Relative genetic diversity of the rare and endangered Agave shawii ssp. shawii and associated soil microbes within a southern California ecological preserve

Abstract Shaw's Agave (Agave shawii ssp. shawii) is an endangered maritime succulent growing along the coast of California and northern Baja California. The population inhabiting Point Loma Peninsula has a complicated history of transplantation without documentation. The low effective population size in California prompted agave transplanting from the U.S. Naval Base site (NB) to Cabrillo National Monument (CNM). Since 2008, there are no agave sprouts identified on the CNM site, and concerns have been raised about the genetic diversity of this population. We sequenced two barcoding loci, rbcL and matK, of 27 individual plants from 5 geographically distinct populations, including 12 individuals from California (NB and CNM). Phylogenetic analysis revealed the three US and two Mexican agave populations are closely related and have similar genetic variation at the two barcoding regions, suggesting the Point Loma agave population is not clonal. Agave‐associated soil microbes used significantly more carbon sources in CNM soil samples than in NB soil likely due to higher pH and moisture content; meanwhile, soil type and soil chemistry analysis including phosphorus, nitrate nitrogen, organic matter, and metals revealed significant correlations between microbial diversity and base saturation (p < 0.05, r 2 = 0.3676), lime buffer capacity (p < 0.01, r 2 = 0.7055), equilibrium lime buffer capacity (p < 0.01, r 2 = 0.7142), and zinc (p < 0.01, r 2 = 0.7136). Soil microbiome analysis within the CNM population revealed overall expected richness (H′ = 5.647–6.982) for Agave species, while the diversity range (1 − D = 0.003392–0.014108) suggests relatively low diversity marked by high individual variation. The most prominent remaining US population of this rare species is not clonal and does not seem to be threatened by a lack of genetic and microbial diversity. These results prompt further efforts to investigate factors affecting Agave's reproduction and fitness.


| INTRODUC TI ON
Habitat loss and fragmentation are major threats to ecosystem stability. Habitat loss reduces the resources and area available to species, while habitat fragmentation results in patches of populations and limits interactions among them (Pardini et al., 2018). Both negatively affect biodiversity (Cagnolo et al., 2009;Newbold et al., 2015), foodweb structure (Bartlett et al., 2016;Cagnolo et al., 2009), and the dispersal and reproduction of species (Browne & Karubian, 2018;Lander et al., 2019;Torrenta et al., 2018). Slow growth and low reproductive rates species, such as agaves, are especially susceptible to environmental disturbances (Martínez-Palacios et al., 1999).
Agave shawii ssp. shawii (Shaw's Agave) is a maritime semelparous succulent, small-to-medium plant inhabiting the Pacific Coast from southwestern California to Baja California (Figure 1g). It has green ovate leaves 20-50 cm long and 8-20 cm wide, and spines along the margins of the leaves; rosettes grow 0.08-2 m wide and 1-2.5 m tall (Vanderplank, 2014;Vanderplank & Lombardo, 2015). It reproduces and grows slowly, taking 20-40 years to reach the flowering stage (Gentry, 1978). This taxon is seriously endangered in California, with a global rank between G2 (imperiled) and G3 (vulnerable), and a state rank of S1 (critically imperiled; California Native Plant Society, 2020).
Once abundant across coastal southern San Diego, the Shaw's Agave population has drastically declined since the 1970s primarily due to urban development (Gentry, 1978;Roof, 1971;Vanderplank, 2014). The population remnants consisting of native individuals and transplants over the last several decades exist on the northern edge of Naval Base Point Loma (NB) (Beauchamp, 1976;Gentry, 1978;Vanderplank, 2014;Vanderplank & Lombardo, 2015). Since 1976, Shaw's Agave threatened by eroding cliffs was transplanted from the NB into the neighboring Cabrillo National Monument (CNM) and across coastal San Diego County (Beauchamp, 1976;Vanderplank, 2014). Three total populations of Shaw's Agave in San Diego County were reported in 1978 (Gentry, 1978). In 2008, a native population on the east side of the Border Field State Park (Border Crossing, BC) was reduced to a single clump of rosettes due to the border fence construction. As of 2014, there are six total populations of Shaw's Agave in the United States: four entirely transplanted (CNM, Torrey Pines State Reserve), and two partially transplanted (east of BC and NB). The majority of rosettes survived transplantation (Vanderplank, 2014), but the transplantations may have compromised the plant's ecological fitness. Since 2008, no new germinating individuals of Shaw's Agave have been noted (Vanderplank, 2014;Vanderplank & Lombardo, 2015) and since 2014, fewer than 1,000 individuals grow in San Diego County. The lack of new seedlings, geographic isolation, and translocation contributed to a concern that the individuals at Point Loma are clonal. Mexican agave populations are more contiguous, intact, abundant, have a larger range of natural habitat, and have evidence of seedling recruitment (sexual reproduction) and thus presumably more genetically diverse; US agave populations grow in shallow soil, produce fewer rosettes per individual, and are considered to be almost exclusively clonal Vanderplank & Lombardo, 2015).
In this study, the relative genetic diversity within and between five Shaw's Agave populations from CNM (US), NB (US), BC (US), Rosarito, northern Baja California (RS-MX), and Arroyo Hondo, northern Baja California (AH-MX) populations was evaluated. We utilized DNA barcodes to confirm the species identity of these US agave populations with unclear transplantation origin at the molecular level and assess whether the observed differences between the US and MX populations are due to the US population being mostly clonal or having relatively lower genetic diversity. The standard DNA barcode genes for land plants, rbcL and matK, were sequenced to assess the relative genetic diversity of Shaw's Agave along its range.
rbcL is a highly conserved gene in plants that offers high universality and high sequence quality but low species discrimination power, while matK, a rapidly evolving gene region, allows for high levels of species discrimination. The combination of these loci yields high resolution and discrimination power (Hollingsworth et al., 2009).
Microbial diversity is a biomarker for a plant's health and fitness (Srivastava et al., 1996). Plant fitness depends on inherited genes and microbes living on and in the organism (Rosenberg et al., 2010). Microbiota significantly contribute to plant health by promoting plant growth, nutrient cycling, suppression of soilborne plant diseases (Mendes et al., 2011;Podila et al., 2009;Rosenberg et al., 2010) and can increase plant fitness by limiting abiotic stress, including temperature, drought, and flooding (Grover et al., 2011). A recently transplanted individual may not have established a diverse and stable microbiome compared to older individuals, and soil composition was also noted as possibly affecting microbial diversity. Shaws' Agave-associated soil samples, ranging from rocky, fibrous to fine sandy soil, were collected in CNM and NB ( Figure 1h) to quantify soil microbiota diversity and activity via 16S rRNA sequencing and microbial enzymatic assays. The highly conserved 16S rRNA V4 gene region offers an effective survey of microbial diversity and microbial phylogenies assembly. Coupled with 16S rRNA sequencing data, the metabolic activity assay provides insight into the microbiome's structure, abundance, and  A lack of genetic diversity among Shaw's Agave could leave the population vulnerable to disease, drought, and other stressors (Frankham et al., 2002;Keller & Waller, 2002). Species loss has negative ecological effects on predatorial food sources, parasitic or symbiotic organisms, and even the surrounding landscape (Holt & Loreau, 2002). Urbanization and U.S.-Mexico border fencing have led to the fragmentation of the Agave's natural environment, potentially facilitating inbreeding and reducing gene flow.

| Sampling
Shaw's Agave leaf tissues were collected from 27 individual plants

| Barcoding
Approximately 100 mg of Shaw's Agave spineless leaf tissue were snap-frozen with liquid nitrogen and homogenized using a mortar and pestle. DNA was isolated using a DNeasy 96 Plant kit (Qiagen DNeasy) according to the manufacturer's instructions. Barcoding regions of rbcL and matK were amplified in an Eppendorf Mastercycler using GoTaq Green. Primers generated in Primer3plus for rbcL PCR products were purified using a PCR clean-up kit (LAMDA Biotech), and samples were sequenced by Eton Bioscience Inc.

| Phylogenetic tree construction
DNA sequences and chromatograms for rbcL and matK were visualized in FinchTV for quality checking. Multiple sequence alignments were performed in GUIDANCE2; DNA segments with low confidence and/or ambiguous reads were manually removed. Phylogenetic trees were generated in ClustalOmega using neighbor-joining and edited in the Interactive Tree of Life (iTOL; Letunic & Bork, 2019).

| Characterization of soil pH, moisture, and chemistry
pH: 3 g of each soil sample was diluted in 6 ml of sterile deionized water, vortexed, incubated at room temperature for 30 min, and gently vortexed every 5 min before pH was measured.
Moisture: 1 g soil samples were oven-dried with a Quincy Lab Inc [QC] Model 30GC Lab Oven at 110°C for 7 days. Mass of each sample before and after incubation was compared to determine percent moisture content.
Chemical properties: Pooled soil samples from 12 sites (from CNM and NB) were sent to the Soil, Plant and Water (SPW) Laboratory at University of Georgia, where they were tested for lime buffer capacity (LBC) by measuring the pH before and 30 min after Ca(OH) 2 addition (Sonon et al., 2015), pH (Kissel & Vendrell, 2012), cation exchange capacity (CEC), metallic elements, phosphorus, nitrate nitrogen by Mehlich I sum and Percent Base Saturation (Sonon et al., 2017), and organic matters (OM) by percentage weight loss on ignition for 3 hr at 360°C.

| Characterization of soil functional biodiversity
Soil microbes were isolated and diluted into an EcoPlate to measure microbial metabolism over time. 0.5 g of soil was diluted in 14.5 ml of sterile water and vortexed at maximum speed for 5 min.
Soil particles settled for 2 min, and a 1:30 dilution of the supernatant was added to one section of a 96-well Biolog EcoPlate.
Escherichia coli OP-50 (Caenorhabditis Genetic Center, University of Minnesota, St. Paul, MN) was used as the positive control.
Absorbances at 590 nm were recorded and compared before and after 7 days of incubation at 25°C. Negative values and changes less than 0.25 Abs were considered not biologically relevant and were set to zero (Garland, 1997). Wells were normalized to the maximal observed value for each carbon source. Functional biodiversity was quantified by calculating Shannon evenness and richness (Gryta et al., 2014).

| 16S rRNA microbial sequencing and analysis
Genomic DNA was extracted from soil samples using a Quick-

| Phylogenetics
Seventeen neutral polymorphisms and 65 non-neutral polymorphisms were identified between the US and MX sites for the rbcL locus. One neutral polymorphism and 9 non-neutral polymorphisms were detected between the US and MX sites for the matK locus; no divergences were detected for either locus. Synonymous/nonsyn-  CNM to RS (p = 0.001), and 3. NB to RS (p = 0.035; HAMOVA, Table   S2). Genetic similarities between samples of different geographic sites for the rbcL and matK loci can be visualized in Figure 2.   Table S4.

| Shaw's Agave soil microbes' carbon source utilization varies greatly
Microbial metabolism assay analysis revealed differences in richness across Shaw's Agave individuals and relatively even distributions of utilized sources. Shannon's evenness index near 0.7 was characteristic for CNM 25, CNM 30, and NB 98 samples, clustering within a major "low richness" group (Table 1, Figure 4). Soil from the same F I G U R E 2 Unrooted phylogenetic tree of Shaw's Agave generated in iTOL for rbcL and matK sequences. Sample names consist of geographic source (e.g., "CNM" or "NB"), plant identifier, and biological replicate (e.g., A/B/C). CNM = Cabrillo National Monument, NB = Point Loma Naval Base, BC = Border Fields, RS = Rosarito, and AH = Arroyo Hondo. (a) The suffix "R" in the above figure refers to the gene rbcL, while (b) the suffix "M" in the above figure refers to the gene matK. Phylogenetic trees were generated in Clustal Omega using neighbor-joining and further edited in the Interactive Tree of Life (iTOL). Radial bars represent evolutionary time distances from common ancestors. Synonymous-nonsynonymous analysis of (c) rbcL and (d) matK was performed for all sample pairs using SNAP v2.   (Table S5). The Bray-Curtis beta diversity calculator suggests no specific microbial community structure distinguishes CNM from NB samples ( Figure 6).

| Correlation between soil properties, carbon utilization, and 16S rRNA diversity
Base saturation is significantly correlated with Shannon's Index of soil microbe functional diversity (p < 0.05, r 2 = 0.3676). Lime buffer capacity (p < 0.01, r 2 = 0.7055), equilibrium lime buffer capacity (p < 0.01, r 2 = 0.7142), and zinc content of the soil (p < 0.01, r 2 = 0.7136) are significantly correlated with Shannon's Index of soil microbe structural diversity. There is no correlation between 16S rRNA diversity and carbon source utilization or functional diversity.

| D ISCUSS I ON
We utilized rbcL and matK barcoding genes, microbial enzymatic assays, and 16S rRNA sequencing to assess the genetic diversity of Shaw's Agave and determine the plants' microbiome and its functional diversity. This rare succulent species is not clonal and exhibit genetic diversity across its habitat range. Low enzymatic activity in 36% of Agave soils samples may be indicative of compromised fitness among transplanted plants. 16S rRNA analysis suggests that Agave's microbiota is diverse within and between sites.

| Phylogenetics
Sequenced rbcL and matK loci indicate similar genetic diversity be-  had scores of 0 across evenness, richness, and Shannon's index. From CNM, evenness ranged from 0 to 0.995, richness ranged from 0 to 28, and Shannon's index ranged from 0 to 3.24. From NB, evenness ranged from 0 to 0.997, richness ranged from 0 to 29, and Shannon's index ranged from 0 to 3.31. There is no statistically significant difference between the two sites for any of the three diversity indicators.

| Differences in soil between Cabrillo National Monument and Naval Base Point Loma
CNM soil is significantly more alkaline (p < 0.05) and moist (p < 0.0001) than NB, and CNM soil microbes use significantly more carbon sources then NB soil microbes (p < 0.05). Although both sites are within Point Loma Peninsula, differences in altitude, soil type, and proximity to the ocean may contribute to such variation. The higher pH and moisture of CNM may be a more suitable environment for microbial metabolic carbon utility (Bell et al., 2008;Zhalnina et al., 2014). Our data show no indication of heavy metal contamination in the Point Loma Shaw's Agave habitat (Table S4).
Soil microbial metabolic activity assay shows large variations both within and between sites. In the metabolic profiling of CNM and NB soil, three major clusters arise: two metabolically active and one relatively inactive agave microbiota enzyme activity clusters.  Table S6). Lime buffer capacity (p < 0.01, r 2 = 0.7055) and equilibrium lime buffer capacity (p < 0.01, r 2 = 0.7142) are significantly positively correlated with 16S rRNA structural diversity of soil microbe. As microbe needs to adapt to pH changes (Bååth, 1996;Sheng & Zhu, 2018) structural and enzymatic functions (Andreini et al., 2006;Gielda & Diritaa, 2012). The positive correlation suggests that maintaining the higher detected level of zinc (6.82-9.88 mg/kg) is beneficial for the microbial diversity in this habitat. Shannon's calculated diversity indices range from 6.1832 to 6.7691 across all samples, and Simpson's calculated diversity indices range from 0.0040 to 0.0079 (Table S5,

| LI M ITATI O N S
Phylogenetic analysis of the barcoding gene regions rbcL and matK is a prequel for deeper sequencing of more comprehensive hypervariable loci, such as ycf1. Both rbcL and matK genes encoded in chloroplast are standard DNA barcodes for land plants (Hollingsworth et al., 2009). While rbcL is highly conserved and allows for differentiation between many species, it does not have great discrimination power within species; rbcL performs well in combination with other loci (Hollingsworth et al., 2009). For instance, matK has high levels of discrimination power between angiosperm species and is often used in combination with rbcL (Burgess et al., 2011;Hollingsworth et al., 2009). The matK/rbcL combination is the standard two-locus DNA barcode (Li et al., 2011). While using multiple loci would increase resolution and power, a 2-locus barcode was chosen for cost-effectiveness (Burgess et al., 2011;Hollingsworth et al., 2009).
However, sequencing more expansive loci will yield a statistically more powerful phylogenetic analysis. The barcoding method is also relatively limited in its discrimination resolution when differentiating between individuals within a species, while more powerful genetic markers such as SSRs and SNPs will yield more accurate and conclusive genetic diversity data.
Low sample size and lack of robust morphological markers among geographically distant agave, while do not affect the testing of our

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.