SWEET gene family, specifically aiming to predict and validate putative glucose transporters within this tropical fruit crop.While Arabidopsis and pineapple represent divergent botanical lineages, we leveraged the conserved principles of protein structure and function to bridge this evolutionary gap.Through a comprehensive approach that integrated structural analysis and functional assays, we elucidated potential glucose transporters among the pineapple SWEET proteins, thereby contributing novel perspectives to the broader understanding of sugar transport mechanisms in plants.Finally, we validated these predictions by exploring the glucose transport activity of AcSWEET6, 8, and 10 in a yeast transport assay and functionally complementing the Arabidopsis sweet8 mutant.Previous studies on SWEET structures have provided important information about their putative function in various plants [7,22,32].SWEETs have a single sugar-binding site within the protein and undergo conformational changes to transport the sugar [19,33].By analyzing homologous amino acid sequences and the amino acid configuration of the binding pocket, we identified a set of conserved residues that were indicative of glucose or sucrose transporters (Fig. 2a and Supplementary Data Fig.S1).Based on these conserved residues, we predicted five sucrose transporters (AcSWEET11, 12, 13, 15, and 16) and nine glucose transporters (AcSWEET1, 2, 3, 5, 6, 7, 8, 9, and 10) in pineapple (Fig. 2a).
Arabidopsis SWEET proteins play critical roles in the transport of glucose directly or in the form of sucrose between compartments, cells, and organs [7,24,31].AtSWEET8 is a wellknown plasma membrane glucose transporter, and its loss of function leads to severely defective pollen and a reduction of up to 90% in seed set [22]; therefore, it was utilized as the reference for this study.Clade III members (AcSWEET6, 8 and 10) displayed high similarity in their physicochemical properties to AtSWEET8 and localize to the plasma membrane (Fig. 3).The glu-cose molecule follows a specific path by interacting with residues across the SWEET8 cavity [18].Notably, the same residues for glucose entry and exit in the AcSWEET6, 8 and 10 structures indicated their potential for glucose transport activity (Supplementary Data Fig.S1).Moreover, similar to AtSWEET8, the high expression of AcSWEET6, 8, and 10 in pineapple stamens showed their potential role in sugar supply in male reproductive organs (Fig. 4a and c and Supplementary Data Fig.S2).
Experimental verification of the putative transporters using multiple hexose transporter-deficient yeast showed that only AcSWEET10 could enable robust growth on glucose (Figs 5  and 6).Notably, AcSWEET10 might have retained a conserved structure that facilitates glucose transport, as evidenced by both its preserved functional characteristics and its ability to rescue the growth of the Atsweet8 mutant, which has impaired sugar transport (Figs 2b and 8).Interestingly, the contrasting behavior of AcSWEET6 and AcSWEET8 is intriguing.Despite sharing identical conserved residues, both proteins failed to transport glucose efficiently and were unable to complement the sugar transport deficiency in the mutation.This disparity underscores the critical role of the C-terminal region in determining their functionality (Fig. 7).
Previous reports on rice SWEET2b have revealed the functional significance of specific residues in glucose transport.One such residue, Y61 of TM2, is considered crucial as its mutation disrupts the extra-facial gate of the transporter, thereby impeding the conformational transitions necessary for glucose movement.Further studies by Selvam et al. [18] revealed that Y184 of TM7 in OsSWEET2b, located at the extracellular side, establishes polar and hydrophobic interactions with glucose, thereby facilitating glucose transport towards the center of the transporter.Moreover, they demonstrated that N190 of TM7 in OsSWEET2b is involved in strong polar interactions with other residues, such as Y61 of TM2, enabling the closure of the transporter in an occluded conformation [15,18].We investigated the roles of conserved residues in AcSWEET10 by conducting site-directed mutagenesis experiments.These experiments targeted residues corresponding to key residues in OsSWEET2b that are crucial for glucose recognition and transport [13,34].Specifically, we conducted site-directed mutagenesis on Y60 of TM2 and Y183 of TM7 of AcSWEET10, which align with Y61 of TM2 and Y184 of TM7 of OsSWEET2b, respectively.The results of the mutagenesis experiments on AcSWEET10 were consistent with predictions from the computational analysis, validating the importance of these residues in glucose recognition and transport.Our experimental findings clearly demonstrated that these substitutions compromised the glucose transport activity of both AcSWEET10 and AtSWEET8 (Fig. 6 and Supplementary Data Fig.S5).These findings provide additional evidence supporting the key roles of these conserved residues in recognizing and delivering glucose molecules.Overall, in combination with computational analyses and previous research, our results underscore the significance of the conserved residues, such as Y60 of TM2, Y183 of TM7, and N196 of TM7, in AcSWEET10, in facilitating glucose transport.The 3D structure of AcSWEET6, 8, and 10 proteins using the AlphaFold algorithm showed high similarity overall, but they differed significantly at the C-terminal region (Fig. 2b).The predicted structures suggest that the C-terminal region of AcSWEET6 forms an extra α-helix, which is absent in AcSWEET8 and 10.In addition, the C-terminal region of AcSWEET8 possessed an additional β-sheet which changed the structure leading to hindrance of the pore (Fig. 2b).Previously, it was shown that three mutant alleles of the Atsweet8 gene (Salk_142803, Salk_062567, and Salk_092239) had a T-DNA insertion in the first exon, fifth exon, and first intron, respectively (Supplementary Data Fig.S4).However, the stronger defect in phenotype pollen of the AtSWEET8 gene was observed in the rpg1 mutant, a T-DNA insertion in the last intron.Consequently, in RT-PCR of the Atsweet8 mutant, the T-DNA insertion primarily affected the transcription of the last exon, which corresponds to the C-terminal tail of its protein [22].Altogether this suggests that the C-terminal region of SWEET8 plays an essential role in its function.Besides, several studies indicate the significance of C-terminal regions for the functions of different transporters [35][36][37].The structural difference of the AcSWEET6 and 8 at the C-terminal compared with AcSWEET10 and AtSWEET8 could impact the glucose transport activity of the proteins AcSWEET6 and 8. Consistently, the present study revealed that AcSWEET10 has a structure similar to AtSWEET8 and displays higher glucose transport activity than SWEET6 and 8 (Figs 2b and 5), most likely due to the difference in their Cterminal region, which may be hindering sugar transport.
To summarize, this research has highlighted the significance of the C-terminal region for the appropriate functioning of SWEETs.The results of the study revealed that AcSWEET10 has a conserved function and imitates the overall structure of AtSWEET8, implying that residues in the C-terminal region can have an impact on transporter activity (Fig. 10 and Supplementary Data Fig.S6).
Further investigation is needed to explore the C-terminal regions of SWEET transporters and the possible presence of motifs interacting with glucose, particularly in the case of AcSWEET6 and 8. Overall, these findings have important implications for understanding plant physiology and metabolism and developing strategies to improve crop yield and quality.Further investigation in this area could develop novel tools for manipulating sugar transport in plants, which could have far-reaching benefits for agriculture and food security.

Homology-based modeling
The structure models of pineapple SWEET proteins in inwardfacing states were built from AlphaFold [41] and were visualized and analyzed using the CLC genomics workbench (version 22.0.2).The models were compared by using available structural templates: AtSWEET8 and AtSWEET13.

Plant materials
Arabidopsis mutant (Atsweet8) was grown in a growth chamber in potted soil under a 16-h light/8-h dark regime at 22 ± 2 • C. Transformation of Arabidopsis plants was done with Agrobacterium strain GV3101 using the f loral dip method [42].Pineapple plants (MD2 variety) were acclimated in soil mix [peat moss:perlite = 2:1 (v/v)] in plastic pots in a walk-in growth chamber.The growth chamber was maintained at 25 ± 2 • C with a 16-h light/8-h dark photoperiod and 70% humidity, as reported earlier [43].The pineapple anthers were collected and photographed at six developmental stages as described before [44], quickly frozen in liquid nitrogen, and stored at −80 • C until RNA extraction.

Total RNA isolation and quantitative-real time PCR analysis
Total RNA was isolated using the RNeasy kit (Qiagen, MD, USA) followed by treatment with DNaseI (Thermo Fisher Scientific, CA, USA), and was reverse-transcribed using the ThermoScript RT-PCR kit (Thermo Fisher Scientific, CA, USA).The qRT-PCR reactions were set up with FastStart DNA Master SYBR Green I master mix (Takara, Japan).For each analysis, two technical replicates from three biological replicates were taken, and pineapple and Arabidopsis EF1α genes were used to normalize the mRNA levels.Finally, the fold change of genes was calculated using the 2 − CT method [45].Primers used in this study are listed in Table S3.

RNA-seq analysis
Transcriptome data of pineapple stamen developmental stages were used to investigate the expression level of SWEET genes [44].Brief ly, the sequencing reads were aligned to the pineapple genome using TopHat v2.1.1 with default parameters.Transcript abundance was calculated as fragments per kilobase of exon model per million mapped reads (FPKM).The heat map was generated using pheatmap R software based on log 2 (FPKM + 0.01).

Pollen viability analysis
Anthers were dissected from f lower buds at stage 12 (before anthesis) and fixed in Carnoy fixative (absolute ethanol: chloroform:acetic acid, 6:3:1) for at least 2 h.Afterwards, anthers were stained in Alexander stain [47].For aniline staining, anthers at the tetrad stage were gently fixed overnight in FAA.The anthers were then stained in 0.1% (w/v) aniline blue in 0.1 M sodium phosphate (pH 9.0) and incubated in the dark for 1-2 h.Stained anthers were mounted on 30% glycerol and viewed with UV epif luorescence (365 nm excitation and a 420 nm long-pass emission filter).

Functional complementation of AcSWEETs in Arabidopsis (sweet8) mutant
PCR fragments of the complete AcSWEET6, AcSWEET8, and AcSWEET10 coding sequences were amplified from f lowerspecific cDNA using specific primers.The fragments were cloned into the entry vector pENTR™/D-TOPO ® and then subcloned into the Gateway destination vector pGWB505 [48] using LR Clonase II (Invitrogen).After confirmation by sequencing, the constructs were transformed into Agrobacterium and finally into homozygous Atsweet8/rpg1 Arabidopsis plants using the f loral dip method.
The transgenic plants were selected on media plates containing 50 mg L −1 hygromycin.

Glucose transport assay in yeast
For the yeast transport assays, the Saccharomyces cerevisiae strain EBY.VW4000 [49] was used, as described previously [28].The ORFs of AcSWEET6, AcSWEET8, and AcSWEET10 were amplified using specific primer combinations.To investigate the impact of specific mutations on the glucose transport ability of AcSWEET10, sitedirected mutagenesis was performed at three different positions (Y60A, Y183A, and N196A).As a reference for the site-directed mutagenesis experiment, Arabidopsis SWEET8 was also mutated at three similar amino acids (Y60A, Y185A, and N199A), which have been previously proposed to be responsible for glucose transport [18].Mutagenic primers were designed and PCR-based sitedirected mutagenesis was carried out using high-fidelity DNA polymerase.The amplified fragments were then infused into the modified yeast expression vector pGBKT7 and sequenced to verify the correctness of the clones.The expression clones and an empty vector (without insert) were transformed into the yeast strain EBY.VW4000.Yeast transformants were then selected on a synthetic deficient medium without tryptophan (SD/−Trp) supplemented with 2% (w/v) maltose as a carbon source.Transformed yeast cells were grown in SD/−Trp liquid medium with 2% (w/v) maltose and were incubated overnight at 30 • C until the optical density at 600 nm (OD 600 ) reached 0.5.The adjusted OD 600 (∼0.2) with water was serially diluted (×10, ×100, and ×1000) and spotted onto SD/−Trp solid medium with 2% (w/v) maltose (control) and 2% (w/v) glucose.All transformants were incubated at 30 • C, and growth was documented after 3 days in maltose media and 3-5 days in glucose media.

C
l a d e Ⅱ ( H e x o s e t r a n s p o r t e r s ) C l a d e Ⅰ ( H e x o s e t r a n s p o r t e r s )

Figure 2 .
Figure 2. a Sequence alignment of the substrate-binding pocket residues in SWEET proteins of pineapple (adapted from the work of Han et al.[23]).The alignment includes the protein sequences of AtSWEET8, a glucose transporter, and AtSWEET13, a sucrose transporter, as references.The numbers displayed at the top of the table correspond to the residues of AtSWEET8, while the ones below represent the residues of AtSWEET13.The four conserved residues among di-and monosaccharide-specific SWEETs are highlighted in grey and magenta, respectively.b 3D structures of SWEET proteins.In the figure, the numbers i, ii, iii, and iv represent AtSWEET8, AcSWEET6, AcSWEET8, and AcSWEET10 proteins, respectively.The numbers on transmembrane helices (TMs) are represented as TM1 to TM7.TM4 (linker helix) and the C-terminal tails are represented in blue and red colors, respectively.

Figure 4 .
Figure 4. Pineapple SWEETs show differential expression in the anthers.a Morphological characteristics of the pineapple stamen tissues used for RNA-seq analysis were harvested at six stages (St1-St6).Scale bar: 1 mm.b Heat map of expression profiles of AcSWEETs in stamen tissues at different stages.The numbers below represent the pineapple SWEET genes.The heat map was created based on the log 2 (FKPM +0.01) value of AcSWEETs and normalized by row.FKPM values >30 are marked with an asterisk.Differences in gene expression changes are shown in color, as shown in the scale to the right.c Relative expression levels of AcSWEET6, 8, and 10 genes from six developmental stages of stamens.Vertical bars represent the mean ± standard error of three biological replicate assays.Asterisks denote statistical significance compared with stage 1 as judged by Student's t-test ( * * P < .01,* * * P < .001).

Figure 5 .
Figure 5. AcSWEET10 transports glucose in yeast.Glucose complementation assay of AcSWEET6, AcSWEET8, and AcSWEET10, where AtSWEET8 (positive control) and vector (without insert, as negative control) were expressed in the hexose transport-deficient yeast mutant EBY.VW4000.The transformed yeast colonies were diluted 10-fold and cultured on synthetic deficient medium without tryptophan (SD/−Trp) supplemented with 2% (w/v) maltose (as growth control) and 2% (w/v) glucose.Images were taken after culture plates had been incubated at 30 • C for 3-5 days.

Figure 6 .
Figure 6.Conserved amino acids of AcSWEET10 play a crucial role in glucose transport.A glucose transport assay of the indicated strains and vector (without insert, as negative control) was carried out in the hexose transport-deficient yeast mutant EBY.VW4000.The transformed yeast colonies were diluted 10-fold and cultured on synthetic deficient medium without tryptophan (SD/−Trp) supplemented with 2% (w/v) maltose (as growth control) and 2% (w/v) glucose.Images were taken after culture plates had been incubated at 30 • C for 3-5 days.

Figure 7 .
Figure 7. AcSWEET6 and 8 do not complement Arabidopsis sweet8 function.a Alexander dye-stained anthers showing non-viable pollen in non-complemented plants, similar to the mutant.Scale bar: 50 μm.b Primary inf lorescence stems showing shorter siliques of non-complemented plants.Scale bar: 3 cm.c Graph showing pollen abortion rate.d Graph showing viable seed number.Asterisks above the columns indicate significant differences compared with Col-0.* P < .001.

Figure 8 .
Figure 8. AcSWEET10 functions as a glucose transporter and complements Arabidopsis sweet8 mutant.a Five-week-old plants showing silique growth from the initial stages.Scale bar: 3 cm.b Representative photographs showing variable extents of restored silique lengths in three complemented lines (L1, L2, and L3) with reference to Atsweet8 mutant.Scale bar: 1 mm.c Graph representing viable seed number per silique (harvested from position 5 onwards from the main inf lorescence stem).d Graph showing pollen abortion rate.e Brightfield microscopy of f lowers showing pollen on anthers and stigma of Col-0, Atsweet8, and three complemented lines, the f lowers showing abundant pollen on stigma of complemented lines.Scale bar: 20 μm.f Viable pollen stained bright red was observed in the Alexander-stained anthers.Scale bar: 200 μm.Values represent means + standard error (n = 10 siliques per genotype) and asterisks above the columns (in c and d) indicate significant differences compared with Col-0.* * P < .01,* * * P < .001.

Figure 9 .
Figure 9. AcSWEET10 regulates callose 5 synthase levels and maintains tetrads during microspore development.a DIC images showing wall architecture of tetrads of wild type (Col-0), Atsweet8, and three AcSWEET10 + Atsweet8 complemented lines.b Cytochemical staining with aniline blue of tetrads of in Col-0, Atsweet8, and three complemented lines Scale bar: 20 μm.c Percentage of defective tetrads in Col-0, Atsweet8, and three complemented lines.d Aniline blue staining width quantified for Col-0, Atsweet8, and one complemented line.The statistical test was based on Tukey's honest test.Groups labeled with the same letter are not statistically different from each other (α = 0.05).e Relative expression levels of callose 5 synthase in anthers (harvested at tetrad stage) from Col-0, Atsweet8, and three complemented lines.Values represent means ± standard error of three biological replicates.* * * P < .001.

Figure 10 .
Figure 10.A schematic model representing the glucose transport mechanism in SWEET proteins.The figure shows the inward open conformations of SWEET proteins, including AtSWEET8, AcSWEET6, AcSWEET8, and AcSWEET10.The transport process is symbolized by arrows, color-coded to indicate varying levels of transport activity.Dark green arrows signify robust glucose transport, light green arrows denote weaker transport, and a 'stop' sign is employed to indicate restricted glucose transport.Notably, the direction of the arrows also signifies the exit of glucose molecules through the transporter's channel.