Detection of Frankia in Alnus glutinosa with NIR dataset

The excel sheets provided in this repository contain all the relevant raw data for both of the experimental sets this study. Each excel sheet has a short metadata section explaining what each of the columns refers to. Below you will find the detailed experimental methods for the experiments conducted with these datasets. This study aimed to distinguish between N supplied through fertilization and N symbiotically fixed by Frankia using NIR spectroscopy. Black alder seeds were sourced from the Dutch State Forestry Department, Staatsbosbeheer. To prepare the inoculum for all the experiments, local field nodules were collected from the roots of a single population of A. glutinosa trees growing naturally in a semi-forested area in Leiden, Netherlands (Latitude: 52.171515 N, Longitude: 4.468883 E). Fresh nodules were collected on the same day as each experiment was set up, ensuring that only healthy, intact nodules of mature A. glutinosa trees were used for inoculum preparation. Methods 2.1 Seed germination To minimize external contamination and ensure that A. glutinosa seedlings were only inoculated with Frankia, we germinated seedlings under sterile conditions. To achieve this, A. glutinosa seeds were surface-sterilized by shaking for 20 minutes in a 14% bleach solution and placed on 0.5 Murashige and Skoog (MS) medium agar plates for germination. The plates were sealed with parafilm, and incubated vertically in a controlled growth cabinet with a 16:8 light-to-dark cycle at 20°C during the light phase and 17°C during the dark phase. Seedlings were allowed to germinate and develop for four weeks until they had at least two mature leaves before being used in experiments. To maintain seed surface sterility, plates were inspected daily for signs of bacterial or fungal contamination near the seeds. If contamination was detected, uncontaminated seedlings were carefully transferred to new sterile agar plates under a flow hood, while contaminated seedlings and plates were discarded. Additionally, this method allowed for the selection of seedlings with consistent shoot and root sizes, providing greater uniformity for experimental setups. 2.2 Fertilization vs Frankia derived N experiment An experiment was established to investigate the hypothesis that N derived from Frankia will be reflected via unique spectral peaks in the NIR spectra compared to soil derived N. Four weeks after germination, as previously described, seedlings were transferred into 1 L pots (dimensions: 11x11x12 cm). To ensure that any differences in the spectra were solely due to the N source and not other stress factors, all the pots were filled with the same gamma-sterilized grassland soil for which soil properties were analyzed (see Georgopoulos et al. 2025 for details). This soil was predominantly sandy, comprising 98% sand with only 1.4% fine particles (silt and clay) and a minimal gravel content of 0.6%. Nutrient availability was limited, with moderate to low concentrations of nitrogen (NH₄⁺ at 37.82 ± 1.61 mg/kg and NO₃⁻ at 15.41 ± 0.57 mg/kg). Phosphorus levels were also low, as indicated by a PO₄³⁻ concentration of 3.86 ± 0.41 mg/mg. The soil exhibited a slightly alkaline reaction, with a pH of 7.37 ± 0.02, and contained a low amount of soil organic matter (1.85 ± 0.04%). To create a gradient of N availability, pots were fertilized with solutions containing 0, 1.25, 2.5, 5, 7.5, 10, or 20 mM NH₄NO₃ (Sigma Aldrich). While Frankia provide NH₄⁺ to the plant via N-fixation, in most soils, plants encounter a mixture of NH₄⁺ and NO₃⁻ and can uptake both. As such, both forms of N were supplied, ensuring that non‐nodulated (pure) controls and nodulated (Frankia‐inoculated) plants all have access to the two major soil-available N soils and preventing pH‐driven nutrient imbalances (rhizosphere acidification with NH₄⁺, alkalinization with NO₃⁻) that themselves may alter plant metabolism and hence, potentially, NIR spectra. Each pot received 15 mL of the designated solution twice weekly for a period of 12 weeks. For each nutrient level, 20 pots were prepared, half of which were inoculated with Frankia while the remaining half served as uninoculated controls (n = 10 replicates). The Frankia inoculum, prepared fresh on the day of the experiment, was made by crushing surface sterilized Frankia nodules and homogenizing nodule tissue in autoclaved miliQ H2O (100 mg / 1 ml). After fertilization, 1 mL of the homogenate was pipetted into a small indentation near the roots of the inoculated seedlings. To ensure that the mechanical insertion of inoculum did not influence plant performance, 1 ml of MiliQ H2O was similarly pipetted near the roots of non-inoculated seedlings. Due to space limitations, the experiment was performed in two sets, one set including the 1.25 - 5 mM fertilizers and the other set including the 7.5 – 20 mM fertilizers. The two sets were performed using the exact space in the same growth chamber and the exact light and humidity conditions, but 12 weeks apart, and included their own sets of controls. As such, per experiment set, ten pots with sterilized soil were left untreated (neither inoculated nor fertilized, which we call control pure), while another ten pots received only the Frankia inoculum (henceforth known as nodulated control), bringing the total number of pots to 160. To ensure that differences in seasonal variations and light conditions would not influence the outcome of either experimental set, plants were grown in a climate-controlled room with a relative humidity of 70%, a 16-hour light and 8-hour dark cycle, and temperatures of 20°C during the light phase (LED lights; Valoya LightDNA BX120, NS1+FR, 2-channel) and 17°C during the dark phase. Light intensity (%) changed according to a timed schedule (7:00, 0 %, 8:00, 30 %, 11:00, 80 %, 13:00-17:00, 100 %, 19:00, 80 %, 22:00, 30 %, 23:00, 0%). Pots were watered three times per week to soil saturation. Over the course of 12 weeks, plant growth metrics, including stem height and leaf count were recorded weekly. Leaf chlorophyll was recorded from the fourth until the final week of the experiments using a Chlorophyll Meter SPAD-502Plus (Konica Minolta Sensing Europe B.V.). Prior to unpotting, on the day of the harvest, NIR spectra were measured from the third leaf from the top of each plant (the same leaf used for each chlorophyll measurement) using a Inventech Benelux NIR spectrometer (Oosterhout, Netherlands) with a spectral detection limit of ~330 - 1100 nm. Spectra were taken from fresh, intact leaves on live plants, and no grinding or drying was performed prior to spectral measurements. We selected the third leaf from the apex because it consistently represented the youngest fully expanded leaf across all treatments. Previous studies have shown that once A. glutinosa leaves reach full expansion, their pigment and nitrogen concentrations remain relatively stable for an extended period. By sampling a fixed leaf position, we standardized for developmental stage rather than chronological age, thereby minimizing the potential influence of treatment-induced differences in growth rate on the NIR spectra. Prior to measuring the first leaf and for every 40 measurements after that, a blank measurement was taken. NIR data were pre-processed by correcting based on the blank measurements by removing spectra lower than the blank. In total, one NIR measurement was taken from each fresh leaf of each plant for a total of 160 collected samples (80 from the first and 80 from the second set). At harvest, plants were carefully unpotted, and roots were rinsed under running tap water. Root nodules were counted, excised using a razor blade, and oven-dried at 40°C to determine the total root nodule biomass. The remaining plant biomass was divided into stems, leaves, and roots. All components were oven-dried under the same conditions to determine aboveground, belowground, and nodule biomass. For leaf nitrogen content analysis, oven-dried leaf samples were ground using a QIAGEN TissueLyser II Bead Mill (Hilden, Germany) at 370 rpm for 5 minutes. Leaf nitrogen percentages were quantified using the dry combustion method (Matejovic, 1997) with a Thermo Scientific FLASH 2000 CN analyzer (Milan, Italy). Soil NH₄+-N and NO₃--N concentrations were determined using a standard 1M KCl extraction and spectrophotometric analysis, while PO₄³--P concentrations were measured using an extraction method with 0.01M CaCl₂. Nutrient concentrations were expressed as mg of NH₄+-N, NO₃--N, and PO₄³--P per kg of soil.