Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Experimental Procedures
Site description and sample collection
Soils for this study were collected from six drained thaw lake basins (DTLB) in the Arctic
Coastal Plain near Barrow, AK. The locations were centered around 71.24°N 156.48°W. The
sites included three DTLB each of old (300-2000 y) and ancient (2000-5500 y) age classes
(Hinkel et al., 2003, Hinkel et al., 2005). These soils are classified as Histoturbels, with organic
layers of 20 cm or greater overlying silt-rich mineral materials (Bockheim et al., 2001). At each
DTLB, two soil cores were collected from a single low centered polygon: one from the rim and
one from the center. For four of the six DTLB, samples were collected using a gas powered drill
and 7.5 cm diameter SIPRE core to a depth of 40 cm. These cores were supplemented with
smaller cores collected from two additional DTLB using an electric drill and a masonry hole bit
(3 cm diameter, 24 cm deep) (Lipson et al., 2013). The soil cores were collected in June 2011,
when the majority of the profile was still frozen. Soils were immediately returned to the
laboratory and stored at -40°C. The soil cores were sectioned into horizons with a band saw (610 cm, 16-20 cm, 26-30 cm and 36-40 cm), which were then cut longitudinally to produce two
halves: one for DNA extraction and the other for soil chemistry measurements. For DNA
extraction, the outer layers of each soil sample were stripped away with a sterile knife to obtain
uncompromised subsamples.
DNA extraction and sequencing
Sample processing, sequencing and core amplicon data analysis were performed by the Earth
Microbiome Project (www.earthmicrobiome.org) (Gilbert et al., 2010), using the protocols
available at http://www.earthmicrobiome.org/emp-standard-protocols/. DNA extraction was
based on the MoBio PowerSoil-htp 96-well extraction method (MoBio, Carlsbad CA, USA).
Amplification of 16S rRNA gene fragments was performed using bacteria/archaeal primers
515F/806R and sequenced using the Illumina HiSeq platform (Caporaso et al., 2012). One of the
36 original samples was lost in the process (a 36-40 cm sample from a polygon center). The
QIIME pipeline was used to quality-filter and demultiplex the DNA sequences. UCLUST was
used to pick operational taxonomic units (OTUs). For comparison, the UPARSE pipeline (Edgar,
2013) was also used to pick OTUs. This was followed by chimera-removal using the referencebased UCHIME algorithm with the RDP Gold database. UCLUST (Edgar, 2010) was used to
assign taxonomy. UCLUST has been reported to overestimate OTU relative to other methods
(Schmidt et al., 2014), but this was not the case in our study, possibly because of the closed
reference UCLUST method we used. Chao1 diversity estimates were correlated at R2=0.911
between the two methods, and assignments at the level of phyla (or class in the case of
Proteobacteria) were correlated with R2 ranging from 0.95 to 0.99. QIIME was used to build a
phylogenetic tree and calculate UniFrac distance to characterize diversity between samples
(Caporaso et al., 2010). All amplicon and metadata is available through the European Nucleotide
Archive (ENA) of the European Bioinformatics Institute (EBI) (www.ebi.ac.uk/ena, study ID’s:
ERP010098 and PRJEB9043). Sequence data is also available through Figshare
(http://dx.doi.org/10.6084/m9.figshare.1309249,
http://dx.doi.org/10.6084/m9.figshare.1309248).
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Soil chemistry measurements
Subsamples were used for determination of gravimetric water content, organic matter content (by
loss on combustion at 500°C), and pH (in water saturation paste). Initially frozen soil samples
(~1 g) were extracted in reverse osmosis-treated (18.2 M) water (10 mL) for one hour on ice,
and in 1M HCl (10 mL) at room temperature overnight. Supernatants of water extracts were
analyzed for total dissolved organic C with a colorimetric method (Lipson et al., 2010) and for
organic and inorganic anions by ion chromatography, using a Thermo Dionex ICS 5000+.
Oxidized and reduced forms of Fe in acid extracts were analyzed using 1,10-phenanthroline
(Lipson et al., 2010). Bulk density was estimated from samples using the previously determined
relationship between organic matter and bulk density (Lipson et al., 2013). These calculated bulk
densities were used to convert values to a volumetric basis so less dense shallow layers could be
appropriately compared with denser deeper soil layers. These measurements were only carried
out on the 24 larger SIPRE cores, as little remained of the smaller masonry bit cores after DNA
extraction, and the original depths of the remnants could no longer be resolved.
Statistical analysis
Microbial community differences were not seen between old and ancient DTLB, and so data
were pooled from the six DTLB and analyses focused instead on the effects of microtopography
and depth. Rarefaction analysis of the OTU data was done with EstimateS (Colwell et al., 2012).
PCA using the weighted UniFrac metric (Lozupone et al., 2006) was generated by QIIME
(Caporaso et al., 2010). Multiple regression analysis was used to relate soil chemistry variables
to individual PC. Multivariate analysis of variance (MANOVA) was used to simultaneously
relate all three PC to soil chemistry variables.
To test changes in the distribution of microbial taxa or functional groups we used analysis of
covariance (ANCOVA) with topography (polygon rim vs. center) as a categorical variable and
depth as a continuous variable. To test for non-linear effects of depth, a two-way analysis of
variance (ANOVA) used with both depth and topography coded as categorical variables.
Correlation and regression analysis were used to explore relationships among variables. In these
analyses the sequence data was expressed as the proportion of each OTU out of the total number
of sequences for each soil sample. Variables were log-transformed when necessary to fit the
assumptions of the model. The software package, R (3.1.1) was used for all statistical analysis
(www.R-project.org).
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Supplemental Tables
Table SI-1.Total (µg C cm-3) and individual (nmole cm-3) organic acids (means and standard errors)
in water extracts of soils from varying depths of rims or centers of low-centered polygons.
Depth
Organic
(cm)
Acids
Rims
6-10
6.24±0.73
16-20
6.57±0.87
26-30
5.64a±0.84
36-40
3.42±0.31
Centers 6-10
8.94±1.95
16-20
9.11±2.66
26-30
6.46±1.03
36-40
6.73±1.80
ANCOVA Depth
0.04
p-values Topo
0.024
Int.
ns
a-one outlier omitted from this group
Formate
Oxalate
Acetate
Citrate
Lactate
77.2±67.4
237±43
204±24
138±14
75.3±18.2
123±23
261±57
314±114
0.014
ns
ns
94.5±43.3
84.1±27.3
59.5±5.2
43.6±13.7
51.0±8.9
51.4±10.2
46.3±4.5
64.7±23.2
0.051
0.044
0.093
58.7±24.4
32.5±20.0
56.5±30.7
18.2±10.8
121.1±43.2
84.1±13.8
78.5±10.0
41.1±15.7
0.023
0.019
Ns
18.0±4.0
6.0±2.5
0.6±0.4
0.0±0.0
27.5±6.4
52.2±33.1
1.0±0.6
0.0±0.0
0.031
ns
ns
9.4±3.0
13.7±2.7
10.0a±1.4
7.7±2.9
53.6±15.4
17.2±4.9
7.2±2.5
11.7±5.4
ns
0.003
0.016
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Table SI-2. Sequencing effort for soil samples of different depths (in cm) and topography (high
and low are rims and centers, respectively, of low-centered polygons). OTU were determined by
two alternative clustering alogorithms, UCLUST and UPARSE.
#samples
mean(seqs/sample)
sum(seqs)
18
63124
6-10
6
16-20
26-30
High
36-40
1136227
OTU observed
(UCLUST)
4813
OTU observed
(UPARSE)
7706
72442
434652
3782
6257
6
65507
393044
3158
5592
3
52123
156369
1635
2809
3
50721
152162
1419
2359
17
52453
891693
3570
5912
6-10
6
54520
327122
2714
4750
16-20
6
56270
337619
1957
3895
26-30
3
47023
141070
1496
2873
36-40
2
42941
85882
920
1796
35
57941
2027920
5595
8776
Low
Grand Total
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Table SI-3. Means (and standard errors) of sequences per sample, unique OTU per sample and
the Chao1 estimate of OTU richness for topographically high and low areas and different depths,
using two different clustering methods to assign sequences to OTU. ANOVA results: Chao1 by
UCLUST: high vs. low (P=0.001), depth (P<0.001); Chao1 by UPARSE: high vs. low
(P=0.038), depth (P<0.001).
high
6cm
16cm
Seqs/sample
63124 (2857)
72442 (2685)
65507 (4934)
UCLUST
OTU/sample
1269 (140)
1842 (154)
1209 (213)
Seqs/sample
81381 (3276)
86810 (4305)
85706 (6391)
UPARSE
OTU/sample
2214 (231)
3068 (269)
2198 (363)
Chao1
1708 (163)
2388 (174)
1643 (241)
Chao1
2608 (260)
3597 (314)
2582 (386)
26cm
36cm
low
6cm
16cm
26cm
52123 (3763)
50721 (3797)
52453 (2798)
54520 (6418)
56270 (2041)
47023 (4593)
852 (110)
664 (185)
861 (85)
1091 (178)
772 (73)
766 (215)
1221 (178)
968 (169)
1175 (103)
1487 (213)
1040 (83)
1041 (235)
73136 (8841)
70119 (5175)
69286 (3199)
67415 (6978)
76077 (2258)
64466 (7059)
1586 (308)
1163 (350)
1762 (169)
2071 (366)
1722 (204)
1596 (412)
1866 (340)
1421 (365)
2111 (191)
2532 (402)
2034 (217)
1900 (445)
36cm
42941 (11387)
588 (121)
844 (91)
61754 (13973)
1202 (307)
1395 (362)
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Table SI-4. Percentage of major families (Burkholderiaceae, Comamonadaceae, Gallionellaceae,
Methylophilaceae, Rhodocyclaceae) within the Betaproteobacteria by topography and depth (Int = depth
x topography interaction).
Rim
Center
ANCOVA
p-values
Depth
(cm)
6-10
16-20
26-30
36-40
6-10
16-20
Burkholder
Comamonad
Gallionella
Methylophil
Rhodocycl
0.59±0.13
0.26±0.12
0.16±0.11
0.08±0.04
0.67±0.15
0.03±0.02
0.41±0.10
3.77±1.34
3.92±0.54
3.00±0.25
1.50±0.44
3.91±1.02
2.60±1.30
2.99±1.51
0.50±0.23
0.62±0.55
0.79±0.34
0.51±0.17
0.14±0.05
0.36±0.10
0.38±0.10
0.54±0.16
0.25±0.09
0.23±0.05
0.49±0.32
0.55±0.06
0.63±0.17
0.74±0.34
0.61±0.19
0.43±0.06
26-30
36-40
Depth
Topo
Int
0.09±0.08
0.13±0.01
0.0003
ns
ns
7.70±3.40
3.65±0.23
0.013
ns
ns
0.19±0.09
0.11±0.04
ns
0.038
ns
0.24±0.07
0.33±0.08
0.003
ns
0.074
0.85±0.43
0.38±0.08
ns
ns
ns
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Table SI-5. MANOVA of the three principle components describing variation in microbial
community (weighted UniFrac) on combinations of soil variables. Pillai= Pillai’s trace, dfe =
degrees of freedom (error), Fe2/tot = reduced Fe over total Fe in acid extracts (a measure of
redox), DOC = dissolved organic carbon, OA = organic acids, OM = organic matter.
Model
Pillai
approx. F
dfe
P
Fe2/tot
0.615
9.56
18
0.0005
DOC
0.289
2.44
18
0.0980
Fe2/tot
0.635
10.43
18
0.0003
OA
0.365
3.45
18
0.0386
Fe2/tot
0.648
10.41
17
0.0004
OA
0.369
3.32
17
0.0449
DOC
0.121
0.78
17
0.5207
Fe2/tot
0.633
9.78
17
0.0006
Lactate
0.524
6.24
17
0.0047
Fe2/tot
0.632
10.30
18
0.0004
Acetate
0.338
3.07
18
0.0542
Fe2/tot
0.647
9.77
16
0.0007
Lactate
0.572
7.13
16
0.0030
Acetate
0.192
1.27
16
0.3197
Fe2/tot
0.643
10.79
18
0.0003
pH
0.148
1.04
18
0.3972
Fe2/tot
0.627
10.09
18
0.0004
OM
0.080
0.52
18
0.6730
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Table SI-6. Correlation coefficients between taxa abundance and soil chemical parameters. Significance is color-coded: pink text,
white background, P<0.1; Pink background, P<0.05; yellow, P<0.01; green, P<.001 (OM = organic matter, DOC = dissolved organic
C, OA = organic acids, PO4 = phosphate, Cl- = chloride, Br- = bromide, FeR = Fe reducers, Methano = methanogens, Methylo =
methylotrophs).
Acidobacteria
Actinobacteria
Bacteroidetes
Caldiserica
Chlorobi
Chloroflexi
Crenarchaeota
Cyanobacteria
Elusimicrobia
Firmicutes
Gemmatimonadetes
Nitrospirae
OP10
Planctomycetes
Alphaproteobacteria
Betaproteobacteria
Deltaproteobacteria
Gammaproteobacteria
Spirochaetes
Verrucomicrobia
FeR
Methano
Methylo
Burkholderiaceae
Comamonadaceae
Gallionellaceae
Hydrogenophilaceae
Methylophilaceae
Oxalobacteraceae
Rhodocyclaceae
OM
0.391
-0.227
-0.161
-0.404
0.068
0.215
0.142
-0.130
0.411
-0.499
0.027
0.320
-0.216
0.097
0.426
-0.138
0.181
0.275
-0.315
0.066
0.063
-0.079
0.230
0.381
-0.354
-0.097
-0.565
-0.291
-0.004
-0.333
pH
-0.531
0.152
0.368
0.562
0.007
0.009
-0.057
0.143
-0.358
0.729
-0.322
-0.039
0.139
-0.250
-0.351
-0.164
-0.132
-0.384
-0.026
-0.475
0.101
0.309
-0.025
-0.200
0.182
-0.233
0.384
0.326
-0.268
0.336
Fe2/tot
-0.726
0.549
0.506
0.611
0.032
0.230
0.145
0.033
-0.521
0.675
-0.321
-0.030
0.318
-0.519
-0.635
-0.282
-0.053
-0.559
0.153
-0.405
-0.156
0.224
-0.187
-0.701
0.446
-0.327
0.358
0.603
-0.491
0.321
DOC
0.324
-0.343
-0.154
-0.354
0.298
0.080
0.269
-0.056
0.216
-0.478
0.080
0.018
-0.275
0.055
0.148
0.304
0.389
0.037
0.061
0.141
0.290
0.202
0.389
0.345
-0.109
0.138
-0.374
-0.292
0.104
-0.006
OA
0.107
-0.257
-0.018
-0.211
0.328
0.108
0.431
-0.143
0.116
-0.285
-0.153
-0.097
0.000
-0.154
0.024
0.180
0.370
-0.007
0.387
0.100
0.137
0.280
0.214
0.243
0.076
-0.072
-0.265
-0.308
0.062
0.017
Sulfate
-0.098
-0.072
-0.064
0.139
-0.032
-0.207
-0.131
0.169
-0.149
0.219
-0.208
-0.351
0.340
0.163
0.207
-0.100
-0.138
0.235
0.269
-0.135
-0.122
0.262
0.042
0.423
-0.120
-0.134
0.139
-0.228
0.409
0.111
PO4
0.087
-0.058
-0.207
0.149
-0.301
-0.506
-0.508
0.038
0.184
0.182
0.203
-0.274
-0.018
0.319
0.253
0.006
-0.344
0.249
-0.014
0.051
-0.266
-0.029
-0.230
0.150
-0.170
0.073
0.296
-0.116
0.285
-0.026
Oxalate
0.252
-0.090
-0.160
-0.013
-0.233
-0.245
-0.294
0.048
0.480
-0.177
0.083
-0.127
-0.053
0.281
0.366
-0.192
-0.155
0.336
-0.078
0.089
-0.385
-0.125
-0.105
0.151
-0.246
-0.136
-0.213
-0.157
0.156
-0.374
Nitrate
0.084
0.129
-0.252
-0.231
-0.199
-0.006
-0.289
-0.302
0.133
-0.086
0.280
-0.022
-0.019
0.126
0.244
-0.025
-0.166
0.451
-0.108
0.086
-0.331
-0.165
-0.161
0.024
-0.153
0.071
0.026
-0.162
0.045
-0.233
Lactate
0.102
-0.141
-0.070
-0.285
0.441
0.372
0.388
0.027
0.057
-0.301
-0.143
0.289
-0.276
-0.152
0.181
0.124
0.382
0.067
0.007
-0.268
0.567
0.513
0.739
0.587
-0.215
-0.031
-0.339
-0.023
0.057
0.039
Formate
-0.361
0.171
0.383
0.472
0.003
-0.005
0.276
0.083
-0.313
0.185
-0.205
-0.355
0.486
-0.260
-0.463
-0.090
0.144
-0.378
0.670
-0.102
-0.170
0.088
-0.151
-0.432
0.246
-0.063
0.176
0.247
-0.049
0.210
Citrate
0.245
-0.396
-0.110
-0.360
0.356
0.109
0.356
-0.153
0.199
-0.358
-0.024
0.153
-0.220
-0.059
0.073
0.213
0.350
-0.008
0.056
0.345
0.138
-0.013
0.146
0.166
0.057
0.004
-0.350
-0.281
0.040
-0.064
Cl-0.045
0.154
0.045
-0.234
0.051
0.236
0.150
-0.295
-0.252
-0.115
-0.017
-0.278
0.104
-0.151
-0.283
0.038
0.175
-0.046
0.224
0.256
-0.119
-0.127
-0.221
-0.318
0.175
0.141
0.077
-0.070
-0.092
0.008
Br0.264
-0.384
-0.092
-0.367
0.320
0.102
0.341
-0.172
0.172
-0.368
-0.056
0.061
-0.232
-0.057
0.010
0.169
0.382
-0.058
0.079
0.447
0.057
-0.044
0.069
0.068
0.064
0.005
-0.347
-0.241
0.072
-0.065
Acetate
0.123
-0.200
-0.176
-0.303
0.218
-0.017
0.111
-0.151
0.034
-0.071
-0.051
-0.076
-0.142
-0.016
0.184
0.299
0.096
0.098
0.001
-0.070
0.382
0.391
0.185
0.497
0.047
0.026
0.020
-0.481
0.038
0.172
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Supplemental Figures
Fig. SI-1. Image of low centered polygons (dark, wet areas) rimmed by higher, drier areas
pushed up by ice wedges (photo: D. Lipson).
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Fig. SI-2. Relative abundance of major taxa by topography and depth. Values are means and
standard errors. Topography (topo): black bars = polygon rims, white bars = centers.
Significance of ANCOVA: *** = significant at p<0.001, ** = p<0.01, * = p<0.05, ? = P<0.1; int
= topo X depth interaction.
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
Figure SI-3. Relative abundance of functional groups by topography and depth. Values are
means and standard errors. Topography (topo): black bars = polygon rims, white bars = centers.
Significance of ANCOVA: *** = significant at p<0.001, ** = p<0.01, * = p<0.05, ? = P<0.1; int
= topo X depth interaction. See main text for interpretation of function.
Supplemental Information
Lipson et al. (2105) Environmental Microbiology Reports
References Cited
Bockheim JG, Hinkel KM, Nelson FE (2001) Soils of the Barrow region, Alaska. Polar Geography, 25, 163181.
Caporaso JG, Kuczynski J, Stombaugh J et al. (2010) QIIME allows analysis of high-throughput community
sequencing data. Nature Methods, 7, 335–336.
Caporaso JG, Lauber CL, Walters WA et al. (2012) Ultra-high-throughput microbial community analysis
on the Illumina HiSeq and MiSeq platforms. ISME J, 6, 1621-1624.
Colwell RK, Chao A, Gotelli NJ, Lin S-Y, Mao CX, Chazdon RL, Longino JT (2012) Models and estimators
linking individual-based and sample-based rarefaction, extrapolation, and comparison of
assemblages. Journal of Plant Ecology, 5, 3-21.
Edgar RC (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics, 26, 24602461.
Edgar RC (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nature
Methods, 10, 996-998.
Gilbert J, Meyer F, Antonopoulos DA et al. (2010) Meeting Report: The Terabase Metagenomics
Workshop and the Vision of an Earth Microbiome Project.
Hinkel K, Eisner W, Bockheim J, Al. E (2003) Spatial extent, age, and carbon stocks in drained thaw lake
basins on the Barrow Peninsula, Alaska. Arctic Antarctic Alp Res, 35, 291-300.
Hinkel K, Frohn R, Nelson F, Eisner W, Beck R (2005) Morphometric and spatial analysis of thaw lakes
and drained thaw lake basins in the western Arctic Coastal Plain, Alaska. Permafrost and
Periglacial Processes, 16, 327-341.
Lipson DA, Jha M, Raab TK, Oechel WC (2010) Reduction of iron (III) and humic substances plays a major
role in anaerobic respiration in an Arctic peat soil. J Geophys Res-Biogeosci (G), 115, G00I06,
doi:10.1029/2009JG001147.
Lipson DA, Raab TK, Goria D, Zlamal J (2013) The contribution of Fe(III) and humic acid reduction to
ecosystem respiration in drained thaw lake basins of the Arctic Coastal Plain. Global
Biogeochemical Cycles, 27, 1–11.
Lozupone C, Hamady M, Knight R (2006) UniFrac - An online tool for comparing microbial community
diversity in a phylogenetic context BMC Bioinformatics, 7, 371.
Schmidt TS, Matias Rodrigues JF, Mering C (2014) Limits to robustness and reproducibility in the
demarcation of operational taxonomic units. Environmental Microbiology.