Characterizing the role of ATP-binding cassette transporters in triterpenoid wax bloom production A THESIS SUBMITTED TO THE FACULTY OF THE UNIVERSITY OF MINNESOTA BY Taylor Abrahamson IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE Advisor: Luke Busta 2024 i ACKNOWLEDGMENTS I am very fortunate to have been given the opportunity to complete my Master’s degree in the Integrated Biosciences program at the University of Minnesota Duluth. Completing this thesis has provided me with invaluable research experience and interpersonal skills. First and foremost, I would like to extend my gratitude to my research advisor, Dr. Luke Busta for all his support and guidance during my time at UMD. I wouldn’t have attended graduate school if it weren’t for his encouragement and confidence in my abilities as a scientist. Additionally, I would like to acknowledge the members of my research lab and those graduate students with whom I shared a working space for making my days more enjoyable. I would also like to thank my committee members, Dr. Amanda Grusz and Dr. Marshall Hampton, for their insight and helpful suggestions. A special thanks to Nicole Groth, who not only provided help with experiments and shared her knowledge of bacterial and plant transformations but was an unwavering source of support inside and outside of the laboratory. Lastly, I would like to thank my family and friends for their patience and support throughout my studies. ii ABSTRACT Some plants have developed lineage-specific mechanisms to increase water retention during drought conditions, including thick epicuticular wax coatings called wax blooms. However, the process underlying synthesis and especially the transport of wax bloom compounds to the leaf surface have yet to be fully elucidated. Numerous ATP-binding cassette (ABC) transporters have been characterized in model systems and crops, and some members of the G transporter subfamily are linked to the production of epicuticular wax blooms. Many studies utilizing model systems focus on long-chain aliphatic wax compounds, even though numerous species produce wax bloom composed primarily of polycyclic triterpenoid compounds. Relative to aliphatic compounds, next to nothing is known about the production of triterpenoid wax blooms, including whether their transport to the plant surface is mediated by ABC transporters with specificity for triterpenoids rather than aliphatics. Kalanchoe is a genus of succulent plants that vary in their ability to produce wax blooms, with some species producing the triterpenoid wax blooms of interest. Here, I investigated the role of ABC transporters in triterpenoid wax bloom production. RNA samples from three Kalanchoe species that produce triterpenoid-rich wax blooms to varying degrees (K. thrysiflora, K. fedtschenkoi, and K. blossfeldiana) were collected and sequenced. Gene expression data and wax chemistry were compared among replicates, varying: (i) plants grown in high light intensity v. low light intensity, (ii) bloom-producing v. bloomless plants, and (iii) leaf epidermal tissue v. mesophyll tissue. I assembled transcriptomes for each species and identified candidate ABC transporter genes using the comparisons described above. I then characterized the transporter candidate genes using heterologous expression in Nicotiana benthamiana and subsequent gas chromatography-mass spectrometry analysis. The data indicated that the ratio of triterpenoids to aliphatic compounds increased when certain iii transporters are co-infiltrated with a triterpenoid synthase gene. This strongly suggests that Kalanchoe has triterpenoid-specific transporters, and that other triterpenoid-rich wax bloom producing species could as well. iv TABLE OF CONTENTS List of Tables……………………………………………………………………………………...v List of Figures……………………………………………………………………………………vi CHAPTER 1. Literature review and analysis of public data……...………………………..……..1 1.1 Plant cuticles, drought tolerance, and chapter overview…….………………….….….1 1.2 Wax bloom composition and synthesis………………………………………….….…3 1.3 Wax transport…………………………………………………………………….…....6 1.4 Exploring ABC transporter evolution using publicly available genomes…………...11 CHAPTER 2. Identifying and characterizing potential triterpenoid-specific ABC transporters using RNA sequencing and heterologous expression……………………………………………22 2.1 Introduction…………………………………………………………………………..22 2.2 Methods………………………………………………………………………………23 2.2.1 RNA extraction and transcriptome assembly………………………………23 2.2.2 Bioinformatic analysis……………………………………………………...24 2.2.3 Heterologous expression in Nicotiana benthamiana……………………….26 2.3 Results & Discussion………………………………………………………………...29 2.3.1 Candidate identification……………………………………………………30 2.3.2 Candidate expression……………………………………………………….39 2.3.3 GC/MS data…………………………………………………………………43 2.4 Conclusion…………………………………………………………………………...47 Future Directions…………………...……………………………………………………………47 Bibliography……………………………………………………………………………………..50 Appendix………………………………………………………………………………………...58 v LIST OF TABLES Table 1: Dataset of transcriptomes for bloom-producing and bloomless plants used to research transporter evolution ………….…………………………………………………………………13 Table 2: Accessions of interest from each wax type….…………………………………………16 Table 3: ABC transporter candidates from K. fedtschenkoi and K. thyrsiflora…………………………………………………………………………………......…36 Table 4: List of transformation combinations…………………………………...………………42 Supplemental table 1: R packages and versions used for data analysis and visualization R packages and versions used for data analysis and visualization ………………………….…….64 Supplemental table 2: Detailed annotations for unique Arabidopsis genes identified in the candidate gene analysis………………………………………………………...………….…….64 vi LIST OF FIGURES Figure 1: Leaf sheath with wax coating on Sorghum bicolor..……………………………………3 Figure 2: Diversity of wax bloom compounds……………………………………………………4 Figure 3: Simplified biosynthesis of two wax compound classes: aliphatics and triterpenoids..…6 Figure 4: Hierarchical structure of ABC transporters…..…………………………………………8 Figure 5: Transport of wax compounds synthesized in the ER...………………………………..10 Figure 6: Number of predicted ABCG proteins per 1000 total predicted proteins identified in 9 flowering plant species of interest based on wax type.…………………………………………..15 Figure 7: Computational pipelines for RNA sequencing data processing for Kalanchoe species and ABC transporter candidate gene identification in K. fedtschenkoi (light inducible bloom) and K. thyrsiflora (constitutive bloom)……………………………………..………………………..26 Figure 8: Workflow of bacterial and plant transformations.…………………………...………..28 Figure 9: Treatment conditions for comparative wax bloom groups.…………….……………..30 Figure 10: Volcano plots of differentially expressed transcripts in K. fedtschenkoi (light- inducible bloom)………………………………...………………………………………………33 Figure 11: Differentially expressed K. fedtschenkoi amino acid sequences containing an ABC transporter domain………………………………………………………………………………34 Figure 12: Homologous transporters between K. fedtschenkoi (light-inducible bloom), K. thyrsiflora (constituitve bloom), and K. blossfeldiana (no bloom)…………..…………………38 Figure 13: Schematic representation of transporter assays………..……………………………40 Figure 14: N. benthamiana leaves under RB-GO UV light………….…………………………41 Figure 15: Bacterial transformation results.……………………………………………………42 Figure 16: Plot representing GC/MS data for each transformation combination ………..……44 vii Supplemental figure 1: Sequences with high similarity to characterized Arabidopsis thaliana ABC transporters (BLASTp)………………….…………………………………………………58 Supplemental figure 2: Sequence alignment of WBC group ABC transporters…………………59 Supplemental figure 3: Sequence alignment of subfamily G ABC transporters as predicted by hmmscan…………………………………………………………………………………………60 Supplemental figure 4: Sequence alignment of sequences from K. thyrsiflora and K. blossfeldiana that have homology to K. fedtschenkoi ABC transporters………………..………61 Supplemental figure 5: pEAQ vector map……………………….………………………...……62 Supplemental figure 6: Wax coverage data for each treatment group from heterologous expression assays…...……………………………………………………………………………63 1 CHAPTER 1 – Literature review and analysis of public data 1.1 Plant Cuticles, drought tolerance, and chapter overview About 500 million years ago, plants experienced evolutionary adaptations that facilitated their transition from aquatic to terrestrial environments. Land plants had to evolve mechanisms to protect themselves from a number of new stressors including desiccation, UV radiation, physical damage, and pathogens (Yeats, 2013). These mechanisms include but are not limited to stomata, deep roots, UV-absorbing phytochemicals, and specialized cell types (Kumar et al., 2022). One key innovation had evolved to control the interface between the plants above ground surfaces and the external environment: the cuticle. The cuticle is a multifunctional, layered chemical coating that is present on plants’ above ground surfaces. The cuticle protects against non-stomatal water loss, UV damage, herbivory, and keeps the plant surface clean and free of pathogens (Dominguez et al., 2017). The cuticle has a complex structure with two primary parts: cutin and wax. Cutin is a polyester scaffold composed of long-chain fatty acids. Waxes consisting of aliphatic or cyclic lipid compounds are embedded within and extruded on top of the cutin matrix (Figure 1) (Kong, 2020). The production of the cuticle is a highly conserved process among essentially all terrestrial plant lineages. Genome sequencing spanning many major land plant families and subsequent phylogenomic analysis confirm that cuticle biosynthetic machinery arises from the most recent common ancestor of embryophytes, the first land plants (Kong, 2020). While land plants have successfully adapted to terrestrial new environments through the formation of the cuticle and other morphological changes, prolonged periods of drought still pose 2 a challenge to them, including crops. Luckily, many plants have evolved mechanisms to further aid in water retention under drought conditions, such as changes in root structure (Bengough et. al., 2011), stomatal regulation (Cominelli et al., 2010), and osmotic adjustments (Chen & Jiang, 2010). Drought tolerant crops, such as Oryza sativa and Sorghum bicolor often utilize more than one of these mechanisms. Both crops are reported to employ a deep root structure to access water in deeper soil layers, where increased root length improves water uptake and leads to increased yield (Wasaya et al., 2018). Many plants can also mitigate stomatal water loss due to transpiration by regulating the production of abscisic acid, which controls stomatal closure. In Arabidopsis thaliana, transcription factors such as ABI3, have been implicated in modulating this process, where abi mutants have improper stomatal regulation and increase water loss (Cominelli, 2010). Although virtually all plant species’ epidermal cells have a ubiquitous layer of hydrophobic lipids that also contribute to drought tolerance, some plants develop specialized wax layers that are extruded through the cuticle matrix and deposited as wax crystals to further protect against a number of abiotic stressors (Figure 1) (Singh, 2018). When these wax layers accumulate to the point that they are visible with the naked eye, they are referred to as epicuticular wax blooms. In addition to being a drought tolerance aid, epicuticular wax blooms serve other functions as well such as protection against UV damage, pathogens, and herbivory (Patwari et al., 2019). The multifunctional nature of epicuticular wax blooms drives interest in studying how they are synthesized and transported on the aerial surface of plants. These two processes, synthesis and transport, will be discussed in detail in subsequent sections (1.2, 1.3, 1.4). 3 Figure 1: Top panel: Leaf sheath with wax coating on Sorghum bicolor. Bottom panel: Leaf cross section displaying membrane, cell wall, cuticle, and epicuticular wax crystals. 1.2 Wax bloom composition and biosynthesis Wax blooms vary in composition among diverse plant taxa. In a previous study that focused on identifying wax compounds in a wide range of bloom-producing plants, it was determined that wax blooms contain a variety of compounds (Nguyen et al., 2023). The main compounds identified were aliphatic alkanes, alcohols, and fatty acids, as well as cyclic triterpenoid compounds (Figure 2). Wax blooms can comprised of almost entirely aliphatic compounds, almost entirely cyclic compounds, or mixtures of the two. These wax bloom 4 constituents form crystals on the surface of epidermal cells that have notable differences in crystal structure depending on whether they are made up of aliphatic compounds or triterpenoid compounds (Liu, 2017; Jenks, 2000). Epicuticular wax blooms composed of aliphatic wax constituents tend to be more common than primarily triterpenoid waxes, with the most ubiquitous compounds being fully saturated unbranched hydrocarbon chains of 18 carbons or greater, called ‘very long chain’ (VLC) compounds (Nguyen, 2023; Buschhaus and Jetter, 2011) (Figure 2B). Triterpenoid compounds seem to preferentially accumulate in the intracuticular wax layer, but predominantly triterpenoid epicuticular waxes have also been observed in different plant genera, such as Vaccinium and Kalanchoe (Buschhaus and Jetter, 2011). 5 Figure 2: Diversity of wax bloom compounds. A. Occurrence of wax blooms in multiple plant lineages. Wax bloom samples were collected as a citizen scientist project. B. Surface wax chemistry of plants sampled by Nguyen (2023). Unique compounds in each wax bloom class are shown on x-axis. Y-axis represents the count of plant species sampled that contain each compound. The synthesis of aliphatic and triterpenoid epicuticular wax compounds follow two distinct biosynthetic pathways. Aliphatic wax compound synthesis begins in the plastid, where C16-C18 precursors are modified into fatty acyl-CoA molecules by long-chain acyl-CoA synthetases (LACS). Fatty acyl-CoA molecules are subsequently elongated in the endoplasmic reticulum (ER) by a fatty acid elongation complex that utilizes a variety of fatty acid elongases to create ‘very long chain (VLC) fatty acyl-CoAs’ (Yeats and Rose, 2013). From here, the pathway diverges: VLC acyl-CoAs are modified into either primary alcohols or esters, while others are converted into aldehydes, alkanes, secondary alcohols, or ketones (Figure 3A). Precursors to triterpenoid wax compounds, in contrast, are first synthesized in the cytoplasm and peroxisomes by the mevalonate pathway that converts acetyl-CoA into sterol isoprenoids, namely 2,3- Oxidosqualene. 2,3-Oxidosqualene can be cyclized by a variety of oxidosqualene cyclase enzymes (OSCs) in the ER to form triterpenoid compounds (Figure 3B). For example, the pentacyclic triterpenoid beta-amyrin found in some wax blooms is cyclized from 2,3- Oxidosqualene by an OSC called beta-amyrin synthase (Thimmappa et al., 2014). Some species, such as Artemisia annua and Kalanchoe thyrsiflora, are capable of further oxidizing triterpenoid alcohols into triterpenoid ketones using cytochrome P-450s, which are also found in epicuticular waxes, such as beta-amyrone (Moses, 2015; Nguyen et al., 2023). 6 Figure 3: Simplified biosynthesis of two wax compound classes: aliphatics and triterpenoids. A. Synthesis of very long-chain fatty acid compounds that further differentiate via addition of functional groups. Aliphatic compounds shown are able to contribute to the total cuticular wax mixture and are found in varying amounts between species. B. Synthesis of cyclic triterpenoid compounds found in cuticular waxes. 2,3-Oxidosqualene can give rise to many different triterpenes found in waxes, with some species further oxidizing them using cytochrome P450s. Example shown is oxidation of beta-amyrin to form beta-amyrone. Asterisks indicate compounds that can be transported to the surface and contribute to the total cuticular wax mixture. Arrows indicate modification of each compound. Abbreviations: LACS: long-chain acyl-coA synthetase, FAE: fatty-acid elongation complex, FPP: farnesyl pyrophosphate, CYP450: cytochrome P-450. 1.3 Wax Transport Once wax compounds are synthesized, they must be transported from within the cells to the plant surface where they can crystallize into wax blooms. The mechanisms by which molecules can be imported or exported out of cells can be divided into three main categories: 7 passive transport, active transport, and bulk transport. Passive transport requires no energy and can be achieved either by small, non-polar molecules, such as carbon dioxide, diffusing through the cell membrane, or facilitated by transport proteins or ion channels. Active transport is also made possible by transport proteins and requires energy, usually in the form of ATP (primary active transport) but can also utilize the coupled electrochemical potential from two interacting molecules as a mechanism for transport (secondary active transport). Lastly, macromolecules such as nucleic acids and proteins are transported using bulk transport, where they are either engulfed via phagocytosis to transport them inside the cell, or their movement outside the cell is mediated by vesicles fusing to the plasma membrane (Philippe, 2022). It is reasonably well known that primary active transport is the mechanism by which leaf wax compounds are trafficked to the aerial surface of plants and this mode of transport is the main focus of this section (Li et al., 2016). Many of the proteins involved in wax transport in the model plant Arabidopsis thaliana have been functionally characterized and belong to a protein superfamily referred to as ATP-binding cassette (ABC) proteins (Figure 4) (Lefevre and Boutry, 2018). 8 Figure 4: Hierarchical structure of ABC transporters. ABC transporters are divided into subfamilies based on structure and substrate transport. Subfamilies are further nested into subgroups that correspond to the protein domain content. Modified from Theodoulou (2000). The ABC transporter protein superfamily consists of full (pleiotropic/multi-drug resistance) and half (white-brown complex) integral membrane transporters that are conserved in many different organisms (Lefevre and Boutry, 2018). Typical membrane bound ABC transporters have four domains: two transmembrane domains, and two nucleotide-binding domains. However, half transporters only contain one transmembrane domain and one nucleotide-binding domain, thus requiring dimerization in the ER to be functional (McFarlane, 2010). Membrane-bound ABC transporters consist of both importers and exporters, though prokaryotic cells seem to only contain importers (Naaz, et al., 2023). ABC transporters are 9 further organized into subfamilies (A-I) based on domain structure and have different expression sites within biological membranes, as well as substrates (Figure 4). ABC transporters have expanded greatly in plants and are functionally diverse, where some are essential for proper growth and development, while others transport xenobiotics and specialized metabolites (Hwang et al., 2016). Plant family-specific expansions have also been demonstrated, indicating that novel ABC transporters can be expressed to transport specialized metabolites. The expansion and characterization of ABC transporters is primarily focused on the largest subfamilies, ABCB, ABCC, and ABCG which mainly transport secondary metabolites and phytohormones (Banasiak and Jasinski, 2022). However, the diversity of these proteins makes it difficult to assign substrates. In plants, few ABC transporters have been tested for substrate specificity using direct transport assays and those that have been studied vary in degree of specificity (Lefevre and Boutry, 2018). In A. thaliana, there are many ABC transporters that transport compounds that are chemically unrelated, while others, namely those of the ABCB subfamily, only have a high affinity for auxinic compounds (Lefevre and Boutry, 2018). In addition to the challenges that come with structural and functional diversity, the characterization of ABC transporters and their substrates is mostly confined to model plants and crops, especially in the context of wax production. Many studies have reported that epidermal cells in plants utilize ABCG transporters specifically to export wax compounds that make up the cuticle and epicuticular waxes (Figure 5). The first transporter found to be involved in wax transport was discovered in A. thaliana ‘eceriferum’ or cer5 (ABCG12 transporter) gene mutants displaying a waxless phenotype. In addition to cer5, wbc11 (ABCG11 transporter) mutants exhibit a significant decrease in wax load 10 compared to wildtype (Bird et al., 2016). Another Arabidopsis transporter, ABCG32, is reported to be involved in cuticle formation. Atabcg32 mutants exhibited altered cuticle structure and reduction of specific cutin monomers (Fabre et al., 2015). The ABCG9 (osabcg9) gene in Oryza sativa, a putative ortholog of the AtWBC12, is also involved in wax transport. It was reported that osabcg9 mutants displayed greater than 53% reduction in total leaf wax load compared to wildtype. These ABCG transporters involved in wax transport make up a distinct clade in the phylogeny of A. thaliana and O. sativa transporters, where many homologs have been identified in other bryophyte species as well, including the moss, Physcomitrella patens (Nguyen, 2018; Buda, 2013). With nearly all of the reported transporters involved in wax transport belonging to the largest and substrate diverse subfamily G, it seems highly likely that transporters involved in wax bloom biosynthesis also belong to subfamily G. If we are to understand wax bloom synthesis in more detail, we need more fundamental knowledge of wax transport particularly with respect to subfamily G. In this thesis, I will utilize similarity searching, RNA sequencing, transcriptome assembly, and heterologous gene expression to identify homologous transporters and their substrates in non-model species to better understand epicuticular wax production. Figure 5: Transport of wax compounds synthesized in the ER. ABC transporters provide transport through plasma membranes, while lipid transfer proteins (LTPs) are thought to aid 11 wax compounds in passing through the cell wall. Some compounds are extruded through the cuticle and deposited as epicuticular waxes. 1.4 - Exploring ABC transporter evolution using publicly available genomes To better understand the diversity of ABC transporters and those that relate to the production and composition of epicuticular wax blooms, I analyzed publicly available genomes of three plant groups: plants that produce triterpenoid wax blooms, plants that produce aliphatic wax blooms, and plants that do not produce wax blooms. By exploring ABC transporter gene family evolution and using characterized transporter sequences, my aims are threefold: 1. To determine if there are differences in the number of ABCG transporters between plants that produce wax blooms and those that do not 2. To determine if there are differences in ABC transporter protein domain structure for triterpenoid wax producers and aliphatic wax producers 3. To determine if plants without wax blooms lack specific ABC transporters involved in wax production. Accomplishing these aims will provide valuable information about ABC transporters from different species in regard to wax production and will allow me to better address the overall goal of this thesis, to identify and characterize triterpenoid-specific wax transporters. 12 Methods All data processing and analyses were completed in R/RStudio using packages described in Supplemental Table 1. Data acquisition Plant families that produce triterpenoid blooms were identified in previous work by analyzing the surface wax composition for a wide range of plant taxa (Nguyen, 2023). From this, I assembled a dataset of primary transcript protein sequences found in distinct species within these plant families (Table 1) by retrieving publicly available data from JGI Phytozome v13. These datasets contain protein sequences from all predicted genes in each species, regardless of whether those genes are actually expressed. Protein sequences were concatenated into a single FASTA file and the ‘hmmscan’ function in HMMER (v3.4) was used to compare the domain structure predicted by the queried amino acid sequences to the Pfam database (Mistry, et al., 2021). Wax type identifiers and species names were added to the resulting data frame for downstream analyses. Data processing The results of the HMM domain prediction were filtered to contain only the sequences with predicted ABC transporter domain structure. Each sequence identified were retained and grouped according to the type of wax bloom produced by that species: aliphatic, triterpenoid, or none. To count the number of ABCG transporters present in each species and wax type, a Pfam accession number for G subfamily ABC transporters (PF19055.3) was retrieved from InterPro (v97.0). Any sequences without this accession number were excluded from the count. This count was normalized to account for differences in the total number of predicted proteins from each 13 species. To assess statistical significance, ANOVA was used to compare the normalized count across wax types followed by Tukey’s HSD post-hoc analysis. Using the hmmscan output, the domain structure of ABC transporters in aliphatic and triterpenoid-bloom producing plants was compared. The data were filtered to retain and group only the sequences with predicted ABC transporter domains. Data were subsampled according to wax type, and unique domain accession numbers with high sequence similarity (bit-score > 50) to Pfam domains were retained for each wax type category. Common accessions between wax type categories were compared, including those in the “none” wax type representing bloomless species. The 7 unique ABC transporter domain accessions were queried against the Pfam database and descriptions of each domain were recorded. Table 1. Dataset of transcriptomes for bloom-producing and bloomless plants used to research transporter evolution. All data retrieved from JGI Phytozome. Species Bloom type Brassica rapa Aliphatic Sorghum bicolor Aliphatic Daucus carota Aliphatic Vaccinium darrowii Triterpenoid Kalanchoe fedtschenkoi Triterpenoid Vitis vinifera Triterpenoid Zea mays No bloom Solanum lycopersicum No bloom Glycine max No bloom 14 To annotate transporters found in the species of interest, an HMM profile was compiled using sequences from characterized ABC transporters spanning all subfamilies and run against the query sequences using ‘hmmsearch’. Query sequences with high sequence similarity (bit- score > 100) to the characterized ABC subfamily G transporters were retained for further analysis. These sequences were used as the query in a protein-protein BLAST (word size 2, e- value cutoff 1) against the well-annotated Arabidopsis proteome to find transporter homologs that could indicate if plants without wax blooms lack wax transporters. Results & Discussion Using computational methods, my goal was to use predicted proteomes to reveal any detectable differences in the ABCG content of plants that vary in their production of wax blooms. It was determined that the number of ABCG transporters is not significantly different in plants that produce wax blooms and those that do not across the species studied, confirmed by ANOVA and Tukey testing (Figure 6). It is possible that the analyses completed here are not sufficiently sensitive enough to detect minor differences, if present. We must also consider that there are ABCG transporters present in all wax types that are involved in transporting other secondary metabolites unrelated to wax bloom production. This would influence our ability to relate any differences in ABCG transporter number to wax bloom production specifically. Another limitation of my analysis is that half ABC transporters dimerize in different combinations to transport different substrates. It is possible that species without wax blooms could contain the same number of transporters as wax producers, with distinct patterns of 15 dimerization to carry-out specific functions. The results of my analysis do not support a connection between the number of ABCG transporters present in a genome and the presence or absence of an epicuticular wax bloom on for a given species. Figure 6: Number of predicted ABCG proteins per 1000 total predicted proteins identified in 9 flowering plant species of interest based on wax type. ANOVA and Tukey’s HSD post-hoc analysis to confirm that the protein count between wax types is not statistically different (p > 0.05). The domain structure among wax types was compared to identify any species-specific ABC transporter domains or domain combinations in triterpenoid versus aliphatic blooms. Kalanchoe fedtschenkoi, which produces a triterpenoid bloom, has a distinct ABC transporter domain (PF14510.9) that has also been found in S. cerevisiae PDR5, a full ABCG transporter that transports a wide range of compounds, including sterols (Rutledge et al., 2011). Aliphatic bloom-producing species and triterpenoid bloom-producing species have 6 common ABC domains: PF00664.26, PF00005.30, PF01061.27, PF19055.3, PF12698.10, PF06472.18. There are no unique ABC domains for any of the aliphatic bloom-producing plants. It also appears that Glycine max and Solanum lycopersicum, two bloomless species, share the unique domain 16 identified in Kalanchoe fedtschenkoi, as well as the 6 common accessions between aliphatic bloom-producers and triterpenoid bloom-producers (Table 2). The domain representing the intracellular N-terminus of ABC transporters could indicate how the domains are arranged within the plasma membrane. This information paired with other identified domains in the same sequence could provide an approximation of which subfamily the transporter belongs to. For example, ABCC transporters have an extracellular N-terminus, thus eliminating the possibility of any transporters with this domain belonging to the C subfamily. Lack of transmembrane domains could also hint at the possibility of the transporter belonging to subfamilies E or F that only contain nucleotide binding domains (Kretzschmar, 2011). As a result, there are more similarities in domain content between wax types than there are differences, given that many of the ABC transporter domains identified are shared among species. However, species with aliphatic blooms do indeed have transporters with intracellular N-terminal domains (ABCA, ABCB, ABCD, and ABCG), but perhaps these sequences were excluded due to the domain score (bit-score) threshold. Table 2. Accessions of interest from each wax type. All accessions correspond to Pfam ABC transporter protein domains residing within the InterPro database. Unique ABC accessions InterPro description Bloom type(s) PF00664.26 ABC transporter unit of six transmembrane helices Triterpenoid, aliphatic, none PF00005.30 ABC transporter nucleotide binding domain Triterpenoid, aliphatic, none 17 PF01061.27 Transmembrane region of ABC-2 type transporters Triterpenoid, aliphatic, none PF19055.3 Subfamily G ABC transporter domain Triterpenoid, aliphatic, none PF12698.10 Related to ABC-2 type transporters Triterpenoid, aliphatic, none PF06472.18 Transmembrane region for small subfamily (D) of ABC transporters, sequence similarity to PF00664 Triterpenoid, aliphatic, none PF14510.9 Intracellular N-terminus of ABC transporters Triterpenoid, none Finally, I also sought to determine if plants without wax blooms lack specific transporters that have been reported to be involved in wax bloom production. The hmmscan analysis identified 12,608 ABC transporter domains across all nine species, of which there are 4,212 unique transporter sequences with a domain score greater than 0 (bits). Interestingly, hmmscan identified 619 sequences with ABCG transporter domain structure, but of the 1,377 ABC transporter sequences homologous to A. thaliana, 1,296 of them are annotated as subfamily G when annotating using a protein BLAST (94.1% of all identified transporters). Nearly all of the ABCG transporters are annotated as ABCG3 and match the same A. thaliana accession number, with only 5.8% of transporters belonging to other ABC subfamilies and 5.7% of ABCG transporters identified being annotated as something other than ABCG3. Transporters with essential functions related to growth and development that should certainly be present in the proteomes studied, such as ABCG11, or transporters linked to wax production e.g. ABCG12, were absent from the protein-protein BLAST. Using HMMER, BLAST, and Pfam, I was not able to annotate amino acid sequences with enough sensitivity to draw any conclusions about any 18 specific wax-related transporters in bloomless plants. BLAST alone cannot accurately distinguish among transporter types based solely on sequence similarity, especially if ABC transporters within the same clade or subfamily have a high degree of similarity. In this case, a protein-protein BLAST isn’t sensitive enough to accurately distinguish between subfamilies or members of the same subfamily. It is likely that transporters such as ABCG11 are indeed present, but are annotated as a closely related transporter (ABCG3). In fact, ABCG3 and ABCG11 belong to the same clade in A. thaliana, but other G subfamily members identified in this analysis comprise distinct clades within the A. thaliana transporter phylogeny. The small percentage of other subfamilies identified in my analyses could be because the G subfamily is the large or because the transporters did not share enough similarity to the Pfam ABC domains to meet my inclusion criteria. While some information regarding transporter subfamilies can be inferred from the domain structure, it does not provide adequate information to assign specific transporters to amino acid sequences. Another limitation I experienced was rounding of extremely small e-values in the BLAST results that equated Arabidopsis transporters identified as excellent matches, or an e- value of 0.00, indicating that the sequences are identical. Excellent matches are not unexpected among homologs across 9 species, however, there is no way to “rank” these matches to provide an annotation when they are all equally similar. Due to these limitations, I am unable to determine if bloomless plants lack specific transporters reported to be involved in wax production. Another method that could be used to better understand the diversity of ABC transporters is phylogenetic analysis. There are published phylogenetic trees of ABC transporters (Andolfo et 19 al., 2015; Buda et al., 2013; Nguyen et al., 2014), many of which are unrooted neighbor-joining trees. I attempted to align all ABC transporters present in nine species of interest, as predicted by hmmscan, however the sequence alignment generated for phylogenetic analysis was of poor quality (high gap percentage, extreme length differences among sequences), and required extensive trimming to the point where a tree could not be constructed using the trimmed alignment (Supplemental Figure 1). To account for the differences in sequence length, I attempted to sort the transporter sequences according to subgroups (WBC, PDR), but these issues persisted (Supplemental Figure 2). To generate a better alignment, I worked to discern between subfamilies of transporters and make phylogenetic trees according to subfamilies as in Andolfo et al. (2015) (Supplemental Figure 3). However, this relies accurately annotating sequences according to subfamilies. Based on the protein BLASTs run previously (see above), BLAST cannot reliably distinguish between subfamilies as 94% of all detected transporters were annotated as subfamily G. Some subfamily information can be inferred from the domains present in each sequence, but this is not a comprehensive approach for assigning subfamilies to the sequences. Finally, individual or concatenated ABC domains from a given protein have the potential to be aligned for phylogenetic analysis, which was attempted using density-based clustering of ABC domains, but it was difficult to set a threshold for an ABC domain based on the scoring provided by hmmscan, and the number of domains for each sequence vary. While it is possible to construct phylogenies for ABC transporters (Andolfo et al., 2015; Buda et al., 2013; Nguyen et al., 2014), published sources do not provide enough information to repeat their methodology or to assess the reliability of published alignments or subfamily classifications. Based on my work in this area and my experience with ABC protein alignments, generating a 20 high-quality ABC transporter phylogeny is outside the scope of this project. Although a phylogeny would provide valuable insights about ABC transporter evolution, a phylogenetic perspective is not required to identify and characterize triterpenoid-specific transporters, which is the primary goal of this study. Conclusions In exploring ABC transporter evolution using publicly available data, I aimed to use statistical analyses and computational tools to identify any differences in ABC transporter number, content, and type among species with different wax types. Specifically, I examined ABCG transporters due to their previous characterization in wax production. I determined that there were no significant differences in the number of ABCG transporters between wax types. To better understand domain structure and substrate specificity of wax transporters, the domain content among wax types was compared. I found that there are many similarities among the ABC transporter domains identified in each species, with the exception of one domain (PF14510.9) that is not present in aliphatic bloom-producing species. One question I posed at the outset of this chapter was: “Do plants without wax blooms lack specific ABC transporters involved in wax production?” I am unable to answer this question because available similarity searching methods are not sufficiently sensitive enough to differentiate between specific transporter types. These limitations make it difficult to identify transporter candidates involved in wax production by linking sequences to specific transporters, or even transporter subfamilies. In Chapter 2 of this thesis, transporter candidates are identified using Hidden Markov profiles to predict protein 21 domains from highly expressed transcripts, which can only be annotated as ABC transporters in a general sense due to the constraints of similarity searching tools. 22 CHAPTER 2 – Identifying and characterizing potential triterpenoid-specific ABC transporters using RNA sequencing and heterologous expression 2.1 Introduction Many plants, including the Kalanchoe and Vaccinium are known to produce epicuticular wax blooms that are made up almost entirely of triterpenoids. But some important details about triterpenoid wax bloom production (as opposed to aliphatic wax bloom production), including both biosynthesis and transport to the plant surface, are unknown (van Marrseveen, 2009; Chu, 2017). Oxidosqualene cyclase enzymes play a role in the biosynthesis of a wide range of triterpene compounds, (Ghosh, 2016) but the transporters involved in trafficking triterpenoids through the plasma membrane represent a gap in our knowledge. ABCG11/12 and other subfamily G proteins have been implicated in transporting aliphatic wax compounds in many other plants (Elejalde-Palmett, 2021). However, there are no studies investigating the role of these ABC transporters in triterpenoid wax production. Previous studies describe plant ABCG transporters that are capable of transporting terpene compounds smaller than triterpenoids (C30), including sesquiterpenes (C15) like capsidiol (Nicotiana benthamiana) and beta-caryophyllene (Artemisia annua) as well as diterpene (C20) compounds such as sclareol (Nicotiana tabacum) (Fu et al., 2017; Shibata et al., 2016; Crouzet et al., 2013). Furthermore, ABC transporters related specifically to triterpenoid transport have been characterized, including the human ABCG3 multidrug resistance transporter that aids in cholesterol efflux (Vasiliou, 2009) and the glycosylated triterpenoid transporter PDR3 in Panax ginseng (Zhang et al., 2013). These studies provide precedent for triterpenoid transport in the ABC protein superfamily, but there are no studies investing such in the transport of wax-related triterpenoid compounds. Due to the gap in 23 knowledge regarding wax-related triterpenoid transport in plants, this thesis aims to explore triterpenoid transport by utilizing a plant system with triterpenoid-rich epicuticular wax blooms. Based on published findings to date, the lineage that produces the most triterpenoid-rich wax blooms is the genus Kalanchoe (Nguyen, 2023). Kalanchoe is a dicotyledonous genus of succulent plants in the Crassulaceae family that vary in their ability to produce wax blooms (van Marrseveen, 2009). Some Kalanchoe species constitutively produce triterpenoid-rich wax blooms, like Kalanchoe thyrsiflora, which generates a wax bloom primarily consisting of alpha/beta-amyrin and alpha/beta-amyrone (Busta Lab, unpublished data). Meanwhile, other Kalanchoe species, e.g. Kalanchoe fedtschenkoi, produce triterpenoid alcohol-rich wax blooms in a light-inducible manner, and some never produce epicuticular wax blooms, as in Kalanchoe blossfeldiana (Busta Lab, unpublished data). The variability in wax production between species makes Kalanchoe an ideal study system as gene expression and wax chemistry can be compared between closely related taxa. The methods used to investigate triterpenoid transport in this lineage, as well as the results obtained, and the discussion thereof are described below. 2.2 Methods 2.2.1 RNA extraction and transcriptome assembly Tissue samples were frozen in liquid nitrogen and homogenized using a mortar and pestle for RNA extraction using the CTAB method (Carra et al., 2006). After RNA extraction, samples with a concentration greater than 1 μg and a quality score (RIN) greater than 6.0 were sequenced by Illumina (250bp, 20M paired reads per each sample) (Figure 7A). Sequence data were 24 processed and assembled using a modified version of the transXpress pipeline as described in Fallon et al. (2023) (Figure 7A). FastQ reads were trimmed with Trimmomatic and assembled using Trinity (v2.13.2). Transcriptome assembly was performed in a shell environment similar to https://github.com/transXpress/transXpress. Assembled sequences were further processed to assess the quality of the assembly using BUSCO (v5.5.0), to obtain data on the expression level of transcripts using Kallisto (v0.48.0), and to identify open-reading frames (ORFs) with TransDecoder (v5.7.0). The resulting data were used for downstream analyses including differential expression, protein domain prediction, and ABC transporter candidate gene identification. The percentage of sequences that contain rRNA was calculated to exclude any highly expressed transcripts that may have been misassembled to include rRNA, but the rRNA content was low enough not to require any normalization of expression level beyond that of a differential expression analysis. 2.2.2 Bioinformatic analysis Kalanchoe fedtschenkoi A differential expression analysis was completed using DESeq2 (v3.18) on the assembled sequences to compare gene expression between light treatment groups. A FASTA file containing the longest ORFs for the differentially expressed transcripts was created and used as a query for an hmmscan analysis (HMMER v3.4) to find sequences with ABC transporter domains (Figure 7B). A protein BLAST using the annotated A. thaliana proteome as a subject was performed to identify transporter homologs (word size 2, e-value cutoff 1). Custom commands used to complete these analyses can be found in https://thebustalab.github.io/integrated_bioanalytics/book/similarity-searching.html, and all https://github.com/transXpress/transXpress https://thebustalab.github.io/integrated_bioanalytics/book/similarity-searching.html 25 relevant arguments are available in the supplemental code file. A. thaliana sequences were retrieved from JGI Phytozome and annotation information was obtained from the TAIR database (Berardini et al., 2015). Kalanchoe thyrsiflora Expression data from Kallisto were used to generate an average of the transcripts per million (TPM) for each sample and a TPM cutoff of 1000 was set. A FASTA containing the longest ORFs for each preserved transcript was generated and input into the hmmscan analysis (Figure 7B). Highly expressed transcripts with ABC domains were subjected to a protein BLAST using the A. thaliana proteome as a subject using the same parameters from the previous section. A similar analysis was repeated for K. blossfeldiana. Transporter sequences from both species were run through a protein BLAST using the assembled K. fedtschenkoi transcriptome as a subject. 26 Figure 7: Computational pipelines for RNA sequencing data processing for Kalanchoe species and ABC transporter candidate gene identification in K. fedtschenkoi (light inducible bloom) and K. thyrsiflora (constitutive bloom). A. RNA sequencing data was assembled using the transXpress pipeline (B). K. fedtschenkoi transcripts were passed to a differential expression analysis where sequences with high expression and ABC transporter protein domain content as identified by hmmscan (HMMER), were used to mine for candidates. K. thyrsiflora candidates were identified by degree of expression and K. blossfeldiana (no bloom) transcripts were processed using a similar pipeline. Modified transXpress pipeline developed by Fallon et al., 2023. FastQ, Trimmomatic, Trinity, kallisto/BUSCO/TransDecoder ORFs were the programs used in assembly and annotation. Figure made in Biorender. 2.2.3 Heterologous expression in Nicotiana benthamiana a. Bacterial transformation Transporter candidates identified in the bioinformatic analysis were synthesized at Twist Biosciences, inserted into a pEAQ plasmid vector, and expressed in Agrobacterium tumefaciens (Hereafter “Agro”, Supplemental Figure 5). pEAQ contains a pBINPLUS backbone in addition to a modified 5’-untranslated region and a 3’ UTR from Cowpea mosaic virus that flanks the gene insert and increases expression as detailed in Sainsbury et al. (2009). Electrocompetent Agro (GV3101 strain) cells were prepared by multiple washes with deionized water to remove any contaminants. Plasmid DNA was added to Agro cells suspended in luria broth (LB) with an OD600 of 0.6-1.0, then transferred to a sterile electroporation chamber. Cells were electroporated at 1600V for plasmid uptake and immediately transferred to recovery media (LB) (Figure 8). Transformed cells were grown for 1 hour at 28℃, shaking, then plated on 15% (w/v) agar in LB media, containing the appropriate antibiotics: rifampicin, gentamicin, and kanamycin 27 (vector-conferred) and grown overnight at 28℃. Liquid cultures (10mL) were prepared for plant transformation. This process was repeated for eight unique vectors containing: GFP as a reporter gene, a beta-amyrin synthase gene from Glycyrrhiza glabra (GgBAS), the two K. fedtschenkoi transporter candidate genes, and the four K. thyrsiflora transporter candidates. b. Plant transformation Nicotiana benthamiana plants were grown for 4-6 weeks in Pro-Mix BX Mycorrhizae soil under a grow light. For the first two weeks of growth, plants were grown under a humidity chamber until visible sprouts formed. Freshly plated Agro cells containing transporter candidates were harvested via centrifugation and resuspended in infiltration buffer (5mM MES, 5mM MgCl2, 100uM acetosyringone, pH 5.7). The OD600 of each cell suspension was measured using a spectrophotometer prior to transformation and leaves were inoculated via syringe infiltration and incubated for 36-48 hours (Figure 8). Post-incubation, leaves were checked for presence of GFP using an RB-GO UV light. Transformations followed the vector combinations outlined in Table 4. For each unique vector construct, 3 biologically independent replicates were transformed and prepped for GC/MS analysis. 28 Figure 8: Workflow of bacterial and plant transformations. Agrobacterium tumefaciens is electroporated to facilitate vector uptake then grown on selection plates containing rifampicin, gentamicin, and kanamycin antibiotics. Cells are grown in liquid media (LB+RGK) in preparation for the infiltration buffer cell suspension used to inoculate N. benthamiana leaves. Asterisks represent OD600 measurements collected. GC/MS is used to detect changes in wax coverage and ratio of wax compounds post-transformation. Figure credit Nicole Groth, 2023. GC/MS sample preparation and analysis Tissue samples from the transformed plants were analyzed using GC/MS. Two 2x2 cm leaf discs were sampled from the transformed plants for analysis. A 24-carbon alkane internal standard was added to the samples and chloroform was used to extract wax from the leaves, then evaporated overnight. A 1:1 mixture of BSTFA to pyridine was added to derivatize the wax compounds present in the samples followed by a 45-minute incubation period at 70℃. Wax 29 compound peaks were identified and integrated to extract the peak area. The peak area was used to compute the ratio of a representative aliphatic compound found in all N. benthamiana leaf waxes to the triterpenoid compound of interest, beta-amyrin, which is predicted to be detectable in the GgBAS transformed leaves. Ratios and wax coverage values were tested for outliers (Grubbs test), then averaged across replicates and differences were subjected to significance testing (Shapiro, Levene, Kruskal-Wallis, and Dunn tests). 2.3 Results & Discussion The goal of this chapter was to test the hypothesis that Kalanchoe plants have triterpenoid-specific ABC transporters. This involved processing the transcriptomes from Kalanchoe species to predict ABC domain structure and thus to identify potential triterpenoid- specific ABC transporter gene candidates (Section 2.3.1) for heterologous expression in Nicotiana benthamiana (2.3.2). The variety of species chosen represents their varying ability to produce wax blooms where K. fedtschenkoi produces a light-inducible weak bloom, K. thyrsiflora constitutively produces a strong bloom, and K. blossfeldiana does not produce a bloom. Three plants from each of these three species were grown in varying light intensity conditions: no shade cloth, 1 layer, 2 layers, and 3 layers, with a total of 12 plants per species (Figure 9). The goal of growing these plants in varying light conditions was (i) to control light conditions for all species and (ii) to induce blooms in K. fedtschenkoi for later comparison. Tissue samples were taken from these plants, allowing us to compare the abundance of 30 transcripts encoding ABC transporters during different wax bloom production states: (i) constitutive blooms vs no bloom, (ii) whole leaf vs epidermal tissue, and (iii) low light intensity vs high light intensity. The assembled K. fedtschenkoi transcriptome contained 157,117 contigs, K. thyrsiflora contained 137,080, and K. blossfeldiana contained 218,036. A transcriptome assembly resulting in ~100,000 contigs is expected, however greater than 200,000 contigs is unusual. In the future, the assembly should be further analyzed using BLAST to check for the presence of foreign transcripts that may have been incorporated. A BUSCO report for all three species generated the following values: 95.3%, 91.8%, and 93.2% respectively. Figure 9: Treatment conditions for comparative wax bloom groups. From left to right, K. fedtschenkoi (low and high light intensity, light inducible bloom), K. blossfeldiana (bloomless), K. thyrsiflora (constitutive bloom) whole leaf, and K. thyrsiflora epidermis only. Figure made in Biorender. 2.3.1 Candidate identification K. fedtschenkoi 31 Because Kalanchoe fedtschenkoi produces a primarily aliphatic wax bloom in a light- inducible manner, I expected that aliphatic transporters will be upregulated in epidermal tissues, especially those exposed to high intensity light. A differential expression analysis was performed to determine which genes, if any, were significantly upregulated when comparing different tissue types and light conditions, and to identify aliphatic transporters. In the low light vs. high light experimental group, I determined that there were a total of 3085 differentially expressed genes, with 1196 being upregulated when exposed to more intense light conditions (Figure 10). Among these, 44 sequences contained at least one ABC transporter domain. For the epidermal vs. mesophyll tissue experimental group, there were many more differentially expressed genes (34,495 total), with 21,083 of them being upregulated in the epidermal tissue. Of the 21,083 upregulated sequences, 439 contained ABC transporter domain structure as predicted by hmmscan (Figure 10). Finally, among the 44 apparently light-induced ABC transporters in K. fedtschenkoi, four were ABC transporters that belonged to both groups. To identify and eventually assay aliphatic transporters in K. fedtschenkoi, I selected ABC transporters that were upregulated and exhibited homology to characterized aliphatic transporters (ABCG11/12, ABCG3). One of the four ABC transporters that belonged to both groups was synthesized as a candidate (TRINITY_DN2286_c0_g1_i24; hereafter “Fe2”) (Figure 11). The candidate had a much greater change in expression in the tissue group (log2FC = 10.336) than the light group (log2FC = 2.7117). Another candidate gene that I synthesized (TRINITY_DN4858_c1_g1_i24, log2FC = 9.1359; hereafter “Fe1”) was significantly upregulated in the light experimental group and was similar to the A. thaliana ABCG11 transporter. Both candidates were annotated as ABCG transporters by a protein BLAST against the A. thaliana proteome, however the specific 32 transporter type could not be inferred from similarity searching alone due to the limitations described in Chapter 1. Despite the limited information gathered from similarity searching, it is expected that these transporters are homologous to those in the Arabidopsis ABCG11/ABCG12/ABCG3 wax-related transporter clade. The BLAST results provided an ABCG3 annotation with a high degree of similarity based on bitscore. Notably, when sequences were uploaded to InterProScan to analyze the protein domain content, the annotations provided by this database were inconsistent with those of BLASTp. 33 Figure 10: Volcano plots of differentially expressed transcripts in K. fedtschenkoi (light-inducible bloom). The horizontal boundary formed by light gray points shifting to dark gray represents a p-value cutoff of 0.05. Large points are those transcripts that are significantly differentially expressed and have the presence of ABC transporter domains as confirmed with hmmscan. The candidates synthesized for heterologous expression are labeled. A. 34 Differentially expressed transcripts in low light versus high light treatment group. B. Differentially expressed transcripts in epidermal versus mesophyll tissue treatment group. Figure 11: Differentially expressed K. fedtschenkoi amino acid sequences containing an ABC transporter domain. Pink shading contains sequences from the high light vs. low light condition group while blue shading contains sequences from the epidermis vs. mesophyll condition group. The transporters that the two conditions have in common are shown above the diagram. Of these, TRINITY_DN2286_c0_g1_i24 and TRINITY_DN4858_c1_g1_i24 were identified as a candidate and synthesized. K. thyrsiflora Kalanchoe thyrsiflora has a triterpenoid-rich wax bloom that is produced constitutively, rather than in response to light, so I assessed the expression level of candidate transporters identified from the transcriptome assembly. To identify potential triterpenoid-specific ABC 35 transporter candidates in K. thyrsiflora, the average expression level across replicates was computed in addition to an hmmscan analysis to tease out sequences with ABC transporter domains. More than 2000 transcripts with an average TPM greater than 1000 were obtained and of these, 60 contained at least one ABC transporter domain. I selected four top candidates (hereafter “Th1-4”) due to their especially high level of expression and high sequence similarity to that of ABC domains in the Pfam database (Table 3). Of those four, TRINITY_DN98_c0_g1_i36 (“Th4”) had a much greater expression level than that of the other three candidates (14,147 TPM), as well as the highest bit score (172.0). Despite the annotation limitations of a protein BLAST, all candidates showed the highest similarity to ABCG3 transporters. However, the two most highly expressed candidates (“Th1”and “Th4”) suffered from rounding of very low e-values so the degree of similarity to the Arabidopsis ABCG3 transporter could not be accurately determined. Additionally, when sequences were analyzed using InterProScan, the annotations were again inconsistent with those of BLASTp. It is expected that the four candidates with the highest average expression level are involved in transporting wax substrates to the epidermis, with some potentially demonstrating substrate specificity for triterpenoid compounds that are abundant in K. thyrsiflora epicuticular waxes. As a secondary candidate identification analysis, the expression level of ABC transporters between K. thyrsiflora and K. blossfeldiana were compared (Figure 12). Despite having a rather weak wax bloom, many ABC transporters in K. blossfeldiana had a high average TPM, with some being expressed at a greater level than that of the K. thyrsiflora transporters. A direct comparison was made by identifying mutual homologs to K. fedtschenkoi transporters, 36 which resulted in 1065 homologs shared between the two species. “Best” homologs based on BLAST scoring criteria could not be determined due to sensitivity limitations, but I found many homologs for each K. fedtschenkoi transporter (57 total) from both species. A multiple sequence alignment was attempted for these transporter homologs, but the issues mentioned in Chapter 1 persisted (Supplemental Figure 4). Among those, three transporter homologs identified in K. blossfeldiana had the greatest degree of similarity to the same K. fedtschenkoi transporter (TRINITY_DN10061_c0_g1_i11) based on bit score. However, these candidates were not synthesized as potential triterpenoid-specific transporters due to the lack of a strong bloom, and the bloom that is present is primarily made of aliphatic compounds (Busta Lab, unpublished data). All four of the synthesized K. thyrsiflora candidates were also homologous to the same K. fedtschenkoi transporter, but with varying degrees of similarity based on bitscore. There are many transporters in K. thyrsiflora that are highly expressed and show homology to other K. fedtschenkoi transporters, but they produced quite low domain scores in the hmmscan analysis, indicating that the domain structure was not very similar to any ABC transporter domains in the Pfam database. However, these highly expressed candidates could be synthesized and tested in heterologous expression in the future. Table 3. ABC transporter candidates from K. fedtschenkoi and K. thyrsiflora. Base mean values were determined by DESeq2 and TPM values were determined by Kallisto and ABC domain information was retrieved from InterPro. Trinity accession Average expression level Domains Domain description TRINITY_DN4858_c1_g1_i24 30.96 (base mean) PF00005.30, PF19055.3 ABC transporter nucleotide binding domain 37 ABC-2 type transporters TRINITY_DN2286_c0_g1_i24 125.3 (base mean) PF01061.27, PF00005.30, PF14510.9, PF19055.3 Transmembrane region of ABC-2 type transporters ABC transporter nucleotide binding domain N-terminus of ABC transporters ABC-2 type transporters TRINITY_DN98_c0_g1_i36 14,147 (TPM) PF01061.27, PF19055.3, PF00005.30, PF12698.10 Transmembrane region of ABC-2 type transporters Subfamily G ABC transporter domain ABC transporter nucleotide binding domain Domain related to ABC-2 type transporters TRINITY_DN1950_c0_g1_i146 3,976 (TPM) PF00005.30, PF00664.26 ABC transporter nucleotide binding domain ABC transporter unit of six transmembrane helices TRINITY_DN3292_c1_g1_i19 3,922 (TPM) PF01061.27 Transmembrane region of ABC-type transporters TRINITY_DN562_c0_g1_i43 3,814 (TPM) PF01061.27, PF06422.15 Transmembrane region of ABC-2 type transporters PDR/CDR region subgroup of ABC transporters 38 39 Figure 12: Homologous transporters between K. fedtschenkoi (light-inducible bloom), K. thyrsiflora (constituitve bloom), and K. blossfeldiana (no bloom). K. fedtschenkoi transporters identified by differential expression and HMM and their corresponding “best” Arabidopsis thaliana transporter homolog are presented on the x-axis. Homologs to K. fedtschenkoi transporters found in K. thyrsiflora and K. blossfeldiana transcriptomes are presented on the y-axis. The number of tiles represents the number of sequences in each species that are homologous to the same K. fedtschenkoi transporter. Transporter homolog count shown has a max of 40, despite some transporters having > 40 homologs, but they were excluded due to a low degree of expression. Red arrow indicates K. fedtschenkoi accession number that is homologous to transporter candidates identified in K. thyrsiflora. 2.3.2 Candidate expression GFP and empty Agro cells (no vector) were infiltrated in N. benthamiana as negative controls to verify that the plant system does not naturally produce beta-amyrin and that these two transformations do not alter the wax load. Additionally, GgBAS was infiltrated alone to determine if endogenous transporters in N. benthamiana were capable of transporting beta- amyrin to the surface (Figure 13). It is possible that N. benthamiana may have a transporter that does not demonstrate substrate specificity for type of wax compound (aliphatic/triterpenoid), or it has an endogenous transporter capable of transporting triterpenoid-like compounds. Literature reports that N. benthamiana utilizes ABCG transporters (ABCG1 and ABCG2) to transport defense compounds like capsidiol and antimicrobial diterpenes (Shibata et al., 2016), which may be capable of triterpenoid transport to some degree. K. fedtschenkoi candidate genes that show homology to aliphatic transporters in A. thaliana were expressed in N. benthamiana to monitor transport activity of aliphatic compounds (Figure 13). These same candidates were also co- filtrated with GgBAS to determine if they were also capable of transporting triterpenoid substrates. Similar to the K. fedtschenkoi candidates, K. thyrsiflora candidates were expressed alone to observe any changes in the amount of aliphatic compounds, and with the addition of GgBAS to monitor the transport activity of beta-amyrin. The results of these assays were assessed by computing the ratio of beta-amyrin to the 495.5 m/z. 40 Figure 13: Schematic representation of transporter assays. N. benthamiana has a wax profile containing aliphatic compounds. The transport of triterpenoids was measured in each treatment group by computing the ratio of beta- amyrin (triterpenoid) to a representative aliphatic compound in the leaf surface wax. Diagrams shown are the predicted behavior of candidate transporters in different experimental conditions. Blue transporters have aliphatic substrates, green transporters are promiscuous and transport both substrates, and yellow transporters have triterpenoid substrates. Bacterial transformation & Plant transformation To observe the effects of the candidate genes identified in a heterologous expression system, transgenic Agro was used to induct expression in N. benthamiana via the transfer of T- DNA to plant cells (Guo et al., 2019). Agro cells without a vector were transformed and grown 41 on vector-conferred resistance (kanamycin) plates as a negative control, while cells containing plasmid vectors with candidate genes were grown on rifampicin, gentamicin, and kanamycin resistance plates. The bacterial transformations were successful as Agro cells containing the candidate gene vectors had growth on LB + RGK plates and there was no growth on only kanamycin plates, indicating no additional conferred resistance (Figure 15). New plates were grown on a regular basis and the transformed Agro maintained the vector-conferred kanamycin resistance. The plant transformations were assessed using the presence of GFP as both a control and a positive indicator for gene expression when co-infiltrated with the candidate gene vectors (Figure 14). In addition to GFP, the results of the transformations were evaluated by comparing the ratio of a representative aliphatic compound (C29 alkane) to beta-amyrin, the compound synthesized by the GgBAS gene for a variety of transformation combinations. Based on the demonstrated fluorescence of GFP under the RB-GO UV light and the results obtained via GC/MS, the plant transformations were successful. The results and analysis of the GC/MS data is detailed in the next section. Figure 14: N. benthamiana leaves under RB- GO UV light. Left: leaf without any bacterial infiltration or plasmid vector. Right: leaf transformed with transgenic A. tumefaciens containing GFP gene. 42 Figure 15: Bacterial transformation results. A. Negative control containing the GV3101 strain of A. tumefaciens on kanamycin resistance plate. B. Transformed A. tumefaciens containing pEAQ vector with gene insert on rifampicin, gentamicin, and kanamycin resistance plate. Example shown is K. thyrsiflora candidate TRINITY_DN1950_c0_g1_i146. Table 4. List of transformation combinations. Fe1 = K. fedtschenkoi ABCG11 homolog, Fe2 = K. fedtschenkoi ABCG3 homolog, Th1 = K. thyrsiflora candidate 1 (TRINITY_DN1950_c0_g1_i146), Th2 = K. thyrsiflora candidate 2 (TRINITY_DN3292_c1_g1_i19), Th3 = K. thyrsiflora candidate 3 (TRINITY_DN562_c0_g1_i43), Th4 = K. thyrsiflora candidate 4 (TRINITY_DN98_c0_g1_i36). Vector None GFP GgBAS Fe1 Fe2 GgBAS + Fe1 GgBAS + Fe2 Th1 Th2 Th3 Th4 GgBAS + Th1 GgBAS + Th2 GgBAS + Th3 GgBA S + Th4 Replic ates 3 3 9 6 9 6 6 3 3 3 3 9 9 9 9 43 2.3.3 GC/MS data To determine if the plants transformed with the transporter candidates resulted in differences in triterpenoid production and transport when compared to the control, leaf wax samples were subjected to GC/MS analysis to identify any trends in the ratio of compounds produced and the total wax coverage. Samples infiltrated with the GgBAS vector alone and co- infiltrated with transporter vectors produced a peak between 37.5-37.7 minutes that contained the 218.2 m/z ion, characteristic of beta-amyrin. A representative aliphatic compound (C29 alkane) found in all samples containing a 495.5 m/z eluted at approximately 37.5 minutes. The ratio was computed by dividing the peak area for the 218.2 molecular ion by the peak area of the 495.5 molecular ion, where a larger ratio represents more beta-amyrin found on the leaf surface. Total wax coverage (ug/cm2) was determined for each sample (Supplemental Figure 6). Additional replicates were collected for samples in the same treatment groups that demonstrated high variability in peak area (Table 4). Outliers were removed from the data according to Grubb’s method. Non-parametric testing was performed to assess the significance in the differences in ratios between treatment groups (Kruskal-Wallis, Dunn post-hoc, Figure 16). Multiple approaches for p-value correction were considered, one of which being the Bonferroni correction method that is very conservative to limit the occurrence of Type I error (false positives). The False Discovery Rate (FDR) method was also considered because it provides a balance between false positives and false negatives and has more statistical power for detecting true effects. Due to the high noise level in our heterologous expression system and the relatively small effect size of beta-amyrin being detected on the leaf surface, it was determined that the FDR correction method was best. 44 Figure 16: Plot representing GC/MS data for each transformation combination. Each N. benthamiana transformation has several replicates (Table 4). Values on the y-axis represent the ratio of beta-amyrin peak area (218.2 m/z) to C29 alkane peak area (495.5 m/z). Treatments and transformation combinations are represented on the x-axis. Letters shown on the bottom of each subplot indicate significant differences in each metric for each treatment group. Letters that are shared among treatments do not have significantly different means. Ratio To test the ability of Kalanchoe transporters to accept triterpenoids as substrates, we used heterologous expression of gene candidates in N. benthamiana. Vectors containing the GFP gene and empty Agro cells (no vector) were infiltrated into N. benthamiana as controls (Figure 13). No beta-amyrin peaks were observed in these samples and the C24-alkane standard peak and the 495.5 m/z (representative aliphatic compound) were present, indicating that beta-amyrin is not synthesized natively. Beta-amyrin was detected to some degree in all samples that were 45 transformed with a vector containing GgBAS, including co-infiltrations, indicating that GgBAS is successfully synthesizing beta-amyrin in leaf cells. GgBAS when expressed alone resulted in a significantly greater ratio of beta-amyrin (0.8936) when compared to the controls (GFP/Agro; 0), indicating that when GgBAS is expressed, some beta-amyrin is transported to the surface. The control transformations behaved according to our hypotheses and the GgBAS assays suggest that N. benthamiana may have endogenous transporters capable of transporting small amounts of beta-amyrin (Figure 13). After confirming that the heterologous expression system is expressing the gene inserts from the vectors, the transporter candidates can be assayed. The K. fedtschenkoi candidates have homology to known aliphatic transporters and these plants have a variety of aliphatic compounds in their epicuticular wax. These candidates were assayed to elucidate their substrates and to serve as a comparison against the K. thyrsiflora candidates. Beta-amyrin was detected in assays containing both the candidate K. fedtschenkoi transporters and GgBAS, but the ratio did not significantly increase when compared to the controls, and the transporter candidates when expressed alone had no change in ratio. So, even when beta-amyrin is being produced there is no change in triterpenoid transport, and therefore these candidates likely transport other compounds like aliphatics (Figure 15). Inspecting the change in total amount of aliphatic compounds on the leaf surface compared to the controls could confirm whether or not the K. fedtschenkoi transporters do indeed have aliphatic transport capabilities. Additional replicates or further characterization of the K. fedtschenkoi candidate gene sequences could provide information about the transporter type and substrate specificity. Wax coverage data regarding the transport 46 activity of aliphatic compounds for the K. fedtschenkoi transporter assays can be found in Supplemental Figure 6. K. thyrsiflora has a strong wax bloom containing a high abundance of triterpenoids and likely contains triterpenoid-specific transporters. A transcriptome assembly resulted in four highly expressed ABCG transporters that were selected as candidates. All four of the transporter candidates from K. thyrsiflora when co-infiltrated with GgBAS produced a significantly higher ratio of beta-amyrin (Th1 + GgBAS; 4.642, Th2 + GgBAS; 3.740, Th3 + GgBAS; 7.341, Th4 + GgBAS; 2.898) when compared against the controls (GFP/Agro; 0, Figure 16). Nearly all of these candidates (Th1, Th2, and Th3) + GgBAS have a greater ratio of beta-amyrin present on the leaf surface than just the transporter assays alone, as the plant is not actively synthesizing beta-amyrin without GgBAS infiltration. Based on this data, the transporter candidates from K. thyrsiflora seem to have a preference for triterpenoid substrates, especially compared to the aliphatic transporter candidates from K. fedtschenkoi (Figure 13). The average ratios of all four candidates + GgBAS are significantly different from that of the suspected K. fedtschenkoi ABCG11 homolog (“Fe1”, 0.05138), even when co-infiltrated with GgBAS (0.4314; Figure 16). This furthers the notion that the K. thyrsiflora candidates demonstrate a preference for triterpenoids, while the K. fedtschenkoi candidates do not share this preference. It should be noted that the amount of beta-amyrin detected on the surface in the K. thyrsiflora transporters assays is not statistically different from that of the GgBAS assays alone, however, this comparison is more relevant to transporter activity in general rather than substrate specificity. These data strongly suggest that the candidates identified from K. thyrsiflora have specificity for triterpenoids (Figure 13). 47 2.4 Conclusion The goal of this chapter was to identify and characterize any potential triterpenoid- specific transporters from Kalanchoe plants. Despite challenges in my ability to annotate transporter sequences with a high degree of certainty, I was able to identify 6 high quality ABC transporter candidates from Kalanchoe based on domain content and expression level. The heterologous expression of these candidates from both species was successfully carried out and provided valuable insights on substrate specificity. In particular, the K. thyrsiflora candidates seem to have a much greater specificity for triterpenoid substrates relative to transporters typically associated with aliphatic substrates (K. fedtschenkoi candidates). It was also learned that N. benthamiana may have a promiscuous endogenous transporter that can also transport small amounts of triterpenoids, despite having a mainly aliphatic wax profile. Additional replicates of the assays described in 2.3.3 could further confirm transport activity of these transporters and their substrate specificity by increasing the signal to noise ratio in the GC/MS data. Controlling the leaf age for each assay presented difficulties which could be reflected in these data. Aliphatic wax compounds are naturally being synthesized in N. benthamiana as leaves grow, and their transport could be the rate limiting step and that influences what is observed on the leaf surface. However, once the leaf is fully expanded, aliphatic compounds aren’t being synthesized as readily and transport should not be affected. FUTURE DIRECTIONS While this thesis provides some foundational knowledge for how triterpenoid compounds found in epicuticular waxes may be transported, there is still much to be considered. For 48 example, the method of extracting waxes from N. benthamiana leaves could be altered to better represent the compounds being transported to the surface. Using the current method of obtaining a standard size leaf cutting (2x2), any compounds being synthesized within the leaf could be released into the solvent upon cutting. This would affect any conclusions about the transport activity of transporter candidates, as we cannot be certain that they are transported to the surface or simply released from the epidermal cells. In future assays, a “leaf dip” or chloroform rinsing method should be utilized to avoid cutting into the leaves. Many studies aimed at characterizing ABC transporters are interested in identifying specific substrates and if the transporter is monospecific or polyspecific. The assays carried out in this thesis can only provide information about what compound class is preferred by the transporter candidates, or about beta-amyrin specifically. It would be interesting to determine if these triterpenoid transporters can discern between different triterpenoid substrates, as we now have some evidence supporting the notion that they can differentiate between aliphatic compounds and triterpenoids. To do this, more assays could be done using different triterpenoid substrates with the same candidate transporters to get an idea of specificity. Triterpenoid transport in general, or by characterized aliphatic transporters has yet to be tested in a model system, which could provide meaningful insights in future studies. It would also be worth studying whether or not other terpene biosynthesis pathways are downregulated in Kalanchoe plants with a large abundance of triterpenoids in their epicuticular wax. It would also be valuable to test the ability of the transporter candidates identified in K. fedtschenkoi to transport aliphatic compounds, either those natively found on N. benthamiana leaves, or those commonly found in other epicuticular waxes. A wax coverage analysis was 49 completed in this thesis by quantifying all major peaks in each sample to determine if transport increased when transporters are overexpressed, but a more comprehensive analysis may be beneficial. It should be noted that the wax coverage for each treatment group may not be directly comparable as leaf age is a factor when analyzing surface waxes. 50 BIBLIOGRAPHY Andolfo, G., Ruocco, M., Di Donato, A., Frusciante, L., Lorito, M., Scala, F., & Ercolano, M. R. (2015). Genetic variability and evolutionary diversification of membrane ABC transporters in plants. BMC Plant Biology, 15(1), 51. https://doi.org/10.1186/s12870- 014-0323-2 Banasiak, J., & Jasiński, M. (2022). ATP-binding cassette transporters in nonmodel plants. New Phytologist, 233(4), 1597–1612. https://doi.org/10.1111/nph.17779 Bengough, A. G., McKenzie, B. M., Hallett, P. D., & Valentine, T. A. (2011). Root elongation, water stress, and mechanical impedance: A review of limiting stresses and beneficial root tip traits. Journal of Experimental Botany, 62(1), 59–68. https://doi.org/10.1093/jxb/erq350 Berardini, T. Z., Reiser, L., Li, D., Mezheritsky, Y., Muller, R., Strait, E., & Huala, E. (2015). The arabidopsis information resource: Making and mining the “gold standard” annotated reference plant genome. Genesis, 53(8), 474–485. https://doi.org/10.1002/dvg.22877 Buda, G. J., Barnes, W. J., Fich, E. A., Park, S., Yeats, T. H., Zhao, L., Domozych, D. S., & Rose, J. K. C. (2013). An ATP Binding Cassette Transporter Is Required for Cuticular Wax Deposition and Desiccation Tolerance in the Moss Physcomitrella patens. The Plant Cell, 25(10), 4000–4013. https://doi.org/10.1105/tpc.113.117648 Buschhaus, C., & Jetter, R. (2011). Composition differences between epicuticular and intracuticular wax substructures: How do plants seal their epidermal surfaces? Journal of Experimental Botany, 62(3), 841–853. https://doi.org/10.1093/jxb/erq366 https://doi.org/10.1186/s12870-014-0323-2 https://doi.org/10.1186/s12870-014-0323-2 https://doi.org/10.1186/s12870-014-0323-2 https://doi.org/10.1111/nph.17779 https://doi.org/10.1111/nph.17779 https://doi.org/10.1093/jxb/erq350 https://doi.org/10.1093/jxb/erq350 https://doi.org/10.1093/jxb/erq350 https://doi.org/10.1002/dvg.22877 https://doi.org/10.1002/dvg.22877 https://doi.org/10.1002/dvg.22877 https://doi.org/10.1105/tpc.113.117648 https://doi.org/10.1105/tpc.113.117648 https://doi.org/10.1093/jxb/erq366 https://doi.org/10.1093/jxb/erq366 51 Carra, A., Gambino, G., & Schubert, A. (2007). A cetyltrimethylammonium bromide-based method to extract low-molecular-weight RNA from polysaccharide-rich plant tissues. Analytical Biochemistry, 360(2), 318–320. https://doi.org/10.1016/j.ab.2006.09.022 Chen, H., & Jiang, J.-G. (2010). Osmotic adjustment and plant adaptation to environmental changes related to drought and salinity. Environmental Reviews, 18(NA), 309–319. https://doi.org/10.1139/A10-014 Chu, W., Gao, H., Cao, S., Fang, X., Chen, H., & Xiao, S. (2017). Composition and morphology of cuticular wax in blueberry (Vaccinium spp.) fruits. Food Chemistry, 219, 436–442. https://doi.org/10.1016/j.foodchem.2016.09.186 Cominelli, E., Galbiati, M., & Tonelli, C. (2010). Transcription factors controlling stomatal movements and drought tolerance. Transcription, 1(1), 41–45. https://doi.org/10.4161/trns.1.1.12064 Crouzet, J., Roland, J., Peeters, E., Trombik, T., Ducos, E., Nader, J., & Boutry, M. (2013). NtPDR1, a plasma membrane ABC transporter from Nicotiana tabacum, is involved in diterpene transport. Plant Molecular Biology, 82(1), 181–192. https://doi.org/10.1007/s11103-013-0053-0 Domínguez, E., Heredia-Guerrero, J. A., & Heredia, A. (2017). The plant cuticle: Old challenges, new perspectives. Journal of Experimental Botany, 68(19), 5251–5255. https://doi.org/10.1093/jxb/erx389 Elejalde-Palmett, C., Martinez San Segundo, I., Garroum, I., Charrier, L., De Bellis, D., Mucciolo, A., Guerault, A., Liu, J., Zeisler-Diehl, V., Aharoni, A., Schreiber, L., Bakan, B., Clausen, M. H., Geisler, M., & Nawrath, C. (2021). ABCG transporters https://doi.org/10.1016/j.ab.2006.09.022 https://doi.org/10.1016/j.ab.2006.09.022 https://doi.org/10.1139/A10-014 https://doi.org/10.1139/A10-014 https://doi.org/10.1139/A10-014 https://doi.org/10.1016/j.foodchem.2016.09.186 https://doi.org/10.1016/j.foodchem.2016.09.186 https://doi.org/10.4161/trns.1.1.12064 https://doi.org/10.4161/trns.1.1.12064 https://doi.org/10.4161/trns.1.1.12064 https://doi.org/10.1007/s11103-013-0053-0 https://doi.org/10.1007/s11103-013-0053-0 https://doi.org/10.1007/s11103-013-0053-0 https://doi.org/10.1093/jxb/erx389 https://doi.org/10.1093/jxb/erx389 https://doi.org/10.1093/jxb/erx389 52 export cutin precursors for the formation of the plant cuticle. Current Biology, 31(10), 2111-2123.e9. https://doi.org/10.1016/j.cub.2021.02.056 Fabre, G., Garroum, I., Mazurek, S., Daraspe, J., Mucciolo, A., Sankar, M., Humbel, B. M., & Nawrath, C. (2016). The ABCG transporter PEC1/ABCG32 is required for the formation of the developing leaf cuticle in Arabidopsis. New Phytologist, 209(1), 192– 201. https://doi.org/10.1111/nph.13608 Fallon, Timothy. (2023). transXpress: A Snakemake pipeline for streamlined de novo transcriptome assembly and annotation | BMC Bioinformatics | Full Text. https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-023-05254-8 Fu, X., Shen, Q., Yan, T., Li, L., & Tang, K. (2017). AaPDR3, a PDR Transporter 3, Is Involved in Sesquiterpene β-Caryophyllene Transport in Artemisia annua. Frontiers in Plant Science, 8. https://doi.org/10.3389/fpls.2017.00723 Ghosh, S. (2016). Biosynthesis of Structurally Diverse Triterpenes in Plants: The Role of Oxidosqualene Cyclases. Proceedings of the Indian National Science Academy, 82(4). https://doi.org/10.16943/ptinsa/2016/48578 Guo, M., Ye, J., Gao, D., Xu, N., & Yang, J. (2019). Agrobacterium-mediated horizontal gene transfer: Mechanism, biotechnological application, potential risk and forestalling strategy. Biotechnology Advances, 37(1), 259–270. https://doi.org/10.1016/j.biotechadv.2018.12.008 Hwang, J.-U., Song, W.-Y., Hong, D., Ko, D., Yamaoka, Y., Jang, S., Yim, S., Lee, E., Khare, D., Kim, K., Palmgren, M., Yoon, H. S., Martinoia, E., & Lee, Y. (2016). Plant https://doi.org/10.1016/j.cub.2021.02.056 https://doi.org/10.1016/j.cub.2021.02.056 https://doi.org/10.1111/nph.13608 https://doi.org/10.1111/nph.13608 https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-023-05254-8 https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-023-05254-8 https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-023-05254-8 https://doi.org/10.3389/fpls.2017.00723 https://doi.org/10.3389/fpls.2017.00723 https://doi.org/10.16943/ptinsa/2016/48578 https://doi.org/10.16943/ptinsa/2016/48578 https://doi.org/10.16943/ptinsa/2016/48578 https://doi.org/10.1016/j.biotechadv.2018.12.008 https://doi.org/10.1016/j.biotechadv.2018.12.008 https://doi.org/10.1016/j.biotechadv.2018.12.008 53 ABC Transporters Enable Many Unique Aspects of a Terrestrial Plant’s Lifestyle. Molecular Plant, 9(3), 338–355. https://doi.org/10.1016/j.molp.2016.02.003 Jenks, M. A., Rich, P. J., Rhodes, D., Ashworth, E. N., Axtell, J. D., & Ding, C.-K. (2000). Leaf sheath cuticular waxes on bloomless and sparse-bloom mutants of Sorghum bicolor. Phytochemistry, 54(6), 577–584. https://doi.org/10.1016/S0031- 9422(00)00153-9 Kong, L., Liu, Y., Zhi, P., Wang, X., Xu, B., Gong, Z., & Chang, C. (2020). Origins and Evolution of Cuticle Biosynthetic Machinery in Land Plants. Plant Physiology, 184(4), 1998–2010. https://doi.org/10.1104/pp.20.00913 Kretzschmar, T., Burla, B., Lee, Y., Martinoia, E., & Nagy, R. (2011). Functions of ABC transporters in plants. Essays in Biochemistry, 50, 145–160. https://doi.org/10.1042/bse0500145 Kumar, P., Kumar, P., Verma, V., Irfan, M., Sharma, R., & Bhargava, B. (2022). How plants conquered land: Evolution of terrestrial adaptation. Journal of Evolutionary Biology, 35(5), 5–14. https://doi.org/10.1111/jeb.14062 Lefèvre, F., & Boutry, M. (2018). Towards Identification of the Substrates of ATP-Binding Cassette Transporters. Plant Physiology, 178(1), 18–39. https://doi.org/10.1104/pp.18.00325 Li, N., Xu, C., Li-Beisson, Y., & Philippar, K. (2016). Fatty Acid and Lipid Transport in Plant Cells. Trends in Plant Science, 21(2), 145–158. https://doi.org/10.1016/j.tplants.2015.10.011 https://doi.org/10.1016/j.molp.2016.02.003 https://doi.org/10.1016/j.molp.2016.02.003 https://doi.org/10.1016/S0031-9422(00)00153-9 https://doi.org/10.1016/S0031-9422(00)00153-9 https://doi.org/10.1016/S0031-9422(00)00153-9 https://doi.org/10.1104/pp.20.00913 https://doi.org/10.1104/pp.20.00913 https://doi.org/10.1042/bse0500145 https://doi.org/10.1042/bse0500145 https://doi.org/10.1042/bse0500145 https://doi.org/10.1111/jeb.14062 https://doi.org/10.1111/jeb.14062 https://doi.org/10.1104/pp.18.00325 https://doi.org/10.1104/pp.18.00325 https://doi.org/10.1104/pp.18.00325 https://doi.org/10.1016/j.tplants.2015.10.011 https://doi.org/10.1016/j.tplants.2015.10.011 https://doi.org/10.1016/j.tplants.2015.10.011 54 Liu, D., Tang, J., Liu, Z., Dong, X., Zhuang, M., Zhang, Y., Lv, H., Sun, P., Liu, Y., Li, Z., Ye, Z., Fang, Z., & Yang, L. (2017). Fine mapping of BoGL1, a gene controlling the glossy green trait in cabbage (Brassica oleracea L. Var. Capitata). Molecular Breeding, 37(5), 69. https://doi.org/10.1007/s11032-017-0674-0 McFarlane, H. E., Shin, J. J. H., Bird, D. A., & Samuels, A. L. (2010). Arabidopsis ABCG Transporters, Which Are Required for Export of Diverse Cuticular Lipids, Dimerize in Different Combinations. The Plant Cell, 22(9), 3066–3075. https://doi.org/10.1105/tpc.110.077974 Mistry, J., Chuguransky, S., Williams, L., Qureshi, M., Salazar, G. A., Sonnhammer, E. L. L., Tosatto, S. C. E., Paladin, L., Raj, S., Richardson, L. J., Finn, R. D., & Bateman, A. (2021). Pfam: The protein families database in 2021. Nucleic Acids Research, 49(D1), D412–D419. https://doi.org/10.1093/nar/gkaa913 Moses, T., Pollier, J., Shen, Q., Soetaert, S., Reed, J., Erffelinck, M.-L., Van Nieuwerburgh, F. C. W., Vanden Bossche, R., Osbourn, A., Thevelein, J. M., Deforce, D., Tang, K., & Goossens, A. (2015). OSC2 and CYP716A14v2 Catalyze the Biosynthesis of Triterpenoids for the Cuticle of Aerial Organs of Artemisia annua. The Plant Cell, 27(1), 286–301. https://doi.org/10.1105/tpc.114.134486 Naaz, S., Ahmad, N., & Qureshi, M. I. (n.d.). ATP Binding Cassette (ABC) Transporters in Plant Development and Defense. Nguyen, D., Groth, N., Mondloch, K., Cahoon, E. B., Jones, K., & Busta, L. (2023). Project ChemicalBlooms: Collaborating with Citizen Scientists to Survey the Chemical https://doi.org/10.1007/s11032-017-0674-0 https://doi.org/10.1007/s11032-017-0674-0 https://doi.org/10.1105/tpc.110.077974 https://doi.org/10.1105/tpc.110.077974 https://doi.org/10.1105/tpc.110.077974 https://doi.org/10.1093/nar/gkaa913 https://doi.org/10.1093/nar/gkaa913 https://doi.org/10.1105/tpc.114.134486 https://doi.org/10.1105/tpc.114.134486 55 Diversity and Phylogenetic Distribution of Plant Epicuticular Wax Blooms (p. 2023.11.11.566687). bioRxiv. https://doi.org/10.1101/2023.11.11.566687 Nguyen, V. N. T., Lee, S. B., Suh, M. C., An, G., & Jung, K.-H. (2018). OsABCG9 Is an Important ABC Transporter of Cuticular Wax Deposition in Rice. Frontiers in Plant Science, 9. https://www.frontiersin.org/articles/10.3389/fpls.2018.00960 Nguyen, V. N. T., Moon, S., & Jung, K.-H. (2014). Genome-wide expression analysis of rice ABC transporter family across spatio-temporal samples and in response to abiotic stresses. Journal of Plant Physiology, 171(14), 1276–1288. https://doi.org/10.1016/j.jplph.2014.05.006 Patwari, P., Salewski, V., Gutbrod, K., Kreszies, T., Dresen-Scholz, B., Peisker, H., Steiner, U., Meyer, A. J., Schreiber, L., & Dörmann, P. (2019). Surface wax esters contribute to drought tolerance in Arabidopsis. The Plant Journal, 98(4), 727–744. https://doi.org/10.1111/tpj.14269 Philippe, G., De Bellis, D., Rose, J. K. C., & Nawrath, C. (2022). Trafficking Processes and Secretion Pathways Underlying the Formation of Plant Cuticles. Frontiers in Plant Science, 12. https://www.frontiersin.org/articles/10.3389/fpls.2021.786874 Rutledge, R. M., Esser, L., Ma, J., & Xia, D. (2011). Toward understanding the mechanism of action of the yeast multidrug resistance transporter Pdr5p: A molecular modeling study. Journal of Structural Biology, 173(2), 333–344. https://doi.org/10.1016/j.jsb.2010.10.012 Sainsbury, F., Thuenemann, E. C., & Lomonossoff, G. P. (2009). pEAQ: Versatile expression vectors for easy and quick transient expression of heterologous proteins in https://doi.org/10.1101/2023.11.11.566687 https://doi.org/10.1101/2023.11.11.566687 https://www.frontiersin.org/articles/10.3389/fpls.2018.00960 https://www.frontiersin.org/articles/10.3389/fpls.2018.00960 https://doi.org/10.1016/j.jplph.2014.05.006 https://doi.org/10.1016/j.jplph.2014.05.006 https://doi.org/10.1016/j.jplph.2014.05.006 https://doi.org/10.1111/tpj.14269 https://doi.org/10.1111/tpj.14269 https://doi.org/10.1111/tpj.14269 https://www.frontiersin.org/articles/10.3389/fpls.2021.786874 https://www.frontiersin.org/articles/10.3389/fpls.2021.786874 https://doi.org/10.1016/j.jsb.2010.10.012 https://doi.org/10.1016/j.jsb.2010.10.012 https://doi.org/10.1016/j.jsb.2010.10.012 56 plants. Plant Biotechnology Journal, 7(7), 682–693. https://doi.org/10.1111/j.1467- 7652.2009.00434.x Shibata, Y., Ojika, M., Sugiyama, A., Yazaki, K., Jones, D. A., Kawakita, K., & Takemoto, D. (2016). The Full-Size ABCG Transporters Nb-ABCG1 and Nb-ABCG2 Function in Pre- and Postinvasion Defense against Phytophthora infestans in Nicotiana benthamiana. The Plant Cell, 28(5), 1163–1181. https://doi.org/10.1105/tpc.15.00721 Singh, S., Das, S., & Geeta, R. (2018). Role of Cuticular Wax in Adaptation to Abiotic Stress: A Molecular Perspective. In S. M. Zargar & M. Y. Zargar (Eds.), Abiotic Stress- Mediated Sensing and Signaling in Plants: An Omics Perspective (pp. 155–182). Springer. https://doi.org/10.1007/978-981-10-7479-0_5 Thimmappa, R., Geisler, K., Louveau, T., O’Maille, P., & Osbourn, A. (2014). Triterpene Biosynthesis in Plants. Annual Review of Plant Biology, 65(1), 225–257. https://doi.org/10.1146/annurev-arplant-050312-120229 van Maarseveen, C., & Jetter, R. (2009). Composition of the epicuticular and intracuticular wax layers on Kalanchoe daigremontiana (Hamet et Perr. De la Bathie) leaves. Phytochemistry, 70(7), 899–906. https://doi.org/10.1016/j.phytochem.2009.04.011 Vasiliou, V., Vasiliou, K., & Nebert, D. W. (2009). Human ATP-binding cassette (ABC) transporter family. Human Genomics, 3(3), 281. https://doi.org/10.1186/1479-7364-3-3- 281 Wasaya, A., Zhang, X., Fang, Q., & Yan, Z. (2018). Root Phenotyping for Drought Tolerance: A Review. Agronomy, 8(11), Article 11. https://doi.org/10.3390/agronomy8110241 https://doi.org/10.1111/j.1467-7652.2009.00434.x https://doi.org/10.1111/j.1467-7652.2009.00434.x https://doi.org/10.1111/j.1467-7652.2009.00434.x https://doi.org/10.1105/tpc.15.00721 https://doi.org/10.1105/tpc.15.00721 https://doi.org/10.1007/978-981-10-7479-0_5 https://doi.org/10.1007/978-981-10-7479-0_5 https://doi.org/10.1146/annurev-arplant-050312-120229 https://doi.org/10.1146/annurev-arplant-050312-120229 https://doi.org/10.1146/annurev-arplant-050312-120229 https://doi.org/10.1016/j.phytochem.2009.04.011 https://doi.org/10.1016/j.phytochem.2009.04.011 https://doi.org/10.1186/1479-7364-3-3-281 https://doi.org/10.1186/1479-7364-3-3-281 https://doi.org/10.1186/1479-7364-3-3-281 https://doi.org/10.3390/agronomy8110241 https://doi.org/10.3390/agronomy8110241 https://doi.org/10.3390/agronomy8110241 57 Yeats, T. H., & Rose, J. K. C. (2013). The Formation and Function of Plant Cuticles. Plant Physiology, 163(1), 5–20. https://doi.org/10.1104/pp.113.222737 Zhang, L., Sasaki‐Sekimoto, Y., Kosetsu, K., Aoyama, T., Murata, T., Kabeya, Y., Sato, Y., Koshimizu, S., Shimojima, M., Ohta, H., Hasebe, M., & Ishikawa, M. (2023). An ABCB transporter regulates anisotropic cell expansion via cuticle deposition in the moss Physcomitrium patens. New Phytologist, nph.19337. https://doi.org/10.1111/nph.19337 https://doi.org/10.1104/pp.113.222737 https://doi.org/10.1104/pp.113.222737 https://doi.org/10.1111/nph.19337 https://doi.org/10.1111/nph.19337 https://doi.org/10.1111/nph.19337 58 Appendix Multiple sequence alignment Supplemental Figure 1. Sequences with high similarity to characterized Arabidopsis thaliana ABC transporters (BLASTp). Alignment contains ABC transporter sequences from all nine species described in Table 1. Amino acid position is reported on the x-axis, gap percent and conservation percent are displayed in the top panel. 59 Supplemental Figure 2. Sequence alignment of WBC group ABC transporters. Sequences shown represent all nine species listed in Table 1, where sequences were sorted into WBC groups based on the number of nucleotide binding domains present. Some sequences contained more than one (WBC) or two (PDR) nucleotide binding domain, as reported in literature. Amino acid position is reported on the x-axis, gap percent and conservation percent are displayed in the top panel. 60 Supplemental Figure 3. Sequence alignment of subfamily G ABC transporters as predicted by hmmscan. This alignment contains all nine species described in Table 1. An HMM profile was generated using characterized ABCG sequences and sequences from the nine species with a high degree of similarity to those containing ABCG domains are shown. Amino acid position is reported on the x-axis, gap percent and conservation percent are displayed in the top panel. 61 Supplemental Figure 4. Sequence alignment of sequences from K. thyrsiflora and K. blossfeldiana that have homology to K. fedtschenkoi ABC transporters. Amino acid position is reported on the x-axis, gap percent and conservation percent are displayed in the top panel. 62 Supplemental Figure 5. pEAQ vector map. pBINPLUS backbone with Cowpea mosaic virus promoter, kanamycin resistance, T-DNA region with left and right borders, RK2 replication origin and trfA, and NOS promoter and terminator regions (Sainsbury, 2009). Plasmid map generated in SnapGene Viewer. 63 Supplemental Figure 6. Wax coverage data for each treatment group from heterologous expression assays. Wax coverage is represented by the total area of all major peaks normalized to the area of the C24 standard peak. The K. fedtschenkoi ABCG3 homolog (kfed_unknown) appears to have the greatest increase in wax coverage when compared to the controls. There is no difference in the wax coverage between the ABCG11 homolog alone and when co-expressed with GgBAS, indicating that this candidate does not have specificity for triterpenoid compounds. The K. thyrsiflora candidates (Th1-4) are among the lowest in wax coverage, indicating that they do not transport wax compounds naturally found on the surface of N. benthamiana leaves. Any transporter capable of transporting triterpenoids is not expected to drastically alter the amount of wax present on the leaves, as we are measuring a small beta-amyrin peak against the background of N. benthamiana’s native wax profile. It cannot be certain that the leaves in each sample were analyzed at the same age, so perhaps these treatments are not directly comparable as the stage of leaf development can affect the amount of wax present. 64 Supplemental Table 1. R packages and versions used for data analysis and visualization. All other source code is available at https://github.com/thebustalab/thebustalab.github.io. Package Version tidyverse 1.1.0 dplyr 1.3.0 Bioconductor 3.18 DESeq2 1.43.5 RHmm 2.2.0 tximport 1.31.1 rBLAST 0.99.4 Supplemental Table 2. Detailed annotations for unique Arabidopsis genes identified in the candidate gene analysis (Figure 12). Information was retrieved from the TAIR database (Berardini et al., 2015). TAIR accession Gene Model Name Primary Gene Symbol ATCG00480.1 chloroplast-encoded gene for beta subunit of ATP synthase ATP SYNTHASE SUBUNIT BETA (PB) AT4G12020.2 Encodes a member of the A1 subgroup of the MEKK (MAPK/ERK kinase kinase) family. MEKK is another name for Mitogen-Activated Protein Kinase Kinase Kinase (MAPKKK or MAP3K). This subgroup has four members: At4g08500 (MEKK1, also known as ARAKIN, MAP3Kb1, MAPKKK8), At4g08480 (MEKK2, also known as MAP3Kb4, MAPKKK9), (WRKY19) https://github.com/thebustalab/thebustalab.github.io 65 At4g08470 (MEKK3, also known as MAP3Kb3, MAPKKK10) and At4g12020 (MEKK4, also known as MAP3Kb5, MAPKKK11, WRKY19). Nomenclatures for mitogen-activated protein kinases are described in Trends in Plant Science 2002,7(7):301. Co-regulates with DSC1 basal levels of immunity to root-knot nematodes. AT2G47000.1 Encodes an auxin efflux transmembrane transporter that is a member of the multidrug resistance P-glycoprotein (MDR/PGP) subfamily of ABC transporters. Functions in the basipetal redirection of auxin from the root tip. Exhibits apolar plasma membrane localization in the root cap and polar localization in tissues above and is involved in root hair elongation. ATP- BINDING CASSETTE B4 (ABCB4) AT2G03670.1 CDC48 - like protein AAA-type ATPaseCell. division control protein 48 homolog B CELL DIVISION CYCLE 48B (CDC48B) AT2G25140.1 Encodes ClpB4, which belongs to the Casein lytic proteinase/heat shock protein 100 (Clp/Hsp100) family. Targeted to the mitochondrion, also referred to as ClpB- m. Transcripts of ClpB4 accumulate dramatically at high temperatures, suggesting that it may be involved in response to heat stress. CASEIN LYTIC PROTEINASE B4 (CLPB4) AT2G28070.1 ABC-2 type transporter family protein;(source:Araport11) ATP- BINDING CASSETTE G3 (ABCG3) AT2G41700.1 ATP-binding cassette A1;(source:Araport11) ATP- BINDING CASSETTE A1 (ABCA1) AT5G48600.2 member of SMC subfamily STRUCTURA L MAINTENAN CE OF CHROMOSO ME 3 (SMC3) AT4G26090.1 Encodes a plasma membrane protein with leucine-rich RESISTANT 66 repeat, leucine zipper, and P loop domains that confers resistance to Pseudomonas syringae infection by interacting with the avirulence gene avrRpt2. RPS2 protein interacts directly with plasma membrane associated protein RIN4 and this interaction is disrupted by avrRpt2. The mRNA is cell-to-cell mobile. TO P. SYRINGAE 2 (RPS2) AT2G19490.1 recA DNA recombination family protein;(source:Araport11) A. THALIANA RECA HOMOLOG 2 (RECA2) AT2G47800.1 Encodes a plasma membrane localized ATPase transporter involved in multidrug transport. The expression of this gene is upregulated by herbicide safeners such as benoxacor, fluxofenim and fenclorim. The mRNA is cell-to-cell mobile. ATP- BINDING CASSETTE C4 (ABCC4) AT2G36910.1 Belongs to the family of ATP-binding cassette (ABC) transporters. Also known as AtMDR1.Possibly regulates auxin-dependent responses by influencing basipetal auxin transport in the root. Exerts nonredundant, partially overlapping functions with the ABC transporter encoded by AT3G28860. PGP1 mediates cellular efflux of IAA and interacts with PIN genes that may confer an accelerated vectoral component to PGP-mediated transport. The non-polar localization of PGP1 at root and shoot apices, where IAA gradient-driven transport is impaired, may be required to confer directionality to auxin transport in those tissues. The mRNA is cell-to-cell mobile. ATP- BINDING CASSETTE B1 (ABCB1) AT2G31970.1 Encodes the Arabidopsis RAD50 homologue. It is involved in double strand break repair. Component of the meiotic recombination complex that processes meiotic double-strand-breaks to produce single-stranded DNA ends, which act in the homology search and recombination. Accumulates in the nucleus during meiotic prophase, a process regulated by PHS1. (RAD50) 67 Supplemental files: transporter_evolution – IPYNB file: contains all of the R code to complete the analyses outlined in Chapter 1, section 1.4 - Exploring ABC transporter evolution using publicly available genomes. RNAseq_data – IPYNB file: contains all of the R code used for processing assembly data, completing the differential expression analysis, and identifying expression based candidates as outlined in Chapter 2, section 2.2. Nicotiana_data – IPYNB file: contains all of the R code for processing the GC/MS data and statistical tests as outline in Chapter 2, section 3.3.