Biomass productivity of black spruce trees is strongly limited by soil nitrogen in shallow active layer on permafrost. Trees and mycorrhizal roots are known to absorb amino acids to bypass slow nitrogen mineralization in nitrogen-limited boreal forest soils. However, the amino acid uptake strategy of tree roots cannot fully explain their advantages in the competition for soil nitrogen with other plants and microbes. Here, we provide evidence that some spruce tree roots absorb intact urea. Tree roots develop plasticity to utilize different nitrogen sources, depending on active layer thickness. Urea uptake is limited to soils with shallow permafrost, where urea accumulates due to limited microbial mineralization activity. This contrasts with soils with deep permafrost, where tree roots absorb amino acids and inorganic nitrogen. Allocation of photosynthate to fine roots in colder subsoil above shallow permafrost provides advantages for trees monopolizing urea-nitrogen. Despite lower energy efficiency of urea utilization compared to inorganic nitrogen and amino acids, urea uptake is one of nitrogen acquisition strategies from nitrogen-starved soil. Warming-induced permafrost degradation could reduce an extent of urea-dependent drunken forests with the high potentials of soil carbon storage.

1. Field site

We compared forest tundra sites dominated by black spruce (Picea mariana L.) on clayey soils and sandy soils near Inuvik, Northwest Territories, Canada (N68°03’, W133°30’). Inuvik has a subarctic climate; the mean annual air temperature is –8.8 ºC. The soil surface consists of lichen (Cladonia mitis L. and Cladonia stellaris L.)-covered mounds and moss (Pleurozium schreberi L., Hylocomium splendens L., and Sphagnum fuscum L.)-covered depressions. The wide variation in permafrost table depth was caused by the differences in geological substrates (glaciofluvial sands and fine-grained sediments) (Tarnocai et al., 1993; Fujii et al., 2019). Clayey soils are derived from a mixture of fine-grained sediments (27–37 % clay and 22–30 % sand), while sandy soils are derived from glaciofluvial sands (8–12 % clay and 73–75 % sand). The clayey soil is classified as Glacic Aquorthel due to the presence of ice wedge and redox morphological feature within the profile, whereas the sandy soil is classified as Spodic Psammorthel or Typic Psammorthel due to the weak feature of podzolization and sandy texture (Soil Survey Staff, 2014). Active layer thickness (from the organic layer surface to permafrost table) was thinner in the clayey soil sites (24–56 cm) than in the sandy soil sites (60–182 cm) (Fujii et al., 2020). Volumetric water content was added in Table S1. The mineral soil horizons in the clayey soil sites were wetter than in the sandy soil sites (Table S1) and were saturated by flooding water of permafrost thawing in summer (Fujii et al., 2019).

2. Experimental design

Annual rates of aboveground tree biomass production were estimated from an increase in biomass measured by the tree census of sixteen plots (6 sandy soil plots and 10 clayey soil plots; 10 m × 10 m) between 2015 and 2016. The individual aboveground tree biomass [W (g)] was estimated by applying the tree height [H (cm)] and stem diameter at the tree height of 30 cm [D (cm)] to the regression equations for black spruce trees [W = 1.0366 ×(D² × H)^0.72722].Within ten clayey plots, six plots (1 km apart) were selected for tracer experiments so that annual biomass production is comparable to the sandy soil plots within the limited area where tracer experiment was permitted (Table S1). In August 2015, we conducted two tracer experiments (glutamic acid and mixed N solution) to study (i) potentials of amino acid uptake by black spruce trees at high substrate availability and (ii) contribution of different N sources to the assimilated N at realistic substrate availability. For tracer experiment of glutamic acid solution, two subplots (0.2 m × 0.2 m, one test subplots and one control subplot) were chosen in each plot so that black spruce trees were present within the subplot. For tracer experiment of mixed N solution, five subplots (four test subplots and one control subplot) were chosen in each plot. At the same time (August 2015), we also collected soil samples from six plots of the clayey and sandy soil sites for soil solution extraction and mineralization activity assay, respectively. The mineral soil horizons (up to 30 cm depth from the boundary between organic and mineral soil horizons) were taken in each pit by inserting horizontally using cores (diameter 5 cm, length 7 cm). For the organic horizon, one block (width × length × depth = 20 cm × 20 cm × 5 cm) was excavated after removing the living organic matter (e.g., grass) and immediately sealed in plastic bags and transported to the laboratory. The subsamples were air-dried and sieved (< 2 mm) to eliminate litter, roots, and stones and used for measurements of physicochemical properties.

​​​​​​​3. Tracer experiment of root N uptake

At 12:00 on 1 August 2015, the experiment was started by applying the glutamic acid solution or the mixed N solution to the test subplots and water to the control subplots (0.2 m × 0.2 m). For the tracer experiment of glutamic acid solution, ¹³C, ¹⁵N-labelled glutamic acid (¹³C₅H₉¹⁵NO₄, ¹³C 99 atom%, ¹⁵N 99 atom%) was used as tracer and injected into the mineral soil in the test subplots (corresponding to 0.525 g N m^-2). A total of 1 L 0.5 mM glutamic acid solution or water was injected into the upper mineral soil (0–5 cm depth, 0.2 m × 0.2 m) using a 50-mL syringe fitted with a long needle (8 cm) to minimize substrate loss. The N dose corresponded to the N deposition by precipitation (or throughfall) during the growth season. In the mixed N solution, one of four N forms (glutamic acid, urea, NH₄, and NO₃) was labelled and mixed with the other three non-labelled N forms. We used 0.1 mM glutamic acid, 0.1 mM urea ([¹³CO(¹⁵NH₂)₂], ¹³C 99 atom%, ¹⁵N 99 atom%), 0.1 mM ¹⁵NH₄NO₃ (98 atom%) or 0.1 mM NH₄¹⁵NO₃ (98 atom%) as tracer and injected into the upper mineral soil in the same manner with the tracer experiment of glutamic acid solution. The total N dose (= 0.105 g N m^-2) corresponded to the actual N pool of soil solution (0–5 cm depth), as shown later in Fig. 3. After 24 h, fine roots (less than 2 mm in diameter) were collected by inserting cores and carefully sorted from soil. Root samples were rinsed in water to remove tracer adsorbed on the root surface (Inagaki and Kohzu, 2005).

​​​​​​​4. Isotope analyses

Root samples were oven dried at 70 °C for 48 h, weighed, and milled in a ball mill to a fine powder. A root sample (1 mg) was weighed into a tin capsule, and δ¹³C and δ¹⁵N were measured with an on-line C and N analyser (NC 2500; Thermo Fisher Scientific) coupled with an isotope ratio mass spectrometer (MAT252; Thermo Electron, Bremen, Germany). Data on C and N contents and atom% ¹³C and atom% ¹⁵N in excess of the natural abundances of ¹³C and ¹⁵N of the root sample from the control subplots were used to calculate mol ¹³C excess and mol ¹⁵N excess (substrate-derived ¹³C and ¹⁵N) of the root sample from the amended subplots. The molar ratio of ¹³C/¹⁵N assimilated by roots indicates whether intact amino acid is absorbed by roots or not (Näsholm et al., 1998). Relative contribution of N sources to assimilated ¹⁵N in roots was calculated by dividing each substrate-derived ¹⁵N by sum of glutamic acid-derived ¹⁵N, urea-derived ¹⁵N, ammonium-derived ¹⁵N, and nitrate-derived ¹⁵N measured individually in tracer experiment.

​​​​​​​5. Soil solution extraction

To test availability of inorganic N, urea, and amino acids in soil, soil solution was sampled for both the organic and mineral soil horizons without the addition of water using the centrifugation–drainage technique (Giesler and Lundström, 1993). Fresh soil samples were centrifuged for 30 minutes (10,560 g; ~1.5 MPa; high-speed refrigerated centrifuge; Hitachi CR20G). Concentrations of amino acids in the soil solution were determined after filtering through a 0.6-µm filter (GF/C; Whatman) using an ultra-high-performance liquid chromatography system equipped with a reverse-phase column (Acquity UPLC BEH C18, Waters; particle size, 1.7 μm; column size, 2.1 × 100 mm) and fluorescence detector (X-LC 3120FP, JASCO; excitation and emission wavelengths, 345 and 455 nm, respectively) after derivatization with o-phthalaldehyde (OPA) (Lindroth and Mopper, 1979). Total dissolved N, urea, and inorganic N species (ammonium and nitrate) concentrations were determined using a total organic C and N analyser (TOC-V_CSH, Shimadzu), standard colorimetric technique (QuantiChrom Urea Assay), and high-performance liquid chromatography (Shimadzu), respectively.

​​​​​​​6. Laboratory incubation of soil

To compare microbial activities of amino acid or urea mineralization between the clayey and sandy soils, mineralization assays of ¹⁴C radiolabelled substrates were conducted in the laboratory incubation. In addition to glutamic acid and urea, arginine was selected as a precursor of urea (Fujii et al., 2018). After transferring the soil samples to the laboratory, we added ¹⁴C-labelled substrates [amino acids (glutamic acid, arginine) and urea (100 µL, specific activity 0.17 kBq mL^-1; pH 4.5)] to 1 g of the soil sample in a 50 mL plastic tube, so that N doses equal to those of the field tracer experiment with the mixed N solution (21 mg N kg^-1 soil). After adding the substrate, the soil was shaken gently to ensure mixing and incubated at 15 °C (equals to air temperature at soil sampling in the field) in sealed tubes. The ¹⁴C-CO₂ produced by the biodegradation of the added substrate was collected after 24 h by placing a plastic scintillation vial containing 1.0 mL of 1 M NaOH on top of the soil, separated by a spacer. ¹⁴C-CO₂ trapped in the NaOH was determined by liquid scintillation using alkali compatible scintillation fluid (Hionic-Fluor; Perkin Elmer) to calculate the microbial rates of amino acid or urea mineralization (Fujii et al., 2018). Microbial biomass C in the incubated soil (0 h) was determined using the chloroform fumigation-extraction method (Fujii et al., 2018).

​​​​​​​​​​​​​​7. Statistics

All results were expressed as mean ± standard error (SE). The Tukey’s test was used to detect statistically significant differences of the values (soil solution concentration, microbial biomass, and substrate mineralization rates) between sandy soil sites and clayey soil sites. Two-way analysis of variance was used to evaluate statistical differences in the root assimilated ¹⁵N by main effects (substrate types and soil types) and their interaction. The association between excess ¹³C and excess ¹⁵N in roots were tested using simple linear regression. The differences in the slopes of the linear regression equations were compared using analysis of covariance (ANCOVA) between sandy soil sites and clayey soil sites. Statistics were performed using Sigma Plot 14.5 software (SPSS Inc., Japan). In all cases, P < 0.05 was considered significant.