1 Supplementary information

2

3 Phylogenetic clustering of small low nucleic acid-content bacteria

4 across diverse freshwater ecosystems

5 Caitlin R. Proctor1,2, Michael D. Besmer1, Timon Langenegger1,#, Karin Beck3, Jean-Claude

6 Walser4, Martin Ackermann1,2, Helmut Bürgmann3,*,#, Frederik Hammes1,*,#

7

8

9

10 Contents:

11

12 1. Method Details

13 2. Diversity of total cell concentrations and %LNA content bacteria between and different

14 freshwater ecosystems

15 3. Flow cytometry comparison measurements with different staining methods and flow

16 cytometer models

17 4. Filterability of LNA and HNA content bacteria

18 5. Agreement with qPCR measurement of bacteria

19 6. Relative importance of various factors explaining community variations within each

20 ecosystem.

21 7. Time-series data in a river indicate consistent LNA content and communities within each

22 faction.

23 8. SEM images of possible cross-contamination on filters

24 9. Detailed phylogenetic trees and relative abundances

25 10. Original SEM images

26

1 27 1. Method Details

28 Table S1: Site information

Site - all located in Switzerland Number of Samples Ground water Source 1 (Niederdorf) 1 Source 2 (Oberdorf 1) 1 Source 3 (Oberdorf 2) 1 River Water [indication in Figure 5] Chriesbach (Dübendorf) [A] 15 Rohrbach (Fällanden) [B] 1 Schachenbach (Bonstetten) [C] 1 Sihl (Zurich) [D] 1 Lake water Alpnachersee 12 Rotsee 1 Hallwilersee 1 Baldeggersee 1 Sempachersee 1 Melchsee 1 Sarnersee 1 Tap water Bonstetten 1 Dübendorf 1 Wollishofen 1 Fällanden 1 Wastewater (secondary effluent) WWTP 1 (Dübendorf) 2 WWTP 2 (Fällanden) 1 Pilot Scale Treatment Plant (Dübendorf) 2 WWTP 3 (Niederglatt) 1 29

30 Table S2: Filtration Volumes

Ground- River Lake Tap Waste- 1 2 water water water1 water water

Number of sampling sites 4 5 (+11) 9 (+11) 4 3

Sampled volume (mL) 25’000 500 500 5’000 500

Volume per filter (mL) 5’000 100 100 1’000 100

1 One river and lake each were sampled 12 times to assess temporal dynamics 2 Secondary effluent from a wastewater treatment plant

31

2 32

33 Table S3: Details for Illumina sequencing

Primer Sequence Reference Bakt_341F A. Klindworth, E.

CCTACGGGNGGCWGCAG Pruesse, T. Schweer, (S-D-Bact-0341-b-S-17) J. Peplies, C. Quast, M. Horn and F. O. Bakt_805R Glöckner, Nucleic GACTACHVGGGTATCTAATCC (S-D-Bact-0785-a-A-21) Acids Res., 2013, 41, e1. Nextera adapter tail before TCG-TCG-GCA-GCG-TCA-GAT- forward GTGTAT-AAG-AGA-CAG-GA

Amplicon PCR Primers PCR Amplicon Nextera adapter tail before GTC-TCG-TGG-GCT-CGG-AGA- reverse TGTGTA-TAA-GAG-ACA-GAG

Cycling Kti/Mix and Reaction Assay Holding Cycling Template/Notes Reps Chemistry

2 µL DNA template 95 °C 95 °C 0:30 1U KAPA 2G robust (0.8-50 ng) HotStart Polymerase (KAPA Biosystems, Two sets of frame- Boston, USA), Amplicon 54 °C 0:30 shifted primer sets 19 X 1 x reaction buffer B, PCR were used on each and 0.4 µM of each replicate extraction per

primer in a final volume sample: Sets 0 and 2 5:00 72 °C 0:30 of 25 µL. Sensoquest for replicate A and Labcycler Basic used. sets 1 and 3 for replicate B

PCR Details PCR 95 °C 95 °C 0:30 1 X KAPA HiFi HotStart Ready Mix and 5 µl of Index each of the respective Pooled amplicon PCR 10 X 55 °C 0:30 PCR Nextera index primers in product a total reaction volume of 50 µl 3:00 72 °C 0:30

Step System Protocol

Purification of Amplicon PCR product

Agencort AMPure beads XP sytem (Beckman Supplier's protocol Purification of Index PCR product Coulter)

Quality Control of Index PCR product Agilent Bioanalyzer Supplier's protocol

Additional Step Additional KAPA library Quantification of Index PCR product Supplier's protocol quantification kit 34

3 35

36

37

38

39 Table S4: Details for Bioinformatics

Algorithm/ Step Parameters Citation Version

Quality Control FastQC v.0.11.2

minimum overlap: 15

maximum overlap: 250 (Magoc and Merge Reads FLASH v1.2.9 Salzberg, 2011) max mismatch density: 0.25

Trim adaptor sequences and error rate: 0 Cutadapt v1.5 (Martin, 2011) sort frame full-length shifts

size range: 390-440 bp

minimum mean quality score: 25 (Schmieder and

Bioinformatics Details Bioinformatics PRINSEQ-lite Quality Filtering Edwards, v0.20.4 max 1 ambiguous nucleotides 2011). GC range: 30-70 low complexity filter dust/25

identity cutoff: 97%

usearch OTU clustering abundance sorting: 2 (Edgar, 2010) v7.0.1090

chimera filtering

40

41

42

43

4 44 2. Diversity of total cell concentrations and %LNA content bacteria 45 between and different freshwater ecosystems

46 47 Figure S1: Overview of investigated freshwater ecosystems with respect to flow cytometry total cell 48 concentration and percentage LNA content bacteria. Clusters are only for visualization purposes. 49 Note that the x-axis is on logarithmic scale.

5 50

51 Figure S2: Staining and Typical flow cytometric density from an array of samples from four 52 ecosystems (Lake water, River water, Tap water, and Wastewater) stained with SYBR Green I. 53 Dotted black lines indicate electronic gates separating bacteria from background. Red gates/dotted 54 lines indicate electronic gates separating LNA and HNA content bacteria. 6 55 3. Flow cytometry comparison measurements with different staining 56 methods and flow cytometer models 57

58

59 Figure S3: Comparison of SG staining measurements of TCC measured on two different flow 60 cytometer models with two operators. Panels A and B are from the same river water sample, and 61 panels C and D are from the same wastewater sample. Measurements for panels B and D were 7 62 performed with a BD Accurri with measurement as described (Methods). Measurements for panels 63 A and C were performed as described below. Both operators and instruments identified a clear 64 separation between the two main clusters, although more sophisticated optics revealed more 65 subpopulations. The LNA cluster was similarly distinguished by low values for mean fluorescence. 66 67 Measurements for panels A and C were performed with a BD Influx v7 Sorter USB, (Becton, 68 Dickinson and Company, Franklin Lakes, NJ, USA) equipped with a 488 nm Sapphire OPS laser 69 (400 mW) and a 355 nm Genesis OPS laser (100 mW, both Coherent, Santa Clara, CA, USA). 70 The 488 nm laser light was used for the analysis of the forward scatter (FSC, 488/10) and the side 71 scatter (SSC, 488/10, trigger signal). The SybrGreenI induced fluorescences were collected in the 72 FL1 channel (616/23, green fluorescence) and the FL3 channel, respectively (670/30, red 73 fluorescence). The DAPI-fluorescence was detected in the FL9 channel (460/50). The fluidic 74 system was run at 33 psi using a 70 µm nozzle. The sheath fluid consisted of 0.5 x FACSFlow 75 buffer (BD). For optical calibration of the cytometer in the linear range, 1 µm blue fluorescent 76 FluoSpheres (Molecular Probes, F8815, Eugene, OR, USA) and 2 µm yellow-green fluorescent 77 FluoSpheres (ThermoFisher Scientific, F8827, Waltham, MA, USA) were used. For calibration in 78 the log range, 0.5 µm UV Fluoresbrite Microspheres (Polysciences, 18339, Warrington, PA, USA) 79 were applied. Data were collected in log mode.

8 80

81 Figure S4: Comparative analysis with different dyes and instrumentation. River water and 82 wastewater were stained with SYBR Green I (A, B) and DAPI (C, D) and analyzed on a BD Influx 83 v7 Sorter USB, (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Bacteria fall within 84 the orange gate. The SYBR Green I data show high similarity with the SYBR Green data in Figure 85 2 and Figure S2. Moreover, the data show clear separation of the LNA cluster on the Forward 86 Scatter, suggesting a clear size separation between HNA and LNA cells. While the LNA cluster 87 was clearly discernable on the DAPI stained samples, it is interesting to note that the fluorescence 88 separation with this dye is much less obvious compared to SYBR Green, and in contrast more 89 HNA clusters are observed. Both stains (DAPI and SYBRGreen) similarly identified the LNA cluster 90 by low values of forward scatter, similar to the identification by low fluorescence (Figure S3). The 91 DAPI staining revealed more sub-communities for the HNA cluster. Although forward scatter is not 92 directly related to cell since in the range of 0.5 to 5 µm (Shapiro, 2003), the trend points to similar 93 low estimates for LNA. 94 95 The flow cytometer was with a 488 nm Sapphire OPS laser (400 mW) and a 355 nm 96 Genesis OPS laser (100 mW, both Coherent, Santa Clara, CA, USA). The 488 nm laser light was 97 used for the analysis of the forward scatter (FSC, 488/10) and the side scatter (SSC, 488/10, 98 trigger signal). The SybrGreenI induced fluorescences were collected in the FL1 channel (616/23, 9 99 green fluorescence) and the FL3 channel, respectively (670/30, red fluorescence). The DAPI- 100 fluorescence was detected in the FL9 channel (460/50). The fluidic system was run at 33 psi using 101 a 70 µm nozzle. The sheath fluid consisted of 0.5 x FACSFlow buffer (BD). For optical calibration 102 of the cytometer in the linear range, 1 µm blue fluorescent FluoSpheres (Molecular Probes, F8815, 103 Eugene, OR, USA) and 2 µm yellow-green fluorescent FluoSpheres (ThermoFisher Scientific, 104 F8827, Waltham, MA, USA) were used. For calibration in the log range, 0.5 µm UV Fluoresbrite 105 Microspheres (Polysciences, 18339, Warrington, PA, USA) were applied. Whereas SybrGreenI 106 stained samples were spiked with 0.5 µm UV Fluoresbrite Microspheres of known concentration in 107 order to calculate the absolute number of cells, 1 µm yellow-green fluorescent FluoSpheres 108 (ThermoFisher Scientific, F13081, Waltham, MA, USA) were added to DAPI stained cells. Data 109 were collected in log mode. 110

10 111 4. Filterability of LNA and HNA content bacteria

112

113 114

115 Figure S5: Example of the filterability of LNA and HNA content bacteria. River water was analyzed 116 before (A,C) and after (B,D) filtration with 0.4 µm pore size filters. All samples were stained with 117 SYBR Green I. Dotted red lines indicate electronic gates separating bacteria from background and 118 LNA from HNA content bacteria. FL1-A = green fluorescence intensity; FL3-A = red fluorescence 119 intensity. 120

11 121 5. Agreement with qPCR measurement of bacteria

122

123 Figure S6: Correlation between flow cytometry (FCM) and qPCR data. Total cell concentration 124 determined by FCM with Sybr Green1 staining. qPCR is a measure of 16S rRNA gene copies (see 125 methods below). Symbol and color determined by one of three size groups determined by filtration 126 method, including: ‘All bacteria’ (gray triangles), the total community, directly filtered onto a 0.2 µm 127 filter; Large bacteria (red circles), the HNA-dominated community filtered onto a 0.4 µm filter; and 128 Small bacteria (blue squares), the LNA-dominated filtration of 0.4 µm filtrate onto a 0.2 µm filter. 129 Linear models calculated for each group had distinct slopes, suggesting a difference in the 130 average number of gene copies in each size of cell. With intercept set to 0, slopes for each group 131 were as follows - small: 1.2 gc/cell, large: 4.4 gc/cell, all: 2.4 gc/cell, with R2of 0.34, 0.23, and 0.26 132 respectively. 133 134 Quantitatve polyermase chain reaction (qPCR) for 16S rRNA gene copies were done as 135 described previously (Proctor et al., 2016). Briefly, DNA extracts were diluted either 10 or 50 times 136 prior to qPCR depending on the DNA concentration of the extract. qPCR quantifications were 137 performed on a Roche LightCylcer II with the following temperature program: 10 min at 95°C, 138 followed by 45 cycles of 40 sec at 95°C and 53°C, and extension 1 min at 72°C. Reactions were 139 performed with Light Cycler 480 probes master (Roche), Bact349F/Bact806R or Arch 140 349F/Arch806R primers and probe Bac516F23 (for bacteria) (Takai and Horikoshi, 2000) (Details, 141 below). 16S rRNA gene copy numbers were quantified against standards created from dilutions of

12 142 plasmids with a matching insert. Concentrations were calculated by normalizing over the filtered 143 sample volume. Data were assembled and prepared for further analysis in R using Excel. 144

Assay Primer Sequence Reference Bact 349F AGGCAGCAGTDRGGAAT K. Takai and K. Bact 806R GGACTACYVGGGTATCTAAT Horikoshi, Appl. 16S qPCR FAM-TGCCAGCAGCCGCG Bac516F_FAM Environ. GTAATACRDAG-TAMRA Microbiol., 2000, Arch349F GYGCASCAGKCGMGAAW 66, 5066–72 Arch806R GGACTACVSGGGTATCTAAT 145

146 Proctor CR, Gächter M, Kötzsch S, Rölli F, Sigrist R, Walser J-C, et al. (2016). Biofilms in shower 147 hoses – choice of pipe material influences bacterial growth and communities. Environ Sci 148 Water Res Technol 2: 670–682. 149 Takai K, Horikoshi K. (2000). Rapid detection and quantification of members of the archaeal 150 community by quantitative PCR using fluorogenic probes. Appl Environ Microbiol 66: 5066–72.

13 151 6. Relative importance of various factors explaining community 152 variations within each ecosystem. 153 154 Table S5: Relative importance of various factors in bacterial communities within each ecosystem, 155 calculated by Adonis. For each explanatory variable (Sampling site, size, and interactions between 156 these two factors), the percent of variation explained by the variable and the p-value for statistical 157 significance of the factor are expressed. Sampling site refers to the specific location for each 158 sample (i.e. River A vs River B). Size refers to one of three groups of bacteria determined by 159 filtration method, including: ‘All bacteria’ is the total community, directly filtered onto a 0.2 µm filter. 160 Large bacteria is the HNA-dominated community collected on a 0.4 µm filter, and Small bacteria is 161 the LNA-dominated community in the 0.4 µm filtrate, collected on a 0.2 µm filter. 162 Explanatory factor for Adonis analysis Number in Portion explained by factor (p-value) analysis Sampling (Filters) - Samples Ecosystem Size Site*size Site

Lake water 0.31 (0.001) 0.27 (0.001) 0.07 (1.00) (48) - 16

River water 0.17 (0.001) 0.30 (0.001) 0.07 (0.459) (54) - 18

Groundwater 0.50 (1.00) 0.22 (1.00) 0.28 (1.00) (12) - 4

Wastewater 0.42 (0.001) 0.25 (0.001) 0.14 (0.858) (18) - 6

Tap water 0.57 (1.00) 0.25 (1.00) 0.18 (1.00) (12) - 4

163

14 164 7. Time-series data in a river indicate consistent LNA content and 165 communities within each faction. 166 167 Extended discussion on temporal variation: As a test for temporal stability within a sampling site, 168 samples taken over four months (June - September) from River Site A were further analyzed. Both 169 community composition by size (as measured with 16S amplicon sequencing) and the percentage 170 of LNA-content bacteria (as measured with FCM) remained relatively stable in this period. 171 172 While the community composition shifted moderately through time (Figure S7), it still 173 remained distinct from other sampling sites (Site A, Figure 5). Moreover, the vectors of temporal 174 change in the ordination space are similar across bacterial groups (Figure S7), indicating that the 175 groups responded similarly to environmental conditions. Looking at FCM data, the percentage of 176 LNA-content bacteria remained stable between approximately 50% and 60% with no clear 177 temporal shift (Figure S8). A more detailed time-series dataset from the same sampling site did 178 show dramatic short-term dynamics linked to rainfall, but these varied around a stable baseline 179 (Besmer et al., 2014). Together with the comparable stability of the time-series on lakes (data not 180 shown), this temporal stability indicates that while we often relied on single samples from a site, 181 these were likely representative in terms of their size groups, and our data was not strongly 182 influenced by temporal variations. 183

184

185 Figure S7: Non-metric multidimensional scaling (NMDS) of bacterial communities (characterized 186 with 16S amplicon sequencing) calculated with Bray-Curtis dissimilarity in a river (River Site A in 187 Figure 5) over four months (June, July, August, September) in the three size-based groups based 188 on filter pore-size: ‘All bacteria’ is the total community, directly filtered onto a 0.2 µm filter. Large 189 bacteria is the HNA-dominated community collected on a 0.4 µm filter, and Small bacteria is the 190 LNA-dominated community in the 0.4 µm filtrate, collected on a 0.2 µm filter. 191 15 192 193 Figure S8: Flow cytometry-based comparison of temporal evolution of the percentages of LNA 194 (blue bars) and HNA (red bars) content bacteria as well as total cell concentration (grey circles) in 195 river water over different time scales: (A) time series of river water sampled 11 times at the same 196 sampling site (River water - site A, Figure 5, Figure S7) over 4 months with low frequency and (B) 197 time series of river water sampled approximately 1'400 times at the same sampling site (River 198 water - site A) over 2 weeks with high frequency and influenced by two precipitation events (panel 199 B is adapted from Besmer et al. (2014)). The percentage of LNA content bacteria remained stable 200 between approximately 45% and 55% with no clear temporal shift. 201 202 Besmer MD, Weissbrodt DG, Kratochvil BE, Sigrist JA, Weyland MS, Hammes F. (2014). The 203 feasibility of automated online flow cytometry for in-situ monitoring of microbial dynamics in 204 aquatic ecosystems. Front Microbiol 5. e-pub ahead of print, doi: 10.3389/fmicb.2014.00265.

16 205 8. SEM images of possible cross-contamination on filters

A

B

206

207 Figure S9: Scanning electron microscopy (SEM) image of bacteria from a river (Site A), filtered 208 onto a 0.4 µm pore-size filter. Filter pores are visible as black holes, bacteria are highlighted in 209 blue/purple/orange/pink shades and extracellular filaments are highlighted in green (all colors were 210 artificially added). Note the difference in size between A and B. (A) Several large bacteria and 211 large extra-cellular matrix which visual inspection suggests would not pass through 0.4 µm pores. 17 212 A single small bacterium (blue) is seen entering the pore, but not passing through. (B) A single 213 small bacterium (pink), while of filterable size, is stuck on the filter. Original image without artificial 214 coloring available in Figure S14. 215

18 216 9. Detailed phylogenetic trees and relative abundances 217

218 219 (Figure S10, continued next page) 220

19 221

222 (Figure S10, continued next page) 223 224

20 225 Figure S10: Phylogenetic trees to give more detail about total diversity measured across 226 freshwater ecosystems. Phylogenetic trees were created using FastTree, including maximum 227 likelihood options. (A) Phylogenetic tree used in Figure 6 with phyla marked with color, (B) 228 Phylogenetic tree Figure 6, expanded to include 3,805 “Eliminated” OTUs [no size scaling as in 229 Figure 6]. (C) Figure S10B with phyla marked with color. Phyla are marked only with “old names”. 230 For example, OD1 are now Parcubacteria and OP3 are now Omnitrophia. 21 231

232

Flavobacteriales 80 Bacteroidetes Flavobacteriales 61 Flavobacteriales 84 Flavobacteriales 88 Flavobacteriales Flavobacteriales 92 Flavobacteriales 70 92 79 Flavobacteriales Flavobacteriales Flavobacteriales Flavobacteriales 76 87 Flavobacteriales Flavobacteriales 76 93 Flavobacteriales 87 Flavobacteriales Flavobacteriales Flavobacteriales Flavobacteriales 72 Flavobacteriales Flavobacteriales 93 94 Flavobacteriales Flavobacteriales 86 Flavobacteriales Flavobacteriales Flavobacteriales Flavobacteriales 75 79 Flavobacteriales 92 Flavobacteriales 89 Flavobacteriales 81 Flavobacteriales Flavobacteriales Bacteroidales Flavobacteriales 92 88 Flavobacteriales Flavobacteriales Flavobacteriales 93 59 Flavobacteriales Flavobacteriales 92 Flavobacteriales 80 82 Flavobacteriales 93 91 Flavobacteriales Flavobacteriales Flavobacteriales Flavobacteriales 70 Flavobacteriales 67 Flavobacteriales 87 Flavobacteriales 53 Flavobacteriales Flavobacteriales Flavobacteriales Sample Sphingobacteriales 93 Sphingobacteriales 84 Sphingobacteriales Total_Bipolar 91 Sphingobacteriales 94 Sphingobacteriales Total_Bipolar2 Sphingobacteriales 84 Sphingobacteriales 83 90 Sphingobacteriales Total_DefHNA Sphingobacteriales Sphingobacteriales 94 Total_DefLNA 90 Sphingobacteriales Sphingobacteriales 88 90 83 Sphingobacteriales Total_NoneAbove 87 Sphingobacteriales 92 Sphingobacteriales 75 92 Sphingobacteriales 94 Sphingobacteriales 89 Abundance 94 Sphingobacteriales Sphingobacteriales Sphingobacteriales 0.2 Sphingobacteriales [Saprospirales] 5.0 [Saprospirales] 94 [Saprospirales] 94 [Saprospirales] 92 86 [Saprospirales] [Saprospirales] [Saprospirales] 82 [Saprospirales] 82 93 [Saprospirales] [Saprospirales] [Saprospirales] 87 [Saprospirales] [Saprospirales] [Saprospirales] 93 [Saprospirales] [Saprospirales] 93 [Saprospirales] 85 [Saprospirales] 82 [Saprospirales] 91 [Saprospirales] 63 88 [Saprospirales] [Saprospirales] [Saprospirales] 90 [Saprospirales] Bacteroidales 88 Bacteroidales Bacteroidales 79 Bacteroidales Bacteroidales 55 77 Bacteroidales Bacteroidales 92 Bacteroidales Bacteroidales Cytophagales 88 77 Cytophagales 87 Cytophagales 54 Cytophagales Cytophagales Cytophagales 92 Cytophagales Cytophagales Cytophagales Cytophagales 90 Cytophagales Cytophagales Cytophagales Cytophagales 92 Cytophagales 86 Cytophagales 89 Cytophagales Cytophagales 233

234 (Figure S11, continued next page)

235

22 236 237 (Figure S11, continued next page) 23 Rhodospirillaceae Alphaproteobacteria 87 Rhodospirillaceae 94 Rhodospirillaceae labeled with family Rhodospirillaceae 86 Rhodospirillaceae Rhodospirillaceae (where available) Acetobacteraceae Acetobacteraceae 84 Acetobacteraceae Acetobacteraceae Rhodospirillaceae Rhodospirillaceae 89 Rhodospirillaceae 76 67 Rhodospirillaceae 89 Rhodospirillaceae 84

87 Rhodospirillaceae 90 83 75 94

89 73 Pseudomonadaceae Rhodospirillaceae 59 Rhodospirillaceae

90 Xanthomonadaceae Rhodospirillaceae Sphingomonadaceae 87 Sphingomonadaceae Sphingomonadaceae 8873 92 Sphingomonadaceae 92 Sphingomonadaceae Sphingomonadaceae 85 Sphingomonadaceae 65 Sphingomonadaceae Sphingomonadaceae 88 88 Sphingomonadaceae Sphingomonadaceae Sphingomonadaceae 81 Sphingomonadaceae 90 Sphingomonadaceae 90 Sphingomonadaceae 65 Rhizobiaceae Sample 85 Hyphomicrobiaceae 88 Hyphomicrobiaceae Total_Bipolar2 88 Hyphomicrobiaceae Hyphomicrobiaceae Total_DefHNA 74 Bradyrhizobiaceae Bradyrhizobiaceae Total_DefLNA 87 Beijerinckiaceae

93 76 Total_NoneAbove 76 74 58 Hyphomicrobiaceae 72 94 Hyphomicrobiaceae 64 Abundance 91 Hyphomicrobiaceae Hyphomicrobiaceae Rhizobiaceae 0.2 90 Rhizobiaceae Methylobacteriaceae 5.0 Rhodobacteraceae 83 88 Rhodobacteraceae 87 Rhodobacteraceae Rhodobacteraceae 87 Rhodobacteraceae 94 Rhodobacteraceae 85 Rhodobacteraceae 90 79 Rhodobacteraceae Hyphomonadaceae 91 Hyphomonadaceae 89 Caulobacteraceae 92 93 Caulobacteraceae 88 Caulobacteraceae

86 Hyphomonadaceae 89 84 Rhodospirillaceae 83 67 Rickettsiaceae 93 Rickettsiaceae 84 86 Rickettsiaceae Rickettsiaceae Rickettsiaceae 87 Rickettsiaceae 94 Rickettsiaceae 60 Pelagibacteraceae Rickettsiaceae 76 Helicobacteraceae 85 Campylobacteraceae Rhodospirillaceae 93 Rickettsiaceae 88 Rhodospirillaceae 88 Rhodospirillaceae 78 Rickettsiaceae 85 Thiotrichaceae 90 Rhizobiaceae Rickettsiaceae 90 Cystobacterineae Coxiellaceae 78 94 mitochondria 82 mitochondria Anaplasmataceae 238

239 (Figure S11, continued next page)

24 Deltaproteobacteria Myxococcales 89

76 Myxococcales

93 Myxococcales

89 PB19 Myxococcales 83

80 Myxococcales Desulfarculales 93 Syntrophobacterales Syntrophobacterales Bdellovibrionales 89

54 Bdellovibrionales Bdellovibrionales 60 Bdellovibrionales 79 81 Bdellovibrionales 87 Bdellovibrionales 78 Bdellovibrionales 79 88 Bdellovibrionales Bdellovibrionales FAC87 72

75 FAC87

92 FAC87 Abundance Syntrophobacterales 79 0.04 Spirobacillales 0.20

69 Spirobacillales 1.00 Spirobacillales 5.00 Spirobacillales

78 Spirobacillales 53 Sample Spirobacillales 88 93 Total_DefHNA Spirobacillales Total_DefLNA Myxococcales Total_NoneAbove NB1-j 92

94 NB1-j 61 DTB120 MIZ46 80 MIZ46 77 MIZ46 91 90 MIZ46 MIZ46 MIZ46 79 MIZ46 Bdellovibrionales

66 Bdellovibrionales Bdellovibrionales Bdellovibrionales 52 Bdellovibrionales

85 Bdellovibrionales 94 Bdellovibrionales

90 Bdellovibrionales

86 Bdellovibrionales Bdellovibrionales Bdellovibrionales 240

241 Figure S11: Detailed trees of various phyla (Bacteroidetes, Parcubacteria (OD1), and classes 242 (Alphaproteobacteria, Deltaproteobacteria), including boot-strap values from maximum-likelihood 243 constuction and labeled with order or family. (Further names are not given for Parcubacteria, as it 244 is a candidate phylum.) Colors of OTUs match Figure 6. 245

25 1 Low Abundance Minor Phyla 0.9 Dependentiae (TM6) SR1 0.8 Elusimicrobia Chloroflexi 0.7 Acidobacteria Verrucomicrobia 0.6 Saccharibacteria (TM7) Proteobacteria 0.5 Planctomycetes Omnitrophica (OP3)

Relative Proportion 0.4 Parcubacteria (OD1) Nitrospirae 0.3 Gracilibacteria (GN02) Firmicutes 0.2 Phyla detailed on next page Cyanobacteria Chlamydiae 0.1 Bacteroidetes Actinobacteria 0 Large Small Large Small Large Small Large Small Large Small Lake River Tap Ground- Waste- water water water water water 1

0.9

0.8

0.7 Rare Eliminated 0.6 Unclassifiable 0.5 Non-exclusive (II) Non-exclusive (I) 0.4 Small OTU

Relative Proportion Large OTU 0.3

0.2

0.1

0 246

247 (Figure S12, continued next page)

26 0.08 Legend Samples as above 0.1 0.06

0.04

0.02

Relative Proportion 0 0 Phylum Verrucomicrobia

0.6 0.06 0.1

0.4 0.04

0.02 0.2

0 0 0 Planctomycetes Proteobacteria Saccharibacteria (TM7)

0.05 0.6 0.04 0.1 0.03 0.4

0.02 0.2 0.01

0 0 0 Nitrospirae Parcubacteria (OD1) Omnitrophica (OP3)

0.05 0.05 0.06

0.04 0.04 0.04 0.03 0.03

0.02 0.02 0.02 0.01 0.01

0 0 0 Cyanobacteria Firmicutes Gracilibacteria (GNO2)

0.5 0.04

0.2 0.4 0.03 0.3 0.02 0.2 0.1 0.01 0.1

0 0 0 Actinobacteria Bacteroidetes Chlamydiae 248

249 Figure S12: Detailed relative abundance of each phyla and OTU type, averaged across all Large 250 (0.4µm filters) and Small (0.2µm filters of filtrate) in each ecosystem. OTUs with abundance <20 251 OTUs in <2 filters are left out of analysis, but included in the total relative abundance proportions 252 (Low abundance/Rare). OTU types are used as defined in Figure 1 - -Large OTU, small OTU, non- 253 exclusive (two types, detailed below), unclassifiable, eliminated and rare). Non-exclusive OTUs are 254 further divided here as to whether this occurred in duplicate matching filters (I) or without matching 255 duplicates (II). On the second page, the abundance of each of these OTU types is shown within 256 each phylum.

27 257 10. Original SEM images

200 nm

258

259 Figure S13: Grayscale (original) image of Figure 3. Scanning electron microscopy (SEM) image of 260 bacteria from a stagnant pond sample rich in LNA content bacteria (> 90%), filtered onto a 0.2 µm 261 pore-size filter. Filter pores are visible as black holes. Bacteria are oblong light gray and 262 extracellular filaments as light-gray strands.

28 A

B

263

264 Figure S14: Figure S9 without artificial coloring.

29