Energetic supply regulates heterotrophic nitrogen fixation along a glacial chronosequence

Highlights

• Free-living N fixation (FLNF) rates were studied along a glacier chronosequence.

• FLNF rates markedly decreased with soil depth.

• Photosynthetic cyanobacteria likely contributed the majority of FLNF at the initial stage.

• Heterotrophic FLNF may prevail at later successional stages.

• Energetic supply, rather than nifH gene abundance, regulated heterotrophic FLNF rates.

Abstract

Ice-free areas are expanding on the Tibetan Plateau and in its surroundings, and these recently exposed lands are typically deficient in nitrogen (N). Free-living N fixation (FLNF) represents a vital source of N input, especially when symbiotic N-fixing plants are absent. However, we know little about how FLNF changes with ecosystem succession after glacier retreat. In 2018, we measured the potential rates of FLNF in soil profiles at four sites (exposed since 2012, 1980, 1970 and 1930, respectively) along the Hailuogou glacier chronosequence and at a reference site with well-developed Podosols. We also tested the role of nifH gene abundance and resource availability in regulating FLNF rates. The results showed that photosynthetic cyanobacteria likely contributed the majority of FLNF in the biological soil crusts (exposed since 2012), while heterotrophic FLNF may prevail at later successional stages with closed canopies. FLNF rates were the highest in the litter layer of the soil profiles and decreased with soil depth. In the litter layer, FLNF rates were lower at the site dominated by N-fixing shrubs (exposed since 1980) than at the older sites dominated by non-N-fixing trees (exposed since 1970 and 1930, and the reference site). Carbon (C) availability was the major factor explaining variation in heterotrophic FLNF rates measured in the lab without light, while the effect of nifH gene abundance was nonsignificant. Overall, our results highlighted the energetic control on heterotrophic FLNF rates in relatively young ecosystems, which likely resulted in the poor relationship between nifH gene abundance and heterotrophic FLNF rates. The findings will help in understanding and modeling the coupled C and N cycles in glacier forefield ecosystems which are being exposed at accelerated rates.

Introduction

Glaciers on the Tibetan Plateau and in its surroundings are experiencing shrinkage, with associated impacts on ecosystem functions and human livelihood (Huss et al., 2017; Yao et al., 2012). The newly exposed lands after glacier retreat are typically deficient in nitrogen (N), and the establishment of vegetation largely depends on N input through biological N fixation (BNF) (Chapin et al., 2016; Reed et al., 2011). BNF includes symbiotic N fixation and free-living N fixation (FLNF). Symbiotic N fixation occurs in the root nodules of N-fixing plants and is carried out by Rhizobia and Frankia bacteria that fix N in conditions well moderated by the plant host. FLNF is carried out by autotrophic or heterotrophic diazotrophs without the transfer of C from a formal plant host (Dynarski and Houlton, 2018). Free-living diazotrophs are a highly diverse group from a wide range of bacterial phyla (e.g., Proteobacteria, Firmicutes, Cyanobacteria) and methanogenic archaea, and they may survive in variable and dynamic environments (Smercina et al., 2019). In temperate areas, N-fixing plants are generally constrained to the early stage of ecosystem development (Menge et al., 2019), although some species may persist for centuries in secondary succession (e.g., green alder in upland forests throughout interior Alaska; Mitchell and Ruess, 2009). In contrast, FLNF is ubiquitous in different successional stages and climate zones and can be found in plant litter, soil, and in association with moss or lichen (Dynarski and Houlton, 2018; Reed et al., 2011). In different biomes, the average amount of N input from FLNF ranges from 1.0 to 15.0 kg N ha⁻¹ year⁻¹, which is comparable to or even greater than that from symbiotic N fixation (Cleveland et al., 2010; Reed et al., 2011). Specifically, FLNF represents an essential source of N input when symbiotic N-fixing plants are absent (Pérez et al., 2014). However, the changing patterns of FLNF during ecosystem succession after glacier retreat have received little attention, which limits our ability to adapt to a future world with more ice-free areas.

FLNF is catalyzed by nitrogenases that are metalloenzymes encoded by the N fixation (nif) genes. The nif genes include structural genes of the highly conserved nitrogenase complex and the genes required for nitrogenase biosynthesis and regulation (Dixon and Kahn, 2004). The synthesis and functioning of nitrogenases are affected by climate (e.g., light, temperature, moisture), soil properties (e.g., pH) and diazotroph community, and require an adequate supply of phosphorus (P), iron and molybdenum (or vanadium in the case of alternative nitrogenases) (Dynarski and Houlton, 2018; Reed et al., 2011; Smercina et al., 2019; Vitousek et al., 2002). In addition, nitrogenase-catalyzed N₂ reduction is energetically expensive, and even more energy is needed to protect nitrogenase from inactivation by oxygen (O) (Smercina et al., 2019). Therefore, FLNF is tightly regulated by environmental factors mainly at the transcription level (Dixon and Kahn, 2004; Ryu et al., 2020). The well-studied N regulation networks of FLNF consist of two components including the N sensor proteins typically belonging to the PII superfamily and the nif-gene regulation system, which sense the N status and regulate the transcription of nif genes according to the N status (Dixon and Kahn, 2004; Leigh and Dodsworth, 2007). Nutrient addition experiments reveal that a sufficient N supply depresses FLNF because the uptake of fixed N is less costly than BNF (Dynarski and Houlton, 2018). However, a recent study has demonstrated increased FLNF rates in N-rich late-successional ecosystems compared with relatively N-poor early-successional ecosystems (Zheng et al., 2020). The authors highlight the role of the C:N ratio in explaining the variations in FLNF rates (Zheng et al., 2020). This stoichiometric view is in line with the findings that the N status-sensing PII proteins and the nif-gene regulation systems of some diazotrophs are coregulated by C and N signals (e.g., 2-oxoglutarate and glutamine) and that N signals override C signals to depress N fixation only in situations of N excess (Dixon and Kahn, 2004). In addition to the C:N ratio, C availability alone may be a major predictor of FLNF rates. This is because in contrast to symbiotic organisms, heterotrophic diazotrophs must find their own source of energy, typically by oxidizing organic molecules released during organic matter decomposition or released by other organisms (e.g., low molecular weight C compounds in plant root exudates) to perform the energetically expensive N fixation. Some previous studies report that the heterotrophic FLNF in litter and soil is limited by C quantity and quality (Jones and Bangs, 1985; Vitousek and Hobbie, 2000). However, whether C availability or C:N ratio is the major influencing factor of FLNF rates remains untested.

Another issue associated with the regulators of FLNF is to what degree the abundance of nif genes (gene potential) may determine FLNF rates. A recent study has demonstrated that the activities of C-degrading hydrolytic enzymes in soils are closely and positively related to their gene abundances (Trivedi et al., 2016), indicating the existence of constitutive expression for these hydrolase (Allison et al., 2005). However, nitrogenase activity may deviate from its gene potential because the expression of nif genes is tightly regulated by resource availability and not likely to be constitutive. Among different studies, FLNF rates and nifH gene (encoding Mo-nitrogenase) abundance have been reported to be positively related or unrelated (Brankatschk et al., 2011; Chen et al., 2019; Reed et al., 2011). The discrepancy may arise from the covariation between nifH gene abundance and resource availability. Therefore, it is necessary to conduct a multiple-variable analysis that clarifies how gene potential and resource availability influence FLNF rates.

In this study, we selected four sites (exposed since 2012, 1980, 1970 and 1930, respectively) along the ~130-year Hailuogou glacier chronosequence and a reference site with well-developed Podosols on the eastern Tibetan Plateau to investigate the FLNF rates in successional ecosystems after glacier retreat. Along the chronosequence and at the reference site, light conditions markedly decrease as canopy closes and resource availabilities also vary with soil depth and along successional stages (Wang et al., 2020c). The potential rates of FLNF (nitrogenase activity) in litter and soil were measured in the field (with ambient light) and in the lab (without light). We supposed that the rates measured in the lab were from heterotrophic FLNF because autotrophic FLNF relies on photosynthesis for energy acquisition (Pérez et al., 2017; Stal, 2012). We hypothesized that (1) a shift from photosynthetic FLNF to heterotrophic FLNF would occur as the canopy closes, due to the decreased light intensity and increased resource availabilities associated with plant litter and root exudates; (2) C availability, rather than N or C:N ratio, is the major factor controlling the heterotrophic FLNF rates in the relatively N-poor ecosystems after glacier retreat. In addition, the effect of gene potential on heterotrophic FLNF rates may be weak because the expression of nif genes is tightly regulated by resource availability and is not likely to be constitutive.