Experimental and sampling design
Biological soil crusts, and related environmental data and
samples, were collected at four sites in the Chihuahuan Desert
with varying vegetation and soils. At each of the four sites in
the study, 2 intersecting 30m transects were laid out in an X
shape crossing at the 15m mark. From these transects, replicate
samples of 5 types of biological soil crusts were collected, along
with representative soil samples and data on vegetation cover.
Collections occurred between May and July of 2020 and samples were
subsequently processed and analyzed using the following methods.
Preparation and metagenomic analysis (sequencing) of
biological soil crust samples
At each of the four sites, five soil crust samples were collected
for each of the five crust types (100 samples total). Some
duplicate samples were collected, for a total of 106 samples. All
soil crust samples were collected along the 2 intersecting 30
meter transects. Samples were stored in a -80C freezer until
processing. Samples were homogenized by physically crushing the
biocrust aggregate and shaking the collection bag until well
mixed. A sample weight of ~0.25g biocrust soil was collected from
each homogenized sample.Whole community genomic DNA was extracted
from each crust type per site using DNeasy PowerLyzer Soil Kit by
QIAGEN (Qiagen Laboratories, Carlsbad, CA, USA), with modification
to the bead beating homogenizer step. Samples were subject to two
45sec interval and 5sec rest cycle in the Precellys 24 tissue
homogenizer (Bertin Instruments, Paris, France) at 2500rpm.
Extracted DNA was sent to University of Minnesota Genomics Center
for library preparation, amplification, and Illumina MiSeq.
Illumina MiSeq sequencing occured in 2x300 PE base format
targeting 16S rRNA V4 gene region with the 515F and 806R primers
following the protocol of the Earth Microbiome Project. Polymerase
chain reaction (PCR) was performed with KAPA HiFidelity Hot Start
Polymerase in a two step process. qPCR was carried out for quality
checking with an initial denaturation 95ºC 5min, followed by 35
cycles of: denaturation 98ºC 20 sec, annealing 55ºC 15sec, and
elongation 72ºC 1 min, followed by a final 70ºC 5min elongation
and holding at 4ºC thereafter. Samples were then normalized to
167,000 molecules/ul. PCR1 was performed using the same steps as
in qPCR but with 25 cycles. Samples were then diluted to 1:100 and
5ul were used in PCR2 where again the same steps were performed as
in qPCR but using different forward and reverse indexing primer
combinations and 10 cycles. Samples were pooled and denatured in 8
pM NaOH diluted in Illumina’s HT1 buffer, spiked with 15% PhiX.
Heat denaturation was also carried out immediately before sample
loading for 2min at 96ºC. Sequencing was performed using a MiSeq
600 cycle v3 kit. Post run trimming was carried out with Nextera
adapter sequences. After demultiplexing, a total of 4,924,626
sequence reads were received in a range from 28-79381 per sample.
The raw sequence data can be found at the NCBI repository under
BioProject #PRJNA748083 Line Point Intercept data Site
characterization via line point intercept surveys was carried out
along two transecs at each of the 4 sites (May-July 2020). Surveys
assessed plant and biocrust percent cover. Presence/absence of
plant and/or biocrust type was recorded every 0.2m along 2
intersecting 30m transects (in X shape crossing at 15m). This
information will be used to upscale carbon and nitrogen fixation
rates to the landscape according to biocrust percent cover.
Soil chemistry data
To characterize soil chemical composition., composite soil samples
(250g) were collected for each crust type at each site from 5
places along the 2 intersecting 30m transects.Samples were
collected using sterile technique and stored at 4˚C until shipping
to Ward Laboratories for analysis.
Cyanobacterial abundance
Supercomputing was used to process sequencing data with
standardized bioinformatics pipelines in Qiime2 and AMPtk.
Briefly, the AMPtk program merged forward and reverse reads,
binned samples according to barcode sequences, removed barcodes
and primers, trimmed reads to 300 base pairs, clustered reads into
ASVs with 97% similarity, and removed singleton ASVs.Bacterial
taxonomy was assigned using Qiime2 pipeline and Silva (Bacteria
and Archaea) database. Sequence data analyses were carried out in
the R package phyloseq. Singletons and 9 samples with low read
count were removed from the data set leaving a total of 4,922,472
sequences in a range of 10,753-79,381 sequences merged. ASVs were
rarified to the minimum read count of 10,753. Chloroplast data was
removed and bacterial data was subsequently assessed for relative
abundance of bacteria by Phylum. Cyanobacterial data was subset
and manually reclassified using Cydrasil version 3.
PLFA Biomass
Broad taxonomic and functional group assignments of soil crust
communities were made using phospholipid fatty acid (PLFA)
analysis. Composite samples (250g) were collected for each crust
type at each site from 5 places along the 2 intersecting 30m
transects (in X shape crossing at 15m). Samples were collected
using sterile technique and stored at 4˚C until shipping to Ward
Laboratories for analysis.
