Leaves of visually healthy almond trees growing in an orchard located in Modesto, California (37°42′21.8″N, 120°56′55.1″W) were collected aseptically in July 2019. Leaves were placed in sterile bags and immediately stored at 4°C in a portable cooler and then transported to the laboratory at the University of California, Merced. Within 3 h after collection, leaves were surface sterilized using 8.25% sodium hypochlorite and afterward rinsed twice with sterile water. Removal of epiphytic microbes was confirmed using scanning electron microscopy (SEM; Figure 1). Surface sterilization and SEM procedures were previously described for cottonwood leaves (Saldierna Guzmán et al., 2020).

Immediately after removing epiphytic microbes, 20 g of surface sterilized leaves were blended using bacterial cell extraction buffer (50 mM Tris–HCl [pH7.5], 1% Triton X-100, 2 mM β-mercaptoethanol) for 3 min to enrich for endophytic bacterial cells (Ikeda et al., 2009). We found that filtering through Miracloth and centrifugation at 500 × g was sufficient to remove plant cell debris. Therefore, we excluded the Nycodenz density gradient centrifugation described in Ikeda et al. (2009). The enriched endophytic bacterial cells were resuspended using 1x phosphate-buffered saline (PBS). Ten microliters of this suspension was plated on Lysogeny Broth (LB) agar media (Bertani, 1951) or on Norris Glucose Nitrogen Free Media (HIMEDIA, Mumbai, Maharashtra, India), and then incubated 5–7 days at 28°C. Subsequently, the obtained colonies were repeatedly streaked and incubated on LB agar media (Bertani, 1951) to obtain pure isolates. A total of 100 bacterial isolates were obtained from surface-sterilized almond leaves. Of these bacteria, 22 isolates were randomly selected and tested for their plant growth-promoting abilities (see below under “Plant inoculation”). Only one of these 22 isolates, named A4, was chosen for its plant growth-promoting ability.

In order to identify the isolates, 16S rRNA gene amplification of each sample was conducted using universal primers 27f and 1492r (Lane, 1991). The 50-μl PCR contained 1x complete reaction buffer, 200 μM dNTPs, 0.5 μM forward primer, 0.5 μM reverse primer, 10 ng of template DNA, and 1.25 U DFS-Taq DNA polymerase (BIORON GmbH, Römerberg, Rheinland-Pfalz, Germany). Cycling conditions were 95°C for 2 min, followed by 25 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 2 min, with a final extension of 72°C for 10 min. All samples were purified from 1.5% agarose gels using the Zymoclean Gel DNA Recovery Kit (Zymo research, Irvine, CA, United States) following the recommended procedure of the manufacturer. Subsequently, amplicons were sent for Sanger sequencing at Eton Bioscience (San Diego, CA, United States). The Sanger sequencing results have been uploaded to the NCBI database and can be found under the GenBank accession numbers MZ227353–MZ227374.

In order to transform the strain A4 with a plasmid that carries maker genes for visualization of A4 in plant tissues, competent cells were generated. For this purpose, A4 cells were grown overnight at 28°C until they reached an optical density (OD₆₀₀) of 0.5. They were cooled in an ice bath and washed with sterile, cold water four times after centrifugation steps at 4,000 × g. Afterward, cells were resuspended in 10% glycerol, aliquoted, and snap frozen in liquid nitrogen. Electroporation was used for transformation (Cadoret et al., 2014) of A4 with the pRU1156 plasmid, Addgene catalog 14473 (Karunakaran et al., 2005).

Transformed and untransformed A4 strains were grown on LB liquid media overnight at 28°C. Bacterial suspensions were normalized by measuring OD₆₀₀ of 0.1, and equal amounts of cells were boiled in sample loading buffer (Sigma-Aldrich, Darmstadt, Germany) and separated in SDS-PAGEs. For immunoblotting, anti-GFP (1:5,000, Roche, 11814460001, Darmstadt, Germany) was used to detect free GFP (Bürger et al., 2017).

Arabidopsis thaliana (ecotype Columbia-0) seeds were surface sterilized by hydrochloric fumigation for 3 h and stratified for 3–5 days at 4°C (Lindsey et al., 2017). Subsequently, the sterile seeds were grown in vitro on half-strength Linsmaier and Skoog (LS) medium at pH 5.7 with 7% plant agar. Seedlings were grown on vertically oriented plates in a growth chamber (Percival, Iowa, United States) with continuous white light (100 μmol m^{– 2} s^{– 1}) at 21°C (Qiu et al., 2019). Prior to inoculation, the 22 selected endophytic isolates were grown for 24 h in LB media with 50 μg tetracycline ml^{– 1} and 100 μg ampicillin ml^{– 1} at 28°C. The bacterial cells were centrifuged at 3,500 × g, resuspended in infiltration buffer (10 mM MgSO₄, 10 mM MES-KOH pH 5.5), and adjusted to an OD₆₀₀ of 0.1 in the same buffer. Subsequently, the roots of 7-days old seedlings were inoculated with 2–5 μl of bacterial inoculum per root, being careful not to contaminate shoot tissues, following the protocol for Salmonella enterica and Escherichia coli O157:H7 (Cooley et al., 2003). Infiltration buffer without bacteria served as the negative control (mock treatment).

The effect of the 22 bacterial isolates on Arabidopsis growth was evaluated 2 weeks postinoculation. Three independent experiments were performed with 38–41 seedlings per treatment in each experiment. Fresh mass of seedlings was measured and compared to mock (infiltration buffer)-treated plants, which served as negative control. Statistical analysis was performed by one-way ANOVA with Tukey’s test using GraphPad Prism 7 (GraphPad Software, San Diego California, United States). A difference was considered statistically significant using an a priori determined α level of 0.05.

In order to test colonization of the endophytic strain A4, plant tissues were histochemically stained with X-Gluc (GUS) 7 days after inoculation of A. thaliana seedlings and observed with a stereomicroscope (Leica MZ16, Leica, Germany; Willige et al., 2011). Moreover, seedlings were screened for GFP expression using a Zeiss LSM 710 fluorescence microscope (Carl Zeiss, Germany).

The bacterial strain A4 was sent to Novogene Biotech (Beijing, China) for DNA extraction, library preparation, and whole genome sequencing and assembly using PacBio Single Molecule, Real-Time (SMRT) Sequencing (Eid et al., 2009). Falcon software (Chin et al., 2016) was used for genome assembly, and BUSCO for the assessment of the genome assembly, gene set, and transcriptome completeness (Simão et al., 2015). Prokka software was used for genome annotation (Seemann, 2014), and functional annotation was done by aligning the sequence with sequences previously deposited in diverse protein databases, including the National Center for Biotechnology Information (NCBI) non-redundant protein (Nr) database, UniProt/Swiss-Prot, Kyoto Encyclopedia of Genes and Genomes (KEGG), and Cluster of Orthologous Groups of proteins (COG).

The phylogenetic tree was constructed with SpeciesTreeBuilder v 0.1.0 from the Kbase platform (Arkin et al., 2018). The Artemis Comparison Tool (ACT; Carver et al., 2005) was used to compare our bacterium genome with an already sequenced genome. To display circular comparisons between both genomes and plasmids, the Blast Ring Image Generator (BRIGS) was used (Alikhan et al., 2011). Finally, a nucleotide identity analysis between strain A4 and a publicly available, closely related bacterial genome was performed with the ANI calculator from EZbiocloud (Yoon et al., 2017).

Strain A4 was tested for its ability to solubilize phosphate using NBRIP medium (Nautiyal, 1999). Prior to inoculation, strain A4 was grown at 28°C until they reached an OD₆₀₀ of 0.5. An aliquot of 10 μl was spotted on plates containing National Botanical Research Institute’s Phosphate growth medium (NBRIP; Nautiyal, 1999) and incubated at 28°C for 7 days.

Siderophore production was determined by chrome azurol S (CAS) agar plates. The medium was prepared according to the method described by Schwyn and Neilands (1987). Strain A4 was grown overnight in LB liquid medium at 28°C. A4 bacterial suspension was normalized by adjusting to an OD₆₀₀ of 0.5 with LB medium. Subsequently, 10 μl of the suspension was spotted on CAS agar plates and incubated for 7 days at 28°C.

Twenty-two strains were isolated from surface-sterilized almond leaves (Figure 1). They were tested for their plant growth-promoting ability, and only the strain A4 had a positive effect on plant growth. Therefore, A4 was selected for further characterization. Sanger sequencing of the 16S rRNA gene revealed its phylogenetic relationship to the genera Pantoea and Erwinia. Three independent experiments revealed that endophyte-treated plants had about 30% higher fresh mass than control plants (p ≤ 0.001; Figures 2A,B). Additionally, the A4 treatment resulted in increased root hair length and abundance (Figures 2C,D). These results suggest that inoculation with an almond-derived endophyte could promote growth of Arabidopsis seedlings.

To test plant colonization by A4, we transformed A4 to express the marker genes GFP and gusA in order to track the colonization of A4 within internal plant tissues. Successful transformation of A4 was confirmed by Western blotting to visualize expression of GFP encoded in the transformed plasmid (Figure 3A).

Arabidopsis roots were inoculated with endophytic bacteria A4 labeled with GFP to determine whether intracellular colonization was occurring by live-cell imaging. Confocal observation revealed GFP-expressing A4 colonizing internal root tissues 7 days after inoculation. Many of these bacterial cells were observed within the root vasculature (Figure 3B). Furthermore, analyses 14 days postinoculation showed that A4 had colonized shoot tissues such as leaves and flowers, as visualized by GUS staining (Figures 3C,D). These observations indicated that the root-inoculated bacteria were able to spread inside the plant and colonize above ground tissues. The mock-inoculated plants did not show any blue staining or fluorescent bacteria.

The whole-genome sequence analysis of A4 was conducted in order to obtain reliable taxonomic classification and identify genes or pathways that could potentially contribute to the plant growth-promoting effects. The A4 genome consisted of a single circular chromosome of 3,858,052 base pairs (bp), with an average GC content of 55%. The A4 strain had one plasmid we named pA401, with a size of approximately 576,382 bp, and also had an average GC content of 55% like the chromosome. The chromosome contained 3,552 genes, including genes for 78 tRNAs, for 22 rRNAs, and 3,451 protein-coding sequences (CDS). Among these CDSs, 2,929 genes were classified into clusters of orthologous group (COG) families composed of 23 categories. KEGG pathway annotation resulted in the functional annotation of 3,414 genes (96%). Of these annotated genes, most were grouped into the two categories: “metabolism” and “environmental information processing” (Table 1). The A4 plasmid pA401 had 521 protein-coding sequences and 1 contig.

We used Tree builder to search for closely related genomes and to assess the phylogenetic relationship between A4 and publicly available genomes (Figure 4 and Supplementary Table 1). This analysis revealed that A4 grouped separately from strains of Pantoea, but together with Erwinia strains and in particular very next to E. gerundensis strain EM595 (Rezzonico et al., 2016). Therefore, we compared chromosomes and plasmids of A4 and EM595 (Figure 5). While the chromosomes of both strains share about 90% of their genes (3175), A4 had 224 unique genes (Supplementary Table 2). Furthermore, EM595 carried two plasmids (pEM01 and pEM02), and pEM01 shared the greatest overlap with pA401 (71%, 463 genes); 20 of the common genes were also present on pEM02. Additionally, the A4 plasmid had 48 unique genes that were not present in both plasmids of the previously sequenced Erwinia strain EM595 (Supplementary Table 3). Taken together, these results suggest A4 represents a novel E. gerundensis strain.

In order to identify genes responsible for growth promotion, we searched the genome and plasmid of A4 for genes involved in nitrogen fixation, and organic acid, siderophore, and spermidine synthesis. Even though A4 was isolated on bacterial medium that is denoted as nitrogen free, we could not detect any genes encoding for nitrogenase subunits. However, we identified a pqq gene cluster (pqqABCDEF) as well as for a gcd gene, indicating that A4 carries all necessary enzymatic components to solubilize phosphorus by organic acid synthesis. Moreover, we found entABCDEF and two entS genes in A4, suggesting that this strain is able to produce and export the siderophore enterobactin. Furthermore, the strain carries all necessary genes (i.e., metK and speABDE) to produce the polyamine spermidine. In addition, we identified several genes involved in spermidine and putrescine transport (i.e., potA, potB, potD, potG, potH, and potI; Figure 6).

Because the genome of A4 encodes for proteins that are involved in the production of organic acids and siderophores, we tested if A4 is able to release insoluble phosphate and synthesize iron chelators. A4 was able to produce a clear zone around the colony, suggesting phosphate solubilization activity by the production of organic acids (Figure 7A). Additionally, production and secretion of siderophores were visualized by a color change of Chrome Azurol S (CAS). In association with ferric ions, CAS appears blue, while the removal of the ferric ions by siderophores changes the color of the media from blue to yellow. Consistent with the presence of genes encoding for siderophore-synthesizing enzymes, A4 was able to cause the color change of the CAS media (Figure 7B). In contrast, a non-growth-promoting Frigoribacterium strain that we isolated from almond leaves was not able to solubilize phosphate or induce a yellow color of the CAS media (Supplementary Figure 1). This indicates that A4 could supply its plant host with essential nutrients such as phosphate and iron to promote growth.