diversity
Article
Microbial Community in Hyperalkaline Steel
Slag-Fill Emulates Serpentinizing Springs
J. Ingemar Ohlsson 1 , Jay T. Osvatic 2 , Eric D. Becraft 3 and Wesley D. Swingley 1, *
1
2
3
*
Department of Biological Sciences, Northern Illinois University, DeKalb, IL 60115, USA
Centre for Microbiology and Environmental Systems Science, Division of Microbial Ecology, University of
Vienna, 1090 Vienna, Austria
University of North Alabama, Florence, AL 35632, USA
Correspondence: wswingley@niu.edu; Tel.: +1-209-777-4653
Received: 21 May 2019; Accepted: 21 June 2019; Published: 30 June 2019
Abstract: To date, a majority of studies of microbial life in hyperalkaline settings focus on
environments that are also highly saline (haloalkaline). Haloalkaline conditions offer microbes
abundant workarounds to maintain pH homeostasis, as salt ions can be exchanged for protons by
dedicated antiporter proteins. Yet hyperalkaline freshwater systems also occur both naturally and
anthropogenically, such as the slag fill aquifers around former Lake Calumet (Chicago, IL, USA).
In this study, 16S rRNA gene sequences and metagenomic sequence libraries were collected to
assess the taxonomic composition and functional potential of microbes present in these slag-polluted
waterways. Relative 16S rRNA gene abundances in Calumet sediment and water samples describe
community compositions not significantly divergent from those in nearby circumneutral conditions.
Major differences in composition are mainly driven by Proteobacteria, primarily one sequence cluster
closely related to Hydrogenophaga, which comprises up to 85% of 16S rRNA gene abundance in
hyperalkaline surface sediments. Sequence identity indicates this novel species belongs to the
recently established genus Serpentinomonas, a bacterial lineage associated with natural freshwater
hyperalkaline serpentinizing springs.
Keywords: hyperalkaline; culture-independent; serpentinization; microbial diversity; Proteobacteria
1. Introduction
The Earth offers many natural habitats that appear very strange and inhospitable, at least to human
observers. Microbes are routinely found surviving or even thriving in all extremes we encounter, short
of the calderas of active volcanoes [1–3]. Industrial resource extraction and processing is the most
common cause of anthropogenic extreme environments, yielding a variety of novel conditions, such as
those associated with mine tailings [4], limestone processing [5], and slag piles [6]. We study natural
extreme biomes, which typically establish over geological timescales, to learn how microbiomes evolve
under these extremes. Anthropogenic environments, on the other hand, are established over mere
years or decades, which allows us to observe how microbes and microbial communities rapidly adapt
to persistent changes in their environment.
The former Calumet Wetlands abut the southwestern shore of Lake Michigan in the central US,
encompassing parts of southern Chicago, IL and western Gary, IN [7]. Surviving parts of the wetlands
include the Calumet River, Calumet Lake, Wolf Lake and Lake George, though very little of the area
can now be described as marsh. As steel mills and other heavy industry expanded operations in the
area, slag from steel production and other solid debris was used to fill in the marshes, simultaneously
providing more land for development and disposing of industrial wastes “safely” by the standards of
Diversity 2019, 11, 103; doi:10.3390/d11070103
www.mdpi.com/journal/diversity
Diversity 2019, 11, 103
2 of 16
the time [8]. Much of this infill is now built over, but exposed berms of slag and coke ash can be found,
such as the slag pile next to Wolf Lake.
The Calumet region has attracted some research in the decades following the cessation of industrial
activities, with some work addressing concerns for human health and safety [9,10], as residential
neighborhoods cover much of the infilled land. This has, in turn, promoted an interest in the microbiology
of the unique soils and waters in the Calumet Wetlands [8,11–13]. Whereas many natural hyperalkaline
freshwater systems are found in remote locations [14,15] or deep in subsurface aquifers [16,17], Calumet
is easily accessible from the Chicago area. This makes the site ideal for studying adaptations that
permit microbial colonization of highly alkaline and heavily polluted environments, as well as the
potential for enlisting indigenous microbes in the remediation of Calumet and similar sites. In addition
to such potential ecological applications, determining the metabolic potential of microbes inhabiting
extreme environments will inform not only our understanding of the physical limits of currently
extant life, but also of the geological conditions that could give rise to and support life on Earth, and
potentially on other moons and planets.
In this study, culture-independent techniques were used to investigate the microbiome of Indian
Creek, near Lake Michigan in southern Chicago, IL, USA, which features hyperalkaline spring seeps
fed by weathering of large tracts of exposed steel slag [12]. Metagenomic sequencing and 16S rRNA
gene sequencing were performed on horizontal and vertical sediment transects near a shore featuring
a high pH spring source, as well as several soil and water samples from nearby sites, including a
lakeside recreational area. Bioinformatics analyses revealed that transects host many microbial taxa
found in nearby circumneutral sediments up- and downstream along Indian Creek, though relative
abundances vary by alkalinity. Phylum Proteobacteria is overrepresented in aerobic interfaces at the
most alkaline site, with a bacterium closely related to Hydrogenophaga [18] and Serpentinomonas [14]
dominating in these oxic, littoral sediments.
2. Materials and Methods
2.1. Site Description and Sampling
Previous research in the area of former Lake Calumet [7,12,13,19] has indicated a significant
microbial presence in hyperalkaline waters near large tracts of steel slag infill. This study uses sample
material from four sites previously described in [19], identified as sites 1, 4, 5, and 7 (Figure 1A).
Site 1 is a hyperalkaline pond in Big Marsh Park, on the east shore of Lake Calumet. A loose
precipitate of fine, loose calcium carbonate (hereafter referred to simply as calcite) covers the bottom
of the pond, caused by the precipitation of atmospheric CO2 dissolved in the shallow water column
due to the high Ca(OH)2 alkalinity. Reeds and shrubs grow along the pond banks, though leaves in
direct contact with the water tend to bleach, and water plants do not survive within the abundant
carbonate flocculent. Extreme concentrations of lead, zinc and calcium ions have been recorded from a
groundwater observation well about 100 m away from the Site 1 sampling location. Groundwater pH
under the slag is over 13, while surface pH in the pond is ~11 [19].
Site 4 is located on the south bank of Indian Creek, an artificial canal that drains Wolf Lake into
Calumet River to the west (Figure 1B). The canal cuts through an undeveloped area that features
72 acres of exposed slag berms. Meteoric water filters through this slag bed, weathering the calcium-rich
minerals and resulting in large amounts of Ca(OH)2 leaching into the canal through several spring
seeps. Near the springs, calcite flocculent like that found at Site 1 fills Indian Creek to within 5–10 cm of
the water surface, depending on season and rainfall (Figure 1B). Alkalinity at Site 4 also varies slightly
with seasonal water levels, dropping in cases of heavy rains or snowmelt, but the canal water and
sediment maintains a pH ≥ 9.5 throughout the year [19]. Groundwater under the slag bed maintains
a pH at or above 13. As with Site 1, very high levels of lead, zinc and calcium are found in the
groundwater. Since the alkalinity is caused by calcium hydroxide seeping out of the slag pile on the
south bank, it attenuates quickly as spring seeps mix with the canal flow and atmospheric gases. The
Diversity 2019, 11, 103
3 of 16
nearby vegetation effectively illustrates this, as lily pads can be found growing along the north canal
bank, at which point the pH has decreased to circumneutral levels and calcite flocculent no longer
covers the canal bottom (Figure 1B).
Figure 1. Indian Creek within the Calumet Wetlands in south Chicago, IL, USA. (A) Locations of
sample sites around Indian Creek addressed in this work. Site 4 lies on the south bank of the canal, next
to a 72-acre tract of exposed slag mounds. Hyperalkaline fluids seep out of the slag bank groundwater
into the canal, raising pH above 9.5 year-round. Site 7 and Site 5 represent upstream and downstream
control sites. Site 1, a hyperalkaline pond in Big Marsh Park, lies ca. 4 km to the northwest. Map data:
Google. (B) View from the south bank at Site 4. Loose, yellow-white calcium carbonate flocculent fills
the canal to a depth of >1 m near the spring seeps.
Site 5 is located along the north shore of Indian Creek, approximately 500 m downstream of Site 4,
adjacent to a baseball field. Some calcite precipitation can still be seen, likely washed downstream
from the slag infill, but it is far enough downstream that plants and wildlife appear unaffected by the
upstream slag bed, and pH and metal concentrations have normalized [19].
Site 7 lies upstream of Site 4 on the western side of Wolf Lake in the William W. Powers State
Recreational Area, near the mouth of Indian Creek. The lake’s water is circumneutral (pH 7–9) and
negligibly affected by slag fills in the area [19]. The lake is populated with various fish and other
aquatic life that draws visitors for recreational fishing.
Samples were collected over four trips between 2013–2015 (Table S1). In August 2013, soil and
water samples were acquired from all four sites. Sites 5 and 7 were treated as examples of “control
conditions”, with pH and water chemistry relatively unaffected by slag weathering byproducts.
Additionally, horizontal and vertical transects were gathered at Site 4. For the horizontal transect, soil
and sediment was sampled at points 0 cm, 10 cm, 20 cm and 35 cm starting on the shore and continuing
into the waterway adjacent to a surface spring, with the waterline at approximately 15 cm along the
transect. At each transect point, the upper 2–5 cm of sediment was cored with a sterile 15 mL Falcon
tube (Corning Inc., Corning, NY, USA). To limit the risk of cross-contamination, tubes were transported
in factory packaging to the collection sites and handled with sterile procedures during sampling. For
the vertical transect, an alcohol-sterilized 3 × 30 cm circular PVC pipe was driven into the sediment
~40 cm from the bank, capped at both ends, and kept vertical until it could be frozen, sectioned, and
sampled in top, middle and bottom sections (approximately 2 h after sampling). Each section was
sampled from the center of the core, in order to minimize contamination.
Diversity 2019, 11, 103
4 of 16
In April 2014, two horizontal transects were recovered at separate Site 4 spring seeps, as described
above, but sampled at 5 cm intervals along the same 35 cm transect. Another vertical transect was
recovered in the same manner as above, frozen, and sectioned into 3 cm segments before sampling.
In July 2014, another horizontal transect was recovered at the first spring sampled in 2013. In
addition, water samples were gathered from the canal at points up- and downstream of Site 4, at the
springs sampled in April 2014 transects, and at the water-flocculent interface.
In March 2015, a larger vertical transect was collected, using an alcohol-sterilized 7.5 cm × 60 cm
PVC pipe. After freezing, the core was sectioned into inch-thick segments (2.54 cm) before sampling
from the center of the core. Comprehensive site information is available in Table S1.
2.2. DNA Extraction and 16S rRNA Gene Amplification and Sequencing
DNA was extracted using QIAgen PowerSoil kits (QIAgen, Venlo, The Netherlands), following the
standard protocol. The 16S small subunit ribosomal RNA gene V4 hypervariable region was targeted
for PCR amplification using primers 515F and 806R [20]. Template and primers were combined with
QIAgen MasterMix (company info) and MicroPure (MicroPure Filtration Inc., Cleveland, OH, USA)
filtered water according to the standard protocols for MasterMix, to a total volume of 25 µL. Reactions
were conducted in an Eppendorf thermocycler, over 30 cycles of: 30 s denaturation at 95◦ C, 60 s
annealing at 55 ◦ C, and 60 s elongation at 68 ◦ C. Control reactions were prepared for each PCR run,
using previously sequenced soil extract as a positive control template, and replacing the template with
MicroPure water for a negative control. Successful amplification of the targeted fragment was verified
by electrophoresis, before samples were sequenced as 150 bp mate pairs using Illumina MiSeq (Illumina
Inc., San Diego, CA, USA). Samples collected in August 2013 were sequenced by Argonne National
Laboratory (Lemont, IL, USA), while all subsequent sequencing was carried out by the University of
Illinois Chicago Sequencing Core (Chicago, IL, USA).
2.3. 16S rRNA Gene Classification and Analysis
PCR amplified 16S gene reads were classified using a local install of the RDP Classifier v2.7 [21],
with default settings (root confidence threshold 50%). Classifier outputs were parsed and summed
with custom Perl scripts to generate absolute counts of sequences matching each taxon in each sample,
for the taxonomic levels of phylum, class, order, family and genus.
Genus-level taxonomic assignments were used in R v3.6.0 [22] to generate diversity measures
(Genus counts, Pielou’s evenness index, and Shannon diversity index) for eight geochemical categories
of samples, indicating sample type (soil, water or sediment) and origin (transect position). Diversity
measures were generated with the microbiome R package [23]. A principal components analysis (PCA)
plot was generated for the genus abundances (grouped by the same categories) with the R built-in
prcomp function and plotted using ggbiplot [24]. A non-metric multi-dimensional scaling (NMDS)
analysis was also applied to genus-level taxon counts (normalized to relative abundances) using the
metaMDS function in the R vegan package [25].
Pairwise Wilcoxon signed-rank tests were performed for each diversity metric, to determine
statistical similarities between groups. An analysis of similarities (ANOSIM) was performed in R vegan,
to directly test the differences between communities in terms of community composition. Default
parameters were used for all analyses.
2.4. Metagenomic Sequencing, Assembly, and Analysis
Metagenome sequencing was carried out at the University of Illinois Chicago Sequencing Core
(Chicago, IL, USA), on the Illumina MiSeq (Illumina Inc.) platform. Mate-pairs of 150 bp were
sequenced with an average insert size of 250 bp, which resulted in 102 bp paired-end reads after
barcodes and adapter sequences were removed by the sequencing facility.
14 sets of 102 bp paired-end read sets were used for metagenomic assembly. De novo assemblies
were generated for each read set using the SPAdes assembler package v3.1.0 [26], with default
Diversity 2019, 11, 103
5 of 16
parameters (k-mer sizes 21, 33, 55) and automatic coverage estimation. Paired-end reads were included
without overlapping prior to assembly.
After assembly, each scaffold set generated was partitioned into sequence bins with MaxBin
2.0 [27], which uses tetranucleotide frequencies to recover draft genomes. Ribosomal RNA genes were
called from each bin using Barrnap v0.9 [28], and any 16S rRNA genes detected were classified with
the RDP Classifier v2.7, in order to assign putative taxonomic identities to the bins. Any sequence bin
containing more than one identifiable 16S rRNA gene was considered chimeric.
A partial Hydrogenophaga genome was extended by read recruitment from all Site 4 read sets.
Assembled contigs containing 16S rRNA genes classified as Hydrogenophaga, detected by BLASTn
alignment [29] to 16S rRNA gene reads from earlier analyses (see 2.3) were used to train a multilayer
perceptron (MLP) machine-learning algorithm in WEKA [30]. As outlined in Becraft et al. [31], the
trained MLP was used to recruit more reads from all read sets. The extended read set was then
reassembled in SPAdes, with standard parameters, and with the training contig set included as
trusted sequences.
The 16S rRNA gene included in this draft genome was aligned in MEGA 6 [32] with a collection
of representative sequences from families in order Burkholderiales (retrieved from the RDP Hierarchy
Browser [33]), documented mesophilic members of genus Hydrogenophaga (retrieved from NCBI
Gene [34]), and predicted 16S rRNA genes from three Serpentinomonas strains cultured from the Cedars,
CA, USA [14]. Alignment was performed with the MUSCLE algorithm, using UPGMA clustering, with
a gap open penalty of -400 and no gap extension penalty. The alignment was used to create a Maximum
Likelihood phylogenetic tree (also in MEGA 6), to elucidate the placement of this draft genome in
relation to genera Hydrogenophaga and Serpentinomonas. No gaps or mismatches were stripped before
phylogenetic reconstruction. Tree generation used the Maximum Likelihood method with Tamura-Nei
nucleotide substitution model. Initial trees for heuristic search were constructed using the Neighbor
Joining method applied to a matrix of pairwise sequence distances estimated by Maximum Composite
Likelihood [32]. Bootstrap support values were generated over 1000 replications.
The draft genome was submitted to RAST [35] for automated gene annotation. RAST annotations
are accessible from the guest account under ID 6666666.223158—Hydrogenophaga sp.
Basic assembly statistics, including contig and scaffold lengths, total assembly lengths, N50,
and L50, were generated for each assembly with custom Perl, Python and BASH scripts. Genome
completeness and contamination estimates for binned contig sets and the Hydrogenophaga assembly
were generated with CheckM [36], using reference set of 139 conserved single copy genes (CSCGs)
common to all Bacteria. Metagenomic sequence reads are deposited at NCBI under accession
number PRJNA548586.
3. Results
3.1. 16S rRNA Gene Analysis
Relative 16S rRNA gene abundances for both neutral and alkaline sites indicate communities
consist primarily of seven bacterial phyla: Acidobacteria, Actinobacteria, Bacteroidetes, Cyanobacteria
(and eukaryote chloroplasts), Firmicutes, Proteobacteria and Verrucomicrobia (Figure 2, Figures S1 and
S2; Table S2).
Most notably, class Betaproteobacteria are present at minimum 20% in circumneutral sites, and
overrepresented in certain segments at Site 4 (Figures 2 and 3), specifically in sediments near the
waterline. Actinobacteria and Bacteroidetes are also present at significant abundance in most samples,
though Actinobacteria abundance is much lower in neutral soil samples (S5S, S7S), while these samples
have a higher relative abundance of Acidobacteria. At points ca. 5 cm or deeper below the soil surface,
or under at least 10 cm of water column, Firmicutes and Bacteroidetes are most prevalent (Figure 4).
Diversity 2019, 11, 103
6 of 16
Figure 2. Phylum-level 16S rRNA gene relative abundances at sites 4, 5 and 7 around the Calumet
Wetlands. All samples were collected in August 2013. S*W and S*S denote soil and water samples from
the corresponding site.
Figure 3. Phylum- and Proteobacterial class-level 16S rRNA gene relative abundances in three horizontal
transects collected at Site 4: (A) 35 cm transect collected at Spring A in April 2014. (B) 35 cm transect
collected at Spring B in April 2014. (C) 45 cm transect collected at Spring A in July 2014. Transects
started inland at each spring seep, ca. 15 cm from the waterline. Samples were collected from the top
layer of soil (on land) or upper layer of sediment (in the canal water column).
Diversity 2019, 11, 103
7 of 16
Figure 4. Phylum- and Proteobacterial class-level 16S rRNA gene relative abundances in two vertical
transects collected at Site 4: (A) 3 cm diameter core collected in April 2014; (B) 7.5 cm diameter core
collected in March 2015. Depth indicates the top side of each core segment, starting at the top of the
calcium carbonate flocculent. Water column is not included, but averaged 5–10 cm above sediment.
PCA plots of all samples based on genus abundances (Figure 5) abstract taxonomic detail in favor
of highlighting similarities between samples. Principal components 1 and 2 explain 48.3% and 17.3% of
variance, respectively. Samples generally clustered by sample type, alkalinity, and geographical origin,
with the notable exception of a cluster containing circumneutral soil and water samples together with
Site 4 hyperalkaline water (hereafter called the “neutral cluster”). Horizontal transects samples are
tightly clustered, but spread across PC1, with the submerged sediment samples frequently clustering
with the neutral cluster. The vertical transect samples are the most dispersed, especially along PC2,
with the deepest samples typically clustering with the neutral cluster.
Diversity 2019, 11, 103
8 of 16
Figure 5. PCA plot of genus-level 16S rRNA gene abundance data for all Calumet area samples.
Principle components 1 and 2 shown, which together explain 65.6% of variance.
In the NMDS plot (Figure 6), samples from Sites 4 were grouped by oxygen access, to clarify
important differences within the alkaline environment. Anoxic alkaline samples, from deeper sediments
and soil, show a wide spread, overlapping to some degree with all other groups. Water samples
and oxic alkaline samples each cluster toward separate edges of the anoxic community, with no
overlap between the two. The oxic alkaline samples, from the bank and waterline of Site 4, are very
tightly clustered.
Figure 6. Non-Metric Multi-Dimensional Scaling plot of Calumet sites using Bray-Curtis distance,
grouped by physical conditions of samples. “Oxic Alkaline” corresponds to the Horizontal
Bank/Waterline samples, “Anoxic Alkaline” corresponds to Horizontal Sediment and all Vertical
transect samples. Ellipses represent 95% confidence intervals for the NMDS coordinates.
Shannon diversity distributions (Figure 7A) based on genus identities for soils and waters, along
with middle segments of vertical transects, have a narrow range of diversity with means ca. 3.5.
Other samples have much wider ranges, with lower means ca. 2.5. However, the only statistically
Diversity 2019, 11, 103
9 of 16
significant difference (at 5% significance by pairwise Wilcoxon signed-rank test) was between the
waterline at Site 4 and the open water sample cohort (both circumneutral and alkaline). Water samples
also have significantly greater Pielou evenness than waterline samples (Figure 7B). While evenness is
generally greater in vertical transect middle layers and in neutral soil samples, these differences were
not statistically significant. While there are no statistically significant differences in genus richness
among sample cohorts (Figure 7C), mean richness is highest in neutral soils (ca. 680) and lowest in
the bottom vertical transect layers (ca. 220). Remaining samples have largely overlapping richness
distributions arrayed between these extremes.
Figure 7. Genus-level statistical measurements of Calumet samples: (A) Shannon (alpha) diversity;
(B) Pielou evenness; and (C) richness. Means are marked with a heavy black line, and hinges indicate
the first and third quartiles. Whiskers indicate the range of values outside these quartiles. Outliers
(farther than 1.5 × the interquartile range) are indicated by black circles. Similarity groups at 95%
confidence are indicated by letters a and b.
ANOSIM results (Figure S3) shows that mean difference between samples is higher than the
differences within the oxic alkaline, anoxic alkaline and water sample groups. Oxic alkaline samples
have the least within-group difference. Each group can be considered distinct, though their quartiles
show some overlap. This assessment does not extend to circumneutral soil samples, of which there
were too few to estimate the distribution.
3.2. Metagenomic Analysis
De novo assembly produced a total of 1.6 Gbp in 3.6 million sequence scaffolds between the 14 input
read sets. Assembly statistics are summarized in Table 1. The highest N50 values were from samples
S435cm (N50 = 17,104), S410cm (N50 = 11,177) and S40cm (N50 = 10,449), all horizontal transect
samples from Site 4. Sequences from water samples uniformly produced fragmented assemblies with
N50 ≤ 896, and circumneutral soil metagenomes produced N50 ≤ 488, supporting the suggestion
≤
≤
that these samples had higher genetic diversity (Figure 7A). Total scaffold length was very high
(216–766 million bp) in all circumneutral metagenomes except S7W (4.8 million bp), while the average
for all read sets was 117,890,266.79 bp.
Diversity 2019, 11, 103
10 of 16
Table 1. Metagenomic assembly statistics for Calumet read sets.
Read Set
Total
Sequence
Input (bp)
Total
Scaffold
Length (bp)
Number of
Scaffolds
Maximum
Contig
Length (bp)
N50 (bp)
L50 (bp)
S1S
S1W
S40cm
S410cm
S420cm
S435cm
S4S
S4verticalbottom1
S4verticaltop1
S4W
S5S
S5W
S7S
S7W
4,819,674,420
1,942,236,264
6,321,349,632
1,138,541,340
8,551,723,860
7,864,380,132
8,901,634,656
3,488,100,120
5,195,853,276
781,642,320
7,256,817,336
1,891,476,984
3,027,277,992
1,202,217,492
9,869,266
96,016,654
27,714,568
4,063,552
33,109,460
31,171,314
17,907,967
45,790,429
11,790,385
91,790,252
766,453,974
216,488,361
293,538,298
4,759,255
8842
189,034
7294
2257
39,264
20,943
13,735
148,148
21,636
237,098
1,650,212
534,113
692,250
17,534
70,488
131,675
288,607
62,774
137,610
138,650
135,896
32,268
99,847
42,452
594,813
123,446
139,485
36,970
4776
896
10,449
11,177
3111
17,104
7433
293
6486
363
488
404
450
259
457
15,108
641
86
2415
437
522
44,271
372
58,467
338,228
124,669
164,345
6,864
4549.21
54,063.00
Sum
62,382,925,824 1,650,463,735
3,582,360
2,034,981
Average
4,455,923,273.14 117,890,266.79
255,882.86
145,355.79
Automated scaffold binning results are summarized in Table 2. MaxBin 2.0 generated 23 sequence
bins for read set S5S, 0 bins for read set S7W, and 2–6 bins for each other read set. Barrnap detected 177
16S rRNA genes (2–42 genes per read set).
Table 2. Metagenomic binning statistics for Calumet read sets.
Read Set
Scaffold
Bins
Detected
16S rRNA
Genes
Chimeric
Bins
Average
Completeness
Average
Genome
Size
Average GC
Content
S1S
S1W
S40cm
S410cm
S420cm
S435cm
S4S
S4verticalbottom1
S4verticaltop1
S4W
S5S
S5W
S7S
2
6
4
2
3
5
4
2
2
4
23
6
4
4
11
12
2
19
22
21
7
10
8
42
10
9
1
2
1
0
1
3
3
2
1
3
4
1
2
64.1%
75.2%
36.5%
49.1%
50.2%
81.3%
46.5%
62.6%
74.3%
68.2%
72.9%
58.6%
75.0%
3,945,426
3,808,967
3,072,671
1,787,619
4,217,172
4,979,437
3,947,249
1,663,564
4,420,534
2,177,976
3,830,319
2,675,264
4,072,426
63.5%
55.3%
59.7%
62.7%
56.3%
45.9%
51.6%
50.2%
54.1%
48.3%
53.9%
46.0%
46.4%
Sum
67
177
24
Comparison of classifications of each rRNA gene sequence found 24 bins to be chimeric (0–4 bins
per read set), as evidenced by presence of multiple incompatible classifications within the same bin.
Only the S410cm read set was free of chimeric bins. This read set produced only two bins, one of
which was a near-complete draft Hydrogenophaga genome (97.2% completeness, genome size 2.4 Mbp),
while the other contained a single 16S rRNA gene that could only be confidently assigned to domain
Archaea, with an estimated completeness only 0.9%, and a genome size of 1.2 Mbp.
Read recruitment by multilayer perceptron produced another draft Hydrogenophaga genome with
2.9 Mbp genome size and 100% completeness as reported by CheckM. RAST annotations showed
the presence of putative carboxysome shell proteins and the large subunit of ribulose bisphosphate
carboxylase, indicating the ability to fix inorganic carbon, and hyaA/hyaB hydrogen utilization genes.
Presence of CO2 fixation subsystems without evidence of light-harvesting complexes suggests that
these signatures are genuine, and not the result of cyanobacterial contamination in the assembly. These
Diversity 2019, 11, 103
11 of 16
annotations match similar activity in Serpentinomonas strains reported in Suzuki et al. [14]. Other
annotated functions include Gram-negative cell wall components, capsule- and EPS-related genes,
several ABC transporters, metal compound resistance genes, and systems for chemotaxis and flagellar
motility. Ten prophage genes were annotated, but no transposable elements. Sequence alignment
with representatives of known Hydrogenophaga, Serpentinomonas and neighboring bacterial families
produced a Maximum Likelihood phylogenetic tree showing the draft genome clustering within
the Serpentinomonas clade (Figure 8) with S. mccroryi and S. rachei [14]. Serpentinomonas raichei and
Serpentinomonas mccroryi were recently renamed, and are referred to as Comamonadaceae strains A1/H1
and B1 in the original publication [14].
Figure 8. Maximum Likelihood phylogenetic tree of 16S rRNA genes from select members of order
Burkholderiales. Highlighted in red: Hydrogenophaga species in the NCBI database, typically
not associated with extreme alkalinity. Highlighted in blue: the Calumet draft genomes, clustered
with three Serpentinomonas strains isolated from the Cedars serpentinizing springs in California,
USA. Serpentinomonas raichei was previously identified as Comamonadaceae strains A1/H1, and
Serpentinomonas mccroryi as Comamonadaceae strain B1. Bootstrap support values from 1000 iterations
are given in bold next to associated nodes.
4. Discussion
The microbial community at Calumet reflects striking similarities to nearby natural (or rather,
not directly polluted) sites, suggesting that the microbial library of this damaged environment has
persisted even through nearly 100 years of anthropogenic impact. Strong similarities between the
microbial communities in the Calumet Wetlands’ interconnected hydrological network (Figure 5),
are a likely source of continual microbial input. Even if these organisms cannot take advantage of the
hyperalkaline conditions at Site 4, their genetic material may remain trapped in the calcium carbonate
sediments for long after their arrival. The relatively high rate of input to this system (Chicago receives
>90 cm of rain per year; www.weather.gov) makes the Calumet hyperalkaline sites unique in their
proximity between neutral waters and such extreme alkalinity, and a perfect test bed for understanding
extremophile adaptation and biogeography.
Scientific understanding of freshwater hyperalkaliphiles has much room for improvement,
particularly in terms of shared traits among such organisms. While any bacterium that is able to
survive at the high end of the pH scale can inform us about responses to alkaline stress, bacteria that
thrive in such harsh conditions would be of interest for use in bioremediation of alkaline pollution.
Once members of the microbiome at Indian Creek are established in culture, and their genomes are
sequenced and annotated, it will be possible to investigate their growth behavior in detail and assess
whether they can be used as a delivery vehicle for remediation strategies.
The differences between soil and water samples are, with a few exceptions, clearer than those
between neutral and alkaline sites. Actinobacteria are virtually absent from Site 5 and 7 soil samples
Diversity 2019, 11, 103
12 of 16
and “Cyanobacteria” (representing both Cyanobacteria and chloroplast genomes of eukaryotes, e.g.,
diatoms) are heavily reduced compared to their numbers in water samples from all studied sites. There
is also a clear transition with vertical depth, as Proteobacteria give way to Bacteroidetes and Firmicutes
in the deeper, increasingly anaerobic sediments (Figure 4).
The most striking feature of the 16S rRNA gene survey is the dominance of a Betaproteobacterial
lineage in oxic sediments at Site 4, reaching as high abundance as 75–85% of the entire microbial
community near the waterline around 15 cm along each transect (Figure 3). It appears that this
bacterium, closely related to the genera Hydrogenophaga, is singularly adapted to the harsh conditions
at Site 4. However, its metabolism is still reliant on abundant oxygen, given its disappearance in deeper
samples. It may also require access both to water and the carbonate flocculent to grow. Culture-based
studies of bacteria isolated from the Cedars, CA, USA [14] found three strains of Serpentinomonas
actively colonizing calcium carbonate precipitates, with growth on solid media dependent on added
carbonate. Solid-phase carbonate is generally not considered bioavailable at the pH levels of Indian
Creek, as it is highly insoluble. The Site 4 Betaproteobacteria may also possess the uncommon ability
to utilize solid carbonate precipitates as a carbon source, which would confer a major advantage in this
environment, where other carbon sources are scarce. The mechanism by which Serpentinomonas takes
up carbon from the flocculent is unclear, but all three strains from the Cedars possess carboxysome shell
proteins and components of the RuBisCo enzyme[14], which suggests that they resolubilize carbonate
ions before using traditional carbon fixation pathways to utilize the carbon content of the flocculent.
Putative genes coding for all of these activities were also present in the Calumet metagenome assembly.
Principal component analysis of genus-level 16S rRNA abundances (Figure 5) showed soil and
water samples clustered very tightly together with some deep vertical core segments. Horizontal
transect samples plot along a very well-defined line, but the arrangement of bank, waterline and
sediment sections of the horizontal transects along this vector does not correspond to their physical order.
Considering that Proteobacteria abundances quickly give way to Firmicutes and Bacteroidetes (Figure 3)
over a distance of 10 cm, this linear cluster may be tied to oxygen availability affecting species richness.
In the shallow, oxic sediment, Serpentinomonas is dominant, distancing the community as a whole from
the more diverse, neutral environments. Diversity increases somewhat with depth, as oxygenation
and Serpentinomonas declines. Alkalinity is also mitigated by calcite precipitation, as calcium ions are
pulled out of solution by dissolved atmospheric CO2 [12]. Overall, this PCA reinforces the notion
that microbial communities at Site 4 are related to nearby neutral environments, but combinations of
geochemical conditions promote highly localized concentrations of certain specialized taxa.
NMDS (Figure 6) and ANOSIM (Figure S3) results also support the connection between Calumet
sample groups. Confidence ellipses in the NMDS plot and group differences overlap noticeably
between the oxic and anoxic alkaline samples, and the water sample cohort. Oxic alkaline samples are
more distinct, due to being tightly grouped. This is most likely a result of the very narrow range of
taxa at the Site 4 bank and waterline. This is the most selective segment of the hyperalkaline slag-fill
waterway; in the main waterway, or deeper in the sediment, more varied communities can survive.
Metagenomic assembly of read sets from circumneutral samples produced the greatest total
assembly lengths, but also the worst N50, indicating the assembled scaffolds were highly fragmentary
(Table 1). This is likely due to the greater diversity and evenness of microbial communities at these
sites, which means that “shotgun sequencing” methods are less likely to hit any given species. Water
samples also had poor assembly performance, likely due to the dilute cell concentrations and lower
total read count in these samples.
Metagenomes from the hyperalkaline soil and shallow sediments of Site 4, on the other hand,
produced the three best assemblies, and other alkaline sample assemblies were of similar quality. This
is likely an effect of high alkalinity restricting species richness by suppressing mesophilic species,
which means that alkaliphiles had higher relative sequence coverage. Contrary to the difficulties
of culturing extremophiles in a laboratory setting, culture-independent methods can be particularly
Diversity 2019, 11, 103
13 of 16
effective in highly selective environments, where a small number of community members dominate,
which bodes well for future applications of environmental metagenomics.
Automatic binning produced multiple scaffold bins for all read sets except S7W (Table 2), derived
from water at Wolf Lake, likely due to the highly fragmented assembly caused by the diverse and
dilute community structure in neutral waters. Given the generally diverse populations detected by 16S
rRNA gene sequencing, fragmented and chimeric bins might be expected, as great sequencing depth
may be required to resolve complex metagenomes. However, a majority of these bins contained only
one 16S rRNA gene, or several that had non-conflicting taxonomic assignments. Further analysis is
needed, with direct sequence-based methods, to confirm whether or not these bins are truly chimeric,
but this initial ribosomal RNA genotyping overview is promising.
The extreme overrepresentation of Hydrogenophaga-related species at the Site 4 waterline led
to highly effective sequence coverage. One bin recovered from the Site 4 waterline represents a
highly complete genome (97.2% completeness, 2.4 Mbp) assigned to genus Hydrogenophaga by the
RDP Classifier. MLP read recruitment of Hydrogenophaga-tagged contigs also resulted in a highly
complete draft genome (100% completeness, 2.9 Mbp). Since genus Serpentinomonas was very recently
established [14], the type strain sequences had yet to be added to RDP at the time of this analysis, but
phylogenetic comparison (Figure 8) shows that the dominant species around the waterline at Site 4 is a
close relative of S. mccroryi.
The presence of Serpentimonas, specialized at surviving in naturally occurring serpentinizing
springs, in an anthropogenic Site of industrial waste pollution thousands of miles from the nearest
serpentinizing site, is perhaps surprising, yet the similarity between conditions in these two disparate
environments are similar enough to promote this organism above the community background. Its
presence at such distant sites makes a strong argument for the Baas-Becking hypothesis that states
“everything is everywhere, but the environment selects” [37], at least for these related extreme
ecosystems. Both types of sites host hyperalkaline aquifers, at pH 13 in the case of Calumet and around
pH 11 for most serpentinization sites [17,38,39]. Neither Indian Creek nor terrestrial serpentinizing
springs have high levels of salinity [17,19,39], meaning alkaliphiles in these waters cannot rely on
antiporter proteins exchanging sodium ions for protons across the cell membrane to maintain pH
homeostasis [40]. For haloalkaliphiles in soda lakes (e.g., Mono Lake, CA, USA [41]), sodium exchange
is a readily available tool [42], but in freshwater systems, like the ones studied here, unknown
mechanisms will be required to maintain homeostasis. Both are recognized by large amounts of calcite
precipitation, though in natural sites the precipitates often build up persistent travertine deposits or
carbonate crusts or sinters [15], as opposed to the loose, snowy flocculent in Indian Creek, which
is likely due to the low persistence time of this anthropogenic system. Copious hydrogen is a key
signature of serpentinizing systems [14,39], and studies of the slag weathering phenomenon in Calumet
suggest that slag-water interactions may release hydrogen gas [12], but its presence in the sediments of
Indian Creek has yet to be confirmed.
The slag-water interface in the Calumet area may be similar enough to serpentinizing bedrock
systems to allow colonization by serpentinization-selected species. More studies are needed to
determine whether the elevated levels of heavy metals and other anthropogenic hazards [12,19] have
forced this novel Serpentinomonas genome to develop further adaptations to dominate this unique
hyperalkaline system. The success of Serpentinomonas at Indian Creek has a wide variety of implications,
beyond the evolutionary history of the species itself. Notably, it associates serpentinization-related
bacterial life with anthropogenic deposits (specifically, the slag infill) which are chemically and
geographically distinct from the ultramafic olivine minerals that drive natural serpentinization [12,43].
This affects our understanding as high-pH, hydrogen-rich environments as potential sites for the origin
of life, by broadening the gallery of minerals that can support such conditions beyond forsterite and
fayalite, which in solid solution form olivine.
Diversity 2019, 11, 103
14 of 16
5. Conclusions
This study has shown that, despite the extreme alkalinity, carbonate precipitation and heavy
metal pollution at the Calumet Wetlands Site 4, microbial communities at this Site remain diverse
and strongly correlated with those in nearby circumneutral sites. Despite the apparent similarities
in communities from both alkaline and circumneutral conditions, a few species thrive in certain
hyperalkaline microenvironments at Indian Creek, particularly a Betaproteobacterium related to genus
Serpentinomonas. Like its recently-described sister taxa, this species thrives in close association with the
carbonate flocculent, a rarely used carbon source among bacteria. The two draft genomes established
here provide more resources for investigating the mechanisms that allow Serpentinomonas to survive
extremely alkaline pH without the aid of abundant salt ions, while subsisting on recalcitrant carbon
sources. Understanding these survival mechanisms opens new avenues of research into bioremediation
of hyperalkaline pollution, as well as the detection of microbial life in serpentinization-like conditions
on Earth and other planets.
Supplementary Materials: The following are available online at http://www.mdpi.com/1424-2818/11/7/103/s1,
Figure S1: Phylum- and Proteobacterial class-level 16S rRNA gene relative abundances from water samples at sites
4, 5 and 7; Figure S2: Rarefaction curves for all datasets. Rarefaction for samples using RDP OTUs at the genus
level using 1000 replicates; Figure S3: Analysis of similarities (ANOSIM) plot of Calumet sites, grouped by physical
condition of samples. Non-overlapping notches suggest medians are distinct. Dissimilarity between groups is
relatively high, and Oxic Alkaline (Serpentinomonas-rich) samples have low internal dissimilarity; Table S1: Field
Sample Information, Table S2: Class-level 16S rRNA gene relative abundances from water samples.
Author Contributions: Conceptualization, W.D.S.; methodology, J.I.O., J.T.O., E.D.B., and W.D.S.; software,
J.I.O., J.T.O., E.D.B., and W.D.S.; formal analysis, J.I.O.; writing—original draft preparation, J.I.O. and W.D.S.;
writing—review and editing, J.I.O. and W.D.S.; visualization, J.I.O. and W.D.S.; project administration, W.D.S.
Funding: This research received no external funding
Acknowledgments: We wish to thank Karel Waska and Melissa Lenczewski for preceding geochemical work;
Stefan Green for assistance with sequencing; Sarah Barmann, Karley Davidson, Amy Daly, Katie Didier, Marie
Kroeger, Lori Lovell, and Kathryn Olson for their help with sampling, culturing and analysis.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
Rothschild, L.J.; Mancinelli, R.L. Life in extreme environments. Nature 2001, 409, 1092. [CrossRef] [PubMed]
Pikuta, E.V.; Hoover, R.B.; Tang, J. Microbial Extremophiles at the Limits of Life. Crit. Rev. Microbiol. 2007,
33, 183–209. [CrossRef] [PubMed]
Harrison, J.P.; Gheeraert, N.; Tsigelnitskiy, D.; Cockell, C.S. The limits for life under multiple extremes. Trends
Microbiol. 2013, 21, 204–212. [CrossRef] [PubMed]
Bonilla, J.O.; Kurth, D.G.; Cid, F.D.; Ulacco, J.H.; Gil, R.A.; Villegas, L.B. Prokaryotic and eukaryotic
community structure affected by the presence of an acid mine drainage from an abandoned gold mine.
Extremophiles 2018, 22, 699–711. [CrossRef] [PubMed]
Kalwasińska, A.; Felföldi, T.; Szabó, A.; Deja-Sikora, E.; Kosobucki, P.; Walczak, M. Microbial communities
associated with the anthropogenic, highly alkaline environment of a saline soda lime, Poland. Antonie Van
Leeuwenhoek 2017, 110, 945–962. [CrossRef] [PubMed]
Freitas, D.B.; Lima-Bittencourt, C.I.; Reis, M.P.; Costa, P.S.; Assis, P.S.; Chartone-Souza, E.; Nascimento, A.M.A.
Molecular characterization of early colonizer bacteria from wastes in a steel plant. Lett. Appl. Microbiol. 2008,
47, 241–249. [CrossRef] [PubMed]
Kay, R.T.; Duwelius, R.F.; Brown, T.A.; Micke, F.A.; Witt-Smith, C.A. Geohydrology, Water Levels and Directions
of Flow, and Occurrence of Light-Nonaqueous-Phase Liquids on Ground Water in Northwestern Indiana and the
Lake Calumet Area of Northeastern Illinois; Water-Resources Investigations Report; U.S. Geological Survey:
Indianapolis, IN, USA, 1996; pp. 1–26/88.
Colten, C.E. Industrial Wastes in the Calumet Area, 1869–1970 A Historical Geography; State Water Survey
Division Report; Hazardous Waste Research and Information Center: Champaign, IL, USA; Available online:
https://www.ideals.illinois.edu/handle/2142/1903 (accessed on 18 May 2019).
Diversity 2019, 11, 103
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
15 of 16
Westphal, L.M.; Longoni, M.; LeBlanc, C.L.; Wali, A. Anglers’ Appraisals of the Risks of Eating Sport-Caught
Fish from Industrial Areas: Lessons from Chicago’s Calumet Region. Hum. Ecol. Rev. 2008, 15, 46–62.
Dourson, M.L.; Chinkin, L.R.; MacIntosh, D.L.; Finn, J.A.; Brown, K.W.; Reid, S.B.; Martinez, J.M. A case
study of potential human health impacts from petroleum coke transfer facilities. J. Air Waste Manag. Assoc.
2016, 66, 1061–1076. [CrossRef] [PubMed]
Bayless, E.R.; Greeman, T.K.; Harvey, C.C. Hydrology and Geochemistry of a Slag-Affected Aquifer and Chemical
Characteristics of Slag-Affected Ground Water, Northwestern Indiana and Northeastern Illinois; Water-Resources
Investigations Report; U.S. Geological Survey; Branch of Information Services [distributor]: Denver, CO,
USA, 1998; pp. 1–4/67.
Roadcap, G.S.; Kelly, W.R.; Bethke, C.M. Geochemistry of extremely alkaline (pH > 12) ground water in
slag-fill aquifers. Ground Water 2005, 43, 806–816. [CrossRef] [PubMed]
Roadcap, G.S.; Sanford, R.A.; Jin, Q.; Pardinas, J.R.; Bethke, C.M. Extremely alkaline (pH > 12) ground water
hosts diverse microbial community. Ground Water 2006, 44, 511–517. [CrossRef]
Suzuki, S.; Kuenen, J.G.; Schipper, K.; van der Velde, S.; Ishii, S.; Wu, A.; Sorokin, D.Y.; Tenney, A.;
Meng, X.; Morrill, P.L.; et al. Physiological and genomic features of highly alkaliphilic hydrogen-utilizing
Betaproteobacteria from a continental serpentinizing site. Nat. Commun. 2014, 5, 3900. [CrossRef] [PubMed]
Woycheese, K.M.; Meyer-Dombard, D.R.; Cardace, D.; Argayosa, A.M.; Arcilla, C.A. Out of the dark:
Transitional subsurface-to-surface microbial diversity in a terrestrial serpentinizing seep (Manleluag,
Pangasinan, the Philippines). Front. Microbiol. 2015, 6, 44. [CrossRef] [PubMed]
Meyer-Dombard, D.R.; Woycheese, K.M.; Yargıçoğlu, E.N.; Cardace, D.; Shock, E.L.; Güleçal-Pektas, Y.;
Temel, M. High pH microbial ecosystems in a newly discovered, ephemeral, serpentinizing fluid seep at
Yanartaş (Chimera), Turkey. Front. Microbiol. 2015, 5, 723. [CrossRef] [PubMed]
Fones, E.M.; Colman, D.R.; Kraus, E.A.; Nothaft, D.B.; Poudel, S.; Rempfert, K.R.; Spear, J.R.; Templeton, A.S.;
Boyd, E.S. Physiological adaptations to serpentinization in the Samail Ophiolite, Oman. ISME J. 2019, 1,
1750–1762. [CrossRef] [PubMed]
Willems, A.; Busse, J.; Goor, M.; Pot, B.; Falsen, E.; Jantzen, E.; Hoste, B.; Gillis, M.; Kersters, K.; Auling, G.;
et al. Hydrogenophaga, a New Genus of Hydrogen-Oxidizing Bacteria That Includes Hydrogenophaga
flava comb. nov. (Formerly Pseudomonas flava), Hydrogenophaga palleronii (Formerly Pseudomonas palleronii),
Hydrogenophaga pseudoflava (Formerly Pseudomonas pseudoflava and “Pseudomonas carboxydoflava”), and
Hydrogenophaga taeniospiralis (Formerly Pseudomonas taeniospiralis). Int. J. Syst. Evol. Microbiol. 1989, 39,
319–333.
Waska, K. Hyperalkaline Aquifers of Calumet Wetlands: Environmental Interactions between Geochemistry
and Microbiology. Ph.D. Thesis, Northern Illinois University, DeKalb, IL, USA, 2014.
Caporaso, J.G.; Lauber, C.L.; Walters, W.A.; Berg-Lyons, D.; Lozupone, C.A.; Turnbaugh, P.J.; Fierer, N.;
Knight, R. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc. Natl.
Acad. Sci. USA 2011, 108, 4516–4522. [CrossRef] [PubMed]
Wang, Q.; Garrity, G.M.; Tiedje, J.M.; Cole, J.R. Naive Bayesian Classifier for Rapid Assignment of rRNA
Sequences into the New Bacterial Taxonomy. Appl. Environ. Microbiol. 2007, 73, 5261–5267. [CrossRef]
[PubMed]
R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing:
Vienna, Austria, 2019.
Lahti, L.; Shetty, S. Microbiome R Package. Available online: http://microbiome.github.com/microbiome
(accessed on 16 June 2019).
Vu, V.Q. Ggbiplot: A Ggplot2 Based Biplot. Available online: http://github.com/vqv/ggbiplot (accessed on 16
June 2019).
Oksanen, J.; Blanchet, F.G.; Friendly, M.; Kindt, R.; Legendre, P.; McGlinn, D.; Minchin, P.R.; O’Hara, R.B.;
Simpson, G.L.; Solymos, P.; et al. Vegan: Community Ecology Package. Available online: https://CRAN.Rproject.org/package=vegan (accessed on 15 June 2019).
Bankevich, A.; Nurk, S.; Antipov, D.; Gurevich, A.A.; Dvorkin, M.; Kulikov, A.S.; Lesin, V.M.; Nikolenko, S.I.;
Pham, S.; Prjibelski, A.D.; et al. SPAdes: A new genome assembly algorithm and its applications to single-cell
sequencing. J. Comput. Biol. J. Comput. Mol. Cell Biol. 2012, 19, 455–477. [CrossRef] [PubMed]
Wu, Y.-W.; Simmons, B.A.; Singer, S.W. MaxBin 2.0: An automated binning algorithm to recover genomes
from multiple metagenomic datasets. Bioinformatics 2016, 32, 605–607. [CrossRef] [PubMed]
Diversity 2019, 11, 103
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
16 of 16
Seemann, T. Barrnap v0. Available online: https://github.com/tseemann/barrnap (accessed on 22 January 2019).
Camacho, C.; Coulouris, G.; Avagyan, V.; Ma, N.; Papadopoulos, J.; Bealer, K.; Madden, T.L. BLAST+:
Architecture and applications. BMC Bioinform. 2009, 10, 421. [CrossRef]
Frank, E.; Hall, M.; Holmes, G.; Kirkby, R.; Pfahringer, B.; Witten, I.H.; Trigg, L. Weka-A Machine Learning
Workbench for Data Mining. Data Min. Knowl. Discov. Handb. 2009, 1269–1277.
Becraft, E.D.; Dodsworth, J.A.; Murugapiran, S.K.; Ohlsson, J.I.; Briggs, B.R.; Kanbar, J.; De Vlaminck, I.;
Quake, S.R.; Dong, H.; Hedlund, B.P.; et al. Single-Cell-Genomics-Facilitated Read Binning of Candidate
Phylum EM19 Genomes from Geothermal Spring Metagenomes. Appl. Environ. Microbiol. 2016, 82, 992–1003.
[CrossRef] [PubMed]
Tamura, K.; Stecher, G.; Peterson, D.; Filipski, A.; Kumar, S. MEGA6: Molecular Evolutionary Genetics
Analysis Version 6.0. Mol. Biol. Evol. 2013, 30, 2725–2729. [CrossRef] [PubMed]
Hierarchy Browser. Available online: http://rdp.cme.msu.edu/hierarchy/hb_intro.jsp (accessed on 27 June 2019).
Home-Gene-NCBI. Available online: https://www.ncbi.nlm.nih.gov/gene/ (accessed on 27 June 2019).
Aziz, R.K.; Bartels, D.; Best, A.A.; DeJongh, M.; Disz, T.; Edwards, R.A.; Formsma, K.; Gerdes, S.; Glass, E.M.;
Kubal, M.; et al. The RAST Server: Rapid annotations using subsystems technology. BMC Genomics 2008, 9,
75. [CrossRef] [PubMed]
Parks, D.H.; Imelfort, M.; Skennerton, C.T.; Hugenholtz, P.; Tyson, G.W. CheckM: Assessing the quality of
microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 2015, 25, 1043–1055.
[CrossRef] [PubMed]
Baas-Becking, L.G.M. Geobiologie; of Inleiding tot de Milieukunde; WP Van Stockum & Zoon NV: Den Haag,
The Netherlands, 1934.
Schrenk, M.O.; Brazelton, W.J.; Lang, S.Q. Serpentinization, Carbon, and Deep Life. Rev. Mineral. Geochem.
2013, 75, 575–606. [CrossRef]
Crespo-Medina, M.; Twing, K.I.; Sánchez-Murillo, R.; Brazelton, W.J.; McCollom, T.M.; Schrenk, M.O.
Methane Dynamics in a Tropical Serpentinizing Environment: The Santa Elena Ophiolite, Costa Rica. Front.
Microbiol. 2017, 8, 916. [CrossRef]
Horikoshi, K. Alkaliphiles: Some Applications of Their Products for Biotechnology. Microbiol. Mol. Biol. Rev.
1999, 63, 735–750.
Banciu, H.L.; Muntyan, M.S. Adaptive strategies in the double-extremophilic prokaryotes inhabiting soda
lakes. Curr. Opin. Microbiol. 2015, 25, 73–79. [CrossRef]
Detkova, E.N.; Boltyanskaya, Y.V. Osmoadaptation of haloalkaliphilic bacteria: Role of osmoregulators and
their possible practical application. Microbiology 2007, 76, 511–522. [CrossRef]
Brazelton, W.J.; Morrill, P.L.; Szponar, N.; Schrenk, M.O. Bacterial Communities Associated with Subsurface
Geochemical Processes in Continental Serpentinite Springs. Appl. Environ. Microbiol. 2013, 79, 3906–3916.
[CrossRef] [PubMed]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).