BSCs sampling
According to the coverage of moss, BSCs were artificially defined as four successional stages, the cyanobacterial crust stage (Crust-C), the moss colonization stage (Crust-CM), the moss development stage (Crust-MC) and the moss dominant crust (Crust-M) (Fig.S1). Crust-C represents light grey cyanobacterial crust without moss. The moss occurred on the cyanobacterial crust with coverage less than 10% was assigned to Crust-CM, while mosses grew in cluster on the cyanobacterial crust with coverage of 15-50% was assigned to Crust-MC. The Crust-M represents the crust that moss became dominant and covered more than 95% of soil surface. The coverage of moss (dominated by Bryum sp.) was measured by point sampling frame [1]. Crust-C was inoculated at 2005 and the other three kinds of crusts were inoculated at 2002. Above mentioned four typical successional stages of BSCs were sampled on September 21-27, 2013. For each type of crust, three 1×1 m2quadrats which were at least 2 m far from vascular plants were selected as sampling area. A 7.5 cm (diameter) cycle knife was pressed approximately 5 cm deep into the soil to excise an intact circular crust sample. The underlying loose sand of the circular crust sample was then removed. Five circular samples were collected in each quadrat. Totally 15 (3 quadrats × 5) circular samples were got for each type of crust. Shifting sand dunes without inoculation of Microcoleus vaginatus and Scytonema javanicum (less than 1000 m away from the inoculated dunes) was sampled as control treatment.The shifting sand dunes did not form crust structure, thus we sampled the surface layer with 1 cm deep of the sand dunes.For long-term storage of BSCs samples, dried crust sample was stored in petri dishes in dark at room temperature until experiments.
Recovery of BSCs with rehydration/desiccation cycles
Before conducting the experiment in determining the effect of desiccation duration on the recovery process of chlorophyll fluorescence and NA, dry field samples were treated with 2 rehydration/desiccation cycles to recover physiological activity, as rehydration/desiccation cycles were reported to be able to recover physiological activity of BSCs [2-4]. To conduct one rehydration/desiccation cycle, the crust sample was saturated with nitrogen free distilled water and incubated for 4 or 5 hours under the condition of 30 μmol photos m−2 s−1 cool white fluorescent illumination at 25C and then air-dried. One rehydration/desiccation cycle was lasted for 24 hours, and followed by another cycle.
Chlorophyll fluorescence measurement
When measuring the chlorophyll fluorescence parameters of Crust-CM and Crust-MC, we found that the values of chlorophyll fluorescence were very unstable. This is because moss and cyanobacteria respond differently on rehydration, so we removed mosses from Crust-CM. However, there are too many mosses on Crust-MC, it is impossible for us to remove all the mosses and meanwhile to keep the integrity of the crust. Therefore, we did not measure the chlorophyll fluorescence of Crust-MC. As no Chl a was detected in the control sample (bare sand), so no corresponding value of Fv/Fm (PS II activity) was gained.
After being subjected to rehydration/desiccation cycles as described above, the chlorophyll fluorescence of the recovered crust samples was measured with a Plant Efficiency Analyzer (Handy PEA, Hansatech Instruments Ltd., Norfolk, UK). Then the Crust-C, Crust-CM and Crust-M were desiccated for 2 days, 2 months and 5 months. To recover the BSCs with different desiccation durations, the samples were fully rewetted with sterilized N-free distilled waterand incubated under the condition of 30 μmol photos m−2 s−1 cool white fluorescent illumination at 25C. During the recovery process, the chlorophyll fluorescence of the crust samples was measured. Dark adaptation was at least 10 min and a saturating pulse that to excite the chlorophyll a molecules was set at approximately 1500 μmol photos m−2 s−1 for all the measurements. Then minimal (Fo) and maximal fluorescence (Fm) were recorded. The variable fluorescence (Fv) was calculated as Fv=Fm-Fo. And Fv/Fm is the maximum quantum efficiency of PS II photochemistry, representing the highest solar energy conversion efficiency in PS II reaction center [5].
Nitrogenase activity measurement
NA was measured by modified acetylene reduction assay (ARA)[6-8]. The crust sample of 1.44 cm2 in area was fully hydrated in 20 ml glass cuvette. Then the glass cuvette was sealed with a rubber stopper and 10% volume of air within the cuvette was replaced with acetylene gas. At specified intervals, 0.3 ml volumes of gas samples were removed from headspace of the glass cuvette and injected directly into a Shimadzu GC14C gas chromatograph equipped with a gas chromatography column and a flame ionization detector to measure the content of ethylene. After every 3 gas subsamples were taken, the stopper of glass cuvette was removed and the inside gas mixture was blown away by an air pump, then immediately 10% volume of air within the cuvette was replaced with new acetylene gas to continue the rest hours of ARA incubation. Calibration with ethylene standard was completed at the time of observations. One gram of control bare sand hydrated with distilled water was chosen as control. When crust samples were fully recovered before desiccation at the last rehydration/desiccation cycle, the NA of the recovered crust samples was measured (incubation for 3 hours).
To measure the continuous NA recovery process in the four successional stages of BSCs, 3 replicates of each type of crust sample which were desiccated for 4-6 months (thereafter referred to long-term desiccation) after rehydration/desiccation cycles were conducted ARA incubations for 48 h under the conditions of 30 μmol photos m−2 s−1 cool white fluorescent illumination at 25C. Gas subsamples (0.3 ml) were taken at different time of incubation (2 h, 6 h, 12 h, 18 h, 24 h, 30 h, 36 h, and 48 h) for ethylene analysis.
To determine the effects of desiccation duration on NA recovery process of BSCs in the light, after rehydration/desiccation cycles, 4 sets (3 replicates per set) of each type of crust sample were desiccated for 2 days, 4 days, 8 days and 4-6 months respectively in culture dishes at room temperature. Based on results of the 48 hours continuous NA recovery process, we conducted ARA incubation for 14 h for Crust-C, Crust-CM and Crust-MC, and 48 h for Crust-M. Then these samples were fully hydrated with sterilized N-free distill water and conducted ARA incubations under the conditions of 30 μmol photos m−2 s−1 cool white fluorescent illumination at 25C. Gas subsamples (0.3mL) were taken at different time of incubation (2 h, 4 h, 6 h, 10 h, 14 h for Crust-C, Crust-CM and Crust-MC, and 2 h, 6 h, 12 h, 18 h, 24 h, 30 h, 36 h, 48 h for Crust-M) for ethylene analysis. Two days desiccation showed highest NA in short term desiccation and 4-6 months desiccation also demonstrated high rates of NA. To clearly compare the total amount of nitrogen fixation of the two treatments in each crust, we calculated light summation of nitrogen-fixation from recovery process of NA in the light of the two treatments, which was obtained by summing of nitrogen fixation of every time period, the product of NA rates times corresponding length of time.
To determine the NA recovery process of BSCs in the dark, after being subjected to rehydration/desiccation cycles, two sets of (with triplicate) of each type of crust sample were desiccated for 2 days and 4-6 months (assigned as L-DR and 2 d -DR respectively) in culture dishes at room temperature and then directly rehydrated and conducted ARA incubation in the dark. Another set (with triplicate) of crust sample (assigned as L-6 h-DR) was subjected to desiccation for 4-6 months firstly and followed by rehydration under the conditions of 30 μmol photos m−2 s−1 cool white fluorescent illumination at 25C for 6 hwhenphotosynthetic activity was fully recovered, and then transferred to dark condition and conducted the ARA incubations in the dark at 25C. During the ARA incubation, the gas subsamples (0.3mL) were taken at 4 h, 8 h and 14 h for ethylene analysis. Like light summation of nitrogen fixation, dark summation of nitrogen-fixation of each crust was calculated fromdark recovery process of NA after 2 days and 4-6 months of desiccation.
References
[1]Li XR, He MZ, Zerbe S, et al. Micro‐geomorphology determines community structure of biological soil crusts at small scales. Earth Surface Processes and Landforms, 2010, 35: 932-940
[2]Coxson D, Kershaw K. Rehydration response of nitrogenase activity and carbon fixation in terrestrial nostoc commune from stipa-bouteloa grassland. Canadian Journal of Botany, 1983, 61: 2658-2668
[3]Proctor MC, Ligrone R, Duckett JG. Desiccation tolerance in the moss polytrichum formosum: Physiological and fine-structural changes during desiccation and recovery. Annals of botany, 2007, 99: 75-93
[4]Harel Y, Ohad I, Kaplan A. Activation of photosynthesis and resistance to photoinhibition in cyanobacteria within biological desert crust. Plant Physiology, 2004, 136: 3070-3079
[5]Wu L, Lan S, Zhang D, et al. Recovery of chlorophyll fluorescence and co< sub> 2</sub> exchange in lichen soil crusts after rehydration. European Journal of Soil Biology, 2013, 55: 77-82
[6]Li Z, Yu J, Kim KR, et al. Nitrogen fixation by a marine non‐heterocystous cyanobacterium requires a heterotrophic bacterial consort. Environmental microbiology, 2010, 12: 1185-1193
[7]Cheng J, Hipkin C, Gallon J. Effects of inorganic nitrogen compounds on the activity and synthesis of nitrogenase in gloeothece (nägeli) sp. Atcc 27152. New phytologist, 1999, 141: 61-70
[8]Stewart W, Fitzgerald G, Burris n. In situ studies on n2 fixation using the acetylene reduction technique. Proceedings of the National Academy of Sciences of the United States of America, 1967, 58: 2071
Table S1 Soil characteristics in different succession stages of BSCs and bare sand (control). Except for dominant components, all values are mean ± standard deviation. Different letters in each row denote significant differences at P = 0·05 level.
Crusttype / Dominant components / Thickness
(mm) / pH / Chl a
(mg kg-1) / Organic matters
(%) / Available Nitrogen
(mg kg-1)
S / - / - / 9.17±0.03a / 0a / 0.83±0.09a / 7.76±0.82a
C / Cyanobacterial crusts. / 2.68±0.77a / 8.55±0.07b / 1.50±0.53b / 1.36±0.20b / 26.02±2.92b
CM / 0-10% moss, >90% Cyanobacteria. / 4.23±0.88b / 8.15±0.03c / 2.68±0.88c / 1.78±0.42bc / 51.28±1.73c
MC / 15-50% moss, 50-85% Cyanobacteria. / 6.22±0.76c / 8.05±0.13c / 2.66±0.15c / 2.21±0.20c / 56.53±5.12c
M / Moss crust. / 11.54±0.84d / 7.50±0.09d / 6.17±0.66d / 3.67±0.65d / 92.72±8.84d
Table S2 The rates of NA and Fv/Fm in different succession stages of BSCs before desiccation started. All values are mean ± standard deviation.
Crust type / NA (C2H4 nmolcm-2h-1) / Fv/FmC / 0.09±0.01 / 0.40±0.04
CM / 1.81±1.39 / 0.41±0.06
MC / 1.78±0.46 / -
M / 0.06±0.01 / 0.72±0.03
Fig. S1 Classification of BSCs. a shifting sand dunes, b cyanobacterial crust stage (Crust-C), c moss colonization stage (Crust-CM), d moss development stage (Crust-MC),e moss dominant crust (Crust-M).
Fig. S2 Recovery of Fv/Fm on rehydration after desiccation periods of 2 days, 2 months and 5 months in three types of soil crusts. Values are means (n=3)±SD.
Fig.S3Light (A) and dark (B) summation of nitrogen-fixation quantity of different crusts under 2 days of desiccation and long desiccation (**: p<0.01; *: p<0.05).