Sampling
Endophyte communities were sampled from the same BCs that were surveyed by Jamison-Daniels et al. , (2021). Three tree species that are widely distributed and common across BCs of different sizes (and thus different successional stages) (Jamison-Daniels et al. , 2021) were selected: Euclea crispa subsp. ovata (Burch.) F.White, Searsia chirindensis (Baker f.) Moffett andCanthium inerme Kuntze. Trees of these species were present in 38 of the 40 BCs.
Field sampling took place between 20 and 23 November 2018, to characterise the summer leaf endophyte communities. In every BC containing the host species, up to four trees (> 1.2 m) per host species were selected. The coordinates of each tree that was sampled were recorded using a Garmin Etrex GPS (Garmin Ltd., Olathe, KS, USA) and the height of each tree was estimated using a 1.2m dowel stick. All hosts did not occur in all BCs, and BCs did not always contain four individuals from each host. For BCs that had more than four individuals of a host, the first four individuals that were encountered were sampled. E. crispa , C. inerme and S. chirindensistrees were present in 27, 21 and 28 of the BCs, and 84, 50 and 55 individual trees were sampled from each host species, respectively. From each tree, five leaves from each of the four cardinal directions were collected half-way between the highest and lowest leaves. Only fully unrolled leaves that had reached maturity and had no visible signs of infection or insect damage were selected. Leaves were stored in envelopes within a cooler box, and processed within 8 hours of collection.
Microclimatic conditions within each BC were characterised to determine whether differences in microclimate influenced endophyte community composition and richness (see Jamison-Daniels et al. , 2021 for full details). Light intensity, temperature (average, minimum, maximum and standard deviation measured over 130 days), relative humidity and soil moisture were all previously measured and subsequently extracted (from Jamison-Daniels et al. , 2021).
The species richness (count of tree species per BC), Shannon-Wiener diversity, species composition and the tree basal area of all trees >1.2 meters per BC were obtained from Jamison-Danielset al. , (2021). Only trees >1.2 meters were considered, as these represent established vegetation within the BCs. Additionally, proximity of each sampled tree to the large indigenous forest which acted as a source population for most BC trees (Jamison-Daniels et al. , 2021) was calculated to represent a proxy for potential endophyte inoculum pressure. Within Google Earth Pro (v7.3.2.5776) a polygon was drawn around the indigenous forest at Buffelskloof Nature Reserve and subsequently extracted as a shape file. The distance from each tree sampled to the edge of the indigenous forest within the reserve was obtained using the points to polygon function in Esri™, ArcMap (Esri™, Redlands, CA, USA).
Leaves underwent surface washing (Arnold and Herre, 2003) to reduce and possibly eliminate the epiphytic burden from each sample. All leaves sampled from one tree were washed together successively in 70% EtOH (30 seconds), 2% NaOCl (60 seconds), 70% EtOH (60 seconds) and autoclaved dH2O (60 seconds). The leaves were then placed on paper towel and left to dry. Leaf disks from the dried leaves were cut using a 6 mm cork-borer (sterilised between each sample) and subsequently stored on silica gel in falcon tubes until DNA extraction. For E. crispaand C. inerme 12-18 leaf disks were cut per leaf, while 22-25 disks were cut from larger S. chirindensis leaves.