Microbial Source Tracking to Assess Human Contributions of Fecal Bacteria to a Freshwater Receiving Stream D. A. Graves1, D. E. Chestnut1, E. B. Rabon1, W. J. Jones2, J. G. Moore3, C. Johnston3, C. K. West1 1South Carolina Department of Health and Environmental Control, Bureau of Water, 2University of , Selah Genomics Facility, 3NOAA, National Ocean Service, Center for Coastal Environmental Health and Biomolecular Research

Bacterial Community Structure Abstract Methodology (cont.) Bacterial indicators of human sources

Northern Tributaries Percent Abundance PCR Amplification and Sample Pooling Northern Tributaries 2012 The State of South Carolina, Department of Health and Environmental Control 2012-07 2012-08 2012-09 2012-07 2012-08 2012-09 • The forward primer contained the 454 Life Sciences primer B, the broadly (SCDHEC) is charged with the responsibility to ensure that waters of the State support 1.00 conserved bacterial primer 27F, and a 2-base linker sequence (“TC”). classified uses and are free from pollutants in excess of State standards. Crucial to any • The reverse primer contained the 454 Life Sciences primer A, the bacterial primer Rank2 plan to achieve compliance with E. coli standards is an understanding of the probable p__Acidobacteria 338R, a “CA” inserted as a linker between the barcode and the rRNA primer, and a 100,000 p__Actinobacteria sources of fecal bacteria (i.e., human vs. animal). unique 12-bp error-correcting Golay barcode used to tag each PCR product PCRs 50,000 0.75 p__Armatimonadetes p__Bacteroidetes consisted of 0.25 μL (30 μM) of each forward and reverse primer, 3 μL of template 20,000 p__Chlorobi DNA, and 22.5 μL of Platinum PCR SuperMix (Invitrogen). 10,000 p__Chloroflexi Metagenomics (sequencing and analysis of DNA from environmental samples) and p__Cyanobacteria • A composite sample for pyrosequencing was prepared by pooling approximately p__Firmicutes quantitative polymerase chain reaction (qPCR) are genetic methodologies that can be Universal_Bacteroides 0.50 p__Gemmatimonadetes equal amounts of PCR amplicons from each sample. Human_Bacteroides 1,000 p__Nitrospirae used for microbial source tracking: the identification of microbial organisms specific to Lachnospiraceae Abundance p__OD1 • For each sample, we amplified the 16S rRNA gene using a primer set similar to M._smithii p__OP11 potential sources of fecal bacteria pollution. Metagenomic methods were used first to that described in Hamady et al. (2007) that was found to be well-suited for the Gene Copies/100mL p__Planctomycetes sequence DNA present in receiving stream water samples. The 16S rRNA amplicon was 100 p__Proteobacteria phylogenetic analysis of pyrosequencing reads. p__TM6 analyzed using BLAST, Greengenes, etc., to derive and contrast the microbial community 0.25 • Once quantified, the DNA was precipitated, centrifuged, and resuspended. The p__TM7 composition. Then, qPCR methodology was used to quantify the presence of known p__Verrucomicrobia sample was sent to the Selah Genomics Facility at the University of South Carolina p__WS3 10 genetic markers to substantiate and confirm the presence of potential human sources of (Columbia) for pyrosequencing on a 454 Life Sciences Genome Sequencer FLX fecal bacteria pollution. This assessment strategy was employed on an urban freshwater (Roche) machine. 0.00 stream system, the Lower , from its “origination” from a major impoundment, to its confluence with the . RC01 KC01 SC01 SC02 RC01 KC01 SC01 SC02 RC01 KC01 SC01 SC02 Phylogenetic Analyses Stat RC01 KC01 SC01 SC02 RC01 KC01 SC01 SC02 RC01 KC01 SC01 SC02 • Sequences were processed and analyzed following the procedure described in Sampling Station Lower Saluda River Percent Abundance Hamady et al. (2007). A comprehensive study conducted July – September 2012 included eight Lower Saluda • Only those sequences >200 bp in length with an average quality score >25 and no Lower Saluda River 2012 2012-07 2012-08 2012-09 River sites, 10 sites on tributary streams with municipal wastewater treatment plant 2012-07 2012-08 2012-09 ambiguous characters were included in the analyses. 50,000 1.00 discharges and two sites on tributaries without discharges. A generally increasing trend in • Sequences were assigned to samples by examining the 12-bp barcode. Phylotypes 20,000 human markers downstream in the Lower Saluda River was noted. The most specific were identified by using megablast to identify connected components (nearest Rank2 human marker, M. smithii, was seen immediately downstream of two municipal 10,000 p__Acidobacteria neighbor) sets of similar sequences. p__Actinobacteria wastewater treatment plants and also from an impoundment outfall. Community structure 0.75 p__Armatimonadetes • A representative sequence was chosen from each phylotype by selecting the most p__Bacteroidetes between tributaries appeared to be substantially different. The extent of these differences p__Chlorobi highly connected sequence, i.e., the sequence that had the most hits more and their influence on the Lower Saluda River are being evaluated. 1,000 p__Chloroflexi significant than the BLAST threshold to other sequences in the dataset. p__Cyanobacteria p__Firmicutes • The set of all representative sequences was aligned by using NAST (parameters: Universal_Bacteroides 0.50 p__Gemmatimonadetes minimum alignment length, 190; sequence identity, 70%) with a PH lanemask Human_Bacteroides p__Nitrospirae Lachnospiraceae Abundance p__OD1 (http://greengenes.lbl.gov/) to screen out hypervariable regions of the sequence. A 100 M._smithii

p__OP11 Lower Saluda River watershed - Study Sites Gene Copies/100mL p__Planctomycetes relaxed neighbor-joining tree was built by using Clearcut, employing the Kimura p__Proteobacteria correction. 0.25 p__TM6 p__TM7 • Unweighted UniFrac was run by using the resulting tree and the sequences 10 p__Verrucomicrobia p__WS3 annotated by environment type. Taxonomic identity of the phylotypes was assigned with BLAST against the Greengenes database by using an E value cutoff of 1e-10

0.00 and the Hugenholtz taxonomy. • The statistical significance of differences in microbial community composition

LS01 LS02 LS04 LS05 LS06 LS07 LS08 LS03 LS01 LS02 LS04 LS05 LS06 LS07 LS08 LS03 LS01 LS02 LS04 LS05 LS06 LS07 LS08 LS03 between sample categories was determined by using the G test on relative Stat LS01 LS02 LS04 LS05 LS06 LS07 LS08 LS03 LS01 LS02 LS04 LS05 LS06 LS07 LS08 LS03 LS01 LS02 LS04 LS05 LS06 LS07 LS08 LS03 phylotype abundances. Sampling Station Fourteen Mile Creek Percent Abundance Fourteen Mile Creek 2012 2012-07 2012-08 2012-09 2012-07 2012-08 2012-09

1.00 Summary 100,000

50,000 Rank2 p__Acidobacteria Bacterial Community Structure 20,000 p__Actinobacteria 10,000 0.75 p__Armatimonadetes •In August and September, in the absence of significant rainfall, bacterial community p__Bacteroidetes p__Chlorobi structure appears to remain relatively stable between months for each stream p__Chloroflexi system. p__Cyanobacteria 1,000 p__Firmicutes •In general, there appears to be a decrease in percent of Proteobacteria and an Universal_Bacteroides 0.50 p__Gemmatimonadetes increase in percent of Actinobacteria in Fourteen mile and Twelve Mile Creeks, from Human_Bacteroides p__Nitrospirae Lachnospiraceae

Abundance p__OD1 upstream to downstream, which does not appear to be the case in the Lower Saluda M._smithii 100 p__OP11 River. Gene Copies/100mL p__Planctomycetes p__Proteobacteria •Diversity is less in tributary streams than in the Lower Saluda River. Twelve Mile 0.25 p__TM6 Creek is the largest tributary watershed and appears to be the most similar to the p__TM7 10 p__Verrucomicrobia Lower Saluda River. p__WS3 Sample Collection •Nitrospirae and Planctomycetes are significant throughout the Lower Saluda River and do not appear in tributaries with the exception of SC02 in July. 0.00 Tributary samples were collected from bridge crossings and Lower Saluda River Bacterial Indicators of human sources samples were collected by boat. Each water sample was collected in 2 L Sterile Nalgene FMC01 FMC02 FMC03 FMC04 FMC01 FMC02 FMC03 FMC04 FMC01 FMC02 FMC03 FMC04 •Human Bacteriodes and Lacnospiraceae are present throughout the Lower Saluda FMC01 FMC02 FMC03 FMC04 FMC01 FMC02 FMC03 FMC04 FMC01 FMC02 FMC03 FMC04 TWMC04 TWMC04 TWMC04 TWMC04 TWMC04 TWMC04 Stat sample bottles at surface (0.3 m) depth. Samples were iced and a temperature blank and most tributary streams. Sampling Station Twelve Mile Creek Percent Abundance used to ensure that the samples were kept at the proper temperature. Samples were •There is a general increase of universal Bacteriodes, human Bacteriodes and then delivered to the NOAA CCEHBR laboratory personnel on-site. Twelve Mile Creek 2012 2012-07 2012-08 2012-09 Lacnospiraceae, from upstream to downstream, in the Lower Saluda River. This 2012-07 2012-08 2012-09

1.00 trend is not evident in tributary streams. 100,000 •M. smithii is absent from Fourteen Mile Creek; Human Bacteriodes appears only at 50,000 Rank2 Methodology FMC04 in July and FMC02 in September. 20,000 p__Acidobacteria p__Actinobacteria Water Filtration •M. smithii appears only downstream of waste water treatment plants on Stoops 10,000 0.75 p__Armatimonadetes Creek and Twelve Mile Creek, and below Corley Mill pond (TMC04) in August. p__Bacteroidetes •Water samples (1 L each) were filtered through 0.45 µm nitrocellulose filters to p__Chlorobi capture microorganisms, and filters were placed in 2 mL bead beating vials for DNA p__Chloroflexi Impact of Rainfall on Community Structure 1,000 p__Cyanobacteria extraction. •There was 3.29 inches of rainfall in the 24 hours prior to sample collection in July. Universal_Bacteroides p__Firmicutes Human_Bacteroides 0.50 p__Gemmatimonadetes There was zero inches of rainfall recorded prior to sampling in August and Lachnospiraceae p__Nitrospirae DNA Extraction M._smithii Abundance September. However, light scattered showers were visually noted in August. p__OD1 •Bead beating vials were processed on a FastPrep (MP Biomedicals) instrument to 100 p__OP11 •The dramatic changes in bacterial community structure in Stoops Creek (SC02) Gene Copies/100mL p__Planctomycetes disrupt the filter and attached microbial cells. DNA was extracted using the DNEasy and stations downstream of LS04 in the Saluda River are attributed to heavy rainfall p__Proteobacteria Plant Mini Kit (Qiagen) following the manufacturer’s protocol. 0.25 p__TM6 in July. Most notably, the presence of Nitrospirae and Planctomycetes in Stoops p__TM7 10 p__Verrucomicrobia Creek and the increase of Bacteriodetes in the Lower Saluda River. p__WS3

0.00

TWMC01 TWMC02 TWMC03 TWMC04 TWMC01 TWMC02 TWMC03 TWMC04 TWMC01 TWMC02 TWMC03 TWMC04 TWMC01 TWMC02 TWMC03 TWMC04 TWMC01 TWMC02 TWMC03 TWMC04 TWMC01 TWMC02 TWMC03 TWMC04 Stat Sampling Station