MIAMI UNIVERSITY the Graduate School Certificate for Approving The

Home , Caldimicrobium, Caldisphaera, Hydrogenobacter, Methanomicrobiaceae, Sulfobacillus, Winslow (crater)

MIAMI UNIVERSITY

The Graduate School

Certificate for Approving the Dissertation

We hereby approve the Dissertation

Qiuyuan Huang

Candidate for the Degree:

Doctor of Philosophy

______

Hailiang Dong, Director

______

Yildirim Dilek, Reader

______

Jonathan Levy, Reader

______

Chuanlun Zhang, External examiner

______

Annette Bollmann, Graduate School Representative

ABSTRACT GEOMICROBIAL INVESTIGATIONS ON EXTREME ENVIRONMENTS: LINKING GEOCHEMISTRY TO MICROBIAL ECOLOGY IN TERRESTRIAL HOT SPRINGS AND SALINE LAKES by Qiuyuan Huang Terrestrial hot springs and saline lakes represent two extreme environments for microbial life and constitute an important part of global ecosystems that affect the biogeochemical cycling of life-essential elements. Despite the advances in our understanding of microbial ecology in the past decade, important questions remain regarding the link between microbial diversity and geochemical factors under these extreme conditions. This dissertation first investigates a series of hot springs with wide ranges of temperature (26-92oC) and pH (3.72-8.2) from the Tibetan Plateau in China and the Philippines. Within each region, microbial diversity and geochemical conditions were studied using an integrated approach with 16S rRNA molecular phylogeny and a suite of geochemical analyses. In Tibetan springs, the microbial community was dominated by archaeal phylum Thaumarchaeota and four bacterial phyla (Firmicutes, Proteobacteria, Cyanobacteria, Chloroflexi). In the Philippines hot springs, the archaeal community mainly consisted of phyla Crenarchaeota, Euryarchaeota, and unclassified Archaea, while the bacterial community was predominated by phyla Aquificae and Firmicutes. Sulfur metabolisms appear to be key physiological functions in these hot springs. Saline lake on high elevation represents another extreme environment where various microorganisms thrive. In this study, the taxonomic and functional diversity of the microbial community in a saline lake (Qinghai Lake, China) was characterized using metagenomes and metatranscriptomes techniques. Metagenomic data showed that microbial communities in two water columns of Qinghai Lake were mainly dominated by Bacteria (mainly phyla Cyanobacteria, Proteobacteria, Verrucomicrobia, Actinobacteria, etc), while metatranscriptomics was performed to study the functional gene expression in elemental cycles (e.g., carbon and nitrogen cycles).

Lastly, an artificial extreme environment was studied, where trichloroethylene (TCE) in bedrock aquifer was a major contaminant. In situ chemical oxidation (ISCO) was a commonly used strategy to oxidize TCE, but how strong oxidants such as permanganate diffuses into solid rocks and its interaction with rock matrices remain poorly known. A series of diffusion experiments were conducted that measured the permanganate diffusion and reaction in four different types of sedimentary rocks (dark gray mudstone, light gray mudstone, red sandstone, and tan sandstone). Various Mn minerals formed as surface coatings as a result of the permanganate reduction coupled with total organic carbon (TOC) oxidation, and the extent of permanganate diffusion and reaction depended on rock properties.

GEOMICROBIAL INVESTIGATIONS ON EXTREME ENVIRONMENTS: LINKING GEOCHEMISTRY TO MICROBIAL ECOLOGY IN TERRESTRIAL HOT SPRINGS AND SALINE LAKES

A DISSERTATION

Submitted to the Faculty of

Miami University in partial

fulfillment of the requirements

for the degree of

Doctor of Philosophy

Department of Geology and Environmental Earth Science

Qiuyuan Huang

Miami University

Oxford, Ohio

2014

Dissertation Advisor: Hailiang Dong

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION ...... 1 REFERENCES ...... 5

CHAPTER 2: ARCHAEAL AND BACTERIAL DIVERSITY IN HOT SPRINGS ON THE TIBETAN PLATEAU, CHINA ...... 10 ABSTRACT ...... 11 INTRODUCTION ...... 12 REFERENCES ...... 26

CHAPTER 3: ARCHAEAL AND BACTERIAL DIVERSITY IN ACIDIC TO CIRCUMNEUTRAL HOT SPRINGS IN THE PHILIPPINES ...... 44 ABSTRACT ...... 45 INTRODUCTION ...... 46 REFERENCES ...... 62

CHAPTER 4: METAGENOMIC AND METATRANSCRIPTOMIC ANALYSIS OF TAXONOMIC AND GENETIC DIVERSITY IN SALINE QINGHAI LAKE, CHINA ...... 80 CHAPTER 5: PERMANGANATE DIFFUSION AND REACTION IN SEDIMENTARY ROCKS ...... 110 ABSTRACT ...... 111 INTRODUCTION ...... 112 REFERENCES ...... 125

CHAPTER 6: SUMMARY ...... 142 APPENDIX: PUBLICATIONS ...... 143

LIST OF TABLES CHAPTER 1: INTRODUCTION ...... 1 CHAPTER 2: ARCHAEAL AND BACTERIAL DIVERSITY IN HOT SPRINGS ON THE TIBETAN PLATEAU, CHINA ...... 10 TABLE 1. LOCATION AND DESCRIPTION OF THE INVESTIGATED SOIL AND TEN HOT SPRINGS ON THE TIBETAN PLATEAU, CHINA ...... 33 TABLE 2. ECOLOGICAL ESTIMATES AND MAJOR GROUP AFFILIATION OF THE BACTERIAL 16S RRNA GENE CLONE SEQUENCES RETRIEVED FROM THE SOIL AND TEN HOT SPRINGS ON THE TIBETAN PLATEAU...... 34 TABLE 3. ECOLOGICAL ESTIMATES AND MAJOR GROUP AFFILIATION OF THE ARCHAEAL 16S RRNA GENE CLONE SEQUENCES RETRIEVED FROM THE SOIL AND TEN HOT SPRINGS ON THE TIBETAN PLATEAU...... 35 CHAPTER 3: ARCHAEAL AND BACTERIAL DIVERSITY IN ACIDIC TO CIRCUMNEUTRAL HOT SPRINGS IN THE PHILIPPINES ...... 44 TABLE 1. GEOGRAPHIC PARAMETERS AND WATER GEOCHEMISTRY OF THE SIX INVESTIGATED HOT SPRINGS IN THE PHILIPPINES...... 70 TABLE 2. ABUNDANCE OF ARCHAEAL AND BACTERIAL 16S RRNA GENES FOR THE SIX SPRING SEDIMENT SAMPLES FROM THE PHILIPPINES ...... 71 TABLE S1. DIVERSITY ESTIMATES OF 454 PYROSEQUENCES RETRIEVED FROM SIX HOT SPRINGS IN THE PHILIPPINES ...... 77 CHAPTER 4: METAGENOMIC AND METATRANSCRIPTOMIC ANALYSIS OF TAXONOMIC AND GENETIC DIVERSITY IN SALINE QINGHAI LAKE, CHINA ...... 80 TABLE 1. SAMPLE LOCATIONS AND GEOCHEMISTRY...... 100 TABLE 2. STATISTIC RESULTS OF METAGENOME AND METATRANSCRIPTOME SEQUENCES FROM SITES B AND E IN QINGHAI LAKE...... 101 CHAPTER 5: PERMANGANATE DIFFUSION AND REACTION IN SEDIMENTARY ROCKS ...... 110 TABLE 1. CHARACTERIZATION OF ROCK PROPERTIES ...... 130 TABLE 2. DOBS VALUES AND EFFECTIVE DIFFUSION COEFFICIENTS FOR DIFFERENT ROCK TYPES ...... 131 TABLE 3. DEFF VALUES PRIOR TO OXIDANT EXPOSURE COMPARED TO DEFF AFTER EXPOSURE TO OXIDANT ...... 132 TABLE 4. MINERALOGY ON PERMANGANATE REACTIVE SURFACES OF SEDIMENTARY ROCKS...... 133 CHAPTER 6: SUMMARY ...... 142

iii

LIST OF FIGURES CHAPTER 1: INTRODUCTION ...... 1 CHAPTER 2: ARCHAEAL AND BACTERIAL DIVERSITY IN HOT SPRINGS ON THE TIBETAN PLATEAU, CHINA ...... 10 FIGURE 1...... 37 FIGURE 2A...... 38 FIGURE 2B...... 39 FIGURE 2C...... 40 FIGURE 3...... 41 FIGURE 4A...... 42 FIGURE 4B...... 42 FIGURE S1...... 43 CHAPTER 3: ARCHAEAL AND BACTERIAL DIVERSITY IN ACIDIC TO CIRCUMNEUTRAL HOT SPRINGS IN THE PHILIPPINES ...... 44 FIGURE 1...... 73 FIGURE 2...... 74 FIGURE 3...... 75 FIGURE 4...... 76 FIGURE S1...... 78 FIGURE S2 ...... 79 CHAPTER 4: METAGENOMIC AND METATRANSCRIPTOMIC ANALYSIS OF TAXONOMIC AND GENETIC DIVERSITY IN SALINE QINGHAI LAKE, CHINA ...... 80 FIGURE 1...... 103 FIGURE 2...... 104 FIGURE 3...... 105 FIGURE 4...... 106 FIGURE 5...... 107 FIGURE 6...... 108 FIGURE 7...... 109 CHAPTER 5: PERMANGANATE DIFFUSION AND REACTION IN SEDIMENTARY ROCKS ...... 110 FIGURE 1...... 135 FIGURE 2...... 136 FIGURE 3...... 137 FIGURE 4...... 138 FIGURE 5...... 139 FIGURE 6...... 140 FIGURE S1 ...... 141

ACKNOWLEDGMENTS Firstly, I would like to thank my doctorate advisor Dr. Hailiang Dong for being tireless, patient and inspirational to my research and education throughout my PhD study in the past five years. I really admire his persistence and enthusiasm as a scientist, and will always remember his classic “be productive” philosophy. It was him who gave me as many opportunities as he could to improve myself as a scientist and urged me to overcome many challenges that I would have thought impossible. This dissertation would not be accomplished smoothly without his generous help and careful guidance. I would like to extend my gratitude to everyone who has served on my doctoral committees over the years: Dr. Yildirim Dilek, Dr. Jonathan Levy, Dr. Annette Bollmann, Dr. Mark Krekeler, and Dr. Chuanlun Zhang. I also thank Brandon Briggs and Hongchen Jiang for generously helping me and teaching me the many techniques I needed for my projects. I thank John Morton for his assistance in geochemical analyses. Thanks to Dr. Brian Hedlund and his group for helping me during my visit at the University of Nevada, Las Vegas. I’d like to thank my group members and friends at Miami for their advice and encouragement. The great friendship has made my PhD life delightful and enjoyable. Finally, I am thankful to my family back in China, especially my parents and grandparents, for taking care of themselves without me being around for so many years. And, I offer my deepest gratitude to my husband, Kun Li, whom I met at Miami University, for giving me endless support, faith, and love throughout this process.

CHAPTER 1: INTRODUCTION Extreme environments are niches that are challenging for most of life forms, due to their physical extremes (e.g., temperature, radiation or pressure) and geochemical extremes (e.g., salinity, pH, redox potential) (Rothschild and Mancinelli, 2001). Investigations on extreme environments and organisms that inhabit them have been considered as key clues to study the evolution of life on the early Earth or even extraterrestrial life, and thus are drawing worldwide attentions in the past decade. Terrestrial hot springs and saline lakes are two natural extreme environments with high temperature and salinity, respectively; and constitute an important part of the global aquatic ecosystem (Dong and Yu, 2007). Microorganisms thrive in these habitats, are the major players in mediating the biogeochemical cycling of life- essential elements (e.g., carbon, nitrogen, and sulfur), and thus are linked to global climate change. For example, thermophilic Archaea have been detected as a dominant component of microbial communities performing ammonia oxidation, which is a vital link in the biogeochemical nitrogen cycle, in terrestrial hot springs of Iceland and Kamchatka, Russian (Reigstad et al., 2008). Chapters 2 and 3 will focus on microbial taxonomic diversity and geochemical conditions in two terrestrial geothermal systems: the Tibetan Plateau in China and the east coast of the Philippines. As the highest plateau in the world, the Tibetan Plateau is also called “the Roof of the World”, and hosts a large number of geothermal springs with varying environmental gradients (Hu et al., 2003). High temperature and UV exposure have made Tibetan hot springs a unique place for studying extremophiles that are resistant to heat and UV light. On the other hand, the Philippines harbors a great number of geothermal features, specifically, acidic hot springs (Dolor, 2005). Acidic hot springs are considered extreme environments, yet they contain a diverse array of thermoacidophilic microorganisms capable of surviving and functioning under such extreme conditions (Rothschild and Mancinelli, 2001). It has been extensively studied worldwide on microbial communities in hot spring environments, for example, those from Yellowstone National Park (Barns et al., 1994; Hall et al., 1

2008; Meyer-Dombard et al., 2005; Mitchell, 2009; Pace, 1997), Kamchatka of Russia (Bonch-Osmolovskaya et al., 1999; Reigstad et al., 2009), Iceland (Aguilera et al., 2010; Marteinsson et al., 2001; Reigstad et al., 2009), and the Tengchong area in Yunnan Province of China (Hou et al., 2013; Jiang et al., 2010; Song et al., 2010; Song et al., 2009). Some studies also focused on microbial diversity in Tibetan hot springs (Lau et al., 2009; Lau et al., 2006; Wang et al., 2013) or microbial diversity of specific groups in Philippines hot springs (Jing et al., 2005; Lacap et al., 2007; Lantican et al., 2011). However, there is still a gap in our knowledge of the microbial distribution in a broader temperature/pH range and how they respond to the environmental variables under these extreme conditions. Chapter 4 will focus on the microbial functional diversity in biogeochemical cycles of a saline lake (Qinghai Lake) in China. Qinghai Lake is the largest inland saline lake (12.5g/L, pH 9.4, 3196 m above sea level) in China. Due to the high UV exposure and salinity, it has been considered as a unique ecosystem for studying extremophiles (Dong et al., 2006) and a perfect analogue for the early Earth and extraterrestrial planets (e.g., Mars) (Cabrol et al., 2007; Catling, 1999; Dib et al., 2008). Although several previous studies have investigated the microbial communities in Qinghai Lake, few of them have focused on the microbial functional metabolisms related to biogeochemical cycles (Jiang et al., 2006; Wu et al., 2006; Xing et al., 2009; Zhang et al., 2013). This research, for the first time, studies the microbial taxonomic and functional diverstiy using a metagenomic and metatrancriptomic survey, and will provide useful insights for further understanding the biogeochemical cycles under extreme environmental stresses. Chapter 5 is a subproject focusing on man-made extreme environment, i.e., TCE contaminated groundwater aquifers. The objective was to evaluate the ability of chemical oxidant (such as potassium permanganate) to treat chlorinated solvents residing in the rock matrix. In situ chemical oxidation (ISCO) using strong oxidants (e.g., permanganate, persulfate, ozone, and hydrogen peroxide) has been known as an effective way to remediate chlorinated solvents in fractured bedrock aquifers at several Department of Defense (DOD) and Department of Energy (DOE) sites 2

(Conrad et al., 2002; Krembs et al., 2010; Schnarr et al., 1998; Siegrist, 2011). Permanganate oxidation of organic contaminants, such as trichloroethylene (TCE), tetrachloroethylene (PCE), and styrene in the aqueous phase usually results in the formation of MnO2 solids (Crimi and Siegrist, 2004; Gates-Anderson et al., 2001; Tunnicliffe and Thomson, 2004; USEPA, 1998; Wu et al., 2012; Yan and Schwartz, 1999). It has been suggested that different Mn minerals would form depending on specific experimental conditions (Loomer et al., 2010; Scott et al., 2011); however, the exact oxidation state of Mn and a comprehensive understanding regarding the interaction between chemical oxidants and rock matrices is still lacking. It is currently unclear how implementation of ISCO impacts the bedrock matrix, as the reactions between oxidants and rock matrix at the fracture-rock interface and diffusion of oxidant into rock matrix could impact rock properties through Mn oxide precipitation and subsequently impact contaminant diffusive flux from the matrix to water-bearing fractures. Therefore, an improved understanding regarding diffusion of chemical oxidants into bedrock and oxidant reactions with naturally occurring reductants is needed to assess the potential benefits of ISCO for chlorinated solvents in bedrock aquifers. This study will provide insights for further assessing the feasibility of applying permanganate for treatment of chlorinated compounds within rock matrices. The goal of this dissertation is to systematically investigate the microbial abundance, taxonomic diversity, and metabolic functional expression in terrestrial hot springs and saline lakes using molecular phylogeny techniques and geochemistry analyses. Furthermore, an artificial extreme environment was investigated for the potential of using chemical oxidants to treat TCE. The specific objectives to achieve this goal include: 1) to investigate the microbial diversity and community structure in ten Tibetan hot springs and six acidic hot springs in the Philippines; 2) to assess the environmental variables that help shape microbial community structure and compare the microbial composition in Philippines hot springs with others in the world, such as Tengchong in China and YNP in USA;

3) to investigate the microbial taxonomic and functional diversity in Qinghai Lake water column and assess the response of functional activity to environmental variables. 4) To assess the effectiveness of using strong oxidant such as permanganate to treat TCE in bedrock matrices.

REFERENCES Aguilera, A., Souza-Egipsy, V., Gonzalez-Toril, E., Rendueles, O. and Amils, R., 2010. Eukaryotic microbial diversity of phototrophic microbial mats in two Icelandic geothermal hot springs. Int. Microbiol., 13: 21-32. Barns, S.M., Fundyga, R.E., Jeffries, M.W. and Pace, N.R., 1994. Remarkable archaeal diversity detected in a Yellowstone National Park hot spring environment. Proc. Natl. Acad. Sci. U S A., 91(5): 1609-1613. Bonch-Osmolovskaya, E.A. et al., 1999. Biodiversity of Anaerobic Lithotrophic Prokaryotes in Terrestrial Hot Springs of Kamchatka. Microbiology, 68: 343– 351. Cabrol, N.A., . et al., 2007. Signatures of habitats and life in Earth's high-altitude lakes: clues to Noachian aqueous environments on Mars. The Geology of Mars. Cambridge University Press. Catling, D.C., 1999. A chemical model for evaporites on early Mars: Possible sedimentary tracers of the early climate and implications for exploration. Journal of Geophysical Research: Planets, 104(E7): 16453-16469. Conrad, S.H., Glass, R.J. and Peplinski, W.J., 2002. Bench-scale visualization of DNAPL remediation processes in analog heterogeneous aquifers: surfactant floods and in situ oxidation using permanganate. Journal of Contaminant Hydrology, 58(1-2): 13-49. Crimi, M. and Siegrist, R., 2004. Impact of Reaction Conditions on Genesis during Permanganate Oxidation. Journal of Environmental Engineering, 130(5): 562- 572. Dib, J., Motok, J., Zenoff, V., Ordoñez, O. and Farías, M., 2008. Occurrence of Resistance to Antibiotics, UV-B, and Arsenic in Bacteria Isolated from Extreme Environments in High-Altitude (Above 4400 m) Andean Wetlands. Curr Microbiol, 56(5): 510-517. Dolor, F.M., 2005. Phases of geothermal development in the Philippines. Workshop for Decision Makers on Geothermal Projects and their Management. Naivasha, Kenya, November 14-18, 2005. 5

Dong, H. and Yu, B., 2007. Geomicrobiological processes in extreme environments: a review. Episodes, 30(3): 202. Dong, H. et al., 2006. Microbial Diversity in Sediments of Saline Qinghai Lake, China: Linking Geochemical Controls to Microbial Ecology. Microb Ecol, 51(1): 65-82. Gates-Anderson, D., Siegrist, R. and Cline, S., 2001. Comparison of Potassium Permanganate and Hydrogen Peroxide as Chemical Oxidants for Organically Contaminated Soils. Journal of Environmental Engineering, 127(4): 337-347. Hall, J.R. et al., 2008. Molecular characterization of the diversity and distribution of a thermal spring microbial community by using rRNA and metabolic genes. Appl Environ Microbiol, 74(15): 4910-4922. Hou, W. et al., 2013. A comprehensive census of microbial diversity in hot springs of Tengchong, Yunnan Province China using 16S rRNA gene pyrosequencing. PloS one, 8(1): e53350. Hu, X., Sun, J., Yao, Z. and Rong, F., 2003. Research on the influence of geothermal activities and exploitation to geological environment in Tibet (in Chinese). Journal of Mountain Research, 21(S1): 45-48. Jiang, H. et al., 2006. Microbial Diversity in Water and Sediment of Lake Chaka, an Athalassohaline Lake in Northwestern China. Applied and Environmental Microbiology, 72(6): 3832-3845. Jiang, H. et al., 2010. RNA-based investigation of ammonia-oxidizing archaea in hot springs of Yunnan Province, China. Appl Environ Microbiol, 76(13): 4538- 4541. Jing, H. et al., 2005. Community phylogenetic analysis of moderately thermophilic cyanobacterial mats from China, the Philippines and Thailand. Extremophiles, 9(4): 325-332. Krembs, F.J., Siegrist, R.L., Crimi, M.L., Furrer, R.F. and Petri, B.G., 2010. ISCO for Groundwater Remediation: Analysis of Field Applications and Performance. Ground Water Monitoring & Remediation, 30(4): 42-53.

Lacap, D.C., Barraquio, W. and Pointing, S.B., 2007. Thermophilic microbial mats in a tropical geothermal location display pronounced seasonal changes but appear resilient to stochastic disturbance. Environmental Microbiology, 9(12): 3065- 3076. Lantican, N., Diaz, M., Cantera, J., de los Reyes, F. and Raymundo, A., 2011. Microbial community of a volcanic mudspring in the Philippines as revealed by 16S rDNA sequence analysis and fluorescence in situ hybridization. World Journal of Microbiology and Biotechnology, 27(4): 859-867. Lau, M.C.Y., Aitchison, J.C. and Pointing, S.B., 2009. Bacterial community composition in thermophilic microbial mats from five hot springs in central Tibet. Extremophiles, 13: 139-149. Lau, M.C.Y., Jing, H., Aitchison, J.C. and Pointing, S.B., 2006. Highly diverse community structure in a remote central Tibetan geothermal spring does not display monotonic variation to thermal stress. FEMS Microbiol. Ecol., 50: 80- 91. Loomer, D.B., Al, T.A., Banks, V.J., Parker, B.L. and Mayer, K.U., 2010. Manganese Valence in Oxides Formed from in Situ Chemical Oxidation of TCE by KMnO4. Environmental Science & Technology, 44(15): 5934-5939. Marteinsson, V.T. et al., 2001. Phylogenetic diversity analysis of subterranean hot springs in Iceland. Appl Environ Microbiol, 67(9): 4242-4248. Meyer-Dombard, D.R., Shock, E.L. and Amend, J.P., 2005. Archaeal and bacterial communities in geochemically diverse hot springs of Yellowstone National Park, USA. Geobiology, 3(3): 211-227. Mitchell, K.R., 2009. Controls on microbial community structure in thermal environments: exploring bacterial diversity and the relative influence of geochemistry and geography. Dissertation Thesis, University of New Mexico, Albuquerque, New Mexico, 98 pp. Pace, N.R., 1997. A Molecular View of. Science, 276: 734-740.

Reigstad, L.J., Jorgensen, S.L. and Schleper, C., 2009. Diversity and abundance of Korarchaeota in terrestrial hot springs of Iceland and Kamchatka. ISME J, 4(3): 346-356. Reigstad, L.J. et al., 2008. Nitrification in terrestrial hot springs of Iceland and Kamchatka. FEMS Microbiology Ecology, 64(2): 167-174. Rothschild, L.J. and Mancinelli, R.L., 2001. Life in extreme environments. Nature, 409(6823): 1092-1101. Schnarr, M. et al., 1998. Laboratory and controlled field experiments using potassium permanganate to remediate trichloroethylene and perchloroethylene DNAPLs in porous media. Journal of Contaminant Hydrology, 29(3): 205-224. Scott, D., Apblett, A. and Materer, N.F., 2011. Follow-up study on the effects on well chemistry from biological and chemical remediation of chlorinated solvents. Journal of Environmental Monitoring, 13(9): 2521-2526. Siegrist, R.L., Crimi, M., & Simpkin, T. J. (Eds.), 2011. In situ chemical oxidation for groundwater remediation (Vol. 3). Springer. Song, Z.-Q. et al., 2010. Diversity of Crenarchaeota in terrestrial hot springs in Tengchong, China. Extremophiles 14: 287–296. Song, Z. et al., 2009. Actinobacterial diversity in hot springs in Tengchong (China), Kamchatka (Russia), and Nevada (U.S.A.). Geomicrobiol. J., 26(4): 256-263. Tunnicliffe, B.S. and Thomson, N.R., 2004. Mass removal of chlorinated ethenes from rough-walled fractures using permanganate. Journal of Contaminant Hydrology, 75(1-2): 91-114. USEPA, 1998. Field applications of in situ remediation technologies: chemical oxidant ground water at fractured rock sites. EPA 542-R-01-010. Vick, T.J., Dodsworth, J.A., Costa, K.C., Shock, E.L. and Hedlund, B.P., 2010. Microbiology and geochemistry of Little Hot Creek, a hot spring environment in the Long Valley Caldera. Geobiology 8: 140-154. Wang, S. et al., 2013. Control of temperature on microbial community structure in hot springs of the Tibetan Plateau. PloS one, 8(5): e62901.

Wu, M.Z., Reynolds, D.A., Fourie, A., Prommer, H. and Thomas, D.G., 2012. Electrokinetic in situ oxidation remediation: Assessment of parameter sensitivities and the influence of aquifer heterogeneity on remediation efficiency. Journal of Contaminant Hydrology, 136–137(0): 72-85. Wu, Q.L., Zwart, G., Schauer, M., Kamst-van Agterveld, M.P. and Hahn, M.W., 2006. Bacterioplankton Community Composition along a Salinity Gradient of Sixteen High-Mountain Lakes Located on the Tibetan Plateau, China. Applied and Environmental Microbiology, 72(8): 5478-5485. Xing, P., Hahn, M.W. and Wu, Q.L., 2009. Low Taxon Richness of Bacterioplankton in High-Altitude Lakes of the Eastern Tibetan Plateau, with a Predominance of Bacteroidetes and Synechococcus spp. Applied and Environmental Microbiology, 75(22): 7017-7025. Yan, Y.E. and Schwartz, F.W., 1999. Oxidative degradation and kinetics of chlorinated ethylenes by potassium permanganate. Journal of Contaminant Hydrology, 37(3-4): 343-365. Zhang, R. et al., 2013. Diversity of bacterioplankton in contrasting Tibetan lakes revealed by high-density microarray and clone library analysis. FEMS Microbiology Ecology: n/a-n/a.

CHAPTER 2: Archaeal and Bacterial Diversity in Hot Springs on the Tibetan Plateau, China Running title: Archaeal and Bacterial Diversity in Tibetan Hot Springs

Qiuyuan Huang1^, Christina Z. Dong2^, Raymond M. Dong2, Hongchen Jiang3*, Shang Wang3, Genhou Wang4, Bin Fang5, Xiaoxue Ding4, Lu Niu4, Xin Li4, Chuanlun Zhang6,7, and Hailiang Dong1,3*

1: Department of Geology and Environmental Earth Science, Miami University, Oxford,

OH 45056 2: Badin High School, Hamilton, OH 45013 3: Geomicrobiology Laboratory, China University of Geosciences, Beijing, 100083, China 4: School of Earth Sciences and Resources, China University of Geosciences, Beijing, 100083, China 5: School of Water Resources and Environment, China University of Geosciences, Beijing, 100083, China 6: State Key Laboratory of Marine Geology, Tongji University, Shanghai, 200092, China 7: Department of Marine Sciences, University of Georgia, Athens, GA 30602 ^Equal contribution

*Corresponding authors: Hailiang Dong: [email protected]; Hongchen Jiang: [email protected];

Published in Extremophiles (2011) 15(5):549-63

ABSTRACT The diversity of archaea and bacteria was investigated in ten hot springs (elevation > 4600 meters above sea level) in central and central-eastern Tibet by using 16S rRNA gene phylogenetic analysis. The temperature and pH of these hot springs were 26-81oC and close to neutral, respectively. A total of 959 (415 and 544 for bacteria and archaea, respectively) clone sequences were obtained. Phylogenetic analysis showed that bacteria were more diverse than archaea, and that these clone sequences were classified into 82 bacterial and 41 archaeal operational taxonomic units (OTUs), respectively. The retrieved bacterial clones were mainly affiliated with four known groups (i.e. Firmicutes, Proteobacteria, Cyanobacteria, Chloroflexi), which were similar to those in other neutral-pH hot springs at low elevations. In contrast, most of the archaeal clones from the Tibetan hot springs were affiliated with Thaumarchaeota, a newly proposed archaeal phylum, which may be unique to the Tibetan hot springs. Statistic analysis showed that diversity indices of both archaea and bacteria were not statistically correlated with temperature, which is consistent with previous studies.

Keywords Archaea, Bacteria, Diversity, hot springs, Thaumarchaeota, Tibet

INTRODUCTION Microbial communities in hot springs at low elevations have been extensively studied worldwide, such as those in Yellowstone National Park (Barns et al., 1994; Pace 1997; Meyer-Dombard et al., 2005; Hall et al., 2008; Mitchell, 2009), Long Valley Caldera near Mammoth Lakes, CA, USA (Vick et al., 2010), Kamchatka of Russia (Bonch-Osmolovskaya et al., 1999; Reigstad et al., 2009), Iceland (Marteinsson et al., 2001; Reigstad et al., 2009; Aguilera et al., 2010), Uttaranchal Himalaya (Kumar et al., 2004), the Tengchong area in Yunnan Province of China (Song et al., 2009; Jiang et al., 2010; Song et al., 2010), Indonesia (Aditiawati et al., 2009), and Tunisia (Sayeh et al., 2010). Several studies have focused on cyanobacteria in microbial mats of terrestrial hot springs of moderate temperatures 50-75oC (Ward et al., 1998; Ward and Castenholz, 2000). Diverse microbial communities are present in microbial mats (Lau et al., 2006; 2009); and temperature, pH, dissolved hydrogen sulfide levels, and biogeography are important factors in controlling abundance and diversity (Ward and Castenholz, 2000; Purcell et al., 2007; Whitaker et al., 2003). Among the diverse communities in hot springs, ammonia- oxidizing archaea (AOA) are an important group and globally distributed (Zhang et al., 2008). Based on environmental 16S rRNA gene sequences, these mesophilic archaea are placed with Crenarchaeota. However, based on the first genome sequence of a Crenarchaeote, Cenarchaeum symbiosum, Brochier-Armanet et al., (2008) showed that these mesophilic archaea are different from hyperthermophilic Crenarchaeota and branched deeper than previously thought. The authors proposed a new archaeal phylum, Thaumarchaeota to account for these mesophilic archaea. Thaumarchaeota may play a more important role in hot spring environment than previously thought, however, it is currently unknown if these organisms are widely distributed in terrestrial hot springs.

Despite these extensive studies, little is known about microbial diversity in hot springs at high elevations, especially in the Tibetan area (e.g., Lau et al., 2006). The Tibetan Plateau (> 4000 meters above sea level) is located in the east-central

Mediterranean-Himalayas tectonic zone, and the region hosts one of the most active geothermal areas in the world and possesses many hot springs with varying environmental gradients (Hu et al., 2003). Up to now, only Lau and colleagues have studied microbiology in some hot springs in central Tibet (Lau et al., 2006; Lau et al., 2009). Lau et al., (2006) investigated the microbial community along a thermal gradient (52-83oC) of an isolated geothermal location in central Tibet and found that the response of microbial diversity was not monotonic to thermal stress. In the other study, Lau and colleagues investigated the bacterial diversity in five hot springs (with a temperature range of 60-65oC) in the central Tibet and found that Proteobacteria and phototrophic bacteria (i.e. Chlorobi, Cyanobacteria, Chloroflexi) were ubiquitous (Lau et al., 2009). However, still little is known about how bacterial and archaeal communities are distributed among different hot springs of a larger temperature range.

The objective of this study was therefore to test if microbial diversity responds to thermal stress in Tibetan hot springs of a wide temperature range. We expanded upon the previous studies by investigating archaeal and bacterial diversity in ten Tibetan hot springs over a temperature range of 26.2-81.2oC. The 16S rRNA gene phylogenetic analysis was conducted to assess any correlation between microbial diversity and environmental variables (e.g. temperature, mineralogy).

MATERIALS AND METHODS

Field Measurements and Sampling

In August 2009, a field expedition was made to the central and central-eastern Tibet, and twenty hot springs were surveyed including sediment/mat color, temperature, and pH. Water pH and temperature were measured in the field using a thermometer and pH meter, respectively. Among these ten representative hot springs covering a range of temperatures were sampled at town Jiwa (JW) and Rongma (RM) of Nima County and town Gulu (GL) of Naqu County in the central and central- eastern Tibet (Fig. 1; Table 1). The springs at Jiwa were located between a river

13 valley and a flat terrain, whereas those from Rongma and Gulu were located in a river valley. These springs varied in size from 10 cm to several meters across. Geographical locations and elevations of these springs were determined using a portable GPS unit (eTrex H, Garmin, US). In general temperature and pH were homogeneous, and there were no dramatic changes of sediment/mat color within a given spring. Thus, only one sediment/mat sample was collected from each spring. Microbial mats and surface sediments were collected with a hand trowel into sterile 50 mL Falcon tubes and preserved in the sucrose lysis buffer (Mitchell and Takacs-Vesbach 2008). Hand trowels were sterilized with 75% ethanol and dried after each use. A sympatric soil at Rongma was also collected for comparison with the adjacent hot spring samples. Within one week, the preserved samples were shipped to the laboratory in Beijing and were then stored at -80oC until further analysis.

Powder X-Ray Diffraction (XRD) and Scanning electron microscopy (SEM)

XRD was performed to identify the mineralogy of the collected solids by using a Scintag X1 powder diffractometer system using CuKα radiation with a variable divergent slit and a solid-state detector (Zhang et al., 2005). For the XRD analysis, solids were air dried overnight, ground into powder, and tightly packed into the well of low-background quartz XRD slides (Gem Dugout, Inc., Pittsburgh, Pennsylvania). To facilitate qualitative comparisons among the samples, a similar amount of solid power was packed into a rectangular volume of the same dimensions. The routine power was 1400W (40kV, 35mA). Samples were scanned from 2 to 70 degree in 0.02 two-theta steps with a count time of 2 seconds per step. Search-match software was used to conduct mineral identification. The relative abundance of each mineral was qualitatively assigned to be one of four categories based on relative peak intensity of characteristic peaks: very abundant, abundant, moderate and present.

SEM was performed to determine the morphology of microorganisms and surrounding minerals using a Zeiss Supra 35 VP Field Emission Scanning Electron Microscope. 14

DNA Extraction and PCR

Genomic DNA was extracted from ~500 mg (wet weight) of each sample using FastDNA Spin Kit for Soil (MP Biomedical, US) according to the manufacturer’s instructions. The archaeal 16S rRNA gene from the extracted DNA was amplified with archaeal forward primer Arch21F (5’- TTCYGGTTGATCCYGCCRGA-3’) and universal reverse primer Univ958R (5’- YCCGGCGTTGAMTCCATTT-3’); while the primer set of Bac27F (5’- AGAGTTTGATCMTGGCTCAG-3’) and Univ1492R (5’- CGGTTACCTTGTTACGACTT-3’) was used for bacterial 16S rRNA gene PCR amplification. These primers have been used for hot spring samples (Pearson et al., 2004; Meyer-Dombard et al., 2005; Lau et al., 2009) and are effective in amplifying the 16S rRNA gene of bacteria and archaea. For PCR, a typical mixture (25µL in volume) for both archaea and bacteria consisted of the following reagents: 10 mM

Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2, and 100 µM of each deoxynucleoside triphosphate, 0.8 mM bovine serum albumin (TaKaRa, Dalian, China), 1.25 U of Taq DNA polymerase (TaKaRa, Dalian, China), and ~25 ng of total DNA. The conditions for archaeal 16S rRNA gene PCRs consisted of an initial denaturation at 95oC for 5 min, and 35 cycles of denaturing at 94 oC for 30s, annealing at 54 oC for 30 s, and extension at 72 oC for 2 min, followed by a final extension at 72 oC for 10 min. The conditions for bacterial 16S rRNA genes were as follows: an initial denaturation at 95 oC for 5 min, and 35 cycles of denaturing at 94 oC for 30 s, annealing at 55 oC for 30 s, and extension at 72 oC for 2 min, followed by a final extension at 72 oC for 10 min. The PCR products were purified with Agarose Gel DNA Purification Kit Ver.2.0 (TaKaRa, Dalian, China) according to the manufacturer’s instructions.

Clone Library Construction

The purified PCR products were ligated into cloning vectors and transformed into Escherichia coli Trans1-T1 competent cells using pEASY-T1 cloning kit (TransGen Biotech, Beijing, China) according to the manufacturer’s suggested 15 protocol. The transformants were plated on Luria-Bertani plates containing 100 µg mL-1 of ampicillin, 80µg mL-1 of X-Gal (5-bromo-4-chloro-3-indolyl-β-D- galactopyranoside) and 0.5 mM IPTG (isopropyl-β-D-thiogalactopyranoside). The Luria-Bertani plates were incubated at 37 oC overnight. Twenty-two clone libraries (eleven each for archaea and bacteria) were constructed. Colonies were randomly selected and analyzed for the 16S rRNA gene inserts. Inserts were amplified using forward primer M13-RV (5'-CAG GAA ACA GCT ATG AC-3') and reverse primer M13-47 (5'-GTT TTC CCA GTC ACG AC-3'). The PCR reaction system was same as described above with the following PCR conditions: 94oC for 10 min; 34 cycles of 94oC for 30s, 52oC for 30 s, 72oC for 90s, with a final elongation step of 72oC for 10 min. The randomly selected clones were sequenced (with primers Arch21F and Bac27F for archaea and bacteria, respectively) using the BigDye Terminator version 3.1 chemistry (Applied Biosystems, Foster City, CA, USA) with an ABI 3100 automated sequencer at Shanghai Sangon Biotech.

Phylogenetic Analyses

The raw sequences were trimmed by using Sequencher 4.8. The potential presence of chimeric sequences was examined with Bellerophon (Huber et al., 2004). Potential chimeric sequences were removed. The chimera-free sequences were blasted in the GenBank (http://www.ncbi.nlm.nih.gov). Operational taxonomic units (OTUs) were determined using DOTUR (Schloss and Handelsman 2005) with a 97% cutoff value. Neighbor-joining phylogenetic trees were constructed from dissimilar distance and pairwise comparisons with the Jukes-Cantor distance model using the MEGA (molecular evolutionary genetics analysis) program, version 4.1. Bootstrap value of 1000 replications was assessed in the analysis. The sequences determined in this study have been deposited in the GenBank database under accession numbers HQ287087- HQ287215.

Statistical Analysis

Coverage (C) of the constructed clone libraries was calculated as follows:

C=1-(n1/N), where n1 is the number of phylotypes that occurred only once in the clone library and N is the total number of clones analyzed (Jiang et al., 2009). LIBSHUFF analysis for the difference between any two clone libraries was performed in the same way as described elsewhere (Jiang et al., 2008). With zt software (http://www.psb.ugent.be/~erbon/mantel/), the Mantel test was performed to reveal any correlation between biotic and environmental data sets according to procedures as previously described (Jiang et al., 2009).

RESULTS

Characteristics of the Sampling Sites

The elevations of the ten investigated hot springs were higher than 4600 meters above sea level. The temperature and pH were 26.2-81.2oC and close to neutral (6.48-8.20), respectively (Table 1).

Mineralogy and Microbial Morphology

XRD analysis showed that the mineral composition varied among the samples from different localities: calcite was predominant in the RM26, RM45, and RM64 samples, whereas quartz was rarely found in these samples; calcite and quartz were the two major minerals in the JW41 and JW56 samples; and quartz and albite were the two dominant minerals in GL63 and GL64 (Table 1). Other minerals included minor amounts of layer silicates such as muscovite, illite and biotite. The morphology of microorganisms and surrounding minerals was observed under SEM (Figure S1).

Bacterial 16S rRNA Gene Phylogenetic Analysis

Four hundred and fifteen bacterial clone sequences were obtained, and they could be classified into the following groups: Firmicutes, Proteobacteria, Cyanobacteria, Chloroflexi, Bacteroidetes, Acidobacteria, Nitrospirae, Planctomycetes, Thermodesulfobacteria, Aquificae, and unclassified bacteria (Table 2 and Fig. 2). Among these groups, Firmicutes, Proteobacteria, Chloroflexi and Cyanobacteria were the major components in the bacterial 16S rRNA gene clone libraries, and they accounted for 91% of all bacterial 16S rRNA gene clone sequences (Table 2).

(i) Cyanobacteria and Chloroflexi. Eighty-six clone sequences (20.7%: 86 out of 415) were affiliated with Cyanobacteria and Chloroflexi (Fig. 2A; Table 2), and these sequences were derived from hot springs with relatively low temperatures (<64oC). Most of these sequences were closely related to clones retrieved from hot spring environments, such as those in Tibet (Lau et al., 2009), Yellowstone National Park (Allewalt et al., 2006), Thailand (Portillo et al., 2009), and Bulgarian (Tomova et al., 2010). In the Cyanobacteria, three clone sequences from JW56 were related to Synechococcus sp. TS-91 isolated from Octopus Spring (49-70oC) in Yellowstone National Park (Allewalt et al., 2006). In the Chloroflexi group, four clone sequences were closely related to sequences retrieved from Bor Khlueng Hot Spring (50-57oC) in Thailand (Kanokratana et al., 2004), and forty-one clone sequences from GL63 and GL64 (30 and 11, respectively) were related (94%) to a Chloroflexi bacterium (FM164953) isolated from a geothermal spring (79oC) in Bulgaria (Tomova et al., 2010).

(ii) Proteobacteria. One hundred and thirty-six sequences (33.0%: 137 out of 415) were affiliated with Proteobacteria, and these sequences could be classified into subgroups: Alpha-, Beta-, Gamma-, and Deltaproteobacteria (Fig. 2 B; Table 2). The clone sequences affiliated with Gammaproteobacteria were predominant (75%: 105 out of 136), and most of them were closely (98-100%) related to cultured Gammaproteobacteria and clones retrieved from low-temperature habitats, such as ice, soils, and sediments. One clone sequence from GL52 was closely related (99%) to

Thiofaba tepidiphila, a novel obligately chemolithoautotrophic, sulfur-oxidizing bacterium of the Gammaproteobacteria isolated from a hot spring (45oC and pH 7.0) in Fukushima prefecture, Japan (Mori and Suzuki 2008).

(iii) Firmicutes. One hundred and sixty-six clone sequences (37.1%: 154 out of 415) were affiliated with Firmicutes. The Firmicutes sequences can be classified into two subgroups: Clostridia and Bacillales (Fig. 2 C; Table 2), among which Clostridia sequences were predominant (95%: 157 out of 166). The Clostridia sequences were closely related (92-99%) to clones retrieved from low- or high- temperature environments. Nine sequences from GL64 were related (93%) to bacterium clone TMP-B3 (EU544532), which was obtained from a high temperature mud pool in the Taupo Volcanic Zone, New Zealand (GenBank description). The majority of Bacillales sequences (88.9%: 8 out of 9) were obtained from the soil sample (RMS) and they were closely related (99%) to isolates or clones recovered from soda lakes (Joshi et al., 2008; Wu et al., 2010).

Around 10% (41 out of 415) of the total bacterial clone sequences were affiliated with some minor groups, such as Acidobacteria, Bacteroidetes, Nitrospirae, Planctomycetes, Thermodesulfobacteria, Aquificae, and unclassified bacteria (Fig. 2A; Table 2). Most of these sequences were closely (98-99%) related to clones retrieved from geothermal features, such as Tibetan hot springs (Lau et al., 2009), Yellowstone hot springs (Hugenholtz et al., 1998; Boomer et al., 2009), and Japanese alkaline geothermal pool (Kimura et al., 2010).

Archaeal 16S rRNA Gene Phylogenetic Analysis

Five hundred and forty-four archaeal 16S rRNA gene clone sequences were obtained and the numbers of clones represented 87.2-100% coverage for each clone library (Table 3; Fig. 3). All these archaeal clone sequences were affiliated with Euryarchaeota, Thaumarchaeota (Brochier-Armanet et al., 2008; Spang et al., 2010), and Crenarchaeota (Fig. 3).

The dominant group was Thaumarchaeota, which accounted for around 75% (407 out of 544) of all archaeal clone sequences retrieved in this study. The Thaumarchaeotal sequences were grouped with two ammonia-oxidizing archaea: Candidatus Nitrosopumilus maritimus (Konneke et al., 2005) and Candidatus Nitrososphaera gargensis (Hatzenpichler et al., 2008). The thaumarchaeotal clone sequences can be divided into two subgroups: Group 1 and Group 2, with the fomer including all the clone sequences retrieved from hot springs in this study (Fig. 3, Table 3).

The euryarchaeotal sequences included three subgroups: Methanomicrobiales, MKCS-J (Yan et al., 2006), and Halobacteriales; while the crenarchaeotal sequences consisted of Thermoproteales, Desulfurococcales and uncultured Crenarchaeota (Fig. 3; Table 3). Methanomicrobiales accounted for 17% (85 out of 489) of all archaeal clone sequences (Table 3). In the Methanomicrobiales group, forty-four clone sequences (1, 2, and 41 from GL52, GL63, and GL64, respectively) were closely related (97 - 99%) to archaeal clones (EF198052 and AY297990) retrieved from thermal origins (Chen et al., 2004; Chen et al., 2008). Thirty-five clone sequences were closely related (95-99%) to uncultured crenarchaeal clones retrieved from hot spring environments, such as Obsidian Pool in Yellowstone National Park (Spear et al., 2005), hot springs in Bulgarian (Tomova et al., 2010), and Iceland and Russia (Reigstad et al., 2008). Desulfurococcales contained seven clone sequences from GL63 and GL81 (Table 3; Fig. 3) and were closely related (95-98%) to clones originated from hot springs such as Iceland hot springs and Nevada Great Boiling Spring (Marteinsson et al., 2001; Costa et al., 2009).

Statistical Analysis

At the 97% OTU cutoff value, bacteria were more diverse than archaea (Table 1). The LIBSHUFF analysis showed that the Rongma soil (RMS) was statistically (P<0.05) different from the hot spring samples with respect to bacterial and archaeal communities (Fig. 4). There appeared to be some site-specific grouping. For example, 20 the bacterial sequences from Gulu appeared to be clustered together, and all archaeal sequences from Rongma formed a distinct group which was different from clusters of the Gulu sequences. The Mantel test showed that the temperature was not statistically correlated with the microbial diversity at either the OTU (r=0.241 and P=0.093 for bacteria, and r=-0.046 and P=0.446 for archaea) or the major group levels (r=0.016 and P=0.378 for bacteria, and r=0.226 and P=0.119 for archaea).

DISCUSSION

Microbial Diversity in Tibetan Hot Springs

The phylogenetic analysis showed that the bacterial communities were highly diverse and the predominant groups (Proteobacteria, Firmicutes, Chloroflexi, and Cyanobacteria), were similar to those in hot springs from the central Tibet (Lau et al., 2006; Lau et al., 2009) and from low-elevation and neutral-pH environments (Meyer- Dombard et al., 2005; Spear et al., 2005; Stout et al., 2009; Sayeh et al., 2010). This observation suggested that elevation was not a major factor influencing the bacterial distribution in global geothermal features.

The archaeal communities were less diverse than bacteria (as indicated by number of OTU and diversity indices in Table 2 and Table 3) but more so than those archaea observed in hot springs of central Tibet (Lau et al., 2006). One possible reason may be ascribed to the different techniques employed in the two studies: Lau and colleagues used the technique of 16S rDNA denaturing gradient gel electrophoresis (DGGE); instead, we used 16S rDNA cloning technique. These two techniques have different resolutions when used in characterizing microbial communities (Kisand and Wikner 2003).

The archaeal communities in many Tibetan hot springs studied here were dominated with Thaumarchaeota, showing uniqueness of Tibetan hot springs relative to other springs. This result is different from those of Lau et al., (2006) who did not observe any thaumarchaeotal sequences in hot springs from Tibet. There are several

21 possible reasons for such inconsistency. First, Thaumarchaeota is a newly proposed phylum (Spang et al., 2010) and it was previously considered as part of Crenarchaeota. Thus, it is possible that some of crenarchaeotal sequences reported in the Lau et al., study (2006) may have belonged to Thaumarchaeota. Second, the sampling sites in the Lau et al., (2006) study are different from those in this study, although they are all in central Tibet. It is conceivable that archaeal community may be different between different sites. Indeed, our data (Table 3) showed site-to-site variations in archaeal community structure. Third, the inconsistency between our results and those of Lau et al., (2006) may have been caused by different primers used. Whereas archaeal primers Arch21F/Univ958R were used in this study, the Archaea 344F/905R primers were used in the Lau et al., (2006) study. Although both sets of primers have been used for hot spring samples (Pearson et al., 2004; Meyer-Dombard et al., 2005), it is not clear which set is superior in terms of covering archaeal diversity in Tibetan hot springs. In addition, the dominance of the Thaumarchaeota sequences in our samples might conceivably be caused by PCR-cloning bias. However, the Thaumarchaeota sequences formed sample-specific grouping: all hot spring sequences were clustered within Group 1, and those from the soil sample (RMS) were grouped with Group 2. These results suggested that the dominance of Thaumarchaeota sequences in the Tibetan hot springs should be a true reflection of the archaeal community composition.

To date, the dominance of the Thaumarchaeota sequences in Tibetan hot spring springs has never been observed. Thaumarchaeota is a newly proposed phylum, which is phygenetically at the same level as Euryarchaeota, Crenarchaeota, and Korarchaeota (Brochier-Armanet et al., 2008; Spang et al., 2010). So far, the Thaumarchaeota phylum only contains ammonia oxidizers Nitrosopumilus maritimus and Nitrososphaera gargensis and clone sequences of putative ammonia-oxidizing isolates (Spang et al., 2010). Thus, the dominance of the Thaumarchaeota suggested that archaeal ammonia oxidation may be a key element-cycling process in the Tibetan hot springs. However, future work is necessary to confirm the functional traits of the Thaumarchaeota.

Response of Microbial Diversity to Temperature in the Tibetan Hot Springs

Previous studies have shown that diverse microbial communities can inhabit global hot springs of a wide temperature range (from ambient to boiling) (Barns et al., 1994; Pace 1997; Bonch-Osmolovskaya et al., 1999; Marteinsson et al., 2001; Kumar et al., 2004; Meyer-Dombard et al., 2005; Lau et al., 2006; Hall et al., 2008; Aditiawati et al., 2009; Lau et al., 2009; Mitchell 2009; Reigstad et al., 2009; Song et al., 2009; Aguilera et al., 2010; Jiang et al., 2010; Sayeh et al., 2010; Song et al., 2010; Vick et al., 2010), and that microbial diversity did not show a monotonic response to thermal stress (Lau et al., 2006; Mitchell 2009). Statistical analysis in this study also confirmed that there was no monotonic relationship between microbial communities and temperature in Titetan hot springs. Moreover, certain clone sequences from the Tibetan hot springs were even related to those of low-temperature origin (e.g. soil, lacustrine/marine sediments) (Fig. 2 and 3). This was especially true for the spring of the highest temperature, i.e. GL81 (81oC), where its microbial community was dominated by mesophilic Thaumarchaeota and gammaproteobacteria. This apparent inconsistency has been observed in a actinobacterial study in hot springs (Song et al., 2009), and it could be ascribed to two reasons: One reason may be due to potential contamination from sympatric soils. However, the LIBSHUFF analysis showed that the microbial communities in the investigated hot springs were distinctly different (P < 0.05) from that in the sympatric soil sample (Fig. 4). This distinct separation suggested that the bacterial and archaeal 16S rRNA gene clone sequences of the hot springs were compositionally different from those retrieved in the sympatric soil, and thus potential contamination from the sympatric soil can be ruled out. The other possible reason may be that microorganisms with identical 16S rRNA gene sequences may have distinct physiological properties (Jaspers and Overmann 2004). The calculated G+C contents of the 16S rRNA gene clone sequences for all samples showed an approximate positive correlation with the measured temperatures of the springs from which the mat or sediment samples were collected (graph not shown), as consistent with a previous report (Kimura et al., 2006). This correlation suggests that

23 the thermophilic nature of clones from the high-temperature spring (GL81) cannot be excluded. Apparent “mesophilic” Thaumarchaeota and gammaproteobacteria may actually contain some thermophilic members or may be thermophilic in functions.

Among other factors that may be more important in controlling microbial distribution, mineralogy and site isolation appears to have some influence. For example, all springs from Rongma contained abundant calcite (Table 1), and the proportion of cyanobacteria and Chloroflexi sequences was very low (<4%, Table 2). Whereas abundant calcite may be explained by lack or low levels of photosynthetic activity in the Rongma springs, it is equally possible that this site-specific different in microbial community composition may also be due to the distance effect. Indeed, the LIBSHUFF analyses suggest site-specific clustering of bot bacterial and archaeal sequences. Definitive resolution of these various controlling factors must await future investigations, where a much larger set of both geochemical and molecular data becomes available.

CONCLUSIONS

The bacterial communities in the studied Tibetan hot springs were predominated by Firmicutes, Proteobacteria, Cyanobacteria, and Chloroflexi, which was similar to other hot springs reported in literature. In comparison, archaeal diversities were less diverse than bacteria. The archaeal communities of the studied Tibetan hot springs mainly consisted of Thaumarchaeota clone sequences, which have never been reported to be a dominant group in any other hot springs worldwide. Statistical analysis confirmed that temperature was not significantly correlated with the microbial diversity, suggesting that other factors, such as mineralogy and distance, may be important in controlling microbial distribution in the Tibetan hot springs.

ACKNOWLEDGEMENTS

This research was supported by grants from the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry, the Fundamental Research Funds for the Central Universities (2010ZY16) National Science Foundation of the United States (OISE 0968421) and China (40972211, 41030211, and 41002123), and Chinese Geological Survey (1212010818014). Three anonymous reviewers are acknowledged for their constructive comments which greatly improved the quality of this manuscript.

REFERENCES

Aditiawati P, Yohandini H, Madayanti F, Akhmaloka (2009) Microbial diversity of acidic hot spring (kawah hujan B) in geothermal field of kamojang area, west Java-Indonesia. Open Microbiol J 3: 58-66.

Aguilera A, Souza-Egipsy V, Gonzalez-Toril E, Rendueles O, Amils R (2010) Eukaryotic microbial diversity of phototrophic microbial mats in two Icelandic geothermal hot springs. Int Microbiol 13: 21-32.

Allewalt JP, Bateson MM, Revsbech NP, Slack K, Ward DM (2006) Effect of temperature and light on growth of and photosynthesis by Synechococcus isolates typical of those predominating in the Octopus Spring microbial mat community of Yellowstone National Park. Appl. Environ. Microbiol. 72: 544- 550.

Barns SM, Fundyga RE, Jeffries MW, Pace NR (1994) Remarkable archaeal diversity detected in a Yellowstone National Park hot spring environment. Proc Natl Acad Sci U S A. 91: 1609-1613.

Bonch-Osmolovskaya EA, Miroshnichenko ML, Slobodkin AI, Sokolova TG, Karpov GA, Kostrikina NA, Zavarzina DG, Prokof’eva MI, Rusanov II, Pimenov NV (1999) Biodiversity of Anaerobic Lithotrophic Prokaryotes in Terrestrial Hot Springs of Kamchatka Microbiology 68: 343–351.

Boomer SM, Noll KL, Geesey GG, Dutton BE (2009) Formation of multilayered photosynthetic biofilms in an alkaline thermal spring in Yellowstone National Park, Wyoming. Appl. Environ. Microbiol. 75: 2464–2475.

Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P (2008) Mesophilic crenarchaeota: proposal for a third archaeal phylum, the Thaumarchaeota. Nat. Rev. Microbiol. 6, 245-252.

Chen C-L, Macarie H, Ramirez I, Olmos A, Ong SL, Monroy O, Liu W-T (2004) Microbial community structure in a thermophilic anaerobic hybrid reactor degrading terephthalate. Microbiology 150: 3429-3440.

Chen CL, Wu JH, Liu WT (2008) Identification of important microbial populations in the mesophilic and thermophilic phenol-degrading methanogenic consortia. Water Res 42: 1963-1976.

Chung FH (1974) Quantitative interpretation of X-ray diffraction patterns of mixtures. I. Matrix-flushing method for quantitative multicomponent analysis Journal of Applied Crystallography 7: 519-525.

Costa KC, Navarro JB, Shock EL, Zhang CL, Soukup D, Hedlund BP (2009) Microbiology and geochemistry of great boiling and mud hot springs in the United States Great Basin. Extremophiles 13: 447-459.

Hall JR, Mitchell KR, Jackson-Weaver O, Kooser AS, Cron BR, Crossey LJ, Takacs- Vesbach CD (2008) Molecular characterization of the diversity and distribution of a thermal spring microbial community by using rRNA and metabolic genes. Appl. Environ. Microbiol. 74: 4910-4922.

Hatzenpichler R, Lebedeva EV, Spieck E, Stoecker K, Richter A, Daims H, Wagner M (2008) A moderately thermophilic ammonia-oxidizing crenarchaeote from a hot spring. Proc Natl Acad Sci U S A. 105: 2134-2139.

Hu X, Sun J, Yao Z, Rong F (2003) Research on the influence of geothermal activities and exploitation to geological environment in Tibet (in Chinese). Journal of Mountain Research 21: 45-48.

Huber T, Faulkner G, Hugenholtz P (2004) Bellerophon: a program to detect chimeric sequences in multiple sequence alignments. Bioinformatics 20: 2317-2319.

Hugenholtz P, Pitulle C, Hershberger KL, Pace NR (1998) Novel division level bacterial diversity in a Yellowstone hot spring. J Bacteriol 180: 366-376.

Jaspers E, Overmann J (2004) Ecological significance of microdiversity: identical 16S rRNA gene sequences can be found in Bacteria with highly divergent genomes and ecophysiologies. Applied and Environmental Microbiology 70: 4831-4839.

Jiang H, Dong H, Deng S, Yu B, Huang Q, Wu Q (2009) Response of archaeal community structure to environmental changes in lakes on the Tibetan Plateau, northwestern China. Geomicrobiol J 26: 289-297.

Jiang H, Dong H, Yu B, Ye Q, Shen J, Rowe H, Zhang C (2008) Dominance of putative marine benthic archaea in Qinghai Lake, northwestern China. Environ Microbiol 10: 2355-2367.

Jiang H, Huang Q, Dong H, Wang P, Wang F, Li W, Zhang C (2010) RNA-based investigation of ammonia-oxidizing archaea in hot springs of Yunnan Province, China. Appl. Environ. Microbiol. 76: 4538-4541.

Joshi AA, Kanekar PP, Kelkar AS, Shouche YS, Vani AA, Borgave SB, Sarnaik SS (2008) Cultivable bacterial diversity of alkaline Lonar Lake, India. Microb Ecol 55: 163-172.

Kanokratana P, Chanapan S, Pootanakit K, Eurwilaichitr L (2004) Diversity and abundance of Bacteria and Archaea in the Bor Khlueng Hot Spring in Thailand. J Basic Microbiol 44: 430-444.

Kimura H, Sugihara M, Kato, K, Hanada S (2006) Selective phylogenetic analysis targeted at 16S rRNA genes of thermophiles and hyperthermophiles in deep- subsurface geothermal environments. Appl. Environ. Microbiol. 72: 21-27.

Kimura H, Mori K, Nashimoto H, Hattori S, Yamada K, Koba K, Yoshida N, Kato K (2010) Biomass production and energy source of thermophiles in a Japanese alkaline geothermal pool. Environ Microbiol 12: 480-489.

Kisand V, Wikner J (2003) Limited resolution of 16S rDNA DGGE caused by melting properties and closely related DNA sequences. J Microbiol Methods 54: 183-191.

Konneke M, Bernhard AE, de la Torre JR, Walker CB, Waterbury JB, Stahl DA (2005) Isolation of an autotrophic ammonia-oxidizing marine archaeon. Nature 437: 543-546.

Kumar B, Trivedi P, Mishra AK, Pandey A, Palni LMS (2004) Microbial diversity of soil from two hot springs in Uttaranchal Himalaya. Microbiol Res 159: 141- 146.

Lau MCY, Aitchison JC, Pointing SB (2009) Bacterial community composition in thermophilic microbial mats from five hot springs in central Tibet. Extremophiles 13: 139-149.

Lau MCY, Jing H, Aitchison JC, Pointing SB (2006) Highly diverse community structure in a remote central Tibetan geothermal spring does not display monotonic variation to thermal stress. FEMS Microbiol Ecol 57 : 80-91.

Marteinsson VT, Hauksdottir S, Hobel CFV, Kristmannsdottir H, Hreggvidsson GO, Kristjansson JK (2001) Phylogenetic diversity analysis of subterranean hot springs in Iceland. Appl. Environ. Microbiol. 67: 4242-4248.

Meyer-Dombard DR, Shock EL, Amend JP (2005) Archaeal and bacterial communities in geochemically diverse hot springs of Yellowstone National Park, USA. Geobiology 3: 211-227.

Mitchell KR (2009). Dissertation. University of New Mexico. http://hdl.handle.net/1928/9315

Mitchell KR, Takacs-Vesbach CD (2008) A comparison of methods for total community DNA preservation and extraction from various thermal environments J Ind Microbiol Biotechnol 35: 1139-1147.

Mori K, Suzuki K-i (2008) Thiofaba tepidiphila gen. nov., sp. nov., a novel obligately chemolithoautotrophic, sulfur-oxidizing bacterium of the Gammaproteobacteria isolated from a hot spring. Int J Syst Evol Microbiol 58: 1885-1891.

Pace NR (1997) A Molecular View of Science 276: 734-740.

Pearson A., Huang Z., Ingalls A. E., Romanek C.S., Wiegel J., Freeman K.H., Smittenberg R.H., Zhang C.L. (2004). Nonmarine crenarchaeol in Nevada hot springs. Appl. Environ. Microbiol. 70: 5229–5237.

Portillo MC, Sririn V, Kanoksilapatham W, Gonzalez JM (2009) Differential microbial communities in hot spring mats from Western Thailand. Extremophiles 13: 321-331.

Reigstad LJ, Jorgensen SL, Schleper C (2009) Diversity and abundance of Korarchaeota in terrestrial hot springs of Iceland and Kamchatka ISME J 4: 346-356.

Reigstad LJ, Richter A, Daims H, Urich T, Schwark L, Schleper C (2008) Nitrification in terrestrial hot springs of Iceland and Kamchatka. FEMS Microbiol Ecol 64: 167-174.

Sako Y, Nunoura T, Uchida A (2001) Pyrobaculum oguniense sp. nov., a novel facultatively aerobic and hyperthermophilic archaeon growing at up to 97 degrees C. Int J Syst Evol Microbiol 51: 303-309.

Sayeh R, Birrien JL, Alain K, Barbier G, Hamdi M, Prieur D (2010) Microbial diversity in Tunisian geothermal springs as detected by molecular and culture- based approaches. Extremophiles 14: 501-514.

Schloss PD, Handelsman J (2005) Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl Environ Microbiol 71: 1501-1506.

Song Z-Q, Chen J-Q, Jiang H-C, Zhou E-M, Tang S-K, Zhi X-Y, Zhang L-X, Zhang C-LL, Li W-J (2010) Diversity of Crenarchaeota in terrestrial hot springs in Tengchong, China Extremophiles 14: 287–296.

Song Z, Jiang H, Zhi X, Zhang C, Dong H, Li W (2009) Actinobacterial diversity in hot springs in Tengchong (China), Kamchatka (Russia), and Nevada (U.S.A.). Geomicrobiol J 26: 256-263.

Spang A, Hatzenpichler R, Brochier-Armanet C, Rattei T, Tischler P, Spieck E, Streit W, Stahl DA, Wagner M, Schleper C (2010) Distinct gene set in two different lineages of ammonia-oxidizing archaea supports the phylum Thaumarchaeota. Trends Microbiol 18: 331-340.

Spear JR, Walker JJ, McCollom TM (2005) Hydrogen and bioenergetics in the Yellowstone geothermal ecosystem. Proceedings of the National Academy of Sciences of the United States of America 102: 2555-2560.

Stout LM, Blake RE, Greenwood JP, Martini AM, Rose EC (2009) Microbial diversity of boron-rich volcanic hot springs of St. Lucia, Lesser Antilles. FEMS Microbiol Ecol 70: 402-412.

Tomova I, Stoilova-Disheva M, Lyutskanova D, Pascual J, Petrov P, Kambourova M (2010) Phylogenetic analysis of the bacterial community in a geothermal spring, Rupi Basin, Bulgaria. World J Microbiol Biotechnol 26: 2019-2028.

Vick TJ, Dodsworth JA, Costa KC, Shock EL, Hedlund BP (2010) Microbiology and geochemistry of Little Hot Creek, a hot spring environment in the Long Valley Caldera. Geobiology 8: 140-154.

Wu X-Y, Zheng G, Zhang W-W, Xu X-W, Wu M, Zhu X-F (2010) Amphibacillus jilinensis sp. nov., a facultatively anaerobic, alkaliphilic bacillus from a soda lake. Int J Syst Evol Microbiol 60: 2540-2543.

Yan B, Hong K, Yu ZM (2006) Archaeal communities in mangrove soil characterized by 16S rRNAgene clones. The Journal of Microbiology 44: 566-571.

Zhang G, Dong H, Xu Z, Zhao D, Zhang C (2005) Microbial Diversity in Ultra-High- Pressure Rocks and Fluids from the Chinese Continental Scientific Drilling Project in China Appl. Environ. Microbiol. 71: 3213-3227.

Zhang CL, Ye Q, Huang Z, Li W, Chen J, Song Z, Zhao W, Bagwell C, Inskeep WP, Ross C, Gao L, Wiegel J, Romanek CS, Shock EL, Hedlund BP (2008) Global occurrence of archaeal amoA genes in terrestrial hot springs. Appl. Environ. Microbiol. 74:6417-6426.

Table 1. Location and description of the investigated soil and ten hot springs on the Tibetan Plateau, China. Samples are coded as: RM26=Rongma hot spring with temperature of 26.2oC; RM45=Rongma hot spring with temperature of 45.1oC; RM55=Rongma hot spring with temperature of 55.6oC; RM64=Rongma hot spring with temperature of 64.1oC; RMS = The sympatric soil near RM26; JW41=Jiwa hot spring with temperature of 41.7oC; JW56=Jiwa hot spring with temperature of 56oC; GL52=Gulu hot spring with temperature of 52oC; GL63=Gulu hot spring with temperature of 63.8oC; GL64=Gulu hot spring with temperature of 64oC; GL81=Gulu hot spring with temperature of 81.2oC.

GPS coordinates Elevation Temp wt% mineralogy Sample pH Sample type (N/E) (m) (oC) Calcite Quartz Albite Muscovite Illite Biotite Others

RMS 32 o57' 43.8" 86 o35' 49.2" 4725 NA NA Brown soil UD ++ UD ++ +++ + ++

RM26 32o57' 46.5" 86 o 35' 48.1" 4719 26.2 6.8 Gray sediment ++++ 11 UD + UD UD ++

RM45 32 o57' 45.3" 86 o35' 49.1" 4715 45.1 7.2 Brown sediment ++++ UD UD UD UD UD ++

RM55 32 o57' 44.6" 86 o35' 48.3" 4720 55.6 7.1 Green mat NA NA NA NA NA NA NA

RM64 32 o57' 46.3" 86 o35' 47.9" 4718 64.1 6.9 Calcareous sinter ++++ + UD UD UD UD +

JW41 33 o08' 23.3" 86 o49' 53.6" 4618 41.7 6.8 Green sediment +++ ++ UD UD UD UD +

JW56 33 o08' 23.4" 86 o49' 54.4" 4630 56.0 6.5 Green sediment +++ ++ UD UD UD + ++

GL52 30 o52' 12.9" 91 o36' 44.7" 4735 52.0 NA Green mat NA NA NA NA NA NA NA

GL63 30 o52' 35.6" 91 o36' 35.2" 4715 63.8 NA Green sediment UD +++ +++ UD UD UD ++

GL64 30 o52' 35.4" 91 o36' 35.0" 4711 64.0 NA Brown sediment + +++ ++ + UD UD +

GL81 30 o52' 34.1" 91 o36' 40.0" 4726 81.2 8.2 Black sediment NA NA NA NA NA NA NA

NA: not available; UD: undetectable Mineralogy: ++++: very abundant; +++: abundant, ++, moderate, +: present

Table 2. Ecological estimates and major group affiliation of the bacterial 16S rRNA gene clone sequences retrieved from the soil and ten hot springs on the Tibetan Plateau.

Community RMS RM26 RM45 RM55 RM64 JW41 JW56 GL52 GL63 GL64 GL81 No. of clones 24 49 31 24 50 47 21 48 39 45 37 Coverage (%) 64 77.6 100 79.2 90 97.9 85.7 95.8 100 91.1 100 No. of observed OTUs 11 16 2 10 12 5 5 6 4 10 2 Chaol 25 43.5 2 20 14.5 5 6.5 6.5 4 16 2 (15.5,104) (22.7,129.5) (2,2) (11.9,62.2) (12.4,29) (5,5) (5.1,20.1) (6.0,14.3) (4,4) (11,47.9) (2,2) Shannon (H') 2.18 2.17 0 2.1 1.88 1.43 0.87 0.87 0.8 1.94 0.66 (1.78,2.59) (1.84,2.5) (-0.03,0.03) (1.81,2.40) (1.57,2.19) (1.3,1.6) (0.38,1.35) (0.54,1.20) (0.49,1.11) (1.71,2.18) (0.57,0.76) Major group affiliation Number of clones (Relative percentage in each clone library) Acidobacteria 2(8.3%) 9(18.0%) 1(2.2%) Bacteroidetes 1(2.0%) 12(24.0%) 3(6.3%) 2(5.1%) Proteobacteria 9(37.5%) 8(16.3%) 19(79.2%) 6(12.0%) 10(21.3%) 2(9.5%) 38(79.2%) 8(17.8%) 37(100%) Nitrospirae 1(2.2%) Firmicutes 13(54.2%) 38(77.6%) 29(93.5%) 5(20.8%) 21(42.0%) 22(46.8%) 3(14.3%) 3(7.7%) 20(44.4%) Aquificae 3(6.7%) Cyanobacteria and Chloroflexi 1(2.0%) 2(4.0%) 15(31.9%) 16(76.2%) 7(14.6%) 34(87.2%) 11(24.4%) Planctomycetes 2(6.5%) Thermodesulfobacteria 1(2.2%) Unclassified Bacterium 1(2.0%)

Table 3. Ecological estimates and major group affiliation of the archaeal 16S rRNA gene clone sequences retrieved from the soil and ten hot springs on the Tibetan Plateau.

Community RMS RM26 RM45 RM55 RM64 JW41 JW56 GL52 GL63 GL64 GL81 No. of clones 55 49 47 56 45 48 55 46 47 49 47 Coverage (%) 96.4 98 100 100 88.9 100 100 93.5 87.2 98 100 No. of observed OTUs 6 5 2 1 8 1 1 3 8 4 2 Chaol 7.5 5 2 1 13 1 1 3 13 4 2 (6.1,22.1) (5,5) (2,2) (1,1) (8.8,40.1) (1,1) (1,1) (3,3) (8.8,40.1) (4,4) (2,2) Shannon (H') 0.62 0.76 0.1 0 0.99 0 0 0.28 1.27 0.65 0.46 (0.3,0.9) (0.5,1.1) (0,0.3) (0,0.3) (0.6,1.4) (0,0.03) (0,0.03) (0.04,0.52) 1.0,1.6) (0.4,1.9) (0.3,0.6) Major group affiliation Number of clones (Relative percentage in each clone library) Thaumarchaeota 5(10.2%) 47(100%) 56(100%) 34(75.6%) 48(100%) 55(100%) 43(93.5%) 22(46.8%) 3(6.1%) 39(83.0%) Euryarchaeota MKCS-J 2(4.1%) 1(2.2%) Halobacteriales 4(8.9%) 2(4.3%) Methanomicrobiales 55(100%) 41(83.7%) 1(2.1%) 2(4.3%) 41(83.7%) Crenarchaeota Uncultured Crenarchaeota 1(2.0%) 6(13.3%) 2(4.3%) 18(38.3%) 5(10.2%) 3(6.4%) Desulfurococcales 3(6.4%) 5(10.6%)

Figure captions Figure 1. A geographic map showing the sampling locations on the Tibetan Plateau, China. JW41 and JW56 were from Site 1; RM26, RM45, RM55, RM64, and RMS were from Site 2; and GL52, GL63, GL64, and GL81 were from Site 3. Figure 2. A): Neighbor-joining tree (partial sequences, ~700bp) showing the phylogenetic relationships of bacterial 16S rRNA gene sequences cloned from the hot spring samples on the Tibetan Plateau to closely related sequences from the GenBank database. One representative clone type within each OTU is shown, and the number of clones is shown at the end (after the GenBank accession number). The number of clones is omitted if there is only one clone within a given OTU. Clone sequences from this study are coded as follows for the example of RM64-B001 (HQ287172) 9: RM64, sample name; B, bacterium; 001, number of clone type; HQ287172, GenBank accession number; 9, number of clone sequences. Scale bar indicates Jukes-Cantor distances. Aquifex pyrophilus is used as an outer group, and a single tree showing all bacterial sequences is created. B) and C): subtrees for Proteobacteria and Firmicutes. Figure 3. Neighbor-joining tree (partial sequences, ~700bp) showing the phylogenetic relationships of archaeal 16S rRNA gene sequences cloned from hot spring samples on the Tibetan Plateau to closely related sequences from the GenBank database. The same algorithms as those for the bacterial tree were used. Figure 4. A) Clustering of the different bacterial 16S rRNA gene clone libraries based on ΔCxy values obtained from the LIBSHUFF analysis. Unweighted-pair group method with average linkages in MEGA4.1 was used to construct the tree. The parameter ΔCxy in the LIBSHUFF represents the difference in coverage of two clone libraries (the larger ΔCxy, the greater dissimilarity between the given clone libraries). Clone libraries are coded as follows for the example of GL81B/A: GL, sample site, Gulu; 81, temperature in Celsius; B, bacterium; A, archaea. B) The UniFrac metric tree showing the difference of archaeal 16S rRNA gene clone libraries from ten Tibetan hot springs. The software for the analysis was available at http://whitman.myweb.uga.edu/libshuff.html.

Figure 1.

China Tibet Tibet Nima Anduo Bange Naqu Shenza Lasa

87O 88O 89O 90O 91O 92O

33O Site1 Site2 Nima 32O Anduo

Naqu Bange

31O Shenza Site3

Figure 2A.

9 Proteobacteria

RM26-B040 (HQ287156) 99 Waste water bacterium clone (EU399664) Unclassified Bacteria 91 Germany contaminated soil clone (GU236072) Italy Lake Specchio di Venere clone (FN687043) 99 RM64-B001 (HQ287172) 9 Tibet hot springs thermophilic microbial mats clone (EF205553) 52 99 RMS-B015 (HQ287136) 2 Lead-zinc mine tailing site clone (EF612355) Acidobateria 65 Desert agricultural soil clone (AY493931) 99 GL64-B031 (HQ287212) Russia Kamchatka thermal pools clone (GQ328483) 58 Yellowstone hot spring clone (AF027004) 70 GL52-B022 (HQ287197) 4 99 Thailand Bor Khlueng Hot Spring clone (AY555774) 82 China Yunnan Tengchong hot spring clone (DQ886533) 99 61 Tibet hot springs thermophilic microbial mats clone (EF205562) Chloroflexi 99GL63-B001 (HQ287200) 30 GL64-B011 (HQ287206) 11 52 Bulgarian geothermal springs clone (FM164953)

Firmicutes

99 GL63-B010 (HQ287201) 2 RM64-B011 (HQ287174) 2 78 Tibet hot springs thermophilic microbial mats clone (EF205451) 99 96 RM64-B022 (HQ287178) 4 Tibet hot springs thermophilic microbial mats clone (EF205446) 99 Yellowstone Mammoth Hot springs clone (AF445706) 93 RM64-B012 (HQ287175) 5 99 GL52-B030 (HQ287198) Bacteroidetes 63 Yellowstone Mammoth Hot springs clone (AF445665) 57 Tibet hot springs thermophilic microbial mats clone (EF205443) GL52-B021 (HQ287196) 2 97 99 China water treatment plant clone (GQ844365) 94 RM64-B041 (HQ287180) Chryseobacterium sp. NX12 (EF601827) 99 RM26-B010 (HQ287146) Ireland feces and sewage samples clone (EU573844) 98GL52-B034 (HQ287199) 3 99 Tibet hot springs thermophilic microbial mats clone (EF205557) 99 Costa Rica geothermal springs clone (EF545646) Yellowstone Synechococcus sp. TS-91 (AY884060) 96RM64-B014 (HQ287176) Tibet hot springs thermophilic microbial mats clone (EF205459) 99 Yellowstone alkaline thermal spring clone (FJ206555) 72 GL63-B023 (HQ287203) 4 99 Thailand hot spring mats clone (EU376425) Asian geothermal springs clone (DQ131174) 99RM26-B021 (HQ287151) China Dongping Lake bacterium clone (FJ612245) 81 Lake Kastoria water column and sediment clone (FJ204882) Cyanobacteria 99 JW41-B009 (HQ287188) Mexico lake Texcoco clone (FJ152939) 79 Greenland alkaline, cold ecological niche clone (AJ431339) JW56-B011 (HQ287189) 14 99 99 55 Tibetan Lake clone (HM128346) 83 84 India Assam Gorompani warm spring clone (DQ512815) RM64-B043 (HQ287181) 77 Xenococcus sp. CR_L15 (EF545631) Philippines hot spring Thermophilic microbial mats clone (EF429517) 99 Italy Pantelleria Island Lake Specchio di Venere clone (FN687082) JW56-B011 (HQ287189) 16 99 Tibet Daggyai Tso geothermal field clone (EF208609) 99 99 Hawaiian lava cave microbial mat clone (EF032785) GL64-B020 (HQ287209) Nitrospirae 99 Tibet hot springs thermophilic microbial mats clone (EF205519) RM45-B051 (HQ287160) 2 99 Red Layer Microbial Observatory Sites clone (FJ206986) Planctomycetes 99 Yellowstone alkaline thermal spring clone (FJ206631) GL64-B013 (HQ287207) 54 Caldimicrobium rimae strain DS (EF554596) Japanese alkaline geothermal pool clone (AB462554) 99 Thermodesulfobacteria China Yunnan Tengchong hot spring clone (AY082369) Yellowstone hot spring clone (AF027096) 99 Yellowstone geothermal ecosystem clone (AY862050) GL64-B028 (HQ287211) 3 Indian Ridge deep-sea hydrothermal vent field clone (AY251061) Yellowstone Hydrogenobacter sp. BB4L1B (AJ320216) Aquificae 99 Tibet hot springs thermophilic microbial mats clone (EF205505) Aquifex pyrophilus (M83548) 0.05

Figure 2B.

99 RMS-B003 (HQ287139) 2 53 USA uranium contaminated soil clone (DQ125880) JW41-B010 (HQ287185) 10 99 China Dongping Lake bacterium clone (GU208356) Petroleum refinery clone (FJ439052) 99 RM64-B010 (HQ287173) 2 75 RM26-B013 (HQ287149) Ariake Sea Coastal Sediments clone (AB559989) Betaproteobacteria 99 97 RM64-B005 (HQ287182) 89 Yellowstone Hillside hot spring clone (FJ207030) 99RM26-B001 (HQ287144) Lake Michigan proteobacterium clone (EU640429) 62 Ramlibacter sp. P-8 (AM411936) 52 99 RMS-B021 (HQ287138) 90Cote d'IvoireYmoussoukro lake clone (GU291523) 99 China deep ice sheets clone (GU246952) 62 The oldest ice on the Earth clone (EF127600) JW56-B039 (HQ287193) 99 South Africa Kalahari Shield subsurface water clone (DQ230964) 99 Silanimonas lenta (NR_025815) RM64-B032 (HQ287179) 2 99RMS-B042 (HQ287142) 89 USA Minnesota gamma proteobacterium clone (AY921928) Panama Lake Gatun clone (EU803464) 99 RM55-B009 (HQ287171) 77 USA Wisconsin, Lake Mendota clone (FJ828370) 89 RM64-B008 (HQ287183) 99 Taiwan Taroko proteobacterium clone (AY874095) Lake Michigan uncultured proteobacterium clone (EU639719) 99 Thiofaba tepidiphila (AB304258) GL52-B002 (HQ287195) 78India Khir Ganga Hot Spring Microbial Mat clone (EU037218) Gammaproteobacteria 99 GL64-B001 (HQ287204) 5 73 Acinetobacter lwoffii (HM163482) RM55-B001 (HQ287162) 4 99 GL81-B018 (HQ287215) 14 52 Tibet Qinghai Lake clone (HM127449) RM55-B011 (HQ287164) 4 99 Taiwan saltern soil clone (FJ348430) 97 88 Hawaiian volcanic deposits clone (DQ490334) RM55-B010 (HQ287163) 5 99 Klebsiella sp. cl40 clone (GU003816) RM55-B012 (HQ287165) 3 RM26-B034 (HQ287145) 3 99 GL81-B001 (HQ287214) 23 Aeromonas salmonicida subsp. VA_K2-M7 (GQ996602) GL52-B001 (HQ287194) 37 Arctic streams clone (FJ849482) 99 RMS-B014 (HQ287135) 2 Spain volcanic environments clone (EF447047) RM55-B002 (HQ287167) 99 Greece Marathonas Reservoir water column clone (GQ340337) Alphaproteobacteria HCH Contaminated soil clone (EF494192) 54 RMS-B050 (HQ287143) 99 JW56-B027 (HQ287192) 99 Hawaiian hotspot clone (AF513446) Archangium sp. 565 (AM489541) 99 Ohio River Sediments clone (EF393274) 66 RMS-B034 (HQ287141) RMS-B013 (HQ287134) 99 Urban aerosols harbor clone (DQ129588) RM55-B022 (HQ287168) 53 99 Chile Rapel reservoir artificial lake sediment clone (EF192881) Deltaproteobacteria China Hangzhou rice field soil clone (FM956228) RM26-B043 (HQ287157) 3 89 99 Taiwan uncultured bacterium clone (AF254389) 70 GL64-B023 (HQ287210) 3 99 Russia Kamchakta thermal pools clone (GQ328526) 74 Tibet hot springs thermophilic microbial mats clone (EF205549)

0.02

Figure 2C.

RM64-B015 (HQ287177) 21 99 JW56-B029 (HQ287190) 97 RM26-B025 (HQ287153) 2 54 Japan Tokyo rice paddy soil clone (AB486695) 71 Baltic Sea sediment clone (EF460100) RM26-B011 (HQ287147) 15 54 85 RM26-B051 (HQ287158) 99 Janpan bacterium clone (AB294746) RM26-B024 (HQ287152) 99 China Jidong oil field bacterium clone (EU735637) 70 RM55-B023 (HQ287169) 99 China Lake Taihu water and sediment clone (FJ755751) Northern Norway Spitsbergen soil clone (EF034733) 98 GL64-B010 (HQ287205) 9 63 New Zealand high temperature mud pool clone (EU544532) 89 GL64-B019 (HQ287208) 10 99 Firmicutes bacterium clone (GQ406191) 83 RM26-B052 (HQ287159) 77 Uncultured bacterium clone (EU828369) Andes Puna de Atacama Socompa Volcano clone (FJ592915) 54 Mining waste piles clone (AJ295664) Clostridia JW41-B012 (HQ287187) 8 84 RM26-B032 (HQ287155) 99 Hydrogen producing bioreactor clone (EU828406) 92 RM26-B026 (HQ287154) RM45-B001 (HQ287161) 29 99 Germany Elbe River clone (AF150697) Firmicutes JW41-B001 (HQ287184) 14 92 RM26-B012 (HQ287148) 6 Clostridium algidixylanolyticum strain SPL73 (NR 028726) 99 95 Clostridium sp. U201 (AB114228) 99 RM55-B015 (HQ287166) 3 99 Northwest Europe Scheldt estuary ferric clone (DQ677006) 87 RM26-B015 (HQ287150) 10 78 JW56-B017 (HQ287191) 2 High-temperature North Sea oil-field clone (DQ647100) 99 Japan high-temperature petroleum clone (HM041935) 99 GL63-B012 (HQ287202) 3 53 Mexico lake Texcoco clone (FJ152958) GL64-B042 (HQ287213) Sedimentibacter sp. B4 (AY673993) 99 Sedimentibacter sp. C7 (AY766466) RM55-B024 (HQ287170) 99 NRIC Lactic Acid Bacteria clone (FJ849487) 86 Sporolactobacillus nakayamae (AB362633) RMS-B010 (HQ287133) 2

99 Amphibacillus sp. Y1 (FJ169626) Great Salt Lake clone (HM057164) Bacillales 99 91 RMS-B019 (HQ287137) 6 99 Hungary Kiskunsag soda lake clone (AJ606037) Poland Lower Silesia arsenic resistant clone (EF491962) 69 RMS-B030 (HQ287140) 5 99 Exiguobacterium sp. LLN (DQ333298) 74 Exiguobacterium aurantiacum (DQ019166)

0.02

Figure 3.

GL64-A027 (HQ287127) Anaerobic thermophilic phenol-degrading enrichment clone (EF198052) GL52-A034 (HQ287115) 67 GL64-A001 (HQ287125) 40

99 Methanomicrobiales 99 GL63-A049 (HQ287122) 2 Thermophilic and anaerobic terephthalate-degrading sludge clone (AY297990) 99 Japan Toyama mesophilic digested sludge Methanosarcinales clone (AB353217) 99 RM26-A014 (HQ287094) 2 Israel Lake Kinneret profundal sediment clone (AM182005) 76 99 Japan Shizuoka petroleum contminated soil clone (AB161325) 88 RM26-A001 (HQ287092) 37

94 Long-chain fatty acids-degrading methanogenic consortium clone (AB244305) Euryarchaeota Taiwan rice field soil Methanolinea sp clone (AB447467) 89 99 Methanoculleus sp. (AJ133793) 98 RM26-A015 (HQ287095) 2 97 China Hangzhou anoxic soil Methanomicrobiaceae clone (AM778323) 99 RM64-A009 (HQ287109) RM26-A017 (HQ287096) 2 99 MKCS-J 99 Egypt Lake Manzallah sediment clone (AB355121) 99 China Hainan Island mangrove soil clone (DQ363834) 64 RM26-A013 (HQ287093) 2

93 Egypt hypersaline lakes Wadi An Natrun clone (DQ432475) Halobacteriales 99 Tibetan Plateau Charhan salt lake water clone(FJ155651) RM64-A016 (HQ287104) 2 99 Egypt hypersaline lakes Wadi An Natrun clone (DQ432489) 99 95 GL63-A053 (HQ287123) Mexico Texcoco alkaline-saline soil haloarchaeon clone (EF690637) 97 Turkey Salt Lake clone Haloterrigena sp A82 (DQ309080) 98 GL63-A036 (HQ287121) 61 China Inner Mongolia Baerhu Soda Lake water clone (AB125107) Candidatus Nitrosocaldus yellowstonii Unclassified Thaumarchaeota Unclassified 70Mexico Texcoco alkaline saline soil clone (FJ784305) 99 RMS-A001 (HQ287087) 48 75 RMS-A042 (HQ287090) 3 69 Uncultured archaeon clone (EF022370) 89 Uncultured archaeon clone (EF020664) 99 RMS-A018 (HQ287088) 2 99 Candidatus Nitrososphaera gargensis clone (EU281336) Chile Atacama Desert quartz clone (FJ638286)

99 RMS-A006 (HQ287091) Thaumarchaeota 97 South Africa Witwatersrand Basin Au mine water clone (AY187899) 81 Mexico Texcoco alkaline saline soil clone (FJ784309) Candidatus Nitrosopumilus maritimus (DQ085097) 52 South Pacific Gyre oligotrophic marine sediments clone (FJ487546) 99 Juan de Fuca Ridge CoAxial segment crenarchaeote clone (EF067905) Antarctica Weddell Sea bathypelagic zone sediments clone (EF069358) Nitrosopumilales 91 RM64-A001 (HQ287101) 34 GL81-A001 (HQ287130) 39 RMS-A041 (HQ287089) RM55-A001 (HQ287100) 56 JW41-A001 (HQ287110) 48 99 RM45-A001 (HQ287098) 47 GL64-A028 (HQ287128) 3 RM26-A013 (HQ287093) 5 JW56-A001 (HQ287111) 55 GL63-A010 (HQ287117) 22 GL52-A001 (HQ287112) 43 95 GL81-A017 (HQ287131) 3 85 GL63-A013 (HQ287118) 17 99 Yellowstone National Park Obsidian Pool clone (AY861950) Icelandic hot springs clone (DQ441516) 99 GL64-A011 (HQ287126) 5 GL63-A025 (HQ287119) 99 Bulgarian Velingrad Varvara hot spring clone (FN296155) 73 Japan Toyama mesophilic digested sludge clone (AB353218)

Okhotsk Sea subseafloor sediments clone (AB094546) Uncultured Crenarchaeota 99 Bulgaria Varvara hot spring clone (FM994134) 51 RM64-A039 (HQ287107) RM64-A011 (HQ287102) Yellowstone National Park hydrothermal features clone (EU239997) 99 99 Yellowstone National Park Obsidian Pool clone (AY861963) RM64-A035 (HQ287105) Crenarchaeota 50 RM64-A036 (HQ287106) 51 RM64-A043 (HQ287108) 2 64 83 Russia Uzon Caldera Zavarzin Spring clone (GQ328181) RM26-A043 (HQ287097) 71 Mexico Tamaulipas phreatic limestone sinkholes clone (FJ902276) California Little Hot Creek hot spring sediments clone (EU924237) 99 GL52-A015 (HQ287114) 96 Yellowstone National Park Obsidian Pool clone (AY861957) Yellowstone Heart Lake hot spring sediments clone (EU240001) 99 GL52-A011 (HQ287113) 99 Nevada Mud Hot Springs sediments clone (EU635908) 99 GL63-A007 (HQ287124) Thermoproteales Pyrobaculum oguniense (AB087402) 99 Southern Okinawa Trough Yonaguni Knoll IV chimney structure clone (AB235330) 9782 East Pacific Rise Ridge flank abyssal hills sea floor clone (DQ417485) GL63-A001 (HQ287116) Desulfurococcales 67 55 GL81-A036 (HQ287132) 5 GL63-A029 (HQ287120) 99 Icelandic hot springs clone (AF361213) 81 Nevada Great Boiling Spring sediments clone (DQ490017) Aquifex pyrophilus (M83548) 0.1

Figure 4A.

GL81B GL63B GL64B RM55B JW41B RM26B GL52B RMSB RM64B JW56B RM45B

8 6 4 2 0

Figure 4B.

RM55A RM64A RM26A RM45A JW41A JW56A GL81A GL64A GL63A GL52A RMSA

5 4 3 2 1 0

Figure S1.

CHAPTER 3:

Archaeal and Bacterial Diversity in Acidic to Circumneutral Hot Springs in the Philippines

Running head: Microbial ecology in six hot springs of the Philippines

Qiuyuan Huang1, Hongchen Jiang2*, Brandon R. Briggs1, Shang Wang3, Weiguo Hou3, Gaoyuan Li3, Geng Wu2, Ramonito Solis4, and Carlo A. Arcilla5, Teofilo Abrajano6, and Hailiang Dong1,3*

1: Department of Geology and Environmental Earth Science, Miami University, Oxford, OH 45056 2: State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Wuhan 430074, China 3: State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Beijing 100083, China 4: Resource Management Department, Energy Development Corporation Ortigas Center Pasig 1605, the Philippines 5: National Institute of Geological Sciences, University of the Philippines, Diliman, Quezon 1101, the Philippines 6: Division of Earth Sciences, National Science Foundation, Arlington, VA 22230, USA

*Corresponding authors: Hailiang Dong: [email protected]

Published in FEMS Microbiology Ecology (2013) 85(3): 452-464.

ABSTRACT The microbial diversity was investigated in sediments of six acidic to circumneutral hot springs (Temperature: 60-92oC, pH 3.72-6.58) in the Philippines using an integrated approach that included geochemistry and 16S rRNA gene pyrosequencing. Both bacterial and archaeal abundances were lower in high-temperature springs than in moderate-temperature ones. Overall, the archaeal community consisted of sequence reads that exhibited a high similarity (nucleotide identity >92%) to phyla Crenarchaeota, Euryarchaeota, and unclassified Archaea. The bacterial community was composed of sequence reads moderately related (nucleotide identity >90%) to 17 phyla, with Aquificae and Firmicutes being dominant. These phylogenetic groups were correlated with environmental conditions such as temperature, dissolved sulfate and calcium concentrations in spring water, and sediment properties including total nitrogen, pyrite, and elemental sulfur. Based on the phylogenetic inference, sulfur metabolisms appear to be key physiological functions in these hot springs. Sulfobacillus (within phylum Firmicutes) along with members within Sulfolobales were abundant in two high- temperature springs (>76oC), and they were hypothesized to play an important role in regulating the sulfur-cycling under high temperature conditions. The results of this study improve our understanding of microbial diversity and community composition in acidic to circumneutral terrestrial hot springs and their relationships with geochemical conditions.

Keywords Biodiversity, hot springs, pyrosequencing, thermoacidophiles, Philippines

INTRODUCTION

Acidic hot springs are considered extreme environments, yet they contain a diverse array of thermoacidophilic microorganisms capable of surviving and functioning under such conditions (Brock et al., 1972, Brock 1978; Stetter, 1996, Rothschild & Mancinelli, 2001). Multiple investigations have shown that thermoacidophiles in terrestrial geothermal features are of great importance because of their potential implications for early life on Earth (Konhauser et al., 2003) and extraterrestrial biology (Cavicchioli, 2002), as well as applications in biotechnology (Edwards et al., 2000) and bioremediation (Norris et al., 2000, Gonzalez-Contreras et al., 2012). Over the last decade, several culture-independent studies have investigated microbial communities in acidic hot springs, such as those in Tengchong, China (Hou et al., 2013), Iceland (Kvist et al., 2007), Yellowstone National Park (YNP) (Jackson et al., 2001, Inskeep et al., 2010, Kozubal et al., 2012a, 2012b, Macur et al., 2013) and Lassen Volcanic National Park (Siering et al., 2006, Wilson et al., 2008) in the United States.

The Philippines, located in the Circum-Pacific rim of volcanic systems, harbors a great number of geothermal features and ranks the second to the United States in generating geothermal energy (Dolor, 2005). Previous microbial studies associated with the Philippines hot springs have either focused on the isolation and cultivation of novel thermophiles (Itoh et al., 1999, Itoh et al., 2003), targeted the microbial diversity of specific groups of microorganisms (Jing et al., 2005, Lacap et al., 2005), or used conventional molecular techniques (e.g., DGGE and clone library) (Jing et al., 2005, Lacap et al., 2007, Lantican et al., 2011). Despite these extensive investigations, there is still a lack of comprehensive understanding on the microbial diversity and microbial community structure in the Philippines acidic to circumneutral hot springs. Here we report on the microbial diversity, community composition, and their relationships with environmental variables in six previously unstudied hot springs in the Philippines. We hypothesize that locally environmental conditions are important in shaping microbial community composition in the Philippines hot springs. Globally, important differences in microbial community composition exist among hot springs, and both geographic distance and environmental conditions are important in controlling microbial community composition.

The objectives of this study were (i) to investigate the microbial diversity and community structure of six Philippines hot springs within a range of temperatures (60- 92oC) and pH conditions (3.72-6.58); (ii) to assess the relationships between microbial community composition and environmental conditions (e.g., water geochemistry, mineralogy) in these acidic to circumneutral hot springs; and (iii) to compare the microbial composition in Philippines hot springs with others in Tengchong (China) and YNP (USA). To achieve these objectives and test the above hypotheses, an integrated approach was employed, including 16S rRNA gene-based 454 pyrosequencing and geochemical analyses.

MATERIALS AND METHODS

Site description and sampling

The study area is part of the Bacon-Manito Geothermal Production Field (BGPF), located on the southeastern chain of the Bicol volcanic arc system in the Philippines (Figure 1). This area is structurally controlled by numerous fault splays and dike intrusions in a shear zone known as the Bacman Fault Zone (BFZ) and contains a number of geothermal features (Dimabayao, 2012). The lithology of the study area is composed of three major units: the Gayong Sedimentary Formation (GSF), the Pocdol Volcnics Formation (PVF), and the Cawayan Intrusive Complex (CIC) (Dimabayao, 2012). GSF is the oldest unit that primarily comprises calcareous clastics and sedimentary breccias. The overlying PVF mainly comprises intensively altered andesite lavas and tuff breccias. The CIC is composed of basalt, microdiorite, and plutonic dikes that intrude the GSF and PVF.

Six hot springs in the BGPF were selected for this study: INA-1 (Inang Maharang Town), MAL-1(Malangto Town), NAG-7 (Naghaso Town), BAG-2 (Balbagon Town), BAL-0 (Balasbas Town), and BAL-1(Balasbas Town) (Figure 1). A portable GPS unit (eTrex H, Garmin, USA) was used to determine the geographical locations of these springs. Temperature and pH of hot spring waters were measured using a portable pH/temperature meter (LaMotte 5 Series, Chestertown, MD, USA). Water samples were collected for laboratory measurements by filtering ~50 mL of spring water through sterile 0.22 µm polycarbonate filters (Millipore, USA). At each hot spring, sediment samples

were collected with sterile spatulas and spoons, and homogenized in a pre-sterilized aluminum pan. The homogenization procedure was to ensure that the solid phase geochemical and molecular data were obtained from the same sample to allow subsequent correlation analysis between microbial community composition and environmental conditions. Multiple aliquots of sediment samples were taken from the homogenized sediments in the aluminum pan and placed into 1.5 mL or 50 mL polypropylene centrifuge tubes. Sediment samples for microbial analysis were immediately frozen in dry ice upon collection, remained frozen during transportation, and stored at -80oC in the laboratory until further analysis.

Geochemical and mineralogical analyses

Spring water was filtrated (0.22 µm) in the field and measured with a HACH colorimeter (model CEL 850, HACH Chemical Co., Iowa, USA) for the concentrations - 2+ of dissolved oxygen (DO), total sulfide, nitrite (NO2 ), and ferrous iron (Fe ) according to the manufacturer’s protocols.

Anion concentrations of water samples were determined using high-performance liquid chromatography (HPLC, Dionex DX-500 chromatography, Dionex Co., USA), and cation concentrations were determined using direct current plasma optical emission spectrometry (DCP-OES, Beckman, USA). Salinity was calculated by summing up the major ion concentrations in the unit of mg L-1. Total organic carbon (TOC) and total nitrogen (TN) of the hot spring sediment samples were analyzed by using a Perkin Elmer Series 2400 carbon-hydrogen-nitrogen analyzer (Perkin Elmer, Norwalk, CT, USA).

All the sediment samples were prepared for the analysis of quantitative powder X- ray diffraction (qXRD) as previously described (Eberl, 2003). Briefly, one gram of ground sample was mixed with 0.25 g of an internal standard (corundum). This mixture along with 4 mL ethanol and 2 ceramic beads were vortexed (Vortex Genie, Scientific Ind. Inc., USA) for 10 min. After drying at 65oC overnight, 600 µL of DuPont Vertrel XF (Miller-Stephenson, Sylmar, CA, USA) was added to the mixture, and dried at room temperature for another 10 min. In order to avoid the orientation effect of mineral particles, the powder was side-packed into a holder. Samples were X-ray scanned from 2

to 70 degree two theta with Cu K-alpha radiation (40 kV, 35 mA), 0.02 degree step size, and a count time of 5 seconds per step. The XRD data were analyzed quantitatively and converted into weight percent using the RockJock computer program (detection limit 0.1%) (Eberl, 2003).

DNA extraction and quantitative polymerase chain reaction (qPCR)

DNA was extracted from the sediment samples (0.5-1g wet weight) using the FastDNA Spin Kit for Soil (MP Biomedical, OH, USA) according to the manufacturer’s instructions. DNA extraction was performed in duplicate for each sample, and the resulting eluted DNA was mixed and homogenized to obtain a final volume of 100 μL. DNA concentration was quantified by using a Qubit 2.0 fluorometer (Invitrogen, Carlsbad, CA, USA). DNA samples were divided into 20 μL aliquots and preserved at - 80oC until further processing.

Archaeal and bacterial 16S rRNA gene copies were quantified by qPCR with primer sets Arch349F (5′-GYGCASCAGKCGMGAAW-3′)/Arch806R (5′- GGACTACVSGGGTATCTAAT-3′) and 331F (5′-TCCTACGGGAGGCAGCAGT- 3′)/797R (5′-GGACTACCAGGGTATCTAATCCTGTT-3′), respectively. qPCR amplifications were performed in a reaction volume of 20 μL, containing 10 μL 2× SYBR Green master mix (Applied Biosystems Inc., Foster City, CA, USA), 0.5 mM of each primer, and 1 µL of template DNA (~20 ng DNA). The thermal cycling program was: 10 min at 95oC, followed by 40 cycles of 94oC for 30 s, 54oC (archaeal) or 60oC (bacterial) for 20 s, and 72oC for 60 s. PCR products of archaeal and bacterial 16S rRNA gene fragments from one of the investigated samples were used for clone library construction. Purified 16S rRNA gene plasmids of two randomly selected clones (one each from archaeal and bacterial clone libraries) served as standards for archaeal and bacterial qPCRs, respectively. Serial dilutions of the archaeal and bacterial standards were made in the range of 102-108 16S rRNA gene copies. The data were used to create

standard curves correlating the threshold cycle (Ct) numbers with the 16S rRNA gene copy numbers. The linear correlation coefficients (R2) of the archaeal and bacterial 16S rRNA genes were higher than 0.99. The qPCR amplification efficiencies were in the

range of 96-101%. All qPCR reactions were conducted in triplicate on an ABI 7500 Real- Time PCR system (Applied Biosystems Inc., Foster City, CA, USA). Melting curve analyses were performed after each run to confirm PCR specificity.

Barcoded 454 pyrosequencing and data analysis

The V4-V8 variable regions of the archaeal and bacterial 16S rRNA gene fragments were PCR-amplified with the universal primer set of the modified 515F (5'- GTGYCAGCMGCCGCGGTAA-3') and 1391R (5’-GACGGGCGGTGWGTRCA-3’) (Hou et al., 2013). Multiple samples were pooled into one run for 454 pyrosequencing using a sample tagging approach (Meyer et al., 2008). Appropriate adaptors and 8-bp unique barcodes were added to the 5’-end of both forward and reverse primers (Meyer et al., 2008). The PCR mix (25 µL) contained the following reagents: 2.5 µL of 10X Taq buffer, 0.4 µM of each primer, 200 µM of dNTP, 0.8 mM of bovine serum albumin (TaKaRa, Dalian, China), 1.25 U of Taq DNA polymerase (TaKaRa, Dalian, China), and ~25 ng of DNA template. The PCR reaction started with an initial denaturation step at 95oC for 5 min, followed by 35 cycles of 94oC for 30 s, 55oC for 30 s, and 72oC for 1min, and a final extension at 72oC for 10 min. To obtain enough amplicons, PCRs were run in quadruplicate for each sample. PCR products were pooled and gel-purified as previous described (Huang et al., 2011). The purified amplicons of all six samples were quantified on a Qubit 2.0 fluorometer, and pooled in equi-molar concentrations for 454 pyrosequencing.

The pyrosequencing was performed on a 454 GS FLX platform (454 Life Sciences, Branford, CT, USA) by MininGene Biotechnology (Beijing, China). The sequences from the 515F-end of the amplicons were used for downstream data analysis. The sequences were processed with the Mothur 1.25.0 software (Schloss et al., 2009, Schloss et al., 2011). After assigning sequencing reads to each sample according to their unique barcodes, low quality sequences (quality score < 25, length < 200 bp, ambiguous base ≥ 1, homopolymer ≥ 6, and chimeras) were removed. The number of sequence reads for each sample was normalized to the one with the smallest number of reads (4210) by random sub-sampling. The number of base pairs for each read was trimmed to a uniform

length of 243 bp. Sequence reads were assigned to operational taxonomic units (OTUs) at the 97% similarity level, and one sequence was randomly selected as a representative from each OTU. Representative sequences were assigned taxonomy with the ribosomal database project (RDP) trainset7_112011.pds.fasta (http://www.mothur.org/wiki/RDP_

reference_files) as the reference database using the “classify.otu” script in Mothur (Schloss et al., 2009), followed by a manual verification. All 454 sequences have been deposited to NCBI under the Sequence Read Archive database (accession no. SRA056526).

Clone library construction

To verify the occurrence of Sulfobacillus in sample BAL-0 (90.8oC), the extracted DNA of this sample was subjected to PCR amplification of the full-length 16S rRNA gene with Bacteria-specific primer pair of Bac27F/Univ1492R followed by clone library construction and phylogenetic analysis according to the methods described in a previous study (Huang et al., 2011). The obtained clone sequences have been deposited in the GenBank database under accession numbers KC493660-KC493669.

Statistical Analysis

To evaluate microbial diversity in each sample, alpha diversity indices, including coverage, Chaol, ACE, npShannon, and invSimpson, were calculated at the 97% OTU similarity level using tools implemented in Mothur (Schloss et al., 2011). In order to compare the similarity of the microbial community structures in the investigated hot springs, the hierarchical cluster analysis was performed using the hclust function in the R software (R Development Core Team, 2012). The metaMDS and envfit functions (‘vegan’ package) were used to display any correlations between microbial community composition and environmental variables (e.g., parameters in Table 1 and mineral compositions in Figure S1) (Oksanen et al, 2011). By fitting individual environmental parameters onto a given non-metric multidimensional scaling (NMDS) ordination of the Bray Curtis similarities for genera (95% OTU similarity), the correlations between genera

and geochemical variables were made. NMDS ordination was rotated so that two axes were along the principal components. Spearman test was also performed to confirm the correlation between microbial groups and environmental variables using the R software.

RESULTS

Geochemistry and mineralogy

The temperature and pH of the six investigated hot springs was 60-92oC and 3.72- 6.58, respectively (Table 1). The dissolved oxygen concentration ranged from 1.6 to 3.1 mg L-1. The aqueous concentrations of cations such as calcium, potassium, and sodium varied among the springs, with the highest levels measured in NAG-7. The chloride concentration co-varied with these cations. Salinities of these hot springs showed a negative correlation with distance from the sample sites to the coastline (Figure S2), and two springs near Albay Gulf (NAG-7 and BAG-2) showed much higher salinities than the other inland springs (Table 1).

The TOC concentration in the sediment samples was 1.57-22.95 mg g-1, and TN concentration was < 2.42 mg g-1. Quantitative XRD analysis showed that smectite (20.6- 56.8%), kaolinite (11.8-43.2%), K-feldspar (5.5-12.5%), and cristobalite (3.1-13.8%) were the major minerals in the hot spring sediments. Minor sulfur- and iron-related minerals (i.e., pyrite, elemental sulfur, gypsum, goethite and ferrihydrite) were in lower abundance, and varied among the samples (Figure S1).

Microbial community characteristics in the Philippines hot springs

Archaeal and bacterial 16S rRNA gene abundances were 1.08×107-1.60×108 and 7.50×107-9.93×109 copies per gram of sediments (dry weight), respectively (Table 2). The archaeal 16S rRNA gene abundance was lower in the two high-temperature springs than the moderate-temperature ones, except for BAL-1. Similarly, the bacterial 16S rRNA gene abundance was lower in higher temperature springs than the low temperature ones.

A total of 57,493 raw sequence reads were obtained for both Archaea and Bacteria. After removing low quality sequences, 41,446 reads (22,257 and 19,189 for Archaea and Bacteria, respectively) remained for the six samples. These sequence reads were clustered into 818 OTUs (203 and 615 for Archaea and Bacteria, respectively) at the 97% similarity level. The sequencing coverage was 92.6-99.6%, indicating that the number of sequences was sufficient to capture most taxa in each sample.

The archaeal community mainly consisted of sequence reads closely related (nucleotide identity 92-100%) to phyla Crenarchaeota, Euryarchaeota, and unclassified Archaea (accounting for 95.8%, 3.2%, and 1.0% of total Archaea, respectively) (Figure 2a). The bacterial community was composed of sequence reads closely related (nucleotide identity 90-100%) to 17 phyla, with Aquificae and Firmicutes being dominant (accounting for 36.0% and 33.5% of total Bacteria, respectively) (Figure 2b).

Calculations of alpha diversity indices including Chao1, ACE, npShannon, and invSimpson showed that samples BAL-0, NAG-7, and BAG-2 had higher microbial diversity than the other three samples (Table S1). A bi-plot was overlaid on a NMDS ordination to display the correlations between microbial community composition and environmental variables (Figure 3). The Spearman test suggested that certain thermophilic groups were significantly correlated with temperature, e.g. crenarchaeal orders Thermoproteales (r=+0.94, p-value=0.005) and Desulfurococcales (r=+0.83, p- value=0.042).

DISCUSSION

Microbial community pattern corresponding to environmental variables

The diversity and abundance of all genera of both Archaea and Bacteria varied across these hot springs and correlated with certain environmental conditions. Based on the correlation between the genus-level phylogenetic composition and geochemical conditions, the six hot springs can be classified into three clusters: Cluster 1 (BAL-0 and MAL-1), Cluster 2 (INA-1 and BAL-1), and Cluster 3 (NAG-7 and BAG-2) (Figure 3).

In the Cluster 1 (BAL-0 and MAL-1), the archaeal community was mainly composed of sequence reads closely related (identity 100%) to crenarchaeal orders Sulfolobales and Thermoproteales and bacterial phyla Aquificae and Firmicutes. Within this cluster (Figure 2), major phylogenetic groups were archaeal genera Sulfolobus (identity 99-100%) and Vulcanisaeta (identity 92%), and bacterial genera Sulfobacillus (identity 96-100%) and Hydrogenobaculum (identity 100%). The distribution of these genera positively corresponded to certain environmental variables (i.e., temperature, sulfate, elemental sulfur, pyrite) (Figure 3). Considering the optimal growth temperature of Sulfobacillus (either 45-50oC, Norris et al., 1996 or 55-58oC, Kozubal et al., 2012a), it was surprising to observe its predominance in high temperature spring BAL-0 (90.8oC) and MAL-1 (75.8oC). However, the occurrence of Sulfobacillus was confirmed by the full-length bacterial 16S rRNA gene in these samples, and it suggests that Sulfobacillus may have higher temperature tolerance than currently known.

Despite distinct temperatures of INA-1 (92.4oC, pH 6.08) and BAL-1 (60.5oC, pH 5.20), these two samples formed a cluster, i.e., Cluster 2. The archaeal community composition in Cluster 2 contained high proportions of sequence reads associated (nucleotide identity 92-100%) with archaeal phylum Crenarchaeota (96-100% of total Archaea, mainly unclassified Desulfurococcales). It is notable that photosynthetic microorganisms were not observed in sample BAL-1, where temperature was lower than the upper limit (73-75oC) for photosynthesis (Brock, 1978, Hou et al., 2013). Several studies (Boyd et al., 2010, 2012, Cox et al., 2011, Hamilton et al., 2012) have demonstrated that the upper temperature limit for photosynthesis was pH and sulfide dependent. In acidic environments (<4-5), the upper temperature limit for the distribution of photosynthetic metabolism could be lowered to ~57oC (Boyd et al., 2010; Cox et al., 2011, Hamilton et al., 2012), which may have accounted for the absence of photosynthetic microorganisms in spring BAL-1. In addition, the absence of bacterial photosynthesis in this spring could also result from their competition with algae. For example, Boyd et al., (2012) suggested that a total sulfide concentration of 5 µM was sufficient to suppress algal phototrophs, but this level of sulfide had no effect on bacterial counterparts. At a low sulfide concentration (<5 µM), algal phototrophs were not suppressed (Boyd et al., 2012), and could be competing against bacterial phototrophs for

nutrients and physical space. This competition seems to limit Cyanobacteria to environments with pH >4.0 (Brock, 1973). The pH of spring BAL-1 (5.20) is certainly appropriate to limit Cyanobacteria.

In contrast to the above four hot springs (INA-1, MAL-1, BAL-0, and BAL-1) where the majority of sequences were associated (nucleotide identity 90-100%) with known orders and genera, large proportions of sequence reads in Cluster 3 (NAG-7 and BAG-2) were mainly associated with unclassified Archaea (3-19%), Crenarchaeota (63- 82%) and Bacteria (25-58%), indicating possible occurrence of novel microorganisms in these springs. The negative correlation between salinity and coastal distance (Figure S2), as well as the close locations of these two high-salinity springs to the Albay Gulf, suggest that a marine influence may be a possible reason for the occurrence of novel microorganisms in these springs. Indeed, both archaeal and bacterial community compositions in these two springs were positively correlated with marine geochemical indictor (e.g., calcium) (Figure 3).

Inferred physiological functions

In these springs, there appear to be a variety of microorganisms with putative sulfur metabolisms (Figure 2). Springs in Cluster 1 (BAL-0 and MAL-1) appear to be dominated by sulfur- and sulfide-oxidizing organisms; springs in Cluster 2 (INA-1 and BAL-1) were dominated by sulfate-reducing organism and other organisms with unknown functions; the majority of prokaryotes in Cluster 3 (NAG-7 and BAG-2) were not related to any cultured representatives and potential functions remain unknown.

Potential sulfur and sulfide oxidation in Cluster 1 (BAL-0 and MAL-1). Sequence reads that were closely related (nucleotide identity 96-100%) to sulfur- and sulfide-oxidizing microorganisms, such as archaeal genera Sulfolobus (Brock et al., 1972, Huber & Prangishvili, 2006, Kozubal et al., 2012a) and Metallosphaera (Stetter et al., 1996, Kozubal et al., 2012a) of the order Sulfolobales, bacterial genera Sulfobacillus of Firmicutes (Bogdanova et al., 2006, Watling et al., 2008) and Hydrogenobaculum of Aquificales (D’Imperio et al., 2007, Clingenpeel et al., 2009, Inskeep et al., 2005), were primarily concentrated in springs BAL-0 (90.8oC, pH4.26) and MAL-1 (75.8oC, pH 5.08)

with the highest dissolved oxygen (3.05 ppm) . Both Sulfolobus and Sulfobacillus are capable of growing autotrophically or mixotrophically by oxidizing elemental sulfur and ferrous iron in acidic geothermal environments (Kozubal et al., 2012a). In addition, ferrous iron and sulfide bearing minerals (e.g., pyrite) were detected (1.8%) in these springs, and ferric iron concentration (30.23 mg/L) in spring BAL-0 was higher than other springs. This observation is consistent with the physiological properties of Metallosphaera, which is an iron and sulfur-oxidizing chemolithoautotroph (Kozubal et al., 2012a). Hydrogenobaculum (55-72oC, optimum 65oC) is a gram-negative rod capable of using oxygen and nitrate as electron acceptors, hydrogen and reduced sulfur

compounds as electron donors, and CO2 as the sole carbon source (Stohr et al., 2001, D’Imperio et al., 2008, Reysenbach et al., 2009).

In addition to the above sulfur- and sulfide-oxidizing-like microorganisms, large amounts of sequence reads associated with Vulcanisaeta (16.4-46%) were also found in these two samples. Vulcanisaeta (optimal temperature 85oC and pH 4.5) is an anaerobic, heterotrophic, and hyperthermophilic archaeon that was first isolated from hot springs in eastern Japan, and is able to utilize sulfur, thiosulfate, or sulfate as electron acceptors (Itoh et al., 2002). The presence of putative Vulcanisaeta-related sequence reads was again consistent with the high concentrations of sulfate and elemental sulfur in these two hot springs.

Putative sulfate- and sulfur-reducing microorganisms in Cluster 2 (INA-1 and BAL-1). Despite distinct temperatures, springs INA-1 (92oC, pH 6.08) and BAL-1 (61oC, pH 5.20) were characterized by low dissolved oxygen (1.55-2.10 ppm in spring water but likely anoxic in the sediments) and circumneutral pH. These conditions appear to favor sulfur- and sulfate-reducing microorganisms. For example, the crenarchaeal order Desulfurococcales is a facultative anaerobic hyperthermophile (optimal temperature 85-106oC, pH 5.5-7) that is capable of reducing sulfur or thiosulfate (Huber & Stetter, 2006, Boyd et al., 2007). Other observed bacteria, such as Thermosulfidibacter (optimal pH 5.5-7.5) (Nunoura et al., 2008), Thermodesulfobacterium (optimal pH 6.5) (Jeanthon et al., 2002), and Thermodesulforhabdus (optimal pH 7.0) (Beeder et al., 1995), are all capable of sulfate reduction.

Novel microorganisms in Cluster 3 (NAG-7 and BAG-2). In springs NAG-7 (64.1oC, pH 3.72) and BAG-2 (59.9oC, 6.58), high proportions of unclassified Crenarchaeota (82% and 63%, respectively), Archaea (3 and 18%, respectively), and Bacteria (52% and 23%, respectively) strongly suggest occurrence of novel microorganisms in these habitats. This novelty could be due to the different physicochemical conditions and the geographical locations of these two springs. It was possible that the sea water from the Albay Gulf could have intruded the hydrological system of these springs and thus affected the microbial composition in this area. It was interesting to note that some moderate-temperature Euryarchaeota (e.g., Methanosaeta) and Bacteria (e.g., Thermodesulforhabdus) that were commonly found and isolated from marine environments (Beeder et al., 1995, Mori et al., 2012) were also detected in these samples. This observation suggests that geochemistry as well as microbiology might be influenced by seawater. Indeed, a previous study (Tobler & Benning, 2011) observed different communities within siliceous sinters in five geochemically diverse Icelandic geothermal systems of different salinities, and found that microbial community in neutral, saline geothermal waters were related to marine genera of the Proteobacteria; whereas those in freshwater were dominated by Aquificae, Deinococci, and non-marine genera of Proteobacteria. However, observation of novel organisms was not limited to the freshwater-seawater interaction zone, as novel organisms have been detected in a number of other geothermal environments. For example, large amounts of unclassified microorganisms were detected in other acidic to circumneutral hot springs, such as an acidic hot spring in Colombia where 65% of archaeal sequences could not be classified (Bohorquez et al., 2012), and two neutral hot springs from Little Hot Creek in the Long Valley Caldera, USA, where >85% of the archaeal libraries were related to unclassified Crenarchaeota (Vick et al., 2010).

Microbial communities of the Philippines hot springs in comparison with those of other springs worldwide

In order to better understand the relationship between the microbial community composition and the environmental conditions in these hot springs, a comparison was

made among the studied Philippines hot springs and other acidic to circumneutral hot springs in the world. For a better comparison, five hot springs from Tengchong, China (Hou et al., 2013) and nine sites from Yellowstone National Park, USA (Meyer-Dombard et al., 2005, Spear et al., 2005, Boyd et al., 2009, Inskeep et al., 2010) that have similar temperature and pH ranges were selected as representatives. The major microbial groups used for these comparisons are phyla Crenarchaeota, unclassified Crenarchaeota, and Aquificae; orders Sulfolobales, Desulfurococcales, Thermoproteales; and genera Sulfolobus, Metallosphaera, Vulcanisaeta, Sulfobacillus, Hydrogenobaculum, and Hydrogenobacter (Figure 4). The relative abundance of each group out of either total Archaea or Bacteria was plotted on the temperature-pH space.

At the phylum-level, the abundance of Crenarchaeota (27-100%) (Figure 4a), novel Crenarchaeota (0-83%) (Figure 4b), and Aquificae (0.2-100%) (Figure 4c), found in the investigated Philippines hot springs was consistent with previous observations in other acidic to circumneutral hot springs, including those in Tengchong, China (Hou et al., 2013) and Yellowstone National Park in the USA (Meyer-Dombard et al., 2005, Inskeep et al., 2010). However, novel Crenarchaeota were relatively dominant in some Philippines and YNP hot springs, while it was rarely found in the Tengchong hot springs. This phenomenon could be ascribed to the geochemistry difference (e.g., salinity) of these hot springs. However, the underlying reasons still await further investigation. Aquificae were found abundant in all the three compared areas. However, the distribution patterns of Aquificae were slightly different: although Aquificae-like microorganisms were dominant in the acidic sites of Tengchong (pH 2-5), they appeared to occur in a wider pH range in YNP and the Philippines (pH 3-7.5) (Figure 4c). This difference was likely caused by different composition of Aquificae. For example, in Tengchong acidic springs, Aquificae was mainly composed of Hydrogenobaculum (optimal temperature 65oC and pH 3.5) (Figure 4k) with a minor amount of Hydrogenobacter (optimal temperature 75oC and pH 7.5) ( Figure 4l) (Pitulle et al., 1994, Hou et al., 2013). In contrast, the Aquificae in the Philippines hot springs consisted of Hydrogenobaculum, Hydrogenobacter, Persephonellai, and unclassified Hydrogenothermaceae. In YNP, Aquificae was composed of Thermocrinis, Sulfurihydrogenibium, Hydrogenobacter, and

Hydrogenobaculum (Fouke et al., 2000, Langner et al., 2001, Jackson et al., 2001, Reysenbach et al., 2005, Boyd et al., 2009).

At the order-level, three major archaeal orders were found in the three compared hot spring systems, including Sulfolobales (0-100%) (Figure 4d), Desulfurococcales (0- 62%) (Figure 4e), and Thermoproteales (0-52%) (Figure 4f). Their distribution patterns were also different: most archaeal sequences from Tengchong were related to Sulfolobales, while the majority of those from the Philippines and YNP were related to Desulfurococcales and Thermoproteales. This difference might be ascribed to pH. For example, the Sulfolobales-dominant hot springs tended to have lower pH (2-5), while Desulfurococcales and Thermoproteales were more abundant in those with higher pH (4- 7). This observation is consistent with previous studies that Sulfolobales (optimum pH 2- 3) prefer a lower pH range than those of Desulfurococcales (optimum pH 6-7) and Thermoproteales (optimum pH 4-6) (Garrity and Holt, 2001, Madigan and Martinko, 2006).

At the genus-level, archaeal genera Sulfolobus (0-100%) (Figure 4g) and Metallosphaera (0-92%) (Figure 4h), both within Sulfolobales, were present in the acidic to circumneutral springs of all the three compared geothermal systems. However, Sulfolobus were more abundant in the Tengchong acidic hot springs (pH 2-5), while Metallosphaera preferred YNP acidic hot springs (pH 2-3). Vulcanisaeta (0-46%) (Figure 4i) was present in both YNP and the Philippines, but was completely absent in Tengchong (Hou et al., 2013, Briggs et al., 2013). Sulfobacillus (0-47%) (Figure 4j) was abundant in two Philippines springs (29-47%), in comparison with negligible levels in YNP and Tengchong. Hydrogenobaculum (0-100%) (Figure 4k) occurred in more acidic Tengchong springs (pH 2-5), while Hydrogenobacter (0-92%) (Figure 4l) tended to prefer higher pH springs (pH 5-7.5). Although Sulfobacillus can carry out similar functions as archaeal Sulfolobus and bacterial Hydrogenobaculum (i.e., sulfur- or iron- related metabolisms such as sulfur and iron oxidation), this bacterial genus appears to be more versatile. For example, Sulfobacillus are both autotrophic and mixotrophic (Norris et al., 1996, Bogdanova et al., 2006, Kozubal et al., 2012) and moderately thermophilic with a wide temperature range (28-62oC) (Watling et al., 2008). In comparison, Hydrogenobaculum in YNP is mostly autotrophic (Stohr et al., 2001, Reysenbach et al.,

2005, Ferrera et al., 2007, D'Imperio et al., 2008). Sulfolobus is largely 0 chemolithotrophic that can fix CO2 coupled with oxidation of H2S or S to H2SO4, but some strains are facultative heterotrophs (Huber & Stetter, 2001). The high total organic carbon content in springs BAL-0 (21.14 mg g-1) and MAL-1 (11.86 mg g-1) where abundant Sulfobacillus was observed suggests that Sulfobacillus may be heterotrophic in these springs. In comparison, low organic carbon contents in Tengchong springs (0.9-2.6 mg g-1) (Hou et al., 2013) may have accounted for the absence of Sulfobacillus and the predominance of autotrophic Hydrogenobaculum.

CONCLUSIONS

The microbial communities in the six hot springs of the Philippines were highly diverse, and were dominated by phyla Crenarchaeota, Aquificae, Firmicutes, and Proteobacteria. Based on the phylogenetic analysis and statistical results, temperature, aqueous concentrations of sulfate and calcium, and certain sediment properties (total nitrogen, pyrite, and elemental sulfur) were important environmental variables affecting the microbial abundance, diversity, and community structure. Sulfur metabolisms appear to be the key physiological functions in these Philippines springs; however, novel organisms were abundant in some springs and their functions remain to be elucidated. Despite no major difference among hot springs from the Philippines, Tengchong in China, and YNP in the USA at the phylum-level, important differences were observed at the order- and genus-level. A combination of geographic distance and environmental conditions may have accounted for these distinct differences.

ACKNOWLEDGMENTS

This work was supported by an National Natural Science Foundation grant of China (41030211), the Key Project of International Cooperation of the Ministry of Science & Technology of China) (No. 2013DFA31980), the Program for New Century Excellent Talents in University (NCET-12-0954), the Fundamental Research Funds for National University (China University of Geosciences-Wuhan), and National Science

Foundation grants (OISE-0968421, DBI REU 1005223, and ETBC-1024614). The authors are grateful to all people involved in field work and sample collection, including those from the National Institute of Geological Sciences, University of the Philippines and the Resource Management Department, Energy Development Corporation Ortigas Center, the Philippines. We are grateful to two anonymous reviewers whose comments improved the quality of the manuscript.

REFERENCES

Beeder J., Torsvik T. and Lien T. (1995) Thermodesulforhabdus norvegicus gen. nov., sp. nov., a novel thermophilic sulfate-reducing bacterium from oil field water. Arch. Microbiol. 164: 331-336.

Bogdanova T.I., Tsaplina I.A., Kondrateva T.F., Duda V.I., Suzina N.E., Melamud V.S., Tourova T.P. and Karavaiko G.I. (2006) Sulfobacillus thermotolerans sp. nov., a thermotolerant, chemolithotrophic bacterium. Int. J. Syst. Evol. Micr. 56: 1039- 1042.

Bohorquez L.C., Delgado-Serrano L., López G., Osorio-Forero C., Klepac-Ceraj V., Kolter R., Junca H., Baena S., Zambrano M.M. (2012) In-depth characterization via complementing culture-independent approaches of the microbial community in an acidic hot spring of the Colombian Andes. Microb. Ecol. 63: 103-115.

Boyd E.S., Jackson R.A., Encarnacion G., Zahn J.A., Beard T., Leavitt W.D., Pi Y., Zhang C.L., Pearson A. and Geesey G.G. (2007) Isolation, characterization, and ecology of sulfur-respiring Crenarchaea inhabiting acid-sulfate-chloride- containing geothermal springs in Yellowstone National Park. Appl. Environ. Microbiol. 73: 6669-6677.

Boyd E.S., Leavitt W.D., and Geesey G.G. (2009) CO2 uptake and fixation by a thermoacidophilic microbial community attached to precipitated sulfur in a geothermal spring. Appl. Environ. Microbiol. 75: 4289-4296.

Boyd, E.S., Hamilton, T.L., Spear, J.R., Lavin, M. and Peters, J.W. (2010) [FeFe]- hydrogenase in Yellowstone National Park: evidence for dispersal limitation and phylogenetic niche conservation. ISME J. 4, 1485-1495.

Boyd E.S., Fecteau K.M., Havig J.R., Shock E.L. and Peters J.W. (2012) Modeling the habitat range of phototrophs in Yellowstone National Park: toward the development of a comprehensive fitness landscape. Front Microbiol. 3: 221.

Briggs B.R., Brodie E.L., Tom L.M., Dong H., Jiang H., Huang Q., Wang S., Hou W., Wu G., Huang L., Hedlund B.P. and Zhang C. (2013) Microbial distributions

accross geochemical and temporal conditions in the hot springs of Tengchong, Yunnan Provindent, China. Environ. Microbiol. (submitted).

Brock T.D., Brock K.M., Belly R.T. and Weiss R.L. (1972) Sulfolobus: a new genus of sulfur-oxidizing bacteria living at low pH and high temperature. Arch. Microbiol. 84: 54-68.

Brock T.D. (1973) Lower pH limit for the existence of blue-green algae: evolutionary and ecological implications. Science 179: 480-483.

Brock T.D. (1978) Thermophilic microorganisms and life at high temperatures. Springer, Berlin Heidelberg New York.

Cavicchioli R. (2002) Extremophiles and the search for extraterrestrial life. Astrobiology 2: 281-292.

Clingenpeel S.R., D'Imperio S., Oduro H., Druschel G.K. and McDermott T.R. (2009) Cloning and in situ expression studies of the Hydrogenobaculum arsenite oxidase genes. Appl. Environ. Microbiol. 75: 3362-3365.

Cox A., Shock E.L. and Havig J.R. (2011) The transition to microbial photosynthesis in hot spring ecosystems. Chem.Geol. 280: 344-351.

D'Imperio S., Lehr C.R., Breary M., and McDermott T.R. (2007) Autecology of an arsenite chemolithotroph: sulfide constranits on function and distribution in a geothermal spring. Appl. Environ. Microbiol. 73: 7067-7074.

D'Imperio S., Lehr C.R., Oduro H., Druschel G., Kuhl M. and McDermott T.R. (2008)

Relative importance of H2 and H2S as energy sources for primary production in geothermal springs. Appl. Environ. Microbiol. 74: 5802-5808.

Dimabayao J.T. (2012) Revisiting the volcanic history of Bacman: old and new insights. EDC internal report. Manila, Philippines.

Dolor F.M. (2005) Phases of geothermal development in the Philippines. Workshop for Decision Makers on Geothermal Projects and their Management. Naivasha, Kenya, November 14-18.

Donahoe-Christiansen J., D'Imperio S., Jackson C.R., Inskeep W.P. and McDermott T.R. (2004) Arsenite-Oxidizing Hydrogenobaculum Strain Isolated from an Acid- Sulfate-Chloride Geothermal Spring in Yellowstone National Park. Applied and Environmental Microbiology 70: 1865-1868.

Eberl D.D. (2003) User's guide to RockJock : A program for determining quantitative mineralogy from powder X-ray diffraction data. U.S.G.S. Open-File Report 78: 47.

Edwards K.J., Bond P.L., Gihring T.M. and Banfield J.F. (2000) An archaeal iron- oxidizing extreme acidophile important in acid mine drainage. Science 287: 1796- 1799.

Ferrera I., Longhorn S., Banta A., Liu Y., Preston D. and Reysenbach A.L. (2007) Diversity of 16S rRNA gene, ITS region and aclB gene of the Aquificales. Extremophiles 11: 57-64.

Fouke B.W., Farmer J.D., Des Marais D.J., Pratt L., Sturchio N.C., Burns P.C. and Discipulo M.K. (2000) Depositional facies and aqueous-solid geochemistry of travertine-depositing hot springs (Angel Terrace, Mammoth hot springs, Yellowstone National Park, U.S.A.). J. Sediment. Res. 70: 565-585.

Garrity G.M. and Holt J.G. (2001) Phylum A1. Crenarchaeota phy. nov. In: Bergey's manual of systematic bacteriology, 2nd edn. pp. 169-210. Springer, New York, NY.

Gonzalez-Contreras P., Weijma J. and Buisman C.J. (2012) Kinetics of ferrous iron oxidation by batch and continuous cultures of thermoacidophilic Archaea at extremely low pH of 1.1-1.3. Appl. Microbiol. Biot. 93: 1295-1303.Hamilton T.L., Vogl K., Bryant D.A., Boyd E.S. and Peters J.W. (2012) Environmental constraints defining the distribution, composition, and evolution of chlorophototrophs in thermal features of Yellowstone National Park. Geobiology 10: 236-249.

Hou W., Wang S., Dong H., Jiang H., Briggs B.R., Peacock J., Huang Q., Huang L., Wu G., Zhi X., Li W., Dodsworth J.A., Hedlund B.P., Zhang C., Hartnett H.E.,

Dijkstra P. and Hungate B.A. (2013) A comprehensive census of microbial diversity in hot springs of Tengchong, Yunnan Province, China using 16S rRNA gene pyrosequencing. PLoS ONE 8: e53350.

Huang Q., Dong C., Dong R. Jiang H., Wang S., Wang G., Fang G., Ding X., Niu L., Li X., Zhang C. and Dong H. (2011) Archaeal and bacterial diversity in hot springs on the Tibetan Plateau, China. Extremophiles 15: 549-563.

Huber H. and Stetter K. (2006) Desulfurococcales. The Prokaryotes 3: 52-68.

Huber H. and Prangishvili D. (2006) Sulfolobales. The prokaryotes 3: 23-51.

Huber H., Huber R. and Stetter K. (2006) Thermoproteales. The prokaryotes 3: 10-22.

Huber R. and Stetter K.O. (2001) Discovery of hyperthermophilic microorganisms. Method. Enzymol. 330: 11-24.

Inskeep W.P., Ackerman G.G., Taylor W.P., Kozubal M., Korf S. and Macur R.E. (2005) On the energetics of chemolithotrophy in nonequilibrium systems: case studies of geothermal springs in Yellowstone National Park. Geobiology 3: 297-317.

Inskeep W.P., Rusch D.B., Jay Z.J., Herrgard M.J., Kozubal M.A., Richardson T.H., Macur R.E., Hamamura N., Jennings R., Fouke B.W., Reysenbach A., Roberto F., Young M., Schwartz A., Boyd E.S., Badger J.H., Mathur E.J., Ortmann A.C., Bateson M., Geesey G. and Frazier M. (2010) Metagenomes from high- temperature chemotrophic systems reveal geochemical controls on microbial community structure and function. PLoS ONE 5: 9773.

Itoh T., Suzuki K. and Nakase T. (2002) Vulcanisaeta distributa gen. nov., sp. nov., and Vulcanisaeta souniana sp. nov., novel hyperthermophilic, rod-shaped crenarchaeotes isolated from hot springs in Japan. Int. J. Syst. Evol. Micr. 52: 1097-1104.

Itoh T., Suzuki K., Sanchez P.C. and Nakase T. (1999) Caldivirga maquilingensis gen. nov., sp. nov., a new genus of rod-shaped crenarchaeote isolated from a hot spring in the Philippines. Int. J. Syst. Bact. 49: 1157-1163.

Itoh T., Suzuki K., Sanchez P.C. and Nakase T. (2003) Caldisphaera lagunensis gen. nov., sp. nov., a novel thermoacidophilic crenarchaeote isolated from a hot spring at Mt Maquiling, Philippines. Int. J. Syst. Evol. Micr. 53: 1149-1154.

Jackson C.R., Langner H.W., Donahoe-Christiansen J., Inskeep W.P. and McDermott T.R. (2001) Molecular analysis of microbial community structure in an arsenite- oxidizing acidic thermal spring. Environ. Microbiol. 3: 532-542.

Jeanthon C., L'Haridon S., Cueff V., Banta A., Reysenbach A. and Prieur D. (2002) Thermodesulfobacterium hydrogeniphilum sp. nov., a thermophilic, chemolithoautotrophic, sulfate-reducing bacterium isolated from a deep-sea hydrothermal vent at Guaymas Basin, and emendation of the genus Thermodesulfobacterium. Int. J. Syst. Evol. Micr. 52: 765-772.

Jing H., Aitchison J.C., Lacap D.C., Peerapornpisal Y., Sompong U. and Pointing S.B. (2005) Community phylogenetic analysis of moderately thermophilic cyanobacterial mats from China, the Philippines and Thailand. Extremophiles 9: 325-332.

Konhauser K.O., Jones B., Reysenbach A. and Renaut R.W. (2003) Hot spring sinters: keys to understanding Earth's earliest life forms. Can. J. Earth Sci. 40: 1713-1724.

Kozubal M.A., Macur R.E., Jay Z.J., Beam J.P., Malfatti S.A., Tringe S.G., Kocar B.D., Borch T. and Inskeep W.P. (2012a) Microbial iron cycling in acidic geothermal springs of Yellowstone National Park: Integrating molecular surveys, geochemical processes and isolation of novel Fe-active microorganisms. Front. Microbiol.3: 109.

Kozubal M.A., Romine M., Jennings R.D., Jay Z.J., Tringe S.G., Rusch D.B., Beam J.P., McCue L.A. and Inskeep W.P. (2012b) Geoarchaeota: a new candidate phylum in the Archaea from high-temperature acidic iron mats in Yellowstone National Park. ISME J 7: 622-634.

Kvist T., Ahring B.K. and Westermann P. (2007) Archaeal diversity in icelandic hot springs. FEMS Microbiol.Ecol. 59: 71-80.

Lacap D.C., Barraquio W. and Pointing S.B. (2007) Thermophilic microbial mats in a tropical geothermal location display pronounced seasonal changes but appear resilient to stochastic disturbance. Environ. Microbiol. 9: 3065-3076.

Lacap D.C., Smith G.J., Warren-Rhodes K. and Pointing S.B. (2005) Community structure of free-floating filamentous cyanobacterial mats from the Wonder Lake geothermal springs in the Philippines. Can. J. Microbiol. 51: 583-589.

Langner H.W., Jackson C.R., McDermott T.R. and Inskeep W.P. (2001) Rapid oxidation of arsenite in a hot spring ecosystem, Yellowstone National Park. Environ. Science & Technol. 35: 3302-3309.

Lantican N., Diaz M., Cantera J., de los Reyes F. and Raymundo A. (2011) Microbial community of a volcanic mudspring in the Philippines as revealed by 16S rDNA sequence analysis and fluorescence in situ hybridization. World J.Microb. & Biot. 27: 859-867.

Madigan M.T. and Martinko J.H. (2006) Prokaryotic diversity: the Archaea. In: Brock, biology of microorganisms, 11th edn. Pearson Education Inc., Pearson Prentice Hall, Upper Saddle River, NJ.

Meyer-Dombard D.R., Shock E.L. and Amend J.P. (2005) Archaeal and bacterial communities in geochemically diverse hot springs of Yellowstone National Park, USA. Geobiology 3: 211-227.

Meyer M., Stenzel U. and Hofreiter M. (2008) Parallel tagged sequencing on the 454 platform. Nat. Protc. 3: 267-278.

Mori K., Iino T., Suzuki K.I., Yamaguchi K. and Kamagata Y. (2012) Aceticlastic and NaCl-requiring methanogen "Methanosaeta pelagica" sp. nov., isolated from marine tidal flat sediment. Appl. Environ. Microbiol. 78: 3416-3423.

Norris P.R., Burton N.P. and Foulis N. (2000) Acidophiles in bioreactor mineral processing. Extremophiles 4: 71-76.

Norris P.R., Clark D.A., Owen J.P. and Waterhouse S. (1996) Characteristics of Sulfobacillus acidophilus sp. nov. and other moderately thermophilic mineral- sulphide-oxidizing bacteria. Microbiology 142: 775-783.

Nunoura T., Oida H., Miyazaki M. and Suzuki Y. (2008) Thermosulfidibacter takaii gen. nov., sp. nov., a thermophilic, hydrogen-oxidizing, sulfur-reducing chemolithoautotroph isolated from a deep-sea hydrothermal field in the Southern Okinawa Trough. Int. J. Syst. Evol. Micr. 58: 659-665.

Oksanen J., Blanchet F.G., Kindt R., Legendre P., Minchin P.R., O'Hara B., Simpson G.L., Solymos P., Henry M., Stevens H. and Wagner H. (2011) Vegan: Community ecology package, http://cran.r-project.org/. http://vegan.r-forge.r- project.org/.

Pitulle C., Yang Y., Marchiani M., Moore E.R.B., Siefert J.L., Aragno M., Jurtshuk P. and Fox G.E. (1994) Phylogenetic Position of the Genus Hydrogenobacter. Int. J. Syst. Bact. 44: 620-626.

R Development Core Team. (2012) R: A language and environment for statistical computing. R Found. for Stat. Comput., Vienna.

Reysenbach A.L., Banta A., Civello S. Daly J., Mitchel K., Lalonde S., Konhauser K., Rodman A., Rusterholtz K. and Takacs-Vesbach C. (2005) Aquificales in Yellowstone National Park. Geothermal Biology and Geochemistry in Yellowstone National Park 1: 129-142.

Reysenbach A.L., Hamamura N., Podar M., Griffiths E., Ferreira S., Hochstein R., Heidelberg J., Johnson J., Mead D., Pohorille A., Sarmiento M., Schweighofer K., Seshadri R. and Voytek M.A. (2009) Complete and draft genome sequences of six members of the Aquificales. J. Bacteriol. 191: 1992-1993

Rothschild L.J. and Mancinelli R.L. (2001) Life in extreme environments. Nature 409: 1092-1101.

Schloss P.D., Westcott S.L., Ryabin T., Hall J.R., Hartmann M., Hollister E.B., Lesniewski R.A., Oakley B.B., Parks D.H., Robinson C.J., Sahl J.W., Stres B., Thallinger G.G., Van Horn D.J. and

Weber C.F. (2009) Introducing mothur: open-source, platform-independent, community- supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75: 7537-7541.

Schloss P.D. and Westcott S.L. (2011) Assessing and improving methods used in operational taxonomic unit-based approaches for 16S rRNA gene sequence analysis. Appl. Environ. Microbiol. 77: 3219-3226.

Siering P.L., Clarke J.M. and Wilson M.S. (2006) Geochemical and biological diversity of acidic, hot springs in Lassen Volcanic National Park. Geomicrobiol. J. 23: 129- 141.

Spear J.R., Walker J.J., McCollom T.M. and Pace N.R. (2005) Hydrogen and bioenergetics in the Yellowstone geothermal ecosystem. P. Natl. A. Sci. USA 102: 2555-2560.

Stetter K.O. (1996) Hyperthermophilic procaryotes. FEMS Microbiol. Rev. 18: 149-158.

Stohr R., Waberski A., Volker H., Tindall B.J. and Thomm M. (2001) Hydrogenothermus marinus gen. nov., sp. nov., a novel thermophilic hydrogen-oxidizing bacterium, recognition of Calderobacterium hydrogenophilum as a member of the genus Hydrogenobacter and proposal of the reclassification of Hydrogenobacter acidophilus as Hydrogenobaculum acidophilum gen. nov., comb. nov., in the phylum 'Hydrogenobacter/Aquifex'. Int. J. Syst. Evol. Micr. 51: 1853-1862.

Tobler D. and Benning L. (2011) Bacterial diversity in five Icelandic geothermal waters: temperature and sinter growth rate effects. Extremophiles 15: 473-485.

Vick T.J., Dodsworth J.A., Costa K.C., Shock E.L. and Hedlund B.P. (2010) Microbiology and geochemistry of Little Hot Creek, a hot spring environment in the Long Valley Caldera. Geobiology 8: 140-154.

Watling H.R., Perrot F.A. and Shiers D.W. (2008) Comparison of selected characteristics of Sulfobacillus species and review of their occurrence in acidic and bioleaching environments. Hydrometallurgy 93: 57-65.

Wilson M., Siering P., White C., Hauser M. and Bartles A. (2008) Novel Archaea and Bacteria dominate stable microbial communities in North America’s largest hot spring. Microb. Ecol. 56: 292-305.

Table 1. Geographic parameters and water geochemistry of the six investigated hot springs in the Philippines.

Sample INA-1 MAL-1 BAL-0 NAG-7 BAL-1 BAG-2 Inang Location Malangto Balasbas Naghaso Balasbas Balbagon Maharang GPS coordinates 13o04'12", 13o07'12", 13o06'36", 13o07'48", 13o06'36", 13o07'12", (N, E) 123o54'36" 123o54'36" 123o53'24" 123o54'36" 123o53'24" 123o55'48" Temperature (oC) 92.40 75.80 90.80 64.10 60.50 59.90 pH 6.08 5.08 4.26 3.72 5.20 6.58 Salinity (mg L-1) 82.55 65.50 163.47 4078.14 28.60 271.62 DO (ppm) 2.10 3.05 n.m. 2.75 1.55 1.75 TOC (mg g-1) 2.62 11.86 21.14 22.95 3.04 1.57 TN (mg g-1) 1.87 0.86 2.42 2.17 0.00 0.00 Ferrous iron (mg L-1) 0.22 0.16 n.m. 0.56 3.09 0.04 Total sulfide (mgL-1) 0.22 0.11 n.m. 0.00 0.00 0.04 Nitrite (mg L-1) 0.01 0.02 n.m. 0.00 0.00 0.02 Concentration of major ions (mg L-1) Sodium 7.99 7.52 4.26 1170.75 3.36 86.68 Potassium 1.36 1.02 0.63 131.85 0.30 13.18 Calcium 8.95 4.86 11.66 108.53 0.34 63.15 Magnesium 4.84 4.46 7.73 3.62 1.36 42.92 Barium 0.04 0.01 0.03 0.68 0.02 0.40 Manganese 0.49 0.03 1.19 0.34 0.39 0.28 Strontium 0.08 0.06 0.12 4.95 0.02 0.97 Ferric iron 1.42 1.94 30.23 12.00 9.75 0.11 Chloride 3.21 2.28 1.75 2620.75 4.96 57.98 Sulfate 53.49 42.76 105.27 23.68 5.81 4.75 Nitrate 0.67 0.54 0.59 0.98 2.29 1.19 n.m.: not measured

Table 2. Abundance of archaeal and bacterial 16S rRNA genes for the six spring sediment samples from the Philippines

16S rRNA gene (copies g-1) Sample Archaea StdDev Bacteria StdDev INA-1 2.12×107 6.39×104 1.01×108 2.51×105 MAL-1 9.17×107 2.42×105 7.50×107 1.09×105 BAL-0 6.67×107 1.84×104 1.26×109 2.34×106 NAG-7 1.60×108 1.57×105 5.45×109 2.82×106 BAL-1 1.08×107 2.52×104 9.93×109 2.98×107 BAG-2 1.21×108 3.71×105 3.86×109 7.29×106

Figure captions Figure 1. A geographic map showing the sample locations of hot springs in the Philippines. Figure 2. Hierarchical clustering and the microbial composition of the six Philippines hot spring samples. a) archaeal community, b) bacterial community. Figure 3. A non-metric multidimensional scaling ordination depicting the distributions of six acidic to circumneutral hot spring samples (big solid circles) and major representative genera (small solid circles). A bi-plot is overlaid on the ordination to the displayed geochemical variables that are correlated with the microbial community structure. Axes 1 and 2 represent the highest variance in genera composition. Only variables that have a significant correlation (p<0.05) are shown. The abbreviations are pyrite (Py), elemental 2- 2+ sulfur (S), temperature (Temp), total nitrogen (TN), sulfate (SO4 ), and calcium (Ca ). Figure 4. A comparison of major phylogenetic groups in the Philippines hot springs (blue circles) with those from YNP of the United States (green circles) and Tengchong, China (red circles). These groups were plotted in the pH-temperature space. The relative abundance of individual microbial groups was calculated as a percentage relative to either total Archaea or Bacteria. Data for Tengchong springs are from Hou et al., (2013); whereas those for YNP springs are from Meyer-Dombard et al., (2005), Spear et al., (2005), and Inskeep et al., (2010). YNP and Tengchong hot springs that have similar temperature and pH ranges to those of the studied Philippines hot springs were selected. The yellow circles on the plots are scaled to 10% and located at the optimal temperature and pH for genus-level plots (g-l). For phylum- and order-level plots (a-f), the yellow circles illustrate the relative abundance only with no information on optimal temperature and pH.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

76 Supporting information

Table S1. Diversity estimates of 454 pyrosequences retrieved from six hot springs in the Philippines

No. of No. of Coverage Sample Chao1 ACE npShannon invSimpson sequences OTUs (%) INA-1 6498 116 99.1 235.2 231.9 2.7 6.9 MAL-1 9398 46 99.6 150.1 203.7 1.9 4.3 BAL-0 12299 294 98.1 826.6 1007.5 3.5 9.7 NAG-7 7241 234 97.3 684.6 871.2 3.1 4.3 BAL-1 3587 93 97.1 360.2 472.4 1.8 1.8 BAG-2 3166 196 92.6 860.6 1327.4 4.6 24.9

77 Figure S1. Quantitative mineral compositions in all six samples as determined by X-ray diffraction and modeled using the RockJock computer program.

78 Figure S2. Relationship between salinity and coastal distance (from sample site to the coastline).

4 3.5 3

2.5 2

Log salinity 1.5 1 0.5 0 0 500 1000 1500 2000 2500 3000 3500 4000 Distance (m)

79 CHAPTER 4: Metagenomic and Metatranscriptomic Analysis of Taxonomic and Genetic Diversity in Saline Qinghai Lake, China Running Title: Metagenomic and Metatranscriptomic Analysis of Qinghai Lake

Qiuyuan Huang1^, Brandon R. Briggs1^, Hailiang Dong1,2,3*，Hongchen Jiang2, Geng Wu2, James T. Hollibaugh4, Christian Edwarson4, Iwijn De Vlaminck5

1: Department of Geology and Environmental Earth Science, Miami University, Oxford, OH 45056, USA 2: State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Wuhan 430074, China 3: State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Beijing 100083, China 4: Department of Maine Sciences, University of Georgia, Athens, GA 30602, USA 5: Departments of Bioengineering and Applied Physics, Stanford University and the Howard Hughes Medical Institute, Stanford, CA 94305, USA ^ Equal contribution

*Corresponding author: Hailiang Dong: [email protected]

To be submitted to ISME J.

80 Abstract The effect of global climate change applies stress to microbial communities. The microbial response to this stress is dependent on both the taxonomic and genetic diversity of the community. Therefore, to predict microbial responses, baseline information is needed on the pathways and the potential to withstand stress associated with environmental change. An Illumina metagenomic and metatranscriptomic dataset was created from water collected at two sites (B and E) in the high-elevation (3196 m) saline (1.4 %) Qinghai Lake, Tibetan Plateau, China. Autotrophic Cyanobacteria dominated the DNA samples, while heterotrophic Proteobacteria dominated the RNA samples at both sites. Photosystem II was the most active at site B and was associated with photoprotection and osmotic stress genes. Oxidative phosphorylation was most active at site E and was associated with oxidative stress genes. Assimilatory pathways were dominant in the nitrogen cycle at both sites. Our data also show a positive relationship between genetic diversity and the number of stress response genes. The combination of metagenomics and metatranscriptomics techniques indicate that the microbial communities in Qinghai Lake are involved in the carbon and nitrogen cycle, and provide baseline information on the taxonomic and genetic diversity in and ecosystem that is highly sensitive to environmental change.

Keywords Metagenomics/ metatranscriptomics/ biogeochemical cycles/ stress response

81 Introduction Microbially mediated biogeochemical cycles respond to global climate change through climate-ecosystem feedbacks. For example, microbial processes such as photosynthesis, respiration, nitrification, and nitrogen fixation are important for ecosystem functioning and play critical roles in climate change, due to their release and absorption of greenhouse gases, such as carbon dioxide, methane, and nitrous oxides (Schimel 1995; Canfield et al., 2010). The rates of these microbial processes can be accelerated or dampened based on the microbial community response to environmental stress induced by climate change, which in turn will affect the rate of global climate change (Schimel 1995; Bardgett et al., 2008; Heimann and Reichstein 2008; Singh et al., 2010).

The extent of the alteration in microbial processes is likely controlled by the biodiversity in the local ecosystem. Generally, a positive relationship between ecosystem function and biodiversity exists (Griffiths et al., 2000; Degens et al., 2001), because a more diverse community increases the likelihood of finding high-performing key species (taxonomic diversity: the diversity of taxa in the community) or the increased probability of functional redundancy (genetic diversity: diversity of a functional gene) (Yachi and Loreau 1999). For example, in higher taxonomic diversity soils denitrification rates were higher and decreased more slowly with increased salinity compared to soils with lower taxonomic diversity (Hallin et al., 2012). Similar results were found for respiration (Langenheder et al., 2005; Chowdhury et al., 2011; Setia et al., 2011), photosynthesis (Murata et al., 2007; Allakhverdiev and Murata 2008; Takahashi and Murata 2008), aerobic methane oxidation (Sorokin et al., 2000; Heyer et al., 2005; van der Ha et al., 2010), and nitrogen fixation (Severin et al., 2012). Furthermore, genetic diversity also positively correlates to microbial processes because organisms with high genetic diversity can withstand greater stress by enacting more protective mechanisms than organisms with low genetic diversity. For example, Synechococcus has multiple protective mechanisms (i.e. high genetic diversity) to cope with UV stress and can maintain photosynthesis, while Procholorcoccus lacking these protective mechanisms shuts down several key metabolic processes under a similar UV stress (Mella-Flores et al., 2012).

82 The Tibetan Plateau is located in western China and is known as the earth’s Third Pole (next to the Antarctica and the Arctic) due to its high elevation (average >4000 meters above sea level). The region is home to 46,000 glaciers, a vast expanse of permafrost (Guo et al., 2012), and thousands of lakes with endemic macro and microbiota. The Tibetan Plateau has experienced dramatic climate changes in the past 15 millions of years. Today it is both affected by and affects regional and global climates. For example, temperature has increased 0.28°C per decade since the early 1960s (Guo et al., 2012) causing 82% of glaciers to retreat (Qiu 2008). The melting glaciers have caused numerous floods and greatly changed salinity and water levels in most of Tibetan lakes (Lei et al., 2012). The fragileness and sensitivity of the Tibetan Plateau’s ecosystem to these environmental changes have resulted in loss of habitats and extinctions of endemic macrobiota (Yang et al., 2009). Thus, to understand the affect of environmental stress on the microbial communities, baseline knowledge is needed on the taxonomic and genetic diversity of microbes present on the Tibetan Plateau.

Qinghai Lake, located on the northwest margin of the Tibetan Plateau, has been considered a unique ecosystem for studying microbial contributions to gloal climate change because it is a sensitive indicator to environmental change (Dong et al., 2006). Previous studies on Qinghai Lake have detected novel archaea commonly found in marine environments (Jiang et al., 2008), and the microbial diversity, composition, and lipids changed in response to salinity and other gradients (Dong et al., 2006; Jiang et al., 2009; Wang et al., 2013). However, questions remain on (1) what is microbial taxonomic and genetic diversity for Qinghai Lake water columns; (2) what is the metabolic potential and active metabolisms related to carbon and nitrogen cycles, and (3) what stress genes are present in organisms involved in the carbon and nitrogen cycles.

To answer these questions in the water column of Qinghai Lake, an integrated approach including geochemical analysis, metagenomic and metatranscriptomic techniques was used. By comparing the synthesized cDNA with genomic DNA retrieved from the same sample, the relative activity of different populations and functional gene expression in the microbial community was assessed. This is the first metagenomic and metatrancriptomic survey of microbial community in Qinghai Lake, and will provide

83 useful insights on the effect of global climate change on microbial mediated biogeochemical cycles.

Materials and Methods

Field sampling and geochemical analysis

Qinghai Lake is a perennial lake located on the Tibetan Plateau at an elevation of 3196 m above sea level. A description of the lake has been published previously (Jiang et al., 2009). Water samples were collected from sites B and E (Figure 1) in Qinghai Lake, with depths of 12.5 m and 13.6 m, respectively. A Horiba multi-parameter meter (W- 20XD Series, HORIBA, Kyoto, Japan) was used to measure the in-situ environmental parameters temperature, pH, conductivity, dissolved oxygen (DO), depth, chloride (Cl-), and salinity. Lake water was pumped and filtered through a polycarbonate filter with a pore size of 0.22 µm (Supor polyethersulfone; Pall life Sciences, Ann Arbor, MI, USA). Filtered water was collected to measure the concentrations of the major ions (e.g., sulfide, sulfate, nitrite, nitrate, ferrous iron, and ammonia) using a HACH colorimeter (model CEL 850, HACH Chemical Co., Iowa, USA) as previously described (Huang et al., 2013). The filters containing microbial biomass were placed on dry ice immediately after filtering 10-12 L of lake water for each site, and stored at -80°C in the laboratory until further DNA and RNA extraction. A portable GPS unit (eTrex H, Garmin) was used to determine the location of each site.

RNA and DNA extraction

Total RNA was extracted from one half of the filter for each site following a modified version of the RNeasy kit (Qiagen, Valencia, CA, USA) as previously described (Poretsky et al., 2009; Gifford et al., 2011). Briefly, half of a frozen filter was thawed and vortexed for 10 min with 8 ml of RLT lysis buffer and 3 g of RNA PowerSoil beads (Mo- Bio, Carlsbad, CA, USA). RNA was then extracted using the RNeasy kit according to manufacturer’s instructions. A Turbo DNA-free kit (Ambion, Austin, TX, USA) was used to remove any residual DNA. RNA was then purified and concentrated using the

84 RNeasyMinElute Cleanup kit (Qiagen, Valencia, CA, USA) (Hollibaugh et al., 2011). Genomic DNA was extracted from the other half of the filter using the FastDNA Spin Kit (MP Biomedical, OH, USA) as described previously (Huang et al., 2013).

amoA amplification

The amoA gene of Qinghai Lake samples were amplified using AOB specific primer set: amoA-1F (5’-GGGGTTTCTACTGGTGGT-3’) and amoA-2R (5’- CCCCTCKGSAAAGCC -TTCTTC-3’) (Rotthauwe et al., 1997). All PCR amplifications were performed using the same conditions as described previously (Rotthauwe et al., 1997) with FailSafe PCR System (Epicentre Biotechnologies, Madison, WI). The amplicons were stained with EtBr and visualized on a 1% agarose gel.

Metatranscriptomic sample preparation

Ribosomal RNA (rRNA) in total RNA extracts was removed via a subtractive hybridization process using sample-specific biotinlated rRNA probes (Stewart et al., 2010). rRNA-subtracted RNA was amplified linearly using the MessageAmp II-Bacteria kit (Ambion, Austin, TX, USA). The amplified RNA was then converted to double- stranded cDNA using the Universal RiboClone cDNA synthesis system (Promega, Madison, WI, USA) using random hexamers primers (Stewart et al., 2010). The synthesized cDNA was purified with the QIAquick PCR purification kit (Qiagen, Valencia, CA, USA).

Library preparation and sequencing

The extracted DNA and synthesized cDNA was sheared to 500 bp using a Covaris ultrasonicator (Covaris Inc., Woburn, MA, USA) according to manufacture’s recommendations. The sheared DNA was end repaired, adaptor ligated with multiplexing, and purified using the Ovation SP ultralow DR multiplex system (NuGEN Technologies Inc., CA, USA). The prepared library was then sequenced using an Illumina MiSeq with 250 base pairs (bp) paired end reads in the Quake laboratory at Stanford University.

85 Data analysis

The sequences were paired, trimmed, and filtered using the CLC Genomics Workbench version 6 (CLC Bio, Aarhus, Denmark). Paired reads were assembled together if 25 bp overlapped. Reads were trimmed based on the length (minimum length 50 bp) and quality (quality score ≥ 20) (Jiménez et al., 2012). Sequences were then uploaded to the metagenomics RAST (MG-RAST) server (Meyer et al., 2008) for annotation and are available under MG-RAST ID 4532866.3 4532865.3, 4532826.3, 4532825.3 for the samples B_DNA, B_RNA, E_DNA, E_RNA, respectively.

The datasets for both DNA and RNA libraries were taxonomically and functionally assigned using BLASTX (Altschul et al., 1997) on MG-RAST v3.3.6 against M5NR database, which integrates multiple databases (e.g., NCBI-nr, KEGG, SEED, and etc.) with a bit score cut-off of 50, E-values of 1×10-5, and a minimum alignment of 50 bp (Jiménez et al., 2012).

DNA and RNA reads that were annotated with the above criteria and contained both taxonomy and function were downloaded from MG-RAST. A custom R script was used to search for annotations of proteins involved in the carbon, nitrogen, and stress response genes (R Development Core Team 2012). A table was created that contained the number of reads that annotated to a function and organism. The number of reads was then transformed into percent of total reads for each sample. This table was then used to calculate the richness of each functional gene by counting the number of species or phyla that contained that functional gene.

Rarefaction curves were created for each of the samples to determine if our sequencing depth was sufficient to detect the number of species containing a functional gene. This was performed using the “rarecurve” function in the R package Vegan (Oksanen et al., 2011).

A network of correlated genes (functional genes found or expressed in the same organisms as stress response genes) was created using the WGCNA package (Langfelder and Horvath 2008) and the igraph package (Csardi and Nepusz 2006) in R (R Development Core Team 2012) then viewed in Cytoscape (Shannon et al., 2003). To reduce the complexity of the network, species that only had one functional gene and

86 functional genes with only one organism were removed (Kara et al., 2013). A spearman correlation between each functional gene and a stress response gene was calculated with a p-value adjusted for multiple comparisons. All correlations were removed that had a p- value greater than 0.05.

Results and Discussion

Water geochemistry

Sites B and E from Qinghai Lake (Figure 1) shared similar geochemical profiles (Table 1): both sites had 1.4% salinity and a pH of 9.1. The dissolved oxygen (DO) was 8.9 and 8.7 ppm for sites B and E, respectively. Site E was 1 m deeper and 2°C warmer than site B. In addition, the ammonia concentration was higher at site E (1 M) than site B (below detection).

Descriptive statistics of DNA and RNA libraries

After quality control, a total of 1,728,111 and 2,893,577 reads were obtained for the RNA libraries for sites B and E, respectively (Table 2). A total of 10,514,407 and 4,035,731 of reads were retrieved in DNA libraries for sites B and E, respectively (Table 2). Reads that were annotated by function and organism ranged between 28-81% of the reads identified to have an open reading frame (Table 2). Protein and rRNA features were predicted and identified for RNA and DNA libraries from both sample sites (Table 2).

Taxonomic diversity

Protein-coding genes, while not as phylogenetically robust as 16S rRNA genes, can be used to identify approximate taxonomic affiliations. A total of 832 genera within 31 phyla were detected in Qinghai Lake. The DNA samples from both sites contained more genera than was detected in the RNA samples. An additional 100 genera were detected at site B_DNA compared to site E_DNA. This is likely because more reads were obtained for the site B_DNA sample (~6 million more reads) (Table 2). Bacterial

87 sequences dominated all sites and samples, with each sample containing < 1% Archaeal sequences. Eukaryota sequences were annotated by MG-RAST; however, these reads were removed from further analysis because of the unreliability of FragGeneScan (prokaryotic gene calling algorithm used by MG-RAST) to identify Eukaryotic open reading frames (Rho et al., 2010).

Phototrophic organisms, mainly Synechococcus within the Cyanobacteria phyla, dominated the B_DNA and E_DNA samples with 52 and 63% of total abundance, respectively (Figure 2). Synechococcus was previously detected in clone libraries from Qinghai Lake (Dong et al., 2006; Jiang et al., 2006; Xing et al., 2009) and are known to be important contributors to carbon fixation in freshwater and marine ecosystems (Joint 1986; Fahnenstiel et al., 1991). The second most abundant phyla in the B_DNA and E_DNA samples were Proteobacteria, mainly the genus Loktanella, with 28 and 24%, respectively (Figure 2). Loktanella are strict aerobic heterotrophs that can tolerate saline conditions, and was first isolated from Antarctic mats (Van Trappen et al., 2004). Other major phyla that were detected in the DNA samples include Actinobacteria, Bacteroidetes, Planctomycetes, and Verrucomicrobia (Figure 2).

In the RNA samples, unlike the DNA samples, Proteobacteria were the most abundant (B_RNA: 47%, E_RNA: 70%) and Cyanobacteria were the second most abundant (B_RNA: 35%, E_RNA: 8%) (Figure 2). Assuming the ratio of RNA to DNA reflects the relative metabolic activity level (Yu and Zhang 2012), Proteobacteria were 1.6 to 2.8 fold more active than Cyanobacteria in samples B_RNA and E_RNA, respectively. Other major phyla that were detected in the RNA samples were Firmicutes, Actinobacteria, Bacteroidetes, and Verrucomicrobia (Figure 2).

Carbon cycle

Functional identification of protein-coding genes identified both photoautotrophic and heterotrophic carbon metabolisms (Figure 3). There are five known pathways for autotrophs to fix carbon (Berg 2011). In Qinghai Lake, carbon fixation was performed by ribulose bis-phosphate carboxylase (RuBisCo) and is indicative of the Calvin-Benson- Bassham (CBB) cycle (Berg 2011). No other carbon fixation pathway was detected.

88 However, activity levels of RuBisCo were low compared to heterotrophic processes (see below), 0.46 and 0.2 RNA:DNA ratios for B and E, respectively. Both photosystems I and II were detected, indicating their reliance on light energy. Photosystem II (1.86 and 0.65 RNA:DNA for B and E, respectively) was more highly expressed than photosystem I (0.52 and 0.07 RNA:DNA for B and E, respectively) (Figure 3). The higher expression of photosystem II than photosystem I indicates that the Cyanobacteria are under UV stress (Campbell et al., 1998), which is expected in Qinghai Lake because of its high elevation and high UV irradiance. In addition, network analysis showed that organisms expressing photosystem I and II were also expressing orange carotenoid protein (Figure 4). These data suggest that the photoautotrophs in Qinghai Lake had low activities and were under UV stress at the time of sampling.

Heterotrophic remineralization of organic carbon proceeded through the glycolysis and tricarboxylic acid cycle (TCA). The key enzymes for glycolysis (phosphofructokinase: 0.67 and 0.76 RNA:DNA for B and E, respectively) and the TCA cycle (pyruvate dehydrogenase: 0.62 and 0.66 RNA:DNA for B and E, respectively) had slightly higher activities than RuBisCo. Oxidative phosphorylation, identified by cytochrome C oxidase, was the most abundant and active (1.53 and 1.99 RNA:DNA for B and E, respectively) energetic pathway. The high concentration of oxygen (8.7-8.9 ppm) suggests aerobic respiration was the most prominent. Network analysis showed that these heterotrophs were also expressing genes involved in osmotic and oxidative stress (Figure 4).

Sites B and E were very similar in the taxonomy and functional genes that were detected; however, slight differences were observed in the photosynthetic and heterotrophic activity. Site E had less photosynthetic activity and more heterotrophic activity compared with site B. A possible reason for this is that Site E was 1 m deeper than site B and would have less photosynthetically active radiation (PAR) because of the water attenuation of PAR (Miller et al., 2004).

Nitrogen cycle

89 Ammonia is the biologically available form of nitrogen and ammonia assimilation into organic molecules using glutamine synthetase was the most active assimilatory pathway (Figure 5), and probably led to the low concentration of ammonia at both sites (≤1 mg/L) (see seasonal dynamics below). In addition, nitrogen fixation genes were detected and active (0.88 and 0.95 RNA:DNA for B and E, respectively). Denitrification pathways had very low abundance; however, had high RNA:DNA ratios (Figure 5). Denitrification requires anoxic conditions, so it was unexpected to detect denitrification transcripts in the oxic water column. It is possible that localized anoxic conditions can exist because of oxygen micro-gradients, such as been shown in marine snow (Ploug 2001) and microbial mats (Cravo-Laureau and Duran 2014). These small patches of anoxic water could harbor a relatively low abundance of anaerobic denitrifiers that are very active.

Biological ammonia oxidation is the first step in the nitrification process (NH3  - NO2 ) and is carried out by ammonia oxidizing bacteria (AOB) and ammonia oxidizing archaea (AOA) that contain the ammonia monooxygenase (amoA) gene (Thamdrup 2012). Putative AOB and AOA were detected in both sites. For example, AOB genera that were detected were Nitrosospira, Nitrosomonas, and Nitrosococcus. AOA that were detected were Nitrosocaldus, and Nitrosopumilus. The AOB were more abundant than AOA in all samples; however, AOB and AOA comprised less than 0.1% and 0.01% of total reads, respectively. Despite detecting putative ammonia oxidizers, the amoA gene was not detected in our metagenomic or metatranscriptomic samples. However, amoA was detected using primers specific for amoA and has been detected in previous studies (Jiang et al., 2009). The abundance of the amoA gene might be too low to be detected in metagenomic surveys.

Seasonal dynamics

Seasonal changes in the microbial community occur in many inland lakes and oceans (Pernthaler et al., 1998; Yannarell et al., 2003; Treusch et al., 2009; Gilbert et al., 2012). It was expected that the microbial community would change in Qinghai Lake with the season because of seasonal variations in physicochemical conditions. Qinghai Lake is

90 frozen for 9 months of the year. This ice cover likely causes stratification of the water column, resulting in gradients of salinity, light availability, and oxygen. In spring, glacier meltwater and thawing permafrost reduces the salinity and increases the flow of organic matter and nutrients into the lake (Li et al., 2008; Qiu 2008; Liu et al., 2009). Previously, the microbial community has been described in samples collected in early summer (July 2005 and 2007) and was near the samples described here (Jiang et al., 2009). The early summer samples had DO concentration of 5.4 ppm at the surface and 7 ppm at the bottom, salinity was 1.25%, and ammonia concentrations were 5.6 mM. The high ammonia and higher DO at depth is indicative of organic matter decomposition. The microbial community in the early summer was dominated by AOA and AOB likely because of the high ammonia concentrations. In contrast, late summer samples described here had higher DO (8.9 ppm), higher salinity (1.4%), and no measureable ammonia. Phototrophs dominated the DNA samples, while heterotrophs were more active, indicating that cyanobacterial blooms occur in Qinghai Lake but our sampling was near the end of the bloom. Therefore, the dominant microbial population in Qinghai Lake shifts throughout the summer in conjunction with physiochemical parameters.

Genetic Diversity

The ability to assign both taxonomy and function to a sequence allows us to identify the number of organisms that contain a particular functional gene. Generally a positive relationship between biodiversity and ecosystem functioning exists because a more diverse community increases the likelihood of functional redundancy or the increased probability of finding high-performing key species (Yachi and Loreau 1999; Griffiths et al., 2000; Degens et al., 2001). For example, when response to increasing salinity, denitrification rates were higher and decreased more slowly in soils with higher 16S rRNA gene diversity than in soils with lower 16S rRNA gene diversity (Hallin et al., 2012). Similar results were found for respiration (Langenheder et al., 2005; Chowdhury et al., 2011; Setia et al., 2011), photosynthesis (Murata et al., 2007; Allakhverdiev and Murata 2008; Takahashi and Murata 2008), aerobic methane oxidation (Sorokin et al., 2000; Heyer et al., 2005; van der Ha et al., 2010), and nitrogen fixation (Severin et al.,

91 2012).

The richness of each functional gene was calculated based on the number of organisms that contained a particular functional gene. The most diverse functional genes were cytochrome C oxidase, glutamine synthetase, phosphofructokinase, and pyruvate dehydrogenase (Figure 6). Photosystem I, photosystem II, and genes involved in the denitrification process had low diversity (Figure 6). Rarefaction curves show that the diversity was sufficiently sampled for most genes.

If the biodiversity hypothesis holds true, then the heterotrophic pathways are the most resilient to stress because they have the highest richness. However, each organism can have a different response to stress. For example, Synechococcus has multiple protective mechanisms to cope with UV stress and can maintain photosynthesis, while Procholorcoccus lacking these protective mechanisms shuts down several key metabolic processes under a similar UV stress (Mella-Flores et al., 2012). Therefore, we assessed which organisms had genes that are used in response to environmental stress, and found genes involved in UV, osmotic, and oxidative stress (Figure 4). It turned out that the number of species with a functional gene corresponded logarithmically to the number of stress response genes (Figure 7), indicating that functional genes with higher richness have more possible responses to stress and would be expected to be more resilient.

Conclusions

A metagenomic and metatranscriptomic survey of two sites from Qinghai Lake water columns identified microbial processes involved in the carbon and nitrogen cycle. While photoautotrophic organisms were the most abundant in the DNA samples, heterotrophic organisms were the most active at both sites. Coupled with previous reports from early summer sampling, our data suggests that successional changes occur in the microbiota throughout the summer. Our sampling was likely at the end of a cyanobacterial bloom and heterotrophic processes were beginning to remineralize the organic carbon. Our data also show a positive relationship between the richness of a functional gene and the number of stress response genes, suggesting an increasing ability to respond to a variety of environmental stresses. This provides a molecular mechanism

92 as to why highly diverse environments are generally more stable. Based on this biodiversity hypothesis, heterotrophs and ammonia assimilation had the highest richness and would be expected to be more resistant to environmental change.

Acknowledgments

This research was supported by the National Natural Science Foundation of China (Grant Nos. 41030211 and 41002123), the Scientific Research Funds for the 1000 “Talents” Program Plan from China University of Geosciences (Beijing), State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences (No. GBL11201), and the Fundamental Research Funds for National University, China University of Geosciences (Wuhan). .

93 References

Allakhverdiev SI, Murata N (2008). Salt stress inhibits photosystems II and I in cyanobacteria. Photosynth Res 98: 529-539.

Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W et al. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389-3402.

Bardgett RD, Freeman C, Ostle NJ (2008). Microbial contributions to climate change through carbon cycle feedbacks. ISME J 2: 805-814.

Berg IA (2011). Ecological aspects of the distribution of different autotrophic CO2 fixation pathways. Appl Environ Microbiol 77: 1925-1936.

Campbell D, Eriksson M-J, Öquist G, Gustafsson P, Clarke AK (1998). The cyanobacterium Synechococcus resists UV-B by exchanging photosystem II reaction-center D1 proteins. Proc Natl Acad Sci USA 95: 364-369.

Canfield DE, Glazer AN, Falkowski PG (2010). The evolution and future of Earth’s nitrogen cycle. Science 330: 192-196.

Chowdhury N, Marschner P, Burns R (2011). Response of microbial activity and community structure to decreasing soil osmotic and matric potential. Plant Soil 344: 241-254.

Cravo-Laureau C, Duran R (2014). Marine coastal sediments microbial hydrocarbon degradation processes: contribution of experimental ecology in the omics’ era. Front Microbiol 5: 39.

Csardi G, Nepusz T (2006). The igraph software package for complex network research. Complex Systems 1695.

Degens BP, Schipper LA, Sparling GP, Duncan LC (2001). Is the microbial community in a soil with reduced catabolic diversity less resistant to stress or disturbance? Soil Bio Biochem 33: 1143-1153.

94 Dong H, Zhang G, Jiang H, Yu B, Chapman LR, Lucas CR et al. (2006). Microbial diversity in sediments of saline Qinghai Lake, China: linking geochemical controls to microbial ecology. Microb Ecol 51: 65-82.

Fahnenstiel GL, Carrick HJ, Iturriaga R (1991). Physiological characteristics and food- web dynamics of Synechococcus in Lakes Huron and Michigan. LIMNOLOGY 36: 219-234.

Gifford SM, Sharma S, Rinta-Kanto JM, Moran MA (2011). Quantitative analysis of a deeply sequenced marine microbial metatranscriptome. ISME J 5: 461-472.

Gilbert JA, Steele JA, Caporaso JG, Steinbrück L, Reeder J, Temperton B et al. (2012). Defining seasonal marine microbial community dynamics. ISME J 6: 298-308.

Griffiths BS, Ritz K, Bardgett RD, Cook R, Christensen S, Ekelund F et al. (2000). Ecosystem response of pasture soil communities to fumigation-induced microbial diversity reductions: an examination of the biodiversity-ecosystem function relationship. OIKOS 90: 279-294.

Guo D, Wang H, Li D (2012). A projection of permafrost degradation on the Tibetan Plateau during the 21st century. J Geophys Res Atmos 117: D05106.

Hallin S, Welsh A, Stenstrom J, Hallet S, Enwall K, Bru D et al. (2012). Soil functional operating range linked to microbial biodiversity and community composition using denitrifiers as model guild. PLoS One 7: e51962.

Heimann M, Reichstein M (2008). Terrestrial ecosystem carbon dynamics and climate feedbacks. Nature 451: 289-292.

Heyer J, Berger U, Hardt M, Dunfield PF (2005). Methylohalobius crimeensis gen. nov., sp. nov., a moderately halophilic, methanotrophic bacterium isolated from hypersaline lakes of Crimea. Int J Syst Evol Microbiol 55: 1817-1826.

Hollibaugh JT, Gifford S, Sharma S, Bano N, Moran MA (2011). Metatranscriptomic analysis of ammonia-oxidizing organisms in an estuarine bacterioplankton assemblage. ISME J 5: 866-878.

95 Huang Q, Jiang H, Briggs BR, Wang S, Hou W, Li G et al. (2013). Archaeal and bacterial diversity in acidic to circumneutral hot springs in the Philippines. FEMS Microbiol Ecol 85: 452-464.

Jiang H, Dong H, Zhang G, Yu B, Chapman LR, Fields MW (2006). Microbial diversity in water and sediment of Lake Chaka, an athalassohaline lake in northwestern China. Appl Environ Microbiol 72: 3832-3845.

Jiang H, Dong H, Yu B, Ye Q, Shen J, Rowe H et al. (2008). Dominance of putative marine benthic Archaea in Qinghai Lake, north-western China. Environ Microbiol 10: 2355-2367.

Jiang H, Dong H, Yu B, Lv G, Deng S, Berzins N et al. (2009). Diversity and abundance of ammonia-oxidizing Archaea and Bacteria in Qinghai Lake, Northwestern China. Geomicrobiol J 26: 199-211.

Jiménez DJ, Andreote FD, Chaves D, Montaña JS, Osorio-Forero C, Junca H et al. (2012). Structural and functional insights from the metagenome of an acidic hot spring microbial planktonic community in the Colombian Andes. PLoS ONE 7: e52069.

Joint I (1986). Physiological ecology of picoplankton in various oceanographic provinces. Photosynthetic picoplankton Can Bull Fish Aquat Sci 214: 287-309.

Kara EL, Hanson PC, Hu YH, Winslow L, McMahon KD (2013). A decade of seasonal dynamics and co-occurrences within freshwater bacterioplankton communities from eutrophic Lake Mendota, WI, USA. ISME J 7: 680-684.

Langenheder S, Lindstrom ES, Tranvik LJ (2005). Weak coupling between community composition and functioning of aquatic bacteria. Limnol Ocean 50: 957-967.

Langfelder P, Horvath S (2008). WGCNA: an R package for weighted correlation network analysis. BMC bioinformatics 9: 559.

Lei Y, Yao T, Yi C, Wang W, Sheng Y, Li J (2012). Glacier mass loss induced the rapid growth of Linggo Co on the central Tibetan Plateau. J Glaciology 58: 177-184.

96 Li X, Cheng GD, Jin HJ, Kang E, Che T, Jin R et al. (2008). Cryospheric change in China. Global Planet Change 62: 210-218.

Liu J, Wang S, Yu S, Yang D, Zhang L (2009). Climate warming and growth of high- elevation inland lakes on the Tibetan Plateau. Global Planet Change 67: 209-217.

Mella-Flores D, Six C, Ratin M, Partensky F, Boutte C, Le Corguille G et al. (2012). Prochlorococcus and Synechococcus have evolved different adaptive mechanisms to cope with light and UV stress. Front Microbiol 3: 285.

Meyer F, Paarmann D, D'Souza M, Olson R, Glass E, Kubal M et al. (2008). The metagenomics RAST server - a public resource for the automatic phylogenetic and functional analysis of metagenomes. BMC Bioinformatics 9: 386.

Miller A, Chai F, Chiba S, Moisan J, Neilson D (2004). Decadal-scale climate and ecosystem iInteractions in the North Pacific Ocean. J Oceanogr 60: 163-188.

Murata N, Takahashi S, Nishiyama Y, Allakhverdiev SI (2007). Photoinhibition of photosystem II under environmental stress. Biochim Biophys Acta 1767: 414-421.

Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O'Hara B et al. (2011). vegan: Community ecology package: http://cran.r-project.org/, http://vegan.r- forge.r-project.org/.

Pernthaler J, Glöckner F-O, Unterholzner S, Alfreider A, Psenner R, Amann R (1998). Seasonal community and population dynamics of pelagic Bacteria and Archaea in a high mountain lake. Appl Environ Microbiol 64: 4299-4306.

Ploug H (2001). Small-scale oxygen fluxes and remineralization in sinking aggregates. Limnol Oceanogr 46: 1624-1631.

Poretsky RS, Hewson I, Sun S, Allen AE, Zehr JP, Moran MA (2009). Comparative day/night metatranscriptomic analysis of microbial communities in the North Pacific subtropical gyre. Environ Microbiol 11: 1358-1375.

Qiu J (2008). China: the third pole. Nature 468: 141-142.

R Development Core Team (2012). R: A language and environment for statistical computing. R Found. for Stat. Comput.: Vienna.

97 Rho M, Tang H, Ye Y (2010). FragGeneScan: predicting genes in short and error-prone reads. Nucleic Acids Res 38: e191.

Rotthauwe J-H, Witzel K-P, Liesack W (1997). The ammonia monooxygenase structural gene amoA as a functional marker: molecular fine-scale analysis of natural ammonia-oxidizing populations. Appl Environ Microbiol 63: 4704-4712.

Schimel DS (1995). Terrestrial ecosystems and the carbon cycle. Glob change biol 1: 77- 91.

Setia R, Marschner P, Baldock J, Chittleborough D, Smith P, Smith J (2011). Salinity effects on carbon mineralization in soils of varying texture. Soil Biol Biochem 43: 1908-1916.

Severin I, Confurius-Guns V, Stal LJ (2012). Effect of salinity on nitrogenase activity and composition of the active diazotrophic community in intertidal microbial mats. Arch Microbiol 194: 483-491.

Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D et al. (2003). Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome research 13: 2498-2504.

Singh BK, Bardgett RD, Smith P, Reay DS (2010). Microorganisms and climate change: terrestrial feedbacks and mitigation options. Nature Rev Microbiol 8: 779-790.

Sorokin DY, Jones BE, Kuenen JG (2000). An obligate methylotrophic, methane- oxidizing Methylomicrobium species from a highly alkaline enviornment. Extremophiles 4: 145-155.

Stewart FJ, Ottesen EA, DeLong EF (2010). Development and quantitative analyses of a universal rRNA-subtraction protocol for microbial metatranscriptomics. ISME J 4: 896-907.

Takahashi S, Murata N (2008). How do environmental stresses accelerate photoinhibition? Trends Plant Sci 13: 178-182.

Thamdrup B (2012). New pathways and processes in the global nitrogen cycle. Annu Rev Ecol Evolution Syst 43: 407-428.

98 Treusch AH, Vergin KL, Finlay LA, Donatz MG, Burton RM, Carlson CA et al. (2009). Seasonality and vertical structure of microbial communities in an ocean gyre. ISME J 3: 1148-1163.

van der Ha D, Hoefman S, Boeckx P, Verstraete W, Boon N (2010). Copper enhances the activity and salt resistance of mixed methane-oxidizing communities. Appl Microbiol Biotechnol 87: 2355-2363.

Van Trappen S, Vandecandelaere I, Mergaert J, Swings J (2004). Algoriphagus antarcticus sp. nov., a novel psychrophile from microbial mats in Antarctic lakes. Int J Syst Evol Microbiol 54: 1969-1973.

Wang H, Liu W, Zhang CL, Jiang H, Dong H, Lu H et al. (2013). Assessing the ratio of archaeol to caldarchaeol as a salinity proxy in highland lakes on the northeastern Qinghai–Tibetan Plateau. Organic Geochem 54: 69-77.

Xing P, Hahn MW, Wu QL (2009). Low taxon richness of bacterioplankton in high- altitude lakes of the Eastern Tibetan Plateau, with a predominance of Bacteroidetes and Synechococcus spp. Appl Environ Microbiol 75: 7017-7025.

Yachi S, Loreau M (1999). Biodiversity and ecosystem productivity in a fluctuating environment: the insurance hypothesis. Proc Natl Acad Sci U S A 96: 1463-1468.

Yang SJ, Dong H, Lei FM (2009). Phylogeography of regional fauna on the Tibetan Plateau: A review. Progr Natural Sci 19: 789-799.

Yannarell AC, Kent AD, Lauster GH, Kratz TK, Triplett EW (2003). Temporal patterns in bacterial communities in three temperate lakes of different trophic status. Microb Ecol 46: 391-405.

Yu K, Zhang T (2012). Metagenomic and metatranscriptomic analysis of microbial community structure and gene expression of activated sludge. PLoS ONE 7: e38183.

99 Table 1. Sample locations and geochemistry. Parameters B E GPS location (N, E) 36.66, 100.60 36.74, 100.69 Temperature (oC) 13.7 15.6 pH 9.1 9.2 Depth (m) 12.5 13.6 Conductivity (s/m) 2.28 2.26 DO (ppm) 8.9 8.7 Salinity (%) 1.4 1.4 Cl- (mg/L) 2930 3420 + NH4 (mg/L) ND 1 - NO3 (mg/L) 0.39 0.4 - NO2 (mg/L) 0.13 NA Si (mg/L) 1.43 NA S2- (mg/L) 0.05 ND 2- SO4 (mg/L) >80 >80 Fe2+ (mg/L) 0.065 NA ND: not detected. NA: not available

100 Table 2. Statistic results of metagenome and metatranscriptome sequences from sites B and E in Qinghai Lake. B E RNA DNA RNA DNA

Number of reads 1,728,111 10,514,407 2,893,577 4,035,731

Mean sequence length (bp) 211±60 194±55 200±59 185±59

Total Mbp 365 2,049 581 749

Mean GC content (%) 49±8 60±8 44±9 63±5

Reads with ORFa 1,463,714 10,161,795 2,098,251 3,901,755 Identified protein features 288,106 3,233,826 304,159 945,021

Identified rRNA features 413,195 10,573 113,464 2,307 Identified functional categories 177,169 2,467,216 152,624 729,594 % annotated readsb 65.0 81.2 28.3 77.3 a Open reading frame

b % of reads identified to have an open reading frame that were annotated by function and taxonomy.

101 Figure captions: Figure 1. A geographical map showing the sampling sites in Qinghai Lake, China. Figure 2. Distribution of phyla detected in the DNA or RNA samples from sites B and E determined by taxonomic assignment of metagenomic or metatranscriptomic reads. Phyla with <1% abundance were grouped into “other”. Figure 3. The carbon cycle depicted by a generalized autotroph and heterotroph in Qinghai Lake. The numbers in boxes represent either the percentage or the RNA:DNA ratio of reads that were annotated within each metabolic pathway for sites B and E. The key genes used to identify a pathway was Ribulose-bisphosphate carboxylase (RuBisCo): Calvin-Benson-Bassham cycle (CBB), D-glucose 6-phosphotransferase: glycolysis, pyruvate dehydrogenase: tricarboxylic acid cycle (TCA), and cytochrome C oxidase: oxidative phosphorylation. Figure 4. Network showing correlations between carbon and nitrogen cycle-related genes to stress response genes for DNA and RNA samples. Lines represent genes that are directly correlated (i.e. detected (DNA) or expressed (RNA) in the same species). Only significant correlations (p<0.05) are shown. The type of stress that each gene is involved in is also listed. Figure 5. DNA and RNA reads detected and annotated within the nitrogen cycle. The numbers in boxes represent either the percent or the RNA:DNA ratio of reads that were annotated as genes within each pathway for sites B and E. Figure 6. Rarefaction curves for the detected genes related to carbon and nitrogen cycle. Highly saturated curves (e.g. photosystem I and II) indicate that sequencing depth was deep enough to capture most species containing the corresponding gene. Figure 7. Plot showing the number of species that contained a particular functional gene in the carbon and nitrogen cycle and the number of stress genes that were found in all species with that functional gene. Triangle symbols represent RNA samples and circle symbols represent DNA samples. Best fit line followed a logarithmic function.

102 Figure 1.

103 Figure 2.

104 Figure 3.

105 Figure 4.

106 Figure 5.

107 Figure 6.

108 Figure 7.

109 CHAPTER 5:

Permanganate Diffusion and Reaction in Sedimentary Rocks

Qiuyuan Huang1, Hailiang Dong1*, Rachael M. Towne2, Timothy B. Fischer1, and Charles E. Schaefer2,

1: Department of Geology and Environmental Earth Science, Miami University, Oxford, OH 45056 2: CB&I, Inc., Lawrenceville, NJ 08648

Corresponding author: Hailiang Dong Department of Geology and Environmental Earth Science Miami University Oxford, OH 45056 [email protected]

Published in Journal of Contaminant Hydrology (2014) 159: 36-46

110 ABSTRACT In situ chemical oxidation using permanganate has frequently been used to treat chlorinated solvents in fractured bedrock aquifers. However, in systems where matrix back-diffusion is an important process, the ability of the oxidant to migrate and treat target contaminants within the rock matrix will likely determine the overall effectiveness of this remedial approach. In this study, a series of diffusion experiments were performed to measure the permanganate diffusion and reaction in four different types of sedimentary rocks (dark gray mudstone, light gray mudstone, red sandstone, and tan sandstone). Results showed that, within the experimental time frame (~2 months), oxidant migration into the rock was limited to distances less than 500 microns. The observed diffusivities for permanganate into the rock matrices ranged from 5.3×10-13 to 1.3×10-11 cm2/s. These values were reasonably predicted by accounting for both the rock oxidant demand and the effective diffusivity of the rock. Various Mn minerals formed as surface coatings from reduction of permanganate coupled with oxidation of total organic carbon (TOC), and the nature of the formed Mn minerals was dependent upon the rock type. Post-treatment tracer testing showed that these Mn mineral coatings had a negligible impact on diffusion through the rock. Overall, our results showed that the extent of permanganate diffusion and reaction depended on rock properties, including porosity, mineralogy, and organic carbon. These results have important implications for our understanding of long-term organic contaminant remediation in sedimentary rocks using permanganate.

Keywords Diffusion, Permanganate, Remediation, Rock matrix

111 INTRODUCTION

Remediation of chlorinated solvents in fractured bedrock aquifers has been a challenging task at several Department of Defense (DOD) and Department of Energy (DOE) sites. In situ chemical oxidation (ISCO) using strong oxidants (e.g., permanganate, persulfate, ozone, and hydrogen peroxide) has been effectively applied for treatment of chlorinated solvents (Conrad et al., 2002; Krembs et al., 2010; Schnarr et al., 1998; Siegrist et al., 2011). Several studies have demonstrated that uptake and back-diffusion of contaminants from the bedrock matrix can have a substantial impact on contaminant fate and treatment potential (Goldstein et al., 2004; Lipson et al., 2005; Sterling et al., 2005; West and Kueper, 2010). While there have been studies examining the effectiveness of chemical oxidants on the removal of dense non-aqueous phase liquids (DNAPLs) in fractured rocks (Schaefer et al., 2012; Smith et al., 1998; Tunnicliffe and Thomson, 2004), published peer-reviewed studies evaluating the ability of chemical oxidant to treat chlorinated solvents residing in the rock matrix are, to the best of our knowledge, not available except for one TCE treatability study in silty clay soil (Struse et al., 2002).

Permanganate has been applied to treat several contaminants, such as trichloroethylene (TCE), tetrachloroethylene (PCE), and styrene in the aqueous phase (Gates-Anderson et al., 2001; Tunnicliffe and Thomson, 2004; USEPA, 1998; Wu et al., 2012; Yan and Schwartz, 1999). For example, in a coupled reaction between - permanganate and TCE, MnO4 is reduced to form a solid Mn oxide precipitate

(birnessite, MnO2), while chlorinated ethylenes break down to produce CO2 (Li and Schwartz, 2004; Yan and Schwartz, 1999).

The success of this ISCO technology depends on a number of considerations, among which rock properties and naturally occurring reactive reductants are important ones. For example, porosity and permeability are important parameters affecting the susceptibility of contaminant to chemical oxidation in aquifer sediments (Bogan and Trbovic, 2003; Hønning et al., 2007a). Besides, naturally present reactive reductants such as total organic carbon (TOC) and reduced forms of iron, manganese, and sulfur, can also impact the effectiveness of the ISCO technology because these reductants may compete with TCE (or other contaminants) for permanganate and may influence the stability and

112 mobility of permanganate (Mumford et al., 2005; Reynolds et al., 2008; Urynowicz, 2007; Xu and Thomson, 2009). Therefore, permanganate diffusion is likely to vary in different rock types due to the reactions between permanganate and different concentrations of naturally present reductants.

Permanganate oxidation of organic contaminants (e.g., TCE) typically results in

the formation of MnO2 solids (Crimi and Siegrist, 2004). A detailed mineralogical study of these solids further revealed semi-amorphous potassium-rich birnessite as a typical product (Li and Schwartz, 2004); however, the exact oxidation state of Mn in such material was not determined. In a field oxidation test with potassium permanganate, Scott et al., (2011) observed birnessite formation as well as an unidentified, poorly crystalline material that was believed to be a manganese oxide phase with a higher oxidation state of manganese. Loomer et al., (2010) studied the Mn oxidation state in manganese oxides formed by oxidation of TCE with permanganate using electron energy loss spectroscopy (EELS) and X-ray photoelectron spectroscopy (XPS). The authors found that the valence of Mn in such manganese oxides can vary from 2.2 to 3.6 depending on the amount of permanganate, pH, and aging (Loomer et al., 2010; Scott et al., 2011). This study raises the possibility that different Mn minerals can form depending on specific experimental conditions; however, the exact nature of such Mn solids remains unknown, especially in different rock types (e.g., mudstones vs. sandstones).

As described above, despite the importance of rock properties on the effectiveness of ISCO, a comprehensive understanding regarding the interaction between chemical oxidants and rock matrices is still lacking. It is currently unclear how implementation of ISCO impacts the bedrock matrix, as the reactions between oxidants and rock matrix at the fracture-rock interface and diffusion of oxidant into rock matrix could impact rock properties through Mn oxide precipitation and subsequently impact contaminant diffusive flux from the matrix to water-bearing fractures. Therefore, an improved understanding regarding diffusion of chemical oxidants into bedrock and oxidant reactions with naturally occurring reductants is needed to assess the potential benefits of in situ chemical remediation of chlorinated solvents in bedrock aquifers.

113 The objective of this study was to understand the mechanisms and controlling factors of permanganate diffusion and reaction in a variety of sedimentary rocks. Specifically, the distances and rates of permanganate diffusion into the rocks were measured, the factors controlling permanganate diffusion were assessed, the nature of the interaction between the oxidant and rock were evaluated (including identification of resulting manganese precipitates), and post-oxidation impact on the rocks (with respect to alteration of the effective diffusion coefficient due to Mn precipitate clogging) were assessed. Insights gained as part of this study provided information useful for assessing the feasibility of applying permanganate for treatment of chlorinated compounds within rock matrices.

MATERIALS AND METHOD

Sample collection and rock characterization

Bedrock materials were collected from the former Naval Air Warfare Center (NAWC) in Trenton, New Jersey. The NAWC site is underlain by the Lockatong and Stockton formations on opposing sides of a geologic fault (Lewis-Brown et al., 2006). Water bearing fractures occur through each of these units, with some units more highly fractured than others. As summarized in Table 1, several different rock types, which were present at the NAWC site, were collected for this study. Four representative rock types were collected, including red sandstone (Stockton Formation), tan sandstone (Stockton Formation), dark gray mudstone (Lockatong Formation), and light gray mudstone (Lockatong Formation). Details of the collection process, which was performed to obtain minimally disturbed (chemically and physically) rock cores, were described previously (Schaefer et al., 2012; Schaefer et al., 2013). To limit oxygen exposure of the rocks, samples were stored in heat-sealed plastic bags under water-saturated and anoxic conditions. Once in the laboratory, samples were cut into 1cm thick slices using a diamond blade electric saw and subsequently stored inside an anaerobic glove box. Initial characterization of rocks, including rock porosity, ferrous iron content, total iron, and effective diffusion coefficient, were performed as described in our previous paper (Schaefer et al., 2012).

114

TOC and oxidant demand experiments

As a potential reductant to reduce permanganate, TOC within rock matrices was measured with 1 gram of crushed powdered rock. A slurry was created with ultra pure water, and the pH was lowered to less than 2 with 85% phosphoric acid to remove inorganic carbon. Samples were then dried at 105 oC overnight. The samples were run on an Apollo 9000 TOC analyzer with the 183 Boat Sampling Module according to EPA Methods 9060 (Ghosh et al., 2000).

To measure permanganate oxidant demand for each rock type, 2 grams of crushed rock were placed in a 25 ml amber glass serum bottle with 2000 ppm potassium permanganate solution (Haselow et al., 2003; Mumford et al., 2005). Controls were prepared with permanganate solution only, and all samples were run in duplicate. The time course decrease of aqueous permanganate concentration was measured using a Mattson Genesis II spectrophotometer at a wavelength of 526 nm (Apha, 1998). Measurements were taken at time zero and once a week thereafter. Once the aqueous permanganate concentration stabilized (about 1 month), oxidant demand was calculated in terms of the mass of permanganate reduced per kilogram of crushed rock.

Permanganate diffusion experiments into rock

A reaction cell was constructed using a plastic dish (16.5 cm×6.3 cm×4.5 cm) that was non-reactive with the oxidant (verified by control tests). A stainless steel frame was used to mount the rock and divide the cell into two halves. Rock slices (~1.0 cm thick) were epoxied onto the frame (covering the opening) using Masterbond EP41S-1HT. The stainless steel/rock divider was then epoxied into the plastic dish (Figure 1). Control experiments were performed to confirm that the oxidant demand of the divider and epoxy were negligible.

Synthetic groundwater containing 120 ml/L CaCl2-2H2O, 0.1 mg/L NaNO3, 70 mg/L NaHCO3, 4 mg/L K2CO3, and 60 mg/L MgSO4-7H2O was prepared based on the groundwater geochemistry in the vicinity of the collected rocks (Schaefer et al., 2012).

115 The reservoir (source) and sink sides of the reaction cell were prepared with potassium permanganate solution (2000 ppm in synthetic groundwater) and the synthetic groundwater, respectively. An equivalent molarity of potassium nitrate was applied to ensure ionic equivalency on each side of the divider. All experiments were performed in duplicate for each of the four rock types, and incubated in the diffusion cell for a period of 7 to 8 weeks. The permanganate concentration on both sides of the cell was measured spectrophotometrically to monitor change of Mn concentration as a function of time.

To determine the extent to which exposure to the oxidants impacted solute diffusion through the rocks, an iodide tracer experiment was subsequently performed on the oxidant impacted rocks. Diffusion cells were prepared in a similar fashion as the permanganate diffusion experiments (using the rock slices previously exposed to the oxidant), except that a 2000 ppm potassium iodide solution was added to the reservoir side instead of the oxidant. Iodide concentrations were determined using a Dionex dx-600 Ion Chromatograph. Iodide effective diffusion coefficients were calculated based on the flux of iodide through the rock, as described in our previous work (Schaefer et al., 2012).

X-ray diffraction (XRD) Analysis

The reaction between permanganate and rock matrix is expected to result in formation of various manganese minerals (Li and Schwartz, 2004; Post, 1999). XRD was used to identify Mn precipitates on the rock surfaces that were exposed to permanganate solution using a Scintag X1 powder diffractometer system (Huang et al., 2011). Samples were X-ray scanned from 2 to 70 degree two theta with Cu K-alpha radiation (40 kV, 35 mA), a 0.02 degree step size, and a count time of 2 seconds per step. A combination of search-match software and manual check was used to identify mineral phases.

Scanning electron microscopy (SEM)

SEM was performed to determine the morphology of mineral precipitates using a Zeiss Supra 35 VP Field Emission Scanning Electron Microscope with an EDX Genesis 2000 X-ray Energy Dispersive Spectrometer (EDS). SEM-EDS line scanning technique

116 was used to measure the penetration distance of permanganate over the experimental duration (~2 months). All line scanning profiles were performed in triplicates. To visually observe the distribution of Mn minerals in relation to TOC and various other minerals, elemental maps were collected using a SEM-EDS standardized method, with an accelerating voltage of 15kV in a vacuum and a working distance of 8.5 mm.

Determination of observed diffusivity of permanganate into rock matrices

The observed diffusivity of the oxidant through the rock was determined by assuming one-dimensional diffusion and reaction of permanganate into the rock, which is described as follows:

(1)

where C is the manganese concentration, t is the time [s], Dobs is the observed diffusivity [cm2/s], and x is the distance into the rock [cm]. Eq. (1) was solved assuming an initial condition of C=0 throughout the rock, a constant permanganate concentration at the rock

interface (Ci, determined from the SEM data at the rock interface), and C=0 for all times at x=infinity. This one-dimensional semi-infinite medium solution is well known (Crank, 1995; Cussler, 2009):

(2)

Eq. (2) was regressed to the experimental data to determine the measured value for Dobs for each rock type. 95% confidence intervals for the regressed values were calculated based on the methodology described previously (Smith et al., 1998).

For comparison, a calculated value of the observed diffusivity was determined as follows:

(3)

117 (4)

where R is the retardation factor, θ is the rock porosity (water uptake porosity, measured by Schaefer et al., 2013 for each of the rock types examined), and ρ is the rock density [assumed to be 2.6 kg/L]. The rock oxidant demand, α [g/kg], was determined from the

oxidant demand tests described in Section 2.2. Cox is the permanganate concentration in

the reservoir. Deff is the effective diffusion coefficient (where Deff accounts for the decrease in diffusion due to an increased tortuous pathway (Fetter, 1999)) for 2 permanganate through the rock [cm /s]. Deff values were previously calculated for each rock type for either trichloroethene or iodide (Schaefer et al., 2012; Schaefer et al., 2013); multiplying these values by the ratio in aqueous diffusivity between permanganate (4.0×10-5 cm2/s, (Kyriacou et al., 2006)) and the previously employed solute allowed for

determination of Deff for permanganate. Thus, the calculated values of Dobs,cal from Eqs.

(3) and (4) were compared to the measured value of Dobs from Eq. (2).

RESULTS AND DISCUSSION

Rock properties

Previously measured porosity, ferrous iron, and total iron information for each of the rock types (Schaefer et al., 2013) is provided in Table 1. TOC and 1-month permanganate oxidant demand ranged from 0.17-4.63 g/kg and 2.58-22.5 g/kg, respectively (Table 1). The available reductants (e.g., TOC and ferrous iron) are greater in the mudstone samples than in the sandstone samples. Corresponding to this trend, permanganate oxidant demand was substantially greater in the mudstones than in the sandstones. The measured oxidant demand values for the sandstones (2.58-2.98 g-

KMnO4 per kg of sandstones) compare favorably with those reported for fine/medium

and silty sand aquifer materials (0.77-2.12 g-KMnO4 per kg of aquifer materials) (Xu and - Thomson, 2009) and for glacial meltwater sand (0.5-2 gMnO4 per kg of dry weight sand) (Hønning et al., 2007a). Likewise, the values for the mudstone samples (10.3-22.5 g-

KMnO4 per kg of mudstones) also compare favorably to reported values for clayey till (5- - 20 g MnO4 per kg of dry weight till) (Hønning et al., 2007a).

118 The measured oxidant demand is better correlated with TOC than with ferrous iron (Table 1), suggesting that TOC exerts the greatest oxidant demand. This result is consistent with a previous study where reported oxidant demand values were highly correlated with TOC, but not with amorphous Fe (Xu and Thomson, 2009). Our measured KMnO4/TOC ratio for the sandstones (11.0-15.2) are similar to previously reported values of 9-15 for sandy aquifer material (Mumford et al., 2005); whereas this ratio for the mudstones (4.9-7.8) also falls within the range for clayey till materials (2.6- 9.2) (Hønning et al., 2007a).

Permanganate diffusion in different sedimentary rocks

SEM-EDS line scanning results obtained from cross sections of all rock types (perpendicular to the Mn-exposed surface) showed that there was a Mn concentration gradient from the exposed surface into rock matrix along the direction of Mn diffusion (Figure 2). Overall, the permanganate penetration distance varied from ~50 to 400 µm for

the different rock types. These diffusion profiles were fitted to obtain Dobs values. An example of the regression of Eq. (2) to the experimental Mn diffusion data into the rock is provided in Figure S1 (Supplemental content). Table 2 shows a summary of the

regressed Dobs values for the 4 rock types (in duplicate), along with the calculated values

of Dobs,cal (based on Eqs. (3) and (4)) and effective diffusion coefficient for each rock type. 5 Measured Dobs values are generally on the order of 10 -times less than the previously measured effective diffusion coefficients (measured for TCE or iodide) through the rock

(Schaefer et al., 2012; Schaefer et al., 2013). The calculated Dobs,cal values are on the

same order of magnitude as the measured Dobs values. Considering the extremely small diffusion distances into the rocks, and the natural heterogeneity of the rock matrices, this agreement is reasonable and suggests that the slow diffusional migration of the oxidant into the rock is greatly hindered by the reaction between permanganate and naturally occurring reductants within the rocks. Thus, the reaction between permanganate and the rock matrix is the primary cause for the very low values of Dobs. Based on the limited

penetration distance into the rock and the low value of Dobs, it is unlikely that application of chemical oxidants will provide treatment for target contaminants residing at distances

119 greater than a few hundred microns into the rock. The exception may be rocks that have extremely low oxidant demand (much less than those measured for the rocks used in this study) and longer oxidant persistence (much greater than the 2 month period of this study), which would allow for a more rapid migration of oxidant into the rock.

Our data further illustrated that over the course of permanganate diffusion, permanganate was reduced by naturally present chemical reductants within the rocks. SEM-EDS line scanning results revealed that manganese (Figure 2) and carbon (Figure 3) exhibited a similar diffusion profile, both of which showed a decreased concentration with penetration distance into the rock matrix. This Mn and C coupling was further confirmed by SEM-EDS elemental maps obtained from a cross section of the red sandstone (Figure 4), where the spatial distribution of Mn was approximately correlated with carbon, but not with any other elements. However, in this two-dimensional map, the correspondence between Mn and C was not exact. In fact, Mn was enriched near the upper left corner of the image, whereas C was enriched in the lower left.

It was unexpected to observe a similar, rather than an opposite, diffusion profile between the Mn (as an oxidant) and C (as a reductant), especially in the light gray mudstone and red sandstone (Figure 3). No such carbon enrichment was observed in a control sample (which was not exposed to oxidant), therefore ruling out the possibility of carbon contamination from sample preparation. We speculate that the C profile may be a result of carbon re-distribution. Once organic carbon was oxidized by permanganate, the resulting inorganic carbon may diffuse from within the rock matrix toward the surface. Because the permanganate diffusion process may have wetted the rock and opened rock pores on the reservoir side (Figure 1), inorganic carbon may have diffused towards the permanganate chamber much faster than the opposite direction (e.g., towards the sink side) and may have accumulated near that surface. This process of organic carbon oxidation followed by re-distribution of the resulting inorganic carbon may have accounted for the enrichment of inorganic carbon near the permanganate-exposed surface of the rocks, where Mn enrichment was also observed. This inorganic carbon redistribution hypothesis would explain the correlation between the C and Mn elemental maps (Figure 4), where the correspondence between the Mn and C elemental distribution was not exact but with some offset. Although not directly verified, this carbon

120 enrichment on the rock surface may be either adsorbed or associated with carbonates. For the dark gray mudstone and the tan sandstone, carbon enrichment is less obvious (Fig. 3), possibly because there was little inorganic carbon redistribution. If carbon only changes the form (from organic carbon to inorganic carbon) without redistribution, local enrichment near rock surface is not expected.

The coupling between Mn and inorganic C was also consistent with the positive correlation between permanganate oxidant demand and TOC (Table 1), where more permanganate reduction would oxidize more TOC and produce more inorganic carbon. This correspondence of Mn and C diffusion profile and the positive correlation between the natural oxidant demand values and TOC strongly suggest that organic matter served as a chemical reductant to reduce permanganate (Mn7+) to Mn2+/Mn4+ (e.g., birnessite). These results suggest that the reaction between permanganate and TOC was responsible for the small travel distance of permanganate into rock matrix. This conclusion is consistent with previous studies where organic matter in sedimentary rock was found as the primary component in the reaction with permanganate (Hønning et al., 2007a; Hønning et al., 2007b).

A previous study (Hartog et al., 2002) suggested that the relative importance of organic matter vs. ferrous iron (in pyrite and siderite) in reducing molecular oxygen was dependent on sediment grain size. The oxidation of sedimentary organic matter (SOM) in a fine grain size fraction was less important than the oxidation of pyrite and siderite, while SOM oxidation was very important in a coarse grain size fraction. The authors ascribed this difference to a decreased reactivity of sedimentary organic matter in the fine fractions as a result of physical protection through sorption and complexation of organic matter by clay minerals or to a higher degree of mineralization of the original organic matter during diagenesis. Our results suggest that TOC, not ferrous iron, was responsible for natural oxidant demand. The higher reactivity of TOC than ferrous iron in our samples could be due to two reasons: 1) Fe(II) is largely associated with non-reactive clay minerals (e.g., chlorite) (Schaefer et al., 2013), not with reactive pyrite and siderite; 2) TOC is not protected by clay minerals and may be readily accessible to permanganate. Protection of organic matter by clay minerals is largely through sorption into the interlayer space of expandable clays such as smectite (Kennedy et al., 2002), but in our

121 samples, there is little smectite-like minerals, even in the mudstones (Schaefer et al., 2013).

Results of the iodide tracer testing performed on the permanganate-exposed rocks are provided in Table 3. Results are presented for only three rock types, as the dark gray mudstone was not available for this testing. Iodide tracer results show that exposure to the oxidant for 2 months did not result in a significant (greater than a factor of 2) decrease in the measured Deff. Thus, oxidant exposure and any accumulation of Mn minerals near the surface of the rock did not have a substantial impact on diffusional flux through the rock. This finding suggests that precipitation of Mn minerals along the fracture surface following application of in situ chemical oxidation likely will not have a large impact on mitigating the back-diffusion of chlorinated ethene contaminants emanating from the rock matrix, at least after a 2 month exposure to the oxidant.

Mineralogical changes as a result of permanganate reaction

The mineral composition varied across different rock types. Major minerals, such as quartz (<71.8%), kaolinite (<20.4%), and albite (10.8-38.8%), were dominant in all of these rock samples (Table 4, Figure 5). No distinct Fe(II)-containing minerals were detected, and it is likely that Fe(II) measured by chemical extraction resides within chlorite and muscovite. Abiotic reduction of permanganate resulted in precipitation of various Mn minerals on the Mn-exposed side of the rocks, including birnessite (2.8-7.5%) and manganocalcite (<2.9%) (Table 4). SEM observations confirmed these results and revealed various morphologies/sizes and composition of these precipitates (Figure 6). Unlike the non-permanganate-reacted rocks (Figure 5), most of these Mn minerals were present as sausage-like particles on rock surfaces with various elements, such as O, C, Mn, Si, and Al (Figure 6). Si and Al were likely derived from surrounding silicate minerals. This is consistent with previous studies where semi-amorphous potassium-rich birnessite and MnO2(s) particles were identified during TCE remediation process using permanganate (Crimi and Siegrist, 2004; Li and Schwartz, 2004; Loomer et al., 2010; Zhang and Schwartz, 2000).

122 Our results show that Mn mineralogy resulting from abiotic reduction of permanganate depended on the rock type and mineralogy. For example, manganocalcite

((Ca,Mn)CO3) was only detected in the dark gray mudstone, where TOC content was the highest among the four rock types, and calcite (CaCO3) was present in the original rock. This result suggests that TOC may be readily available in the dark gray mudstone to reduce permanganate and to form manganocalcite. The reduced Mn2+ (ionic radius r=0.80 Å) may have substituted for Ca2+ (r=0.99 Å) within the calcite structure due to their similar charge and ionic radius (Gunasekaran and Anbalagan, 2008; Pingitore, 1978). The formation of the lowest Mn oxidation state (2+) in manganocalcite would increase the amount of permanganate consumption, and may have partially accounted for its highest oxidant demand in this rock.

The nature of these Mn mineral precipitates could have important implications for the development of strategies for controlling the stability of these reaction products because settling of these Mn solid particles has a potential to deposit in the subsurface and to impact the flow regime in and around permanganate injection point. However, our iodide tracer testing result indicated that within 2 months of permanganate exposure, the accumulation of Mn minerals near the surface of the rock, regardless the nature of these minerals, did not have a notable impact on diffusional flux through the rock. Nonetheless, this impact cannot be ruled out if the rocks are exposed to permanganate for a prolonged time.

CONCLUSIONS

Our results show that permanganate diffusion into solid sedimentary rocks was limited to distances less than 500 microns. The observed diffusivities for permanganate into the rock matrices ranged from 5.3×10-13 to 1.3×10-11 cm2/s. Permanganate reaction with naturally present organic carbon within these rocks was responsible for these short diffusion distances. Various Mn mineral precipitates, including birnessite and manganocalcite, formed during this reaction depending on the nature of the rock types. Within short experimental time frame of permanganate diffusion (2 months), these precipitates did not seem to have any impacts on diffusional flux through these rocks.

123 ACKNOWLEDGMENTS

Support for this research was provided in part by the Strategic Environmental Research and Development Program (SERDP) under Project ER-1685. The authors also are appreciative of the support provided by Pierre Lacombe of the United States Geological Survey (USGS) for the rock core collection. We are grateful to two anonymous reviewers whose comments improved the quality of the manuscript.

124 REFERENCES

Apha, A., 1998. Standard methods for the examination of water and wastewater. American Public Health Association, Water Pollution Control Federation, Washington DC: 2-55.

Bogan, B.W. and Trbovic, V., 2003. Effect of sequestration on PAH degradability with Fenton's reagent: roles of total organic carbon, humin, and soil porosity. Journal of Hazardous Materials, 100(1-2): 285-300.

Conrad, S.H., Glass, R.J. and Peplinski, W.J., 2002. Bench-scale visualization of DNAPL remediation processes in analog heterogeneous aquifers: surfactant floods and in situ oxidation using permanganate. Journal of Contaminant Hydrology, 58(1-2): 13-49.

Crank, J., 1995. The mathematics of diffusion, second edition. Oxford Science Publications.

Crimi, M. and Siegrist, R., 2004. Impact of Reaction Conditions on Genesis during Permanganate Oxidation. Journal of Environmental Engineering, 130(5): 562-572.

Cussler, E.L., 2009. Diffusion: mass transfer in fluid systems. Cambridge university press.

Fetter, C.W., 1999. Contaminant hydrogeology. Vol. 500. USA: Prentice hall.

Gates-Anderson, D., Siegrist, R. and Cline, S., 2001. Comparison of Potassium Permanganate and Hydrogen Peroxide as Chemical Oxidants for Organically Contaminated Soils. Journal of Environmental Engineering, 127(4): 337-347.

Ghosh, U., Weber, A.S., Jensen, J.N. and Smith, J.R., 2000. Relationship between PCB Desorption Equilibrium, Kinetics, and Availability during Land Biotreatment. Environmental Science & Technology, 34(12): 2542-2548.

Goldstein, K.J., Vitolins, A.R., Navon, D., Parker, B., Chapman, S., Anderson, G.A., 2004. Characterization and pilot-scale studies for chemical oxidation remediation of fractured shale. Remediation Journal, 14(4): 19-37.

125 Gunasekaran, S. and Anbalagan, G., 2008. Optical absorption and EPR studies on some natural carbonate minerals. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 69(2): 383-390.

Hartog, N., Griffioen, J. and van der Weijden, C.H., 2002. Distribution and Reactivity of O2-Reducing Components in Sediments from a Layered Aquifer. Environmental Science & Technology, 36(11): 2338-2344.

Haselow, J.S., Siegrist, R.L., Crimi, M. and Jarosch, T., 2003. Estimating the total oxidant demand for in situ chemical oxidation design. Remediation Journal, 13(4): 5-16.

Hønning, J., Broholm, M.M. and Bjerg, P.L., 2007a. Quantification of potassium permanganate consumption and PCE oxidation in subsurface materials. Journal of Contaminant Hydrology, 90(3-4): 221-239.

Hønning, J., Broholm, M.M. and Bjerg, P.L., 2007b. Role of Diffusion in Chemical Oxidation of PCE in a Dual Permeability System. Environmental Science & Technology, 41(24): 8426-8432.

Huang, Q. Dong, C. Z., Dong, R. M., Jiang, H., Wang, S., Wang, G., Fang, B., Ding, X., Niu, L., Zhang, C., and Dong, H., 2011. Archaeal and bacterial diversity in hot springs on the Tibetan Plateau, China. Extremophiles, 15(5): 549-563.

Kennedy, M.J., Pevear, D.R. and Hill, R.J., 2002. Mineral Surface Control of Organic Carbon in Black Shale. Science, 295(5555): 657-660.

Krembs, F.J., Siegrist, R.L., Crimi, M.L., Furrer, R.F. and Petri, B.G., 2010. ISCO for Groundwater Remediation: Analysis of Field Applications and Performance. Ground Water Monitoring & Remediation, 30(4): 42-53.

Kyriacou, G., Vadgama, P. and Wang, W., 2006. Characterization of a laminar flow cell for the prevention of biosensor fouling. Medical Engineering & Physics, 28(10): 989-998.

Lewis-Brown, J.C., Carleton, G.B. and Imbrigiotta, T.E., 2006. Hydraulic and Solute- Transport Properties and Simulated Advective Transport of Contaminated Ground Water in a Fractured-Rock Aquifer at the Naval Air Warfare Center, West

126 Trenton, New Jersey, 2003. Scientific Investigations Report. United States Geological Survey, 38.

Li, X.D. and Schwartz, F.W., 2004. DNAPL remediation with in situ chemical oxidation using potassium permanganate: Part I. Mineralogy of Mn oxide and its dissolution in organic acids. Journal of Contaminant Hydrology, 68(1-2): 39-53.

Lipson, D.S., Kueper, B.H. and Gefell, M.J., 2005. Matrix Diffusion-Derived Plume Attenuation in Fractured Bedrock. Ground Water, 43(1): 30-39.

Loomer, D.B., Al, T.A., Banks, V.J., Parker, B.L. and Mayer, K.U., 2010. Manganese Valence in Oxides Formed from in Situ Chemical Oxidation of TCE by KMnO4. Environmental Science & Technology, 44(15): 5934-5939.

Mumford, K.G., Thomson, N.R. and Allen-King, R.M., 2005. Bench-Scale Investigation of Permanganate Natural Oxidant Demand Kinetics. Environmental Science & Technology, 39(8): 2835-2840.

Pingitore, N.E., 1978. The behavior of Zn 2 + and Mn 2 + during carbonate diagenesis; theory and applications. Journal of Sedimentary Research, 48(3): 799-814.

Post, J.E., 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance. Proceedings of the National Academy of Sciences, 96(7): 3447-3454.

Reynolds, D.A., Jones, E.H., Gillen, M., Yusoff, I. and Thomas, D.G., 2008. Electrokinetic Migration of Permanganate Through Low-Permeability Media. Ground Water, 46(4): 629-637.

Schaefer, C.E., Towne, R.M., Lazouskaya, V., Bishop, M.E. and Dong, H., 2012. Diffusive flux and pore anisotropy in sedimentary rocks. Journal of Contaminant Hydrology, 131(1-4): 1-8.

Schaefer, C.E. Towne, R.M., Lippincott, D.R., Lazouskaya, V., and Fischer, T.B., Bishop, M.E. and Dong, H., 2013. Coupled Diffusion and Abiotic Reaction of Trichlorethene in Minimally Disturbed Rock Matrices. Environmental Science & Technology, in press.

127 Schnarr, M. Truax, C., Farquhar, G., Hood, E., Gonullu, T., Stickney, B.,1998. Laboratory and controlled field experiments using potassium permanganate to remediate trichloroethylene and perchloroethylene DNAPLs in porous media. Journal of Contaminant Hydrology, 29(3): 205-224.

Scott, D., Apblett, A. and Materer, N.F., 2011. Follow-up study on the effects on well chemistry from biological and chemical remediation of chlorinated solvents. Journal of Environmental Monitoring, 13(9): 2521-2526.

Siegrist, R.L., Crimi, M., & Simpkin, T. J. (Eds.), 2011. In situ chemical oxidation for groundwater remediation (Vol. 3). Springer Science + Business Media, New York.

Smith, L.H., McCarty, P.L. and Kitanidis, P.K., 1998. Spreadsheet Method for Evaluation of Biochemical Reaction Rate Coefficients and Their Uncertainties by Weighted Nonlinear Least-Squares Analysis of the Integrated Monod Equation. Applied and Environmental Microbiology, 64(6): 2044-2050.

Sterling, S.N., Parker, B.L., Cherry, J.A., Williams, J.H., Lane, J.W., Haeni, F.P., 2005. Vertical Cross Contamination of Trichloroethylene in a Borehole in Fractured Sandstone. Ground Water, 43(4): 557-573.

Struse, A., Siegrist, R., Dawson, H., and Urynowicz, M., 2002. Diffusive Transport of Permanganate during In Situ Oxidation. Journal of Environmental Engineering, 128(4), 327–334.

Tunnicliffe, B.S. and Thomson, N.R., 2004. Mass removal of chlorinated ethenes from rough-walled fractures using permanganate. Journal of Contaminant Hydrology, 75(1-2): 91-114.

Urynowicz, M.A., 2007. In Situ Chemical Oxidation with Permanganate: Assessing the Competitive Interactions Between Target and Nontarget Compounds. Soil and Sediment Contamination: An International Journal, 17(1): 53-62.

USEPA, 1998. Field applications of in situ remediation technologies: chemical oxidant ground water at fractured rock sites. EPA 542-R-01-010.

West, M.R. and Kueper, B.H., 2010. Plume Detachment and Recession Times in Fractured Rock. Ground Water, 48(3): 416-426.

128 Wu, M.Z., Reynolds, D.A., Fourie, A., Prommer, H. and Thomas, D.G., 2012. Electrokinetic in situ oxidation remediation: Assessment of parameter sensitivities and the influence of aquifer heterogeneity on remediation efficiency. Journal of Contaminant Hydrology, 136–137(0): 72-85.

Xu, X. and Thomson, N.R., 2009. A long-term bench-scale investigation of permanganate consumption by aquifer materials. Journal of Contaminant Hydrology, 110(3–4): 73-86.

Yan, Y.E. and Schwartz, F.W., 1999. Oxidative degradation and kinetics of chlorinated ethylenes by potassium permanganate. Journal of Contaminant Hydrology, 37(3- 4): 343-365.

Zhang, H. and Schwartz, F.W., 2000. Simulating the in situ oxidative treatment of chlorinated ethylenes by potassium permanganate. Water Resources Research, 36(10): 3031-3042.

129 Table 1. Characterization of rock properties TOC 1-month oxidant Ferrous iron Total iron Rock type Porosity (g/kg) demand (g/kg) (mmol/g) (mmol/g) Dark gray mudstone 4.63 22.5 0.30 0.96 0.050±0.006 Light gray mudstone 1.32 10.3 0.38 1.23 0.055±0.009 Red sandstone 0.27 2.98 0.01 0.03 0.074±0.004 Tan sandstone 0.17 2.58 0.005 0.09 0.068±0.005

130 Table 2. Dobs values and effective diffusion coefficients for different rock types. Measured (Dobs) and calculated (Dobs,cal) values were obtained from duplicate experiments based on Eqs. (2), (3) and (4). 95% confidence intervals are shown. ± values represent the standard error.

2 2 Effective diffusion Rock type D (cm /s) D (cm /s) 2 obs obs,cal coefficient (cm /s) 9.5 ± 5.2×10-13 Dark gray mudstone 1.1×10-12 2.3±1.4×10-7 1.4 ± 0.6×10-12 7.1 ± 6.6×10-13 Light gray mudstone 1.3×10-12 2.6±1.1×10-7 6.2 ± 5.4×10-13 3.2 ± 0.01×10-12 Red sandstone 7.1×10-12 4.2±3.4×10-7 1.3 ± 0.5×10-11 5.3 ± 3.2×10-13 Tan sandstone 4.4×10-12 2.3±1.0×10-7 1.1 ± 0.4×10-12

131 Table 3. Deff values prior to oxidant exposure compared to Deff after exposure to oxidant (Schaefer et al., 2013).

D (prior to oxidant) D (post oxidant) Rock type eff eff (cm2/s) (cm2/s) Light gray 1.1×10-7 1.3×10-7 mudstone 1.1×10-7 1.2×10-7 Red sandstone 2.0×10-7 5.5×10-7 4.0×10-8 Tan sandstone 1.6×10-7 1.0×10-7 1.6×10-7

132 Table 4. Mineralogy on permanganate reactive surfaces of sedimentary rocks.

Minerals Formula Dark gray Light gray Red Sandstone Tan Sandstone mudstone (%) mudstone (%) (%) (%) Quartz SiO2 8.2 29.1 71.8 Albite Na(AlSi3O8) 21.2 10.8 38.8 22.7 Kaolinite Al2Si2O5(OH)4 10.2 20.4 9.6 Amesite (Mg2Al)(AlSiO5(OH)4) 38.3 Microcline KAlSi3O8 10.3 Calcite CaCO3 4.8 Dolomite CaMg(CO3)2 15.8 Armenite Ca3(Al6Si9O30)(H2O)2 Anorthite CaAl2Si2O8 13.9 Muscovite KAl2(Si3Al)O10(OH)2 9.9 14.9 Chlorite Mg2Al3(Si3Al)O10(OH)8 23.0 Birnessite MnO2 2.8 7.5 7.5 5.5 Manganocalcite (Ca,Mn)CO3 2.9

133 Figure captions: Figure 1. Plastic dish used for permanganate diffusion experiment. Rock slices were place between permanganate solution (on the left) and synthetic water (on the right).The pink color on the sink side is epoxy, not permanganate. Figure 2. Manganese profiles of permanganate diffusion in different sedimentary rocks after 7-8 weeks of exposure to a permanganate solution (2000 ppm). These profiles were obtained from cross sections perpendicular to the permanganate-exposed rock surface using SEM-EDS line scanning. The left edge of a rock slice corresponds to distance of zero. CPS: counts per second. Figure 3. Carbon profiles in different sedimentary rocks after ~2 months of exposure to a permanganate solution (2000 ppm). These profiles are similar to the Mn profiles. Figure 4. SEM-EDS elemental maps showing the distributions of C, O, K, Mn, and Fe, as well as a secondary image for the red sandstone sample. The brighter areas on the photos indicate higher elemental concentrations. These maps do not distinguish different forms of elements (such as organic or inorganic carbon) or different oxidation states and only represent total elemental concentrations. Figure 5. SEM images (left) and EDS spectra (right) of non-permanganate-reacted surfaces on different sedimentary rocks, including dark gray mudstone (a), light gray mudstone (b), red sandstone (c), and tan sandstone (d). The plus signs on the SEM images mark the locations from which EDS spectra were obtained, indicating dominant minerals in these rocks, such as albite, kaolinite, and quartz. CPS: counts per second. Figure 6. SEM images (left) and EDS spectra (right) of permanganate-reacted surfaces on different sedimentary rocks, including dark gray mudstone (a), light gray mudstone (b), red sandstone (c), and tan sandstone (d). The plus signs on the images mark the location from which EDS spectra were obtained, indicating the formation of Mn mineral precipitates, such as birnessite and manganocalcite.

134 Figure 1.

135 Figure 2.

136 Figure 3.

137 Figure 4.

138 Figure 5.

139 Figure 6.

140 Figure S1. Model regression (Eq. 3) for the red sandstone.

20 Mn (counts) Mn 10

0 0 50 100 150 200 Distance from Oxidant Exposed Interface (microns)

141 CHAPTER 6:

SUMMARY

This dissertation was designed to answer the key questions list in Chapter 1, and can be summarized into three parts of scientific study:

1. Investigating the microbial abundance and diversity in Tibetan and Philippines hot springs (Chapters 2 & 3).

2. Understanding the taxonomic and genetic diversities in high-elevation, saline Qinghai Lake (Chapter 4).

3. Evaluating the permanganate diffusion in rock matrix and reactions with naturally occurring reductants in bedrock aquifers (Chapter 5).

The first part of this dissertation was to systematically study the archaeal and bacterial communities in 10 Tibetan and 6 Philippines hot springs using an integrated approach of culture- dependent and -independent methods. A suite of geochemical and mineralogical analyses were also performed to investigate the possible environmental variables that control microbial community structure. Based on the statistical analyses, we have come to the following conclusions:

1. The microbial communities in the studied hot springs of Tibet and the Philippines were highly diverse. The bacterial communities in hot springs from these two areas were similar, both predominated by phyla Firmicutes, Proteobacteria, and Aquificae.

2. While the archaeal communities of the Tibetan hot springs mainly consisted of Thaumarchaeota clone sequences, phyla Crenarchaeota were dominant in the Philippines hot springs.

3. Sulfur metabolisms appear to be the key physiological functions in these Philippines springs; however, novel organisms were abundant in some springs and their functions remain to be elucidated.

4. A combination of geographic distance and environmental conditions may have accounted for characteristics of microbial communities in these Tibetan and Philippines hot springs.

142 The goal of the second part of this dissertation was to apply a metagenomic and metatranscriptomic survey to investigate the taxonomic and functional diversities in the high- elevation, saline Qinghai Lake. This project identified microbial processes involved in the biogeochemical cycles (e.g., carbon and nitrogen cycles) which provide feedback to the global climate change. The comparison between potential occurrence and actual activity of a specific gene within a metabolic process enables us better assess microbially mediated pathways in the biogeochemical cycles. This study draws the following conclusions:

1. Photoautotrophic organisms were the most abundant in the DNA samples, while heterotrophic organisms were the most active in the RNA samples.

2. Ammonia assimilation had the highest richness and this process would be expected to be more resistant to environmental change.

3. Our data along with previous studies suggested that microbial successional changes occurred in the water column of Qinghai Lake throughout the summer, and our sampling was likely at the end of a cyanobacterial bloom.

4. Functional genes with higher richness have more stress response genes and would be expected to be more resilient against environmental stress.

The last part of this dissertation was aimed to evaluate in situ chemical oxidation of chlorinated solvents in fractured bedrock aquifers by studying the permanganate diffusion and reactions in a variety of sedimentary rocks. In addition, the change of existing minerals and the newly formed Mn mineral precipitates were examined for their impacts on diffusional flux through these rocks. The conclusions are list below:

1. Permanganate diffusion into solid sedimentary rocks was limited to distances less than 500 microns, and the observed diffusivities ranged from 5.3×10-13 to 1.3×10-11 cm2/s.

2. Various Mn mineral precipitates formed during this reaction depending on the nature of the rock types, such as birnessite and manganocalcite.

3. Within short experimental time frame of permanganate diffusion (2 months), these precipitates did not seem to have any impacts on diffusional flux through these rocks.

These results have important implications for our understanding of long-term organic contaminant remediation in sedimentary rocks using permanganate. Because of short diffusion

143 distance of permanganate into rocks and its reaction with organic matter and ferrous mineral, its effectiveness in treating organic contaminant may be compromised.

144 APPENDIX: PUBLICATIONS

1. Qiuyuan Huang, Brandon Briggs, Hongchen Jiang, Geng Wu, James Hollibaugh, Hailiang Dong (to be submitted). Metagenomic and metatranscriptomic analysis of taxonomic and functional diversity for saline Qinghai Lake, China.

2. Shang Wang, Hailiang Dong, Weiguo Hou, Hongchen Jiang, Qiuyuan Huang, Brandon R. Briggs, Liuqin Huang (to be submitted). Seasonal dynamic relationship between microbial community and geochemistry in Tengchong hot spring, Yunnan Province, China.

3. Qiuyuan Huang, Hailiang Dong, Rachael Towne, Timothy Fischer, Charles Schaefer (2014). Permanganate diffusion and reaction in sedimentary rocks. Journal of Contaminant Hydrology, 159: 36-46.

4. Qiuyuan Huang, Hongchen Jiang, Brandon Briggs, Shang Wang, Weiguo Hou, Gaoyuan Li, Geng Wu, Ramonito Solis, Carlo Arcilla, Teofilo Abrajano, Hailiang Dong (2013) Archaeal and bacterial diversity in acidic to circumneutral hot springs in the Philippines. FEMS Micro Ecol.,85(3): 452-464.

5. Brandon Briggs, Eoin Brodie, Lauren Tom, Hailiang Dong, Hongchen Jiang, Qiuyuan Huang, Shang Wang, Weiguo Hou, Geng Wu, Liuquin Huang, Brian P. Hedlund, Chuanlun Zhang, Paul Dijkstra, and Bruce A. Hungate (2013). Seasonal patterns in microbial communities inhabiting the hot springs of Tengchong, Yunnan Province, China. Environ Microbiol: doi: 10.1111/1462-2920.12311.

6. Julienne Paraiso, Amanda Williams, Qiuyuan Huang, Yuli Wei, Paul Dijkstra, Bruce Hungate, Hailiang Dong, Brian Hedlund, Chuanlun Zhang (2013). The distribution and abundance of archaeal tetraether lipids in U.S. Great Basin hot springs. Frontiers in Terrestrial Microbiology 08/2013; 4(247):1-14. DOI:10.3389/fmicb.2013.00247.

7. Brian Hedlund, Julienne Paraiso, Amanda Williams, Qiuyuan Huang, Yuli Wei, Paul Dijkstra, Bruce Hungate, Hailiang Dong, Chuanlun Zhang (2013). Wide distribution of autochthonous branched glycerol dialkyl glycerol tetraethers (bGDGTs) in U.S. Great Basin hot springs. Frontiers in Terrestrial Microbiology 4:222. DOI:10.3389/fmicb.2013.00222.

145 8. Liuqin Huang, Hailiang Dong, Shang Wang, Qiuyuan Huang, and Hongchen Jiang (2013). Diversity and abundance of ammonia oxidizing archaea and bacteria in diverse Chinese paddy soils. Geomicrobiology Journal DOI:10.1080/01490451.2013.797523

9. Weiguo Hou, Shang Wang, Hailiang Dong, Hongchen Jiang, Brandon Briggs, Joseph Peacock, Qiuyuan Huang et al. (2013) A comprehensive census of microbial diversity in hot springs of Tengchong, Yunnan Province China using 16S rRNA gene pyrosequencing. PloS One. 8(1): e53350.

10. Chao Peng, Hongchen Jiang, Liuqin Huang, Weiguo Hou, Jian Yang, Shang Wang, Qiuyuan Huang, Shicai Deng, Hailiang Dong (2013) Abundance and diversity of ammonia-oxidizing bacteria and archaea in cold springs on the Qinghai-Tibet Plateau. Geomicrobiology Journal. 30(6): 530-539.

11. Qiuyuan Huang, Christina Z Dong, Raymond M Dong, Hongchen Jiang, Shang Wang, Genhou Wang, Bin Fang, Xiaoxue Ding, Lu Niu, Xin Li, Chuanlun Zhang, Hailiang Dong (2011) Archaeal and bacterial diversity in hot springs on the Tibetan Plateau, China. Extremophiles 15(5):549-63

12. Hongchen Jiang, Christina Z. Dong, Qiuyuan Huang, Genhou Wang, Bin Fang, Chuanlun Zhang, Hailiang Dong (2011). Actinobacterial diversity in microbial mats of five hot springs in central and central-eastern Tibet, China. Geomicrobiology Journal. 29(6):520-527

13. Qiuyuan Huang, Hongchen Jiang, Chuanlun Zhang, Wenjun Li, Shicai Deng, Bingsong Yu, Hailiang Dong (2010) Abundance of ammonia-oxidizing microorganisms in response to environmental variables of hot springs in Yunnan Province, China. ACTA MICROBIOLOGICA SINICA 01/2010; 50(1):132-6.

14. Hongchen Jiang, Qiuyuan Huang, Hailiang Dong, Peng Wang, Fengping Wang, Wenjun Li, Chuanlun Zhang (2010). RNA-based investigation of ammonia-oxidizing archaea in hot springs of Yunnan Province, China. Applied and Environmental Microbiology 76(13): 4538-4541.

15. Hongchen Jiang, Qiuyuan Huang, Shicai Deng, Hailiang Dong, Bingsong Yu (2010). Planktonic actinobacterial diversity along a salinity gradient of a river and five lakes on the Tibetan Plateau. Extremophiles 14(4): 367-376.

146 16. Hongchen Jiang, Shicai Deng, Qiuyuan Huang, Hailiang Dong, Bingsong Yu (2010). Response of aerobic anoxygenic phototrophic bacterial diversity to environment conditions in saline lakes and Daotang River on the Tibetan Plateau, NW China. Geomicrobiology Journal 27(5): 400-408.

17. Hongchen Jiang, Hailiang Dong, Shicai Deng, Bingsong Yu, Qiuyuan Huang, and Qinglong Wu (2009). Response of archaeal community structure to environmental changes in lakes on the Tibetan Plateau, northwestern China. Geomicrobiology Journal. 26(4): 289- 297.

FIELDS OF STUDY

Major Field: GEOMICROBIOLOGY

147