UC Irvine UC Irvine Electronic Theses and Dissertations

Title Role of microbial communities in mediating an ecosystem's response to global change

Permalink https://escholarship.org/uc/item/21z2r9np

Author Matulich, Kristin L.

Publication Date 2015

Peer reviewed|Thesis/dissertation

eScholarship.org Powered by the California Digital Library University of California UNIVERSITY OF CALIFORNIA, IRVINE

Role of microbial communities in mediating an ecosystem’s response to global change

DISSERTATION

submitted in partial satisfaction of the requirements for the degree of

DOCTOR OF PHILOSOPHY

in Biological Sciences

Kristin L. Matulich

Dissertation Committee: Professor Jennifer B.H. Martiny, Chair Professor Diane R. Campbell Associate Professor Steven D. Allison

2015

Chapter 1 © 2015 Nature Publishing Group Chapter 2 © 2015 Ecological Society of America Chapter 3 © 2015 Kristin L. Matulich, Jeniffer Peña, and Jennifer B.H. Martiny All other materials © 2015 Kristin L. Matulich

TABLE OF CONTENTS

Page

LIST OF FIGURES ...... iii

LIST OF TABLES ...... iv

ACKNOWLEDGMENTS ...... v

ABSTRACT OF THE DISSERTATION...... ix

INTRODUCTION ...... 1

CHAPTER 1 Temporal variation overshadows the response of leaf litter microbial communities to simulated global change...... 5

CHAPTER 2 Microbial composition alters the response of litter decomposition to environmental change ...... 35

CHAPTER 3 Phylogenetic conservation of substrate use, but not temperature response, in leaf litter bacteria...... 59

REFERENCES ...... 90

LIST OF FIGURES

Page

Figure 1.1: PERMANOVA results for microbial, plant, and enzyme composition ...... 29

Figure 1.2: NMDS ordination of bacterial and fungal composition...... 30

Figure 1.3: Temperature, precipitation, and diversity across collection dates ...... 31

Figure 1.4: Responses of abundant bacterial and fungal OTUs to environmental treatments...... 32

Figure 1.5: Litter mass loss across environmental treatments...... 33

Figure 2.1: NMDS ordination of bacterial and fungal composition...... 54

Figure 2.2: Respiration for each environmental treatment...... 55

Figure 2.3: Respiration for each of the eight microbial communities ...... 56

Figure 2.4: Correlation of community and functional responses for each environmental treatment ...... 57

Figure 3.1: Theoretical distribution of functional and response traits...... 79

Figure 3.2: Sample Ecoplate and CO2 production data ...... 80

Figure 3.3: Distribution of functional traits...... 81

Figure 3.4: Distribution of CO2 production trait ...... 82

Figure 3.5: Substrate yield and bacterial biomass correlation ...... 83

Figure 3.6: Phylogenetic tree of bacterial strains with functional & response traits ...... 84

Figure 3.7: Functional and response trait correlation...... 85

Figure 3.S1:8 Substrate use richness for each strain ...... 86

Figure 3.S2:9 Substrate use richness at each temperature ...... 87

Figure 3.S3:10 Mineralization rates for each bacterial strain ...... 88

Figure 3.S4:11 Phylogenetic tree of bacterial strains with Ecoplate substrate...... 89

iii

LIST OF TABLES

Page

Table 1.1: List of extracellular enzymes analyzed ...... 25

Table 1.2: Overview of pyrosequencing results of microbial communities ...... 26

Table 1.3: ANOVA results for each experimental factor on microbial diversity ...... 27

Table 1.4: PERMANOVA and ANOVA results for each collection date...... 28

Table 2.1: ANOVA and PERMANOVA results for each treatment ...... 52

Table 2.2: Mean variance and Bartlett's K2 test statistic across treatments ...... 53

Table 3.1: ANOVA results for functional traits ...... 76

Table 3.2: Results for each Biolog Ecoplate substrate ...... 77

ACKNOWLEDGMENTS

This dissertation, and my journey to it, would not have been possible without the incredible guidance, support, and love of my family, friends, and colleagues. Gregory Dolan: your unconditional love and patience has kept me sane through the failed experiments, rejected papers, and other “sub-optimal” experiences that we’ve overcome over the past five years. I feel so honored to have you by my side and can’t wait to see where our post- PhD adventures will lead us. My parents, John and Barbara Matulich: your love and guidance has helped me grow into the woman and scientist I am today. Thank you for teaching me the value of hard work, the imagination to dream big, and the confidence to achieve bigger. My advisor and mentor, Jennifer Martiny: this dissertation would not have been possible without your thoughtful mentorship and unwavering encouragement. Thank you for always offering me a safe space the wonderful support to develop my ideas, both within and outside of science. My dissertation committee, Steve Allison and Diane Campbell: your wise and thoughtful guidance greatly improved the quality of my science. To all members, past and present, of the J. Martiny Lab, Cohort Coffee Hour, Microbial Group, and Ecology Group: when I reflect on my time at UCI, many of my fondest memories are with you all. Thank you for the stimulating conversations, helpful advice, wonderful sense of comradery, and timely doses of comic relief. It has been an absolute privilege working with you all, and I am excited to see where each of your journeys and successes will lead you.

I would like to acknowledge the funding sources that made my dissertation research possible. Fellowship support was provided by a US Department of Education Graduate Assistance in Areas of National Need (GAANN) Fellowship and UC Irvine’s Brython Davis Graduate Fellowship. Research support was provided by grants from the NSF Division of Environmental Biology, and the Office of Science (BER) DOE program in Microbial Communities and Carbon Cycling.

Chapter 1 was published with the permission of Nature Publishing Group. The text of this chapter is a reprint of the material as it appears in the ISME Journal. Chapter 2 was published with the permission of the Ecological Society of America. The text of this chapter is a reprint of the material as it appears in the journal Ecology.

CURRICULUM VITAE

Kristin L. Matulich Department of Ecology & Evolutionary Biology University of California, Irvine • Irvine, CA 92697

EDUCATION 2015 Ph.D. in Biological Sciences: Ecology & Evolutionary Biology University of California, Irvine

2015 Professional Certificate in Grant Writing San Diego State University

2011 Hopkins Microbiology Course Stanford University

2009 B.S. in Environmental Studies and Ecology, Evolution, & Marine Biology University of California, Santa Barbara

APPOINTMENTS 2014-2015 Research Development Intern, Office of Research Development University of California, Irvine

2014-2015 Communications Consultant, Graduate Resource Center University of California, Irvine

2010-2014 Teaching & Administrative Assistant, Ecology & Evolutionary Biology University of California, Irvine

2007-2010 Research Specialist, Ecology, Evolution, & Marine Biology University of California, Santa Barbara

PUBLICATIONS Matulich KL, C Weihe, SD Allison, AS Amend, R Berlemont, M Goulden, S Kimball, AC Martiny, and JBH Martiny. Temporal variation overshadows the response of leaf litter microbial composition to simulated global change. The ISME Journal (in press)

Amend AS, KL Matulich, and JBH Martiny (2015). Nitrogen addition, not phylogenetic diversity, increases decomposition rates of fungal communities in leaf litter microcosms. Frontiers in Microbiology. 6:109.

Matulich KL and JBH Martiny (2015). Microbial composition alters the response of litter decomposition to environmental change. Ecology 96:154–163.

Mouginot C, R Kiwamura, KL Matulich, R Berlemont, SD Allison, AS Amend, and AC Martiny (2014). Elemental stoichiometry of Fungi and Bacteria strains from grassland leaf litter. Soil Biology & Biogeochemistry 76:278-285.

Matulich KL, DJ Mohamed, and JBH Martiny (2013). Microbial Biodiversity. Pages 252-258 in Encyclopedia of Biodiversity. S. Levin, editor. Elsevier.

Shade A, H Peter, SD Allison, DL Baho, M Berga, H Bürgmann, DH Huber, S Langenheder, JT Lennon, JBH Martiny, KL Matulich, TM Schmidt, and J Handelsman (2012). Fundamentals of microbial community resistance and resilience. Frontiers in Microbiology 3:417.

Hooper DU, EC Adair, BJ Cardinale, JEK Byrnes, B Hungate, KL Matulich, A Gonzalez, JE Duffy, L Gamfeldt, and MI O’Connor (2012). A global synthesis reveals biodiversity loss as a major driver of ecosystem change. Nature 486:105-108.

Cardinale BJ, KL Matulich, DU Hooper, JEK Byrnes, JE Duffy, L Gamfeldt, P Balvanera, MI O’Connor, and A Gonzalez (2011). The functional role of producer diversity in ecosystems. American Journal of Botany 98:1–21.

ORAL PRESENTATIONS Matulich KL (2014). Simulated environmental change and temporal variation drivers of microbial community change. International Society for Microbial Ecology (ISME) General Meeting

Matulich KL (2014). All hands on deck: using microbes to battle climate change. UC Irvine Associated Graduate Student Symposium.

Matulich KL (2014). Role of microbial communities in mediating an ecosystem’s response to global change. Ecology and Evolutionary Biology Graduate Recruitment.

Matulich KL, C Weihe, SD Allison, AS Amend, R Berlemont, M Goulden, AC Martiny, and JBH Martiny (2014). Simulated environmental change and interannual variation drivers of microbial community change. Ecology and Evolutionary Biology Graduate Student Symposium.

Matulich KL and JBH Martiny (2012). Importance of microbial community composition in predicting ecosystem response to global change. Ecology and Evolutionary Biology Graduate Student Symposium.

POSTER PRESENTATIONS Penã J, KL Matulich, and JBH Martiny (2014). The phylogenetic conservation of bacterial functional traits. UC Irvine Undergraduate Research Opportunities Program.

Kawamura R, C Mouginot, KL Matulich, and AC Martiny (2013). Stoichiometrical variability within phylum of soil bacteria. UC Irvine Undergraduate Research Opportunities Program.

vii

Matulich KL and JBH Martiny (2013). Importance of microbial community composition in predicting ecosystem response to global change. American Society of Microbiology (ASM) General Meeting

Matulich KL, IT Carroll, and BJ Cardinale (2009). Priority effects: Initial densities alter exponential growth rates of competing species. Ecological Society of America (ESA) General Meeting

Matulich KL, IT Carroll, and BJ Cardinale (2009). Light partitioning does not explain coexistence of the world’s dominant oceanic plankton. UC Santa Barbara Undergraduate Research Colloquium.

AWARDS & HONORS 2012, ‘14, ‘15 Brython Davis Graduate Fellowship 2014 UC Irvine Associated Graduate Student Symposium: Dean’s Prize 2013 Lake Forest Garden Club Scholarship 2012 School of Biological Sciences Graduate Award Fellowship 2011, ‘12 Graduate Assistance in the Areas of National Need Fellowship 2011, ‘12 NSF Graduate Research Fellowship: Honorable Mention 2011 UCI-HHMI Graduate Fellow Travel Award 2009 Environmental Studies Outstanding Academic Achievement Award 2009 UCSB Undergraduate Research Colloquium: Honorable Mention 2008 Worster Award for Undergraduate Research

viii

ABSTRACT OF THE DISSERTATION

Role of microbial communities in mediating an ecosystem’s response to global change

Kristin L. Matulich

Doctor of Philosophy in Biological Sciences

University of California, Irvine, 2015

Professor Jennifer B.H. Martiny, Chair

A central goal of global change biology is to predict the impact of environmental change on ecosystem processes. Currently, most global change models treat the local microbial community as a single, homogenously functioning entity, thereby assuming that the specific microbial composition is functionally irrelevant. However, microorganisms perform key transformations in ecosystems, and recent research demonstrates that microbial communities vary greatly across space and in response to environmental change.

Therefore, parameters describing microbial communities may be key for improving predictions of how future global changes will impact ecosystem processes. For this reason, my dissertation research examined the effect of environmental changes on resident communities and determined how potential shifts in microbial community composition will impact litter decomposition rates. To accomplish this, I gathered litter samples from a chaparral ecosystem undergoing global change manipulations (elevated nitrogen availability or reduced precipitation), and characterized the microbial community using 454 high- throughput sequencing (Chapter 1). While microbial communities are much more variable through time, this research showed that microbial composition will likely shift in response to environmental change. I also examined the role of microbial community composition for a

ix key ecosystem process, litter decomposition, and how that role changes under environmental perturbations. By isolating microbial taxa from the same ecosystem discussed above, I constructed artificial microbial communities with varying composition. I then conducted a laboratory experiment in which I subjected the communities to different global change manipulations and monitored decomposition rates and community composition (Chapter 2). Microbial composition had a main effect on leaf litter decomposition and also interacted with the environmental treatment, suggesting that future shifts in microbial communities will influence the magnitude in which environmental change affects ecosystem processes. Lastly, I investigated the functional and response traits of individual microbial taxa to better predict how microbial communities might respond to global change perturbations, and found that many functional traits displayed a phylogenetic pattern, but a taxa’s response to increased temperature did not (Chapter 3). Ultimately, this set of studies further justifies the need to incorporate microbial communities into models and begins to identify which parameters might be most relevant.

INTRODUCTION

A central goal of global change biology is to predict the impact future environmental change will have on ecosystem processes. Currently, most ecosystem and global change models treat the local microbial community as a single, homogenously functioning entity

(Schimel 2001), thereby assuming that the specific microbial composition is functionally irrelevant. However, microorganisms perform key transformations in ecosystems, and recent research demonstrates that microbial communities vary greatly across space and in response to environmental change (Martiny et al. 2006). Therefore, parameters describing microbial communities may be key for improving predictions of how future environmental change will impact ecosystem processes. For this reason, my dissertation research examined the effect of environmental change on resident microbial communities and determined how potential shifts in microbial community composition will impact one important ecosystem function, litter decomposition.

Over the next fifty years, southern California is expected to experience a range of environmental changes, including increased temperature, nitrogen deposition rates, and precipitation variability (IPCC 2013). Much research has been devoted to determining how natural plant communities will respond to these environmental changes. For example,

Padgett and Allen (1999) determined that increased nitrogen deposition rates caused the composition of the natural system to shift from a shrub to grass dominated system.

Additionally, Zavaleta et al. (2003) experimentally manipulated temperature, nitrogen deposition rates, and precipitation and observed that each environmental treatment resulted in distinct alterations in the plant composition.

While the effects of environmental change on plant communities are currently being investigated, less research has been done to examine how changing plant communities and

1 environmental conditions will impact the composition of resident microbial communities (but see (Castro et al. 2010, Sheik et al. 2011, Yuste et al. 2011). However, microorganisms have relatively short generation times and rapid growth rates, and could be among the fastest components of an ecosystem to respond to environmental change (Prosser et al.

2007). Additionally, microbial communities are extremely diverse (Truper 1992, Hawksworth

2001), and both experimental and observational studies demonstrate that microbial communities often vary through time and across very fine special scales (Martiny et al.

2006).

Not only are microbial communities extremely diverse and variable, but they also play a central role in many ecosystem processes, including the terrestrial carbon cycle thru mediation of decomposition rates. In fact, decomposition is performed almost exclusively by fungi and bacteria (Swift et al. 1979), and is the largest source through which carbon is transferred from the biosphere to the atmosphere (Schlesinger and Andrews 2000).

Although the carbon flux associated with decomposition is very large, it tends to be relatively stable, simultaneously relying on multiple biological processes. Therefore, monitoring changes in decomposition rates can give scientists a conservative estimate of how the whole carbon cycle operates under different environmental conditions.

Determining the role of microbial communities in ecosystem processes like decomposition may be key to predicting future global change, and has been an area of advancing research. Bell et al. (2005) conducted a microcosm experiment with varying bacterial species richness, and found a saturating relationship between the bacterial community richness and decomposition. Similarly, Tiunov and Scheu (2005) found that the rate of organic matter decomposition was positively associated with fungal species richness. Fewer studies have examined the effect of microbial composition containing both fungal and bacterial taxa on decomposition. Griffiths et al. (2001) constructed microbial 2 communities containing both fungal and bacterial species, and found no relationship between microbial composition and decomposition. Conversely, Mille-Lindblom and Tranvik

(2003) found that microbial composition had a significant effect on decomposition, and microcosms containing both cultured fungal and bacterial taxa had a higher decomposition rate than microbial communities containing just fungal or bacterial taxa alone. These conflicting findings exemplify how little is understood about how microbial composition affects decomposition.

While research has shown that microbial composition shifts in response to environmental change, and that microbial communities regulate important ecosystem functions, it is unknown about how the effect of microbial composition on ecosystem process rates will vary as environmental conditions change (Singh et al. 2010). Changes in environmental conditions like temperature and moisture might cause relatively few shifts in the composition microbial community, and may primarily alter decomposition rates directly through physiological responses. Alternatively, other environmental changes, like nitrogen availability, may cause shifts in the microbial community composition, altering decomposition rates indirectly. If this is true, then models that aim to predict how ecosystems will respond to global change should consider microbial community composition in some way (Allison and Martiny 2008, Mcguire and Treseder 2010).

One approach to linking microbial composition, ecosystem function, and environmental change is to identify traits that influence the structure and functioning of microbial communities (Green et al. 2008, Litchman and Klausmeier 2008, Wallenstein and

Hall 2012, Fierer et al. 2014). For instance, plant ecologists have proposed dividing traits into two types: effect and response traits (Lavorel and Garnier 2002). Effect traits (hereafter

“functional” traits as in Srivastava et al. 2012) are characteristics that influence ecosystem properties or processes such as nutrient cycling and trace gas emissions. In contrast, 3 response traits determine how an organism’s abundance changes in the face of new environmental conditions (Lavorel and Garnier 2002, Suding et al. 2008). These traits may be correlated with one another, or be patterned by phylogenic history. For instance, Lennon et al. (2012) found that functional traits like niche breadth and optimum water potential accurately predicted a bacteria’s response to decreased moisture, and that these traits were conserved at relatively coarse taxonomic scales. Overall, determining the distribution and patterning of functional and response traits in microbial communities may provide insights to how composition, and microbial-controlled ecosystem functions, will shift in response to future environmental change.

During my graduate school career, I showed that a natural microbial community shifts in response to global change manipulations in the field (Chapter 1; Matulich et al.

2015). Using a microcosm experiment with isolates from this community, I showed that microbial composition directly influenced a key ecosystem process, litter decomposition, and how that process responded to changing environmental conditions (Chapter 2; Matulich and Martiny 2015). Lastly, I found that carbon use functional traits are phylogenetically conserved in leaf litter bacteria and that information about microbial composition may provide insights to predicting ecosystem function under one set of conditions, but not necessarily relevant to predicting the response of ecosystem functions to environmental change (Chapter 3). Ultimately, this set of studies further justifies the need to incorporate microbial communities into models and begins to identify which parameters might be most relevant.

CHAPTER 1

Temporal variation overshadows the response of leaf litter microbial communities to simulated global change

Abstract

Bacteria and fungi drive the decomposition of dead plant biomass (litter), an important step in the terrestrial carbon cycle. Here, we investigate the sensitivity of litter microbial communities to simulated global change (drought and nitrogen addition) in a

California annual grassland. Using 16S and 28S rDNA amplicon pyrosequencing, we quantify the response of the bacterial and fungal communities to the treatments and compare these results to background, temporal (seasonal and interannual) variability of the communities. We found that the drought and nitrogen treatments both had significant effects on microbial community composition, explaining 2-6% of total compositional variation.

However, microbial composition was even more strongly influenced by seasonal and annual variation (explaining 14-39%). The response of microbial composition to drought varied by season, while the effect of the nitrogen addition treatment was constant through time. These compositional responses were similar in magnitude to those seen in microbial enzyme activities and the surrounding plant community, but did not correspond to a consistent effect on leaf litter decomposition rate. Overall, these patterns indicate that in this ecosystem, temporal variability in the composition of leaf litter microorganisms largely surpasses that expected in a short-term global change experiment. Thus as for plant communities, future microbial communities will likely be determined by the interplay between rapid, local background variability and slower, global changes.

Introduction

Leaf litter microorganisms play an important role in terrestrial ecosystems through their effects on decomposition rates and carbon cycling (Swift et al. 1979, Schimel and

Schaeffer 2012). Microbial communities in general (Allison and Martiny 2008, Singh et al.

2010), and leaf litter communities in particular (Allison et al. 2013), are sensitive to global changes in climate and nutrient availability. The variety of pathways by which such changes can affect leaf litter communities, however, complicates predictions of future responses.

Changes in the abiotic environment can directly alter the growth and survival of particular microbial taxa. For example, moisture determines the physical connectivity of the terrestrial matrix, and desiccation stress can result in population extinction and decreased community diversity (Seifert 1961, Treves et al. 2003, Castro et al. 2010, Sheik et al. 2011). In addition, environmental change might affect the leaf litter community indirectly via changes in the plant community and the quality or quantity of leaf litter (Kominoski et al. 2009, Cleveland et al. 2014). To complicate matters further, shifts in microbial composition are just one route by which environmental change impacts leaf litter decomposition. Decomposition rates are also influenced directly by abiotic conditions and changes in litter quality (e.g., Mcclaugherty et al. 1985, Cotrufo et al. 1994, Cleveland et al. 2014).

To understand the importance of these different response pathways, we investigated leaf litter microbes and their functioning in a well-replicated, global change experiment

(Potts et al. 2012, Kimball et al. 2014). Southern California grasslands are currently experiencing a range of environmental changes. Projections for California over the next 50 years suggest that annual temperature will increase 3°C and that precipitation will decrease

10-25% (CEC 2003). These changes are expected to drastically affect Mediterranean-type ecosystems where water availability is a key environmental constraint (Specht et al. 1983,

Larcher 2000). Indeed, most of the mean annual precipitation falls almost exclusively from

November through April, with a predictable summer drought from May through October

(Arguez et al. 2012). Further, nitrogen deposition rates in Southern California are among the highest in the United States, with deposition ranging from 2.5-4.0 g N m-2 yr-1 in many regions (Fenn et al. 2003b). If these rates continue, nitrogen loading will remain a significant driver of southern California’s ecosystems (Padgett and Allen 1999, Fenn et al. 2003b,

Zavaleta et al. 2003).

In temperate ecosystems, microbial responses to global change must be considered in light of temporal variation and particularly, seasonal variation. While previous studies have examined the response of microbial composition to environmental change over multiple years (e.g., Sheik et al. 2011, Gutknecht et al. 2012), only a handful have considered the effect of seasonal variation on these responses (e.g., Lage et al. 2010, Bell et al. 2014). However, the fast generation times of microbes mean that composition can turnover quickly, even across seasons (Bardgett et al. 1999, Kennedy et al. 2006, Habekost et al. 2008, Cregger et al. 2012, Gutknecht et al. 2012). We hypothesize that this high background variability will have several, important consequences for global change responses. First, the response of microbial communities to an incremental, mean change in the abiotic environment will be relatively difficult to detect compared to slower growing organisms like plants. Specifically, seasonal variation will greatly exceed the variation due to global change (or global change treatments). Second, the response of microbial communities to environmental change will depend on the time of year. Seasonal variation in the traits of the microbial community could alter its response to environmental change

(Bardgett et al. 1999, Niklaus et al. 2003, Cregger et al. 2012). Third, season-dependent responses will be particularly strong when the environmental parameter also varies seasonally. For instance, a drought treatment might shift the phenology of the microbial 7 community (e.g., promote an early appearance of the dry season community during the wet season). To test these hypotheses, we compared the compositional and functional responses of leaf litter microorganisms over two years (years 4 and 5 of the ecosystem manipulation). We characterized bacterial and fungal composition by amplicon pyrosequencing of the 16S and 28S rRNA genes and assayed functioning by potential enzyme activities and litter decomposition rates.

Methods

Experimental design

The study area is a California grassland ecosystem 5 km north of Irvine, CA USA

(33°44’ N, 117°42’ W, 365 m elevation) that is dominated by exotic annual grasses and forbs. For this experiment, we used a subset of plots from an existing field manipulation of precipitation and nitrogen inputs that began in February 2007 (Potts et al. 2012, Allison et al. 2013). We utilized 2 levels of precipitation (ambient and ~50% reduction) applied at the plot scale (6.1 m x 12.2 m) and 2 levels of nitrogen (ambient or 60 kg N ha-1 yr-1 added) applied to subplots within precipitation treatments. This design is replicated in 8 experimental blocks. Within each block, we used subplots with 1) ambient precipitation and ambient nitrogen (“control” plots), 2) reduced precipitation + ambient nitrogen (“drought” plots), and 3) ambient precipitation + added nitrogen (“nitrogen addition” plots). Drought was imposed by covering the plots (using polyethylene sheeting over steel arch frames) only when rainfall was forecast, and the covers were removed soon after rainfall. A previous study at this site reported that litter derived from the nitrogen addition treatment contained significantly more nitrogen, cellulose, and hemicellulose, but lower concentrations of lignin

(Allison et al. 2013). Similarly, litter decaying in the drought plots contained significantly more nitrogen and protein, and significantly lower concentrations of cellulose, 8 hemicellulose, and lignin (Allison et al. 2013). Surface soil moisture (0 cm depth) was also significantly lower in drought plots than plots in the ambient treatment (two sample t-test, p>0.001; unpublished data).

Litter sample collection

Surface litter was collected from 8 replicate plots for each environmental treatment

(ambient, reduced precipitation, and nitrogen addition) four times a year for two years (3 treatments × 8 times × 8 replicates = 192 samples). The collection dates roughly corresponded to seasonal sampling, and occurred on April 14th 2010, August 20th 2010,

December 17th 2010, February 28th 2011, June 10th 2011, September 21st 2011, December

14th 2011, and March 12th 2012. We define the first 4 dates as year 1 and the last 4 dates as year 2. For each collection, litter was randomly sampled three times from a roughly 4 m2 section of a plot, combined and stored on ice for approximately 2 hours. Litter samples were ground using a blade coffee grinder (KitchenAid model BCG111OB) for approximately 1 minute, and 0.05 g of ground litter was flash frozen using liquid nitrogen and stored at -80°C for DNA extraction.

DNA extraction

Microbial DNA from all litter samples was extracted following a previously published procedure (DeAngelis et al. 2009), with a few modifications. Briefly, samples were extracted in lysing matrix E tubes (MP Biomedicals, Santa Ana CA, USA) using a hexadecyl- trimethylammonium bromide (CTAB) extraction buffer, which is 10% CTAB, 500 mM phosphate buffer, and 1 M sodium chloride; a final concentration of 100 mM aluminum ammonium sulfate was also added (Braid et al. 2003), along with phenol-chloroform- isoamylalcohol (25:24:1). Following bead-beating in a FastPrep FP120 (Bio101, Vista CA) at 5.5 m s−1 for 45 s, extracts were purified with chloroform. Nucleic acids were precipitated in 30% polyethylene glycol (PEG-NaCl) 6000, washed once with 70% ethanol, and 9 reconstituted in water. These extracted nucleic acids were then cleaned using an AllPrep kit

(Qiagen, Valencia CA, USA) and then used for PCR amplification.

Microbial community sequencing

For the bacteria, 2 µl of a 1:10 dilution of DNA (median of 7.4 ng DNA; lower and upper quartiles of 4.8 and 10.1 ng DNA, respectively) from each extract was added to a

PCR cocktail containing 1.2 units of HotStarTaq polymerase (Qiagen, Valencia CA, USA),

1x PCR buffer supplied by the manufacturer, 200 µM of each dNTP, 0.7 µM of each primer, and H2O to a final volume of 25 µL. 16S rRNA gene amplification used concatemers containing the universal primer 907R (Lane 1991), an 8-bp multiplex tag, and the 454 ‘B’ adaptor CCTATCCCCTGTGTGCCTTGGCAGTCTCAG (in the reverse direction) and the complimentary primer 515F (Turner et al. 1999) and the 454 ‘A’ adaptor

CCATCTCATCCCTGCGTGTCTCCGACTCAG (in the forward direction). Following an initial denaturation step at 95°C for 5 min, PCR was cycled 35 times at 95°C for 30 sec, 63°C for

45 sec, 72°C for 30 sec, with a final extension at 72°C for 10 min. We acknowledge that 35 cycles is relatively high and may exacerbate amplification bias, but we wanted to follow the

Earth Microbiome protocol (Caporaso et al. 2012) for comparison to their extensive database.

For the fungi, the same PCR cocktail was used as above except the concentration of each primer was reduced to a final concentration of 0.4 µM. 28S rRNA gene amplification used concatemers containing the fungal-specific LROR (Tedersoo et al. 2008), an 8-bp multiplex tag, and the 454 ‘B’ adaptor (in the forward direction) and the complimentary primer LR5F (Tedersoo et al. 2008) with the 454 ‘A’ adaptor (in the reverse direction).

Following an initial denaturation step at 95°C for 3 min, PCR was cycled 35 times at 95°C for 30 sec, 54°C for 45 sec, 72°C for 50 sec, with a final extension at 72°C for 10 min. For both the bacteria and fungi, each aliquot was amplified twice and subsequently pooled to 10 reduce the effect of random PCR amplification.

PCR products were cleaned using the Agencourt AMPure XP PCR Purification Kit

(Beckman Coulter Inc., Indianapolis IN, USA) following the manufacturer’s instructions, quantified using a Quant-iT (Life Technologies, Grand Island NY, USA) assay kit and a

Synergy 4 microplate reader (BioTek, Winooski VT, USA), pooled into equimolar concentrations and pyrosequenced at the Duke University ISGP Sequencing Facility on a

454 Life Sciences FLX sequencer using Titanium chemistry (454 Life Sciences, Branford

CT, USA). 16S and 28S rRNA amplicons were run in separate regions of a gasketed plate.

Additionally, the two sample years were prepared and sequenced separately, which could contribute additional, artificial variation across the sample years.

Characterization of the microbial community using metagenomic sequencing was obtained from Berlemont et al. (2014). Briefly, DNA extracts were pooled such that each collection date and environmental treatment combination contained 2 replicates (8 dates x 3 treatments x 2 replicates = 48 DNA libraries total). Metagenomic libraries were prepared using a Truseq library kit (Illumina, San Diego CA, USA) and sequenced with an Illumina

HiSeq2000 (100bp-paired ends). Taxonomic origin of the sequences down to the genus level was obtained using the M5NR database.

Plant community

Plant community composition data used in this study are from Kimball et al. (2014).

Species composition and fractional cover was determined in all plots by point intercept during early to mid-April of 2010 and 2011, coinciding with late flowering and maximum seed set. Briefly, two 160x60 cm PVC frames with 10-cm interval grids were positioned within each plot. A stiff wire was dropped from each grid point, and the first-intercepted species recorded. The point was recorded as plant litter or bare soil if live plant material was not encountered. The number of interceptions for each species was summed within a plot to 11 calculate fractional cover. Fractional cover data of all species observed (22 total) was used to generate a Jaccard and Bray–Curtis distance matrix. Only plant data collected from plots used for microbial sampling were used in the analyses.

Extracellular enzyme activity

The potential activities of 9 extracellular enzymes involved in carbon and nutrient cycling were assayed as described in Alster et al. (2013)(Table 1.1). Briefly, litter samples were collected seasonally from September 2011 to March 2013 (7 sample dates total) and frozen at -80°C until analysis. Sample homogenates were prepared by mechanically homogenizing 0.1 g of litter in 60 mL of 25 mM maleate buffer (pH 6.0). The homogenates were continuously stirred while dispensing 200 µl per well into 96-well microplates with eight replicate wells per sample per assay. Fluorimetric enzyme assays were performed according to the methods described in German et al. (2011) and Alster et al. (2013), and oxidative enzymes were measured using a colorimetric assay described in Allison and

Jastrow (2006) and Alster et al. (2013). The potential activities (in µmol g-1 h-1) of all 9 enzymes were used to generate a Euclidian distance matrix of all samples.

Leaf litter decomposition

To assess how the environmental treatments influenced leaf litter decomposition rates, we analyzed a subset of the data from a litterbag study previously conducted at this site (Allison et al. 2013). Briefly, litterbags containing 2 g (dry weight) of litter from the drought, nitrogen addition, or ambient plots were placed into nylon membrane bags with

0.45 µm pores and sterilized with at least 22 kGy gamma irradiation. Sterile litter bags were then reinoculated with 50-mg of air dried, ground (Wiley mill, 1 mm mesh) non-sterile litter collected from drought, nitrogen addition, or ambient plots. Litterbags were placed in the field on December 15, 2010, and subsets were retrieved on March 3, 2011, June 14, 2011, and November 14, 2011, for analysis of percent mass loss. Fresh litter in each bag was 12 weighed and a subsample was dried to a constant mass at 65°C to obtain dry weight; all mass losses are reported as percent initial dry mass. For this study, we only analyzed

“control” litterbags whose litter contents, microbial inoculum, and deployment environment were all from the same environmental treatment (8 litterbags x 3 treatments x 3 dates = 72 litterbags total).

Sequence analysis

Pyrosequencing data were processed using the QIIME (version 1.6.0) toolkit

(Caporaso et al. 2010) with the following parameters: quality score >50, sequence length

>300 and <700 for fungi and >200 and <550 for bacteria, maximum homopolymer of 6, 6 maximum ambiguous bases and 0 mismatched bases in the primer. Sequences were denoised using Denoiser (Reeder and Knight 2010) and OTUs were picked at the 97% identity level using UCLUST (Edgar 2010) in QIIME. Fifteen of the 192 samples were excluded from analyses for which quality data were not obtained. Because the number of reads in the remaining samples varied widely (623-42,560, median=6,410), we generated median Bray–Curtis distance matrices based on 100 random sub-sampling of the full fungal and bacterial OTU-by-sample matrices. Briefly, 100 OTU-by-sample matrices were created in QIIME such that each matrix contained samples with the same number of bacterial (623) or fungal (2269) sequences (the number of sequences in the smallest sample). For each of these matrices, a Bray-Curtis distance matrix was created. Then, for each pairwise sample- to-sample comparison, the median value across the 100 distance matrices was chosen to make up the final median Bray-Curtis matrix used in the data analyses.

The taxonomic identity of OTUs was assigned to bacteria by reference to the

Greengenes database (DeSantis et al. 2006) using the naïve RDP classifier within QIIME

(Wang et al. 2007) and to fungi by reference the 28S LSU RDP database using BLAST.

QIIME was also used to determine the Shannon diversity (α diversity metric) of all libraries 13 on the rarefied OTU matrix (Caporaso et al. 2010). Unprocessed sequences are available through NCBI’s Sequence Read Archive (accession number SRP041807).

Data analysis

We analyzed microbial diversity and enzyme activity data using a factorial mixed- model ANOVA (hereafter “overall ANOVA”). The model included plot treatment (ambient, drought, or nitrogen addition) and collection date (hereafter referred to as season) nested within year as fixed, categorical effects and year and block as random effects, along with the season and plot treatment interaction. Plant data were analyzed using the same model but without the season and season and plot treatment interaction variables. Decomposition data were analyzed using a mixed-model ANOVA with plot treatment (ambient, drought, or nitrogen addition) and time (3, 6, and 11 months) as fixed factors and block as a repeated measure, random effect. All analyses were conducted in the R software environment (R

Development Core Team 2011).

Permutational multivariate analysis of variance (PERMANOVA) was used to test the effects of experimental factors on the distribution of microbial enzyme activity and bacterial, fungal, and plant community composition. The microbial and enzyme models included plot treatment (ambient, drought, or nitrogen addition) and season nested within year as fixed effects and year and block as random effects, along with the season and treatment interaction. The plant model included plot treatment as a fixed effect and year and block as random effects. PERMANOVA analyses were conducted using partial sums of squares on

999 permutations of residuals under a reduced model. When the model returned nonsignificant variables, these terms were removed and model was run again until all terms in model had a p-value of <0.05. Terms with lowest mean square were removed first.

Multivariate analysis was conducted using PRIMER6 and PERMANOVA+ (Primer-E Ltd,

Ivybridge, UK). Statistical routines are described in (Clarke and Warwick 2001) and

(Anderson et al. 2008).

To test whether the compositional differences were correlated between the microbial and plant communities, we performed a Mantel test based on 999 permutations using both

Jaccard and Bray-Curtis matrices. We also determined if the overall profile of enzymatic activity was correlated with bacterial or fungal composition. For litter samples where we had both microbial composition and potential enzyme activity data (72 samples), we first used a

Mantel test to compare a Euclidean distance matrix of all enzyme activities and the microbial Bray-Curtis matrix. To further investigate whether overall enzyme activity could be explained by particular microbial taxa, we performed a distanced-based linear model

(DISTLM) test in PRIMER6 using the relative abundance of the top 20 bacterial and fungal

OTUs as predictor variables. Analyses were conducted using a step-wise selection procedure to maximize the adjusted R2 based on 999 permutations.

Results

Leaf litter microbial composition

We characterized 2,199 bacterial and 830 fungal OTUs using a 97% sequence similarity cutoff from roughly 1.2 and 1.3 million high quality sequences, respectively (Table

1.2). Most of the bacterial OTUs were classified as Proteobacteria (28%), Bacteriodetes

(16%) or Actinobacteria (15%), with Planctomycetes (11%) and Firmicutes (10%) comprising notable fractions as well. The vast majority of the fungal OTUs were classified as Ascomycota (92%) or Basidiomycota (8%), but OTUs classified as Chytridiomycota or

Blastocladiomycota were also detected. The majority of bacterial and fungal OTUs were detected more than once (only 30 and 35% were singletons, respectively) and in multiple

15 samples. However, much of this diversity was not evenly distributed across samples; 873 bacterial OTUs (40%) and 281 fungal OTUs (34%) were unique to one sample.

The most diverse bacterial phyla were also the most abundant; Proteobacteria,

Bacteriodetes, and Actinobacteria made up 23%, 26%, and 49% of the sequences, respectively. Notably, four OTUs (from the genera Duganella, Curtobacterium,

Frigoribacterium, and Kineococcus) were observed in all samples, and together represented approximately 40% of all sequences (Table S1). Indeed, despite high total richness, a small number of bacterial taxa dominated the samples; the two most abundant OTUs,

Curtobacterium and Frigoribacterium (both in the Microbacteriaceae family), represented approximately 34% of the sequences and 20 OTUs made up 73% of all the sequences

(Tables 1.1 and S1). This trend was even more striking in the fungal community; six OTUs

(Capnodiales, Davidiellaceae OTU 1, Phaeosphaeriaceae OTU 1, Pleosporaceae OTU 1,

Agaricostilbaceae, and Tremellaceae) were observed in all samples, and together represented 83% of the sequences (Table S1). The top 20 most abundant fungal OTUs made up 94% of the sequences (Tables 1.2,S1).

Although amplicon sequencing is known to suffer from a variety of methodological biases, the broad compositional results observed in our datasets were also observed using metagenomic sequencing of the same samples (Berlemont et al. 2014; Figure S1). The most abundant bacterial and fungal phyla were detected in both datasets. While families

(the finest taxonomic resolution that can generally be assigned to the metagenomic sequences) were distributed more evenly in the metagenomes, their relative abundances were positively correlated across ribosomal DNA and metagenomic datasets (Figure S2).

Additionally, the families only detected in the metagenomes represented just 14% and 11% of the bacterial and fungal metagenomic sequences, respectively.

Treatment and temporal effects on microbial composition

The drought and nitrogen addition treatments had small but significant effects on microbial community composition analysis (white and gray bars in Figure 1.1;

PERMANOVA: p<0.002 in all instances). The main effect of the drought treatment explained approximately 3% of the bacterial and 6% of the fungal community composition.

Additionally, the main effect of the nitrogen treatment was consistent for both the bacterial and fungal communities, explaining roughly 2% of the changes in community composition.

Bacterial and fungal community evenness (mean ± SEM) was not significantly affected by either of the environmental treatments (Table 1.3).

While microbial composition was altered by the environmental treatments, it was even more strongly influenced by annual and seasonal variation (Figures 1.1-1.2,S3). The inter-annual variation can even be seen at the phylum level. For example, 16S rRNA sequences identified as Actinobacteria, Bacteriodetes, and Proteobacteria represented

43%, 28%, and 28% of all sequences in year 1, but shifted to 60%, 25%, and 22% in year 2.

Microbial evenness also varied by season (Table 1.3) and was lower during the summer months than the rest of the year (Figure 1.3c).

Drought, but not nitrogen addition, interacted with season to affect bacterial and fungal composition, explaining 3% of the variation (Figure 1.1). Drought was more likely to affect composition in samples collected during the rainy season (Table 1.4). Further, the drought x season interaction was apparent when we repeated the analyses with just the abundant (top 20 OTUs) or rare taxa (not in the top 20)(Figure S5). Finally, the effect of drought on fungal evenness also varied by collection date (Table 1.3), although a seasonal trend was less apparent (Table 1.4).

Given that the microbial response varied by season, we next asked whether taxa that responded to the drought manipulation responded in a similar way to the dry, summer 17 season. In general, the majority of fungal taxa that decreased in relative abundance in the drought treatment tended to have reduced relative abundances during the summer, although this trend was not statistically significant (χ2=0.98, p=0.32; Figure S4b). The same trend was not apparent for the bacteria; taxa that decreased in relative abundance during the dry season were equally likely to show increased or decreased relative abundances in the drought manipulation (χ 2=0.25, p=0.62; Figure S4a).

Although the degree to which microbial composition responded to the two environmental treatments was similar (comparing panels a and b in Figure 1.1), the direction of responses did not appear to be linked (Figure 1.4). Some taxa, like

Methylobacterium OTU 2, had a higher relative abundance in the nitrogen addition plots but decreased in relative abundance in the drought plots. Other taxa, like Kineococcus and

Pleosporaceae OTU 1, showed a positive response to the drought treatment but no response to the nitrogen addition treatment. Three of the most abundant fungal taxa

(belonging to the Dothioraceae, Phaeosphaeriaceae, and Tricholomataceae families) also had significantly lower relative abundances in both the nitrogen addition and drought treatments (Figure 1.4b).

Extracellular enzyme activities

To investigate potential functioning of the microbial decomposers, we assayed the potential activity of extracellular enzymes in the leaf litter. Like the bacterial and fungal communities, variation in enzyme composition was most strongly influenced by temporal variation (black bars in Figure 1.1), but was also affected significantly by the nitrogen treatment. When each enzyme was analyzed individually, we also observed strong seasonal effects and, in general, activities were lower in the dry summer months (Table S2-

S3). Some enzyme activities varied significantly across environmental treatments. For example, acid phosphatase activity was significantly higher in the nitrogen treatment than 18 the ambient treatment, and polyphenol oxidase activity was significantly lower in the drought treatment (Table S2). The potential activities of three enzymes (polyphenol oxidase, β- xylosidase, and leucine aminopeptidase) also had a significant interaction between drought and season and were generally lower in the dry, summer months (Table S2). While microbial composition and enzyme activities varied similarly through time and in response to the environmental treatments, microbial composition was not correlated with the overall profile of potential enzyme activities (Mantel tests: p>0.05). However, a selection of the most abundant microbial taxa did correlate with enzyme activities. A model of eight bacterial

OTUs or five fungal OTUs explained 13% and 23% of total variation in the overall profile of enzymatic activity, respectively (Table S4).

Comparison to plant community

To put the variation in microbial composition into perspective, we analyzed plant composition from the same plots. Although the plant and microbial datasets are not entirely comparable (plant composition was not measured across seasons), we could examine interannual versus treatment variation. As with the microbial communities, plant composition varied largely through time (checkered bars in Figure 1.1). Drought also altered plant composition (Figure 1.1a), due to a significant reduction of native grasses in drought plots

(Table S5). Unlike the microbial communities, however, nitrogen addition treatment did not affect plant community composition (Figure 1.1b), although it did decrease the relative abundance of native grasses and increase the relative abundance of non-native grasses

(Table S5). When we compared the plant and microbial communities directly, we found a significant positive correlation between plant community composition (binary Jaccard distance) at the end of each growing season with the microbial communities (Bray-Curtis distance) the following winter (rM=0.31, p=0.001 and rM=0.22, p=0.001 for bacterial and

19 fungal communities, respectively). Thus, plant and microbial communities appeared to be similarly sensitive to the treatments and temporal variability in this system.

Litter decomposition rates

Leaf litter mass generally declined through time, with an average of 78-83% of initial litter mass remaining after 11 months (Figure 1.5). Across all time points, the drought and nitrogen addition treatments did not significantly affect the fraction of leaf litter decomposed

(mixed-model ANOVA: χ2=0.14, p=0.71 (drought) and χ2=0.13, p=0.72 (nitrogen)). The response of decomposition to drought varied by season, however. Decomposition was significantly lower in the drought litterbags after 6 months (mixed-model ANOVA: χ2=22.2, p<0.001; Figure 1.5). By the end of the dry season (month 11), mass loss was again similar across treatments.

Discussion

Climate and plant production are highly seasonal in California grasslands. These factors are also known to affect microbial composition either directly or indirectly (Zak et al.

2003, Fierer et al. 2009, de Vries et al. 2012). In light of this background temporal variation, we quantified the compositional and functional responses of grassland litter microbes to two simulated global changes: drought and nitrogen addition. The responses of microbial composition to these treatments were consistent with our three hypotheses. Both bacterial and fungal composition varied strongly over seasons and years, and these effects were larger than either environmental treatment (Hypothesis 1). The response of both microbial groups to drought depended on season, however (Hypothesis 2). In contrast, their responses to nitrogen were consistent among seasons, supporting the idea that seasonal- dependence is stronger for highly seasonal factors like precipitation (Hypothesis 3).

The response of microbial composition to the treatments and temporal factors was mostly concordant with the response of the functional potential of the litter communities, as assayed by the potential activities of extracellular enzymes. Of the variation that could be explained, most of the variation in enzymatic activity could be attributed to seasonal and interannual variation, and less so to the nitrogen treatment. Unlike microbial composition, however, drought did not significantly affect overall enzyme activity, although the activity of many individual enzymes (AG, AP, BX, NAG, and PPO) tended to be lower in the drought plots than the ambient plots. Further, the enzyme response to drought seemed to be less season-dependent than the compositional response. Although three of nine enzymes responded differently to drought over time, the overall enzyme profile did not (i.e., there was no drought x time interaction). These results indicate some degree of functional redundancy amongst the extracellular enzyme-producing microbes among the wet and dry season communities. In contrast, the nitrogen treatment had a small, but similar impact on the variation in microbial composition and its functional potential, suggesting that added nitrogen selects for taxa that differ in their enzymatic activities. Indeed, a selection of some of the dominant bacterial and fungal taxa explained a small (13% or 23%, respectively) but significant amount of variation in the profiles of enzyme activities across our samples.

Like microbial composition and enzyme activity, plant community composition also varied primarily over time and was altered by drought. Nitrogen addition did not affect overall plant species composition, but did alter the relative abundance of native and non- native grasses. In addition, plant community composition was positively correlated with microbial composition on litter collected later in the year. The parallel response patterns of microbes and plants and the direct correlation between their composition both indicate that the plant community might be driving some of the changes in microbial composition and its potential functioning (as seen in Zak et al. 2003, Loranger-Merciris et al. 2006). Indeed, litter 21 chemistry was altered in both treatments (Allison et al. 2013). For instance, drought increased lignin and decreased cellulose and hemicellulose concentrations, whereas added nitrogen had the opposite effects.

Together, however, the plant and microbial responses to the global change treatments did not translate into discernible impacts on litter decomposition. Drought slowed decomposition during the spring; however, neither drought nor nitrogen addition altered total mass loss over a full year. This result is surprising considering the mounting evidence for the influence of microbial composition (Bell et al. 2005, Tiunov and Scheu 2005, Strickland et al. 2009, Matulich and Martiny 2015) as well as that of the plants (Hobbie 1996,

Hattenschwiler et al. 2005). In this system in particular, we used a reciprocal transplant experiment to tease apart the microbial effect on decomposition (Allison et al. 2013).

Microbial communities from the drought environment were slower to decompose litter than communities from the ambient environment. Thus, it appears that, when combined, the various pathways that the treatments alter may work in different directions and buffer any changes on litter decomposition. Given the importance of temporal variability, however, we recognize that additional decomposition measurements are needed to confirm this result over multiple years.

Despite its high temporal variability, the microbial community was dominated by only a handful of highly abundant bacterial and fungal taxa throughout the two-year study. Such an uneven abundance distribution was also observed in oak leaf litter, where approximately

80% of all fungal sequences belonged to only 30 OTUs (Voriskova and Baldrian 2013).

Nevertheless, little is known about the taxa that dominate leaf litter. Broadly speaking, the taxonomic composition observed in this study is similar to that found in other terrestrial litter and soil systems. The majority of fungal sequences in this study were Ascomycota and

Basidiomycota (Barnard et al. 2013, Voriskova and Baldrian 2013, Weber et al. 2013), and 22

Proteobacteria, Bacteriodetes, and Actinobacteria made up most of the bacterial diversity

(Castro et al. 2010, Sheik et al. 2011, Barnard et al. 2013, Kim et al. 2014). In contrast, many soils contain a large fraction of Acidobacteria (Castro et al. 2010, Sheik et al. 2011,

Barnard et al. 2013, Kim et al. 2014), but they represented less than 1% of our litter communities.

Previous studies in California ecosystems have also observed general effects of drought and nitrogen on microbial composition. For example, Yuste et al. (2011) found soil bacterial communities were significantly affected by a simulated drought treatment in both a holm-oak forest and a scrubland. Additionally, microbial composition in a grassland ecosystem varied in response to added nitrogen (Gutknecht et al. 2012). The specific taxonomic responses were not always consistent with previous studies, however. For instance, drought reduced the relative abundance of Proteobacteria in our study, as previously reported in a constructed old-field ecosystem (Castro et al. 2010); two abundant

Proteobacteria OTUs (Methylobacterium 2 and Mitsuaria) had a significantly lower relative abundance in the drought treatment. In contrast, Weber et al. (2013) reported a positive response of Ascomycetes to nitrogen addition, whereas many of the abundant Ascomycete

OTUs decreased in relative abundance in our nitrogen treatment. Also in contrast to prior studies, added nitrogen or drought did not increase the relative abundance of Actinobacteria

(Ramirez et al. 2012, Barnard et al. 2013). Further work is needed to understand whether these taxonomic groups are not consistent in their responses (Philippot et al. 2010), or whether such responses depend on the community and system context.

In sum, the responses of leaf litter microbes to the global change manipulations might have been easily overlooked without consideration of background seasonal variability.

This pattern is consistent with other recent experiments (Cruz-Martinez et al. 2009, Yuste et al. 2011, Cregger et al. 2012) and suggests that future changes in seasonality may be even 23 more important than changes in annual averages. We predict in this system, for instance, that additional summer rain will have a greater impact on litter decomposition than a slight increase in mean annual precipitation. Consequently, information about the seasonal phenology of microbial taxa might be useful to predict longer-term responses to climate variables, as it seems to be for plants (Menzel et al. 2006, Khanduri et al. 2008, Wolkovich et al. 2012). Indeed, the responses of litter fungi to the dry season and drought were correlated in this study. Thus as with larger organisms, the future of microbial communities will likely depend on the interplay between rapid, local background variability and slower, global change.

Acknowledgements

This work was funded by the US Department of Energy Program in Ecosystem

Research; the Office of Science (BER), US Department of Energy (program in Microbial

Communities and Carbon Cycling), the NSF Major Research Instrumentation program, and the US Department of Education Graduate Assistance in Areas of National Need (GAANN) fellowship. We thank Orange County Parks and the Irvine Ranch Conservancy for site access, and Michaeline Nelson, Eoin Brodie, and Kathleen Treseder, and four anonymous reviewers for comments on previous versions of the manuscript.

Tables and Figures

Table 1.1: List of extracellular enzymes analyzed. Extracellular enzymes assayed in litter decaying in a southern California grassland, and their abbreviations, functions, corresponding substrates, and final substrate concentrations.

Substrate Enzyme Abbreviation Function Substrate Concentration α-glucosidase AG Starch degradation 4-MUB-α-D-glucopyranoside 200 µM Acid phosphatase AP Mineralizes organic P into 4-MUB Phosphate 800 µM phosphate β-glucosidase BG Cellulose degradation 4-MUB-β-D-glucopyranoside 400 µM β-xylosidase BX Hemicellulose degradation 4-MUB-β-D-xylopyranoside 400 µM 25 Cellobiohydrolase CBH Cellulose degradation 4-MUB-β-D-cellobioside 200 µM

Leucine LAP Peptide breakdown L-leucine-7-amido-4- 200 µM aminopeptidase methylcoumarin hydrochloide N-acetyl-β-D- NAG Chitin degradation 4-MUB-N-acetyl-β-D- 400 µM glucosaminidase glucosaminide Polyphenol PPO Degrades lignin and other pyrogallol 1000 µM oxidase aromatic polymers Peroxidase PER Catalyzes oxidation reactions pyrogallol 1000 µM

Table 1.2: Overview of pyrosequencing results of microbial communities. Results of 454 pyrosequencing of bacterial and fungal communities across all samples (177 total).

Bacteria Rarefied Variable Raw Total Ambient Drought Nitrogen Total Sequences 1,197,853 110,271 36,134 35,511 38,626 Median Sequences/Sample 5953 623 623 623 623 Total OTUs (97% Similarity) 2199 935 573 621 589 Singleton OTUs 35% 38% 36% 36% 37% Relative Abundance Top 20 OTUs 73% 74% 74% 75% 73% Median OTUs/Sample 211 80 84 77 85 Mean Shannon Diversity (± SE) - 4.46 (0.07) 4.49 (0.13) 4.35 (0.13) 4.54 (0.13) 26 Fungi

Rarefied Variable Raw Total Ambient Drought Nitrogen Total Sequences 1,261,095 401,613 136140 129333 136140 Median Sequences/Sample 6507 2269 2269 2269 2269 Total OTUs (97% Similarity) 830 602 455 412 420 Singleton OTUs 30% 25% 24% 29% 27% Relative Abundance Top 20 OTUs 94% 93% 92% 94% 93% Median OTUs/Sample 97 64 71.5 59 65.5 Mean Shannon Diversity (± SE) - 2.54 (0.04) 2.75 (0.09) 2.37 (0.06) 2.49 (0.07)

Table 1.3: ANOVA results for each experimental factor on microbial diversity. Significant drivers of microbial community evenness (Shannon diversity) from mixed model ANOVA results (reported as p-values). Season and Treatment terms are treated as fixed factors, while Year and Block are treated as random factors. The variable Season is nested within Year.

Drought Nitrogen Addition Bacterial Community df χ2 P-value χ2 P-value Year 1 NS NS NS NS Season 7 44 <0.001 37.8 <0.001 Treatment 1 NS NS NS NS Treatment x Season 7 12.3 0.09 NS NS Block 1 7.21 <0.01 12.4 <0.001 Fungal Community df χ2 P-value χ2 P-value Year 1 NS NS NS NS

27 Season 7 86.7 <0.001 84.8 <0.001

Treatment 1 NS NS NS NS Treatment x Season 7 23.3 <0.01 NS NS Block 1 6.4 0.01 3.66 0.06

Table 1.4: PERMANOVA and ANOVA results for each collection date. PERMANOVA and mixed model ANOVA analysis of bacterial and fungal community for each of the 8 collection dates. Each model included the variables Treatment and Block. The significant and marginal p-values for the variable Treatment are reported.

Community Composition Community Evenness Drought Nitrogen Addition Drought Nitrogen Addition Month Bacteria Fungi Bacteria Fungi Bacteria Fungi Bacteria Fungi

April 2010 NS 0.089 NS NS NS NS NS 0.009 August 2010 NS NS 0.069 NS NS 0.032 NS NS December 2010 0.016 NS NS NS NS 0.001 NS NS Year 1 Year February 2011 0.043 0.042 NS NS 0.044 0.001 NS NS

June 2011 NA NS NS NS NS 0.06 NS NS September 2011 0.079 0.012 0.03 NS 0.098 NS 0.018 NS

28 December 2011 0.07 0.038 0.048 0.048 0.023 0.009 NS 0.019 Year 2 Year

March 2012 NS 0.013 NS 0.019 NS <0.001 NS 0.008

60 (a) Bacteria Fungi 50 Enzymes Plant 40

NA NS NS NA NS 0 60 (b) 50

40 Estimated Variation (%) EstimatedVariation

NA NS NS NS NS NA 0 Year Season Treatment Treatment x Block Residual Season

Figure 1.1: PERMANOVA results for microbial, plant, and enzyme composition. PERMANOVA multivariate components of variation for the bacterial (white), fungal (grey), microbial enzymes (black), and plant (checkered) communities for each the drought (a) and nitrogen addition (b) treatments (average based on 999 unique permutations). Data includes samples from all collection dates (8 total for microbial communities, 7 for enzyme concentrations, and 2 for plant community). The values plotted are the proportional sizes of the estimates of the variance components. All p-values <0.05.

2653&'#!+

/01#23456 !"#$%&'(()#*+*,

:34' $';%'<7'& !*-- "'9'<7'& =0&9> !*-! ?<76'4% "&@35>% A6%&@5'4#?BB6%6@4

/71#809%'&60 !"#$%&'(()#*+-.

Figure 1.2: NMDS ordination of bacterial and fungal composition. Non-metric multidimensional scaling (NMDS) ordination based on Bray Curtis similarities depicting (a) fungal and (b) bacterial community composition in the second sampling year. The samples collected during the same collection date are displayed in the same symbol. The color of the symbols represent the plot treatment: ambient conditions (blue), drought (yellow), or nitrogen addition (green).

30 Figure 3.

<*> %"

#" 20=;08*3C80- ?@08*A0-B*63:=0 !" !" ' & %

2,3*4-5,+3)46 # 7809:;:3*3:,+-<9=> " <9> 5 4 3 2 ()*++,+-.+/01 1 ?;8:4 ?CAE B09E G0HE FC+0 (0;3E B09E 5*89)

#"!" #"!! #"!#

Figure 1.3: Temperature, precipitation, and diversity across collection dates. Average (± SE) maximum daytime temperature (a), total monthly precipitation (b), and (c) shannon index values for bacterial (white) and fungal (grey) communities across all collection dates.

Figure 4.

80 (a) 60 !" 40 !" !"

-20 !" -40 !" !" Actinobacteria Bacteriodetes ! -Proteobacteria " -Proteobacteria -60 Mitsuaria Duganella Variovorax Variovorax Rhizobium Acidovorax Kineococcus Dyadobacter Nocardioides Roseomonas Pedobacter1 Pedobacter2 Sphingomonas Microbacterium Curtobacterium Mucilaginibacter Hymenobacter1 Hymenobacter2 Frigoribacterium Methylobacterium1 Methylobacterium2

100 (b) 80

40 !" Treatment response (%change) Treatment

-20

-40

-60 !" !" !" !" -80 !" !"!" !" !" !" Ascomycota !" !" Basidomycota -100 !" Coriolaceae Capnodiales Dothioraceae Tremellaceae Tremellaceae Helotiaceae1 Helotiaceae2 Davidiellaceae1 Davidiellaceae2 Pleosporaceae1 Pleosporaceae2 Agaricostilbaceae Tricholomataceae Tricholomataceae Ceratobasidiaceae Christianseniaceae Teratosphaeriaceae Teratosphaeriaceae Mycosphaerellaceae Didymosphaeriaceae Phaeosphaeriaceae1 Phaeosphaeriaceae2 Phaeosphaeriaceae3

Figure 1.4: Responses of abundant bacterial and fungal OTUs to environmental treatments. Percent change (mean ± SE) in relative abundance of the 20 most common (a) bacterial and (b) fungal OTUs in nitrogen addition (diagonal) or drought (white) plots relative to ambient plots across all collection dates. Asterisks (*) denote p-value of <0.05 in two sample t-test between the ambient and drought or nitrogen treatments.

Figure 5.

Nitrogen Addition 100 Drought Ambient

* 90

80 Percent Mass Remaining

December 2010 March 2011 June 2011 November 2011 (deployed) (3 months) (6 months) (11 months)

Figure 1.5: Litter mass loss across environmental treatments. Mean (±SE) percent mass remaining over time in each of the three environmental treatments. N=8 for each treatment x sample date combination. Asterisks (*) represent significant differences between treatments, based on a Tukey post-hoc test.

Supporting Information

This chapter contains supporting information that can be found online at http://www.nature.com/ismej/index.html (doi:10.1038/ismej.2015.58).

CHAPTER 2

Microbial composition alters the response of litter decomposition to environmental change

Abstract

Recent studies demonstrate that microorganisms are sensitive to environmental change, and that their community composition influences ecosystem functioning. However, it is unknown whether microbial composition interacts with the environment to affect the response of ecosystem processes to changing abiotic conditions. To investigate the potential for such interactive effects on leaf litter decomposition, we manipulated microbial composition and three environmental factors predicted to change in the future (moisture, nitrogen availability, and temperature). We isolated fungal and bacterial taxa from leaf litter and used them to construct unique communities. Communities were inoculated into microcosms containing sterile leaf litter and exposed to four environmental treatments

(control conditions, increased temperature, decreased moisture, and elevated nitrogen availability). Respiration was tracked over 60 days, and communities were pyrosequenced to assess compositional changes. As hypothesized, composition and environmental treatment interacted to influence respiration rates. In particular, microbial composition interacted more strongly with changing nitrogen availability and less so with changing moisture or temperature. Further, the magnitude of a community’s response to a particular environmental change was partly explained by changes in composition over the course of the experiment; microcosms that showed a large change in respiration rate included more taxa whose relative abundance changed as well. Together, these results suggest that information about microbial composition may be more useful for predicting functional responses to some types of environmental changes than others.

Introduction

Biodiversity (both the richness and composition of taxa) has a direct influence on the rate of ecosystem processes (reviewed in Cardinale et al. 2011). Yet until recently, it was often assumed that microbial communities are so diverse and physiologically flexible that changes in their diversity would not affect ecosystem processes (Schimel 2001). However, recent research demonstrates that microbial diversity directly influences a variety of ecosystem processes such as litter decomposition, CO2 flux, and nitrogen cycling (Mille-

Lindblom and Tranvik 2003, Bell et al. 2005, Tiunov and Scheu 2005, Allison et al. 2009,

Strickland et al. 2009, Reed and Martiny 2013).

In addition to these direct effects on process rates, biodiversity can affect the response of ecosystem processes to changing abiotic conditions. Such interactive effects of biodiversity and abiotic change have been documented in plant communities. A recent study showed that plant species diversity and environmental perturbations (elevated CO2 and N levels) act together to influence plant biomass (Isbell et al. 2013). These results suggest that predictive models need to consider plant diversity when predicting ecosystem functioning under changing environmental conditions. It remains unknown, however, whether microbial diversity similarly affects an ecosystem’s “functional response” (Schimel and Gulledge 1998, Lavorel and Garnier 2002, Allison 2012). Given that microorganisms drive key ecosystem processes, such information may aid in predicting ecosystem functioning under future environmental conditions (Singh et al. 2010).

Leaf litter decomposition is performed largely by fungi and bacteria (Swift et al.

1979), and the specific composition of microbial decomposers influences decomposition rates. For instance, microcosm experiments with bacterial (Bell et al. 2005) or fungal

(Tiunov and Scheu 2005, Allison et al. 2009) isolates found that decomposition rate, as

measured by CO2 production, varies with richness and composition. Similarly, microcosms including both fungal and bacterial taxa had a higher decomposition rate than microbial communities containing just fungal or bacterial taxa alone (Mille-Lindblom and Tranvik

2003). Further, natural variation found in complex microbial communities affected the rate of decomposition when inoculated onto a common leaf litter (Strickland et al. 2009).

To test whether microbial diversity (specifically composition) alters an ecosystem’s functional response to environmental change, we constructed mixed bacterial and fungal microcosms using isolates from natural leaf litter communities from a mediterranean-type ecosystem located in southern California. The use of cultured isolates allowed us to hold richness constant while manipulating the identity of the taxa. We then measured microbial respiration (a proxy for decomposition) and changes in composition over the course of the experiment. Each community was replicated under control conditions and three environmental scenarios (reduced moisture, elevated nitrogen, and elevated temperature).

These particular environment changes are predicted to occur in Southern California in the near future (IPCC 2013) and will likely impact local ecosystems. For example, water availability is a key environmental constraint in this ecosystem due to the combination of high summer temperatures and low rainfall (Specht et al. 1983, Larcher 2000). Additionally,

Southern California ecosystems suffer from chronic nitrogen loading and deposition rates are expected to increase in the future (Fenn et al. 2003a).

Based on previous studies, we expected that respiration rate in the microcosms would depend in part on the direct effects of both the environment and community composition. Abiotic conditions broadly affect the physiology and growth of microorganisms, as well as the activities of microbially-produced extracellular enzymes. For instance, microbial metabolism and enzymatic activity generally increase under warmer temperatures

(Lloyd and Taylor 1994). We also expected that microbial composition would directly

influence decomposition because different taxa (and therefore communities) likely vary in their functional traits, including their ability to decompose leaf litter. In addition to these direct effects, we hypothesized that the community and environment would interact to affect decomposition. Specifically, taxa would vary in the their responses to the environmental changes, resulting in compositional shifts. Such variability in response traits, combined with that in functional traits, would lead to differential functional responses among communities

(Allison and Martiny 2008).

We also predicted that the type of environmental change would affect the importance of microbial composition for a microcosm’s functional response (i.e., the strength of the community-by-environment interaction). In particular, we hypothesized that the effect of the temperature and moisture treatments would depend less on community composition than that of the nitrogen treatment. We expected the response of microbial taxa to increased nitrogen to be highly variable, as the traits involved in microbial nitrogen acquisition are highly diverse (Schimel et al. 2005). In contrast, we expected that the response to moderately elevated temperature and reduced moisture would be relatively similar among taxa. For instance, extracellular enzymes involved in decomposition tend to operate more efficiently at slightly elevated temperatures (Lloyd and Taylor 1994), whereas arid conditions usually result in slower decomposition by limiting physiological performance and the diffusion of nutrients in the soil pore space (Harris 1981).

Methods

Strain isolation and identification

We isolated 32 bacterial taxa and 24 fungal taxa from leaf litter collected from the

Loma Ridge research site located 5 km north of Irvine, California, USA (33 44’N, 117 42’W,

365 m elevation). Briefly, litter was collected in January, April, July, and September of 2011.

During this period, the site was dominated by the annual grass genera Avena, Bromus, and

Lolium the annual forb genera Erodium and Lupinus and the native perennial grass

Nassella pulchra. Once collected, litter was ground and then washed with sterilized deionized (DI) water. Litter fragments (106-212 µm) were suspended in sterile DI water and the resulting solution was passed through a 100 µm filter (Millipore).

For fungal isolation, the 100 µm filter containing the washed litter was collected and placed into a Falcon tube containing 30 mL of 0.6% carboxymethyl cellulose (CMC) solution. The resulting slurry was then added to 96-tube microplates containing malt extract agar, tap water agar, minimal nutrient medium, or cornmeal agar media (modified from

Rossman et al. 1998, unpublished data). Microplates were stored at 21°C for up to two months and resulting hyphal strands were transferred to fresh agar plates at least three times to ensure pure cultures.

For bacterial isolations, 20 ml of the filtered litter solution was collected and mixed with 20 mL of 0.6% CMC solution. The slurry was then plated onto luria broth (LB) agar plates or litter agar plates and allowed to incubate at 21°C for up to two weeks. Individual colonies that appeared were transferred to fresh LB agar plates at least three times to ensure pure cultures. Litter media was created by mixing 10 g ground litter with 1 L DI water on a magnetic stir plate for 24 hours. After the solution was allowed to settle for 24 hours,

800 mL of the solution was decanted and filtered through a 0.8 µm filter (Millipore). The filtered media was diluted with DI water to 1 L and autoclaved with 15 g agar.

We identified each microbial isolate by PCR amplification and Sanger sequencing of the 16S rRNA gene for bacteria (~1500 bp) or the 18S rRNA and internal transcribed spacer (ITS) region (~700 bp) of the rRNA gene for fungi. For the fungal isolates,

approximately 2.5 ng of DNA extract was added to a PCR cocktail containing 1.2 units of

HotMasterTaq polymerase (5PRIME) 1x PCR buffer supplied by the manufacturer, 200 µM of each DNTP, 0.5 µM of each primer (ITS1F (5'-CTTGGTCATTTAGAGGAAGTAA-3') and

TW13 (5'-GGTCCGTGTTTCAAGACG-3'); White et al. 1990, Gardes and Bruns 1993), and

H2O to a final volume of 25 µL. Following an initial denaturation step at 94°C for 1 min, PCR was cycled 34 times at 94°C for 1 min, 51°C for 1 min, 72°C for 1 min, and a final extension at 72°C for 8 min. All PCRs here and below were performed on a PTC-100 thermocycler

(Bio-Rad). DNA extract from each bacterial isolate was added to a PCR cocktail containing

1.5 units of HotMasterTaq polymerase (5PRIME) 1x PreMixF (FailSafe), 0.3 µM of each primer (pA (5'-AGAGTTTGATCCTGGCTCAG-3') and pH (5'-

AAGGAGGTGATCCAGCCGCA-3'); Edwards et al. 1989), and H2O to a final volume of 31

µL. Following an initial denaturation step at 95°C for 4 min, PCR was cycled 30 times at

95°C for 40 sec, 55.5°C for 30 sec, 72°C for 2 min, and a final extension at 72°C for 3 min

30 sec. The taxonomic identity of each fungal and bacterial isolate was determined by matching our sequences to the most closely related cultured representative using the

BLAST tool in the GenBank database (Appendix B: Table B1).

Microcosm construction

To prepare the microcosms, leaf litter from the Loma Ridge site was collected in the summer season of 2010, ground in a Wiley mill to pass a #20 mesh screen, and sterilized with roughly 23 kGy of gamma irradiation. To confirm sterility, irradiated litter was added to liquid LB and monitored for microbial growth. In the sequencing data, we did detect sequences that were not added to the microcosm. However, it is unknown whether these sequences were amplified from residual DNA or viable organisms that escaped irradiation.

Glass vials (40 ml) with a sampling septum were filled with 4 g of sterile sand and autoclaved. Sterilized litter (50 mg) was then added to each microcosm.

To create the communities, each of the 56 cultured taxa was randomly assigned to two communities; this produced eight unique communities that each contained 8 bacterial and 6 fungal isolates (Appendix B: Table B1). Cultures of each isolate were first separated from the solid growth medium by mixing three agar plugs (7x2 mm) with 1.5 ml of 0.9%

NaCl solution and ~30 glass beads (2 mm diameter) in an eppendorf tube and shaken in a

FastPrep-24 Instrument (MP Biomedicals, Solon, OH, USA) for 15 sec at 4.0 ms−1. This procedure was used to standardize the isolates in terms of growth stage rather than by abundance. For each community, equal volumes of the 14 assigned isolates were combined and the community inoculum was dispersed among replicate microcosms. The inoculated microcosms were then divided evenly among four environmental treatments

(control conditions, reduced moisture, elevated nitrogen, and elevated temperature) such that each community (n=8) and treatment (n=4) combination contained six replicate microcosms for a total of 192 microcosms.

Control microcosms were kept at room temperature (22°C) after adding 1 ml H2O to each microcosm. The microcosms were placed in large plastic containers to keep the humidity near 100% and prevent desiccation. Reduced moisture microcosms were kept at ambient conditions except that only 750 µl H2O (25% reduction) was added to each vial.

Similarly, the elevated nitrogen microcosms were at control conditions but supplemented with 5 µl of a 12.5 M NH4NO3 solution (equivalent to 5 mg of NH4NO3), decreasing the C: N ratio from approximately 40:1 to 16:1. Finally, the elevated temperature microcosms were similar to the control treatment but kept at 26°C in an incubator. The environmental treatments were chosen to mimic changes expected in California, including at least a 3°C increase in annual temperature, a decrease of 10-25% in annual precipitation, and increased nitrogen deposition over the next 50 years (IPCC 2013).

Respiration measurements

CO2 concentrations in the microcosm headspace were measured every 5-9 days to estimate cumulative CO2 production over a 60-day incubation period. For each measurement, an 8 ml subsample of headspace gas was withdrawn by syringe and injected into an infrared gas analyzer (PP-Systems EGM-4). To keep the microcosms at atmospheric CO2 concentrations, the vial caps were kept loose except for 24 hours before each sampling. At the end of the experiment, cumulative CO2 production was divided by initial litter mass (50 mg) and total duration of the experiment (60 days) to determine

-1 -1 average daily CO2 production (measured as mg C glitter day ), our metric for litter decomposition.

DNA extraction

At the end of the experiment, 1 g of the sand/litter mixture was collected from half of the microcosms (three from each community composition and treatment combination). DNA was extracted using FastDNA SPIN Kit for Soil (MP Biomedicals, Solon, OH, USA) as described by the manufacturer, with the addition of a freeze-thaw step where samples were frozen in liquid nitrogen and thawed in a 50°C water bath three times prior to bead beating.

Extracted DNA was eluted with 70 µl H2O. DNA from the initial community inoculums (8 total) was also extracted in duplicates.

Community sequencing

Fragments of the 16S and 28S rRNA genes were then amplified and sequenced to monitor changes in the bacterial and fungal community members, respectively. For the bacteria, approximately 2 µl of a 1:10 dilution of each DNA extract was added to a PCR cocktail containing 1.2 units of HotStarTaq polymerase (Qiagen), 1x PCR buffer supplied by the manufacturer, 200 µM of each DNTP, 0.5 µM of each primer, and H2O to a final concentration of 25 µL. 16S amplification used concatemers containing the universal primer

907R (Lane 1991), an 8-bp multiplex tag (Appendix A: Table A1), and the 454 ‘B’ adaptor

CCTATCCCCTGTGTGCCTTGGCAGTCTCAG (in the reverse direction) and the complimentary primer 515F (Turner et al. 1999) and the 454 ‘A’ adaptor

CCATCTCATCCCTGCGTGTCTCCGACTCAG (in the forward direction). Following an initial denaturation step at 95°C for 5 min, PCR was cycled 34 times at 95°C for 30 sec, 62°C for

45 sec, 72°C for 1 min, and a final extension at 72°C for 10 min.

For the fungi, the same PCR cocktail was used as above. 28S amplification used concatemers containing the fungal-specific LROR (Tedersoo et al. 2008), an 8-bp multiplex tag, and the 454 ‘B’ adaptor (in the forward direction) and the complimentary primer LR5F

(Tedersoo et al. 2008) with the 454 ‘A’ adaptor (in the reverse direction). Following an initial denaturation step at 95°C for 3 min, PCR was cycled 34 times at 95°C for 30 sec, 54°C for

1 min, 72°C for 1 min, and a final extension at 72°C for 10 min.

For both the bacteria and fungi, each aliquot was amplified twice to adjust for biases of individual PCR reactions, and negative controls were run on both DNA extractions and

PCR. Polymerase chain reaction products were cleaned using the AMPure XP PCR

Purification Kit (Agencourt) following the manufacturer’s instructions, quantified using a

Synergy 4 microplate reader (BioTek), and pooled into equimolar concentrations. Libraries were pyrosequenced at the Duke University ISGP Sequencing Facility, each using two out of eight regions of a 454 Life Sciences FLX Titanium Shotgun plate (454 Life Sciences).

Most of the taxa used in the experiment were detected by sequencing (42 out of 56); however, 3 fungi and 11 bacteria taxa were not detected even in the inoculum samples

(Appendix B: Table B1), suggesting they were present but not amplified during PCR. While the method detected clear differences in community composition (see Results), this bias may have masked some correlations between functional (respiration rate) and taxonomic responses.

Analyses

We analyzed average daily respiration rate and final community richness using a factorial mixed-model ANOVA. The model included environmental treatment as a fixed effect and community composition as a random effect, and their interaction. To compare the initial and final community richness in the control treatment (hereafter, “inoculum ANOVA”), we used a factorial mixed-model ANOVA with time as a fixed effect and community composition as a random effect. Significant pairwise differences were determined post-hoc using Tukey’s honest significant difference test. We tested for normally distributed residuals

(Shapiro-Wilk test) and homogeneity of variances (Bartlett test) and transformed the data when necessary.

To correlate functional and taxonomic responses, we used a linear model with the number of individual taxa that responded to the treatment as the independent variable and changes in respiration rate as the dependent variable. To assess individual taxon responses to an environmental treatment, we compared each taxon’s relative abundance to that observed in the control treatment (two sample t-test) and considered the response significant if p < 0.10 (we used this more generous cutoff because each community only had three replicates, reducing our statistical power). All analyses and post-hoc tests were conducted in the R software environment using the “stats”, “car”, and “lmerTest” packages

(R Development Core Team 2011).

To test the effect of environmental treatment on overall microbial composition, similarity matrices were generated on rarified abundance data using the Bray–Curtis method in QIIME (Bray and Curtis 1957, Caporaso et al. 2010). The resulting resemblance matrix served as a basis for non-metric multidimensional scaling (NMDS), as well as a permutational multivariate analysis of variance (PERMANOVA; Anderson et al. 2008) to test

the effects of experimental factors on the final composition. PERMANOVA analyses were conducted using partial sums of squares, on 999 permutations of residuals under a reduced model. All multivariate analyses were conducted using PRIMER6 and PERMANOVA+

(Primer-E Ltd, Plymouth, UK). Statistical routines are described in Clarke and Warwick

(2001) and Anderson et al. (2008).

Results

Community responses

One possible outcome of microcosm experiments is that a small fraction of the taxa grow especially well under microcosm conditions, such that the realized diversity of the experimental community differs from the inoculated community (e.g., McGrady-Steed et al.

1997). To examine this possibility, we first compared the composition of the initial and final communities. Although bacterial and fungal composition changed relative to the initial inoculum, most of the variation in composition persisted (Figure 2.1). Both bacterial and fungal composition remained highly distinct in the different communities at the end of the

60-day experiment (a significant community effect; Table 2.1). Further, of the taxa that we could track by sequencing (see Methods), all were still present in the microcosms at the end of the experiment. In particular, taxon richness in the control treatment was similar at the beginning and end of the experiment (average of 9 taxa; inoculum ANOVA: χ2=2.62, p=0.11).

All three environmental treatments – moisture, nitrogen, and temperature – altered both microbial richness and composition in the microcosms. Richness was significantly lower in the three environmental perturbations than under control conditions (Table 2.1;

Figure B1). This reduction was most pronounced in the elevated nitrogen treatment, where

on average 1.25 fewer taxa were observed by the end of the experiment. Further, the response of richness to nitrogen addition varied by community (χ2=6.0, p=0.01); the other environmental treatments did not show an interactive effect (Table 2.1).

Like richness, bacterial and fungal composition (as measured by relative abundance of the taxa) was also influenced by environmental change. However, for all three perturbations, bacterial and fungal composition was not directly affected by the environmental treatment alone. Instead, the response to each environmental treatment varied by community; that is, the community and environment interacted to affect composition (Table 2.1). This interaction was strongest in the elevated nitrogen treatment, as reflected by the variance partitioned (Table 2.1). Nitrogen addition caused the greatest number of taxa to shift in relative abundance. Overall, 36% of the fungal taxa in the elevated nitrogen treatment had a significantly different relative abundance, compared to 11% and

22% of the taxa in the reduced moisture and elevated temperature treatments, respectively.

This trend differed slightly for the bacterial community, where 22% of the bacterial taxa in the nitrogen addition treatment significantly changed in relative abundance, compared to

24% and 15% in the reduced moisture and elevated temperature treatments, respectively.

Functional responses

As expected, microbial composition had a significant effect on decomposition in the microcosms. Average daily CO2 production differed among the eight communities (mixed- model ANOVA: χ2=6.5, p=0.01). Environmental treatment also had a significant effect on respiration (χ2=38.4, p<0.001; Table 2.1); the average rate increased in the three environmental treatments (reduced moisture, elevated nitrogen, and elevated temperature) relative to the control treatment (Figures 2.2, Appendix B: Figure B2). Further, microbial composition altered the functional response to environmental change. Microbial composition interacted with the environmental treatment to affect CO2 production (lumping all

environments: χ2=38.8, p<0.001; Table 2.1; Appendix B: Figure B2).

To test if the influence of microbial composition on the functional response varied with the type of environmental change, we examined the interactive effect for each of the environmental treatments (reduced moisture, elevated nitrogen, elevated temperature) when compared to the control conditions. For each environment, the main effect of microbial composition on average respiration rate was no longer significant, although the effect of environmental treatment remained (Table 2.1). However, microbial composition still showed a significant interaction with both moisture and nitrogen to affect CO2 production, and the composition by temperature interaction was marginally significant (Table 2.1). To quantify the magnitude of these interactive effects, we calculated the variance among communities in the difference between mean respiration rate in the control treatment and the environmental treatment. As hypothesized, composition mattered most to the functional response under changing nitrogen conditions (σ2= 6.65), whereas it mattered less under changing moisture (σ2= 1.38) and temperature conditions (σ2= 1.49)(Table 2.2).

The magnitude of a community’s functional response was also partly explained by changes in community composition over the course of the experiment. Microcosms that showed a large change in respiration rate in the face of an environmental perturbation included higher numbers of taxa whose relative abundance significantly differed from the control treatment

(R2=0.16, p=0.05; Figure 2.4). For example, none of the taxa in Community 7 differed in relative abundance when exposed to elevated temperature, and the average respiration rate of this community did not significantly vary from the control conditions (Figure 2.4). In contrast, 40% of taxa in Community 4 shifted in their relative abundances when exposed to elevated nitrogen, and the average respiration rate of this community was significantly higher than the control treatment. This correlation was largely driven by the nitrogen treatment (R2=0.22, p=0.24; Figure 2.4), suggesting that the compositional shifts resulting

from increased nitrogen availability were especially important for changes in overall respiration. In contrast, changes in physiology or total microbial abundance may be more important in explaining the functional responses to moisture and temperature shifts.

Discussion Here we show that, like for plant biodiversity, composition influences the functional response of model microbial communities. Microbial composition mattered both through a main effect of community on leaf litter decomposition (as measured by respiration rate), as well as a community-by-environment interaction. These results suggest that microbial decomposers vary in two types of traits (Lavorel and Garnier 2002, Allison and Martiny

2008). First, the direct effect of community composition on decomposition indicates these isolated taxa differ in their functional traits. Indeed, it is well known that both bacterial and fungal taxa vary in their rate and ability to decompose particular compounds in leaf litter (as seen in Romani et al. 2006, Allison et al. 2009, Lebauer 2010). Second, the interactive effect of community and environment indicates that the taxa differ in their response traits, or how the new environment alters the taxa’s metabolism and/or abundance through reproduction or death. Only if both functional and response traits vary would one observe an interactive effect of the community and the environment.

Artificially assembled communities in microcosms allow researchers to test questions about microbial diversity that would be infeasible in natural communities; however, their applicability to the ‘real-world’ is a relevant concern (Carpenter 1996). Our experimental communities were highly reduced in diversity compared to natural litter- decomposing communities (Voriskova and Baldrian 2013, Kim et al. 2014). Indeed, in a

PCR survey of litter from the same ecosystem (Matulich et al. 2015), we observed well over

2000 bacterial and 800 fungal taxa (again defined by 97% sequence similarity of 16S rDNA). However, almost all of the bacterial taxa used in the microcosms were detected in

the natural community, where their relative abundance ranged from 0.002% to 18.6% of the samples (Table B1). Further, fungal families containing our cultures represented almost

50% of the natural community, although <1% of the particular fungal taxa were detected.

Finally, our communities were more complex than many model microbial experiments (e.g.,

Mille-Lindblom and Tranvik 2003, Tiunov and Scheu 2005). In particular, our microcosms included natural leaf litter substrate and both bacteria and fungi, likely increasing the complexity of interactions that would otherwise occur in more simplified community. And, while the number of isolates available limited the richness we could include, microbial decomposer communities are often highly uneven and generally dominated by a small fraction of taxa (Voriskova and Baldrian 2013).

As expected, environmental treatment affected decomposition in the microcosms regardless of microbial composition (Figure 2.2). Elevated temperature increased CO2 production for all the communities. Warmer temperatures can increase decomposition rates through a variety of mechanisms, including increased enzymatic activity, metabolism, and growth (increased abundance), or a combination (Ratkowsky et al. 1982, Lloyd and Taylor

1994, Sheik et al. 2011). Likewise, nitrogen addition also increased decomposition, and elevated nitrogen availability often stimulates decomposition by lowering the C:N ratio of the organic material (Taylor et al. 1989, Allison et al. 2009, Bragazza et al. 2012). Contrary to our predictions, however, microcosms subjected to reduced moisture showed significantly higher respiration rates than those in control conditions (Figure 2.2). Arid environments often have slow rates of decomposition because water limits the physiological performance of microbes and the diffusion of nutrients in the soil pore space (Harris 1981). We suspect that the higher respiration rates in the drier microcosms may be caused by increased fungal activity, as many fungi grow well in drier, Mediterranean conditions (Yuste et al. 2011).

Indeed, both the bacteria and fungi used in this study were isolated from a Mediterranean-

like ecosystem where high temperatures and prolonged droughts are common, thus they may not only tolerate, but thrive in such conditions. Alternatively, the moisture level in our control treatment may have been detrimental to some taxa by inhibiting oxygen diffusion in some parts of the microcosm.

We also observed a significant interaction between composition and environmental treatment, showing that at least in this system, the response of litter decomposition to environmental change does depend on microbial community composition (Figure 2.3; Table

2.1). In this study, the interactive effect of microbial composition and environment was strongest in the elevated nitrogen treatment and suggests that, like plants, the response to nitrogen availability for microbes is taxon-specific (Reich et al. 2001). This community-by- environment interaction was also seen on the final microbial composition, suggesting that compositional shifts (i.e., changes in relative abundance) are at least partly responsible for our observed main effect of environmental treatment on CO2 production. Indeed, there was a correlation between taxonomic and functional responses, such that changes in respiration were generally correlated with large changes in the relative abundance of taxa. This result was most evident in the elevated nitrogen treatment, where 11-50% of the taxa in all communities had significantly different relative abundances than in the control treatment

(Figure 2.4). However, functional responses could not be correlated to the presence or absence of particular taxa (i.e., sampling effects). But note this trend could not be tested directly, because each taxon was only present in two of the communities. Moreover, environmental treatment, which had a main effect on respiration rates, did not directly alter microbial composition over the course of the experiment (Table 2.1). Together, these results suggest that some of the changes in respiration rate may be due not only to changes in composition, but also to changes in microbial activity or total abundance, which were not measured here.

To our knowledge, this work is the first to directly demonstrate that microbial composition influences a community’s functional response to environmental perturbations.

The results also confirm previous microcosm experiments finding that microbial composition influences decomposition (Strickland et al. 2009; others) and generally, ecosystem processes (for example, Reed and Martiny 2013). More work is needed, however, to investigate the details of the relationships between microbial composition, ecosystem processes, and environmental change. Previous work has shown that the strength of microbial composition–functioning relationships depends on the particular process (Schimel

1995, Balser and Firestone 2005). Here, composition was more important in explaining respiration rates when nitrogen availability varied than when temperature or moisture did.

Thus, it may also be useful to consider the sensitivity of composition-functioning relationships under different environmental changes, whether for microorganisms or plants and animals.

Acknowledgements

We would like to thank S. Allison and D. Campbell for advice on the design and analysis of the experiment, A. Amend for isolation of the fungal taxa, C. Weihe for assistance sequencing the communities, S. Nguyen for collecting measurements of CO2 production, and two anonymous reviewers for comments that improved the clarity of the manuscript. Funding for this project provided by the Office of Science (BER) US

Department of Energy (program in Microbial Communities and Carbon Cycling); the NSF

Major Research Instrumentation program (1126749); and the US Department of Education.

Tables and Figures

Table 2.1: ANOVA and PERMANOVA results for each treatment. Overall analysis of variance statistics for factors affecting final community richness and average daily respiration rates and PERMANOVA statistics and proportional variation (measured as the sizes of the variance components) explained for factors affecting final bacterial and fungal community structure.

Final Community Bacterial Community Fungal Community Average Daily Richness Composition Composition Respiration Proportional Proportional Source of variation df Chisq P df P Variation df P Variation df Chisq P All Treatments Community 1 10.4 <0.001 7 <0.001 0.89 7 <0.001 0.85 1 6.5 0.010 Environment 3 18.5 <0.001 3 0.030 0.01 3 0.556 NS 3 38.4 <0.001 Community*Environment 1 1.4 0.244 21 <0.001 0.06 21 <0.001 0.11 1 38.8 <0.001

52 Residual 0.05 0.04

Reduced Moisture Community 1 11.9 <0.001 7 <0.001 0.94 7 <0.001 0.94 1 0.7 0.416 Environment 1 6.9 0.009 1 0.116 NS 1 0.63 NS 1 28.6 <0.001 Community*Environment 1 0.0 0.999 7 0.003 0.03 7 <0.001 0.02 1 7.3 0.007 Residual 0.03 0.04 Elevated Nitrogen Community 1 0.2 0.630 7 <0.001 0.83 7 <0.001 0.78 1 0.0 1.000 Environment 1 9.9 0.002 1 0.180 NS 1 0.453 NS 1 16.5 <0.001 Community*Environment 1 6.0 0.010 7 <0.001 0.12 7 <0.001 0.18 1 29.4 <0.001 Residual 0.05 0.03 Elevated Temperature Community 1 8.0 0.005 7 <0.001 0.95 7 <0.001 0.92 1 1.2 0.280 Environment 1 12.2 <0.001 1 0.274 NS 1 0.129 NS 1 15.7 <0.001 Community*Environment 1 0.0 1.000 7 <0.001 0.02 7 <0.001 0.04 1 3.4 0.070 Residual 0.03 0.04

Table 2.2: Mean variance and Bartlett's K2 test statistic across treatments. The variance in mean difference of average respiration rate between control microcosms and microcosms undergoing each of the three environmental treatments, as well as Bartlett's K2 test statistic for homogeneity of variances.

Mean Difference Bartlett's Treatment Variance K-squared Probability Reduced Moisture 1.38 0.01 0.926 Elevated Nitrogen 6.65 5.99 0.014 Elevated Temperature 1.49 2.92 0.087

Note: A smaller K2 value corresponds to more homogenous variances between groups.

G9H !"#$%&'(()#*+*,

E D B !

GIH !"#$%&'(()#*+*,

-./%&.0 1'234'2#5.6(%3&' 70'89%'2#:6%&.;'/ 70'89%'2#<'=>'&9%3&' ?/.4303=

Figure 2.1: NMDS ordination of bacterial and fungal composition. Non-metric multidimensional scaling (NMDS) ordination based on Bray Curtis similarities depicting bacterial (a) and fungal (b) community composition. The samples inoculated with the same initial microbial community are displayed in the same symbol color (8 communities total). The shape of the symbols represent the environmental treatment, either control conditions (filled squares), reduced moisture (open squares), elevated nitrogen (diamonds), and increased temperature (triangles). The initial microbial inoculum (stars) is also plotted but excluded from analyses. N=3 per community and treatment combination.

Figure 2.2: Respiration for each environmental treatment. Daily microbial respiration rates for all microbial communities averaged over 60 days for each environmental treatment. Blue boxes represent microcosms in control conditions, yellow in reduced moisture, green in elevated nitrogen, and red in elevated temperature. N=48 for each treatment. Different letters represent significant (p<0.05) differences between treatments, based on a Tukey post-hoc test.

Community 1 Community 2 Community 3 Community 4 * 1.2 ) -1 1.0 * * * * * day 0.8 -1 0.6 0.4 Community 5 Community 6 Community 7 Community 8

1.2 * * 1.0 * * * 0.8 Respiration (mg C g 0.6 0.4

Figure 2.3: Respiration for each of the eight microbial communities. Average daily respiration for each treatment by the eight microbial communities. Blue boxes represent microcosms in control conditions, yellow in reduced moisture, green in elevated nitrogen, and red in elevated temperature. Inoculations consisted of 1 of 8 unique microbial communities. N=6 for each box. Asterisks significant (p<0.05) differences between treatments, based on a Tukey post-hoc test.

Figure 2.4: Correlation of community and functional responses for each environmental treatment. Each point represents the percent of taxa in a given community that had significantly different relative abundances when exposed to a perturbation treatment and the corresponding percent change in average respiration rate. The percent of significant taxa responses to the environmental treatments were moderately correlated with the corresponding change in average community respiration (R2=0.16, p=0.05), but this was largely driven by the elevated nitrogen treatment (R2=0.22, p=0.24). Each point is labeled by community (1-8) and represents the average of a given community-environment combination.

Supporting Information

This chapter contains supporting information that can be found online at http://www.esapubs.org/archive/ecol/E096/016/.

Appendix A. Supplementary methods: 8-bp multiplex barcodes used to differentiate the 16S and 28S gene sequences across samples.

Appendix B. Supplementary results: microbial composition affects a model ecosystem’s functional response to environmental change

CHAPTER 3

Phylogenetic conservation of substrate use, but not temperature response, in leaf litter bacteria

Abstract

The ability to predict how ecosystems will respond to future environmental conditions is an active area of research. One component of natural ecosystems, the resident microbial community, will undoubtedly be affected by future changes in climate. Environmental change will also influence ecosystem processes regulated by microbial communities, including leaf litter decomposition. To assess how microbial communities and their ecosystem functions might respond to future changes in climate, like increases in temperature, we quantified the distribution of functional and response traits in leaf litter bacteria isolated from a natural grassland ecosystem in Southern California. These bacteria varied substantially in their carbon substrate use, as well as their response to changes in temperature. To better predict the functioning and responses in natural communities, we also examined if the functional and response traits were phylogenetically patterned or correlated with one another. We found that the distribution of functional traits displayed a phylogenetic pattern, but the sensitivity of the traits to changes in temperature did not. We also did not detect any correlations between carbon substrate use and sensitivity to changes in temperature. Together, these results suggest that information about microbial composition may provide insights to predicting ecosystem function under one set of conditions, but not necessarily relevant to predicting the response of ecosystem functions to environmental change.

Introduction

Over the next fifty years, southern California is expected to experience a range of environmental changes, including increased temperature, nitrogen deposition, and precipitation variability (CEC 2003). While much research has been devoted to predicting how natural plant communities will respond to these environmental changes, the response of resident microbial communities remains relatively unknown (but see Castro et al. 2010,

Sheik et al. 2011, Yuste et al. 2011). However, microbial communities play a central role in many ecosystem processes, including the terrestrial carbon cycle thru mediation of leaf litter decomposition rates. Decomposition is the largest source through which carbon is transferred from the biosphere to the atmosphere (Schlesinger and Andrews 2000) and is performed almost exclusively by fungi and bacteria (Swift et al. 1979). Therefore, determining how microbial communities will respond to changing environmental conditions may be key improving predictions of future ecosystem process rates.

For this study, we characterized traits of bacterial taxa isolated from natural leaf litter communities located in southern California (Potts et al. 2012). Bacteria dominate these communities (Allison et al. 2013), and their composition varies seasonally and in response to environmental change (Matulich et al. 2015). Furthermore, variation in microbial composition has been shown to impact the decomposition rate of leaf litter (Strickland et al.

2009, Allison et al. 2013), as well as the response of ecosystem process rates to environmental change (Matulich and Martiny 2015). Despite this knowledge, how and why the composition of a litter bacterial community influences decomposition rates remains relatively unknown.

One approach to linking microbial composition, ecosystem function, and environmental change is to identify traits that influence the structure and functioning of

microbial communities (Green et al. 2008, Litchman and Klausmeier 2008, Wallenstein and

Hall 2012, Fierer et al. 2014). Plant ecologists have proposed dividing traits into two types: response and effect traits (Lavorel and Garnier 2002). Response traits determine how an organism’s abundance changes in the face of new environmental conditions. For example, the physiological response of microorganisms to increased temperature was been well documented, and includes elevated rates of enzymatic reactions and respiration (reviewed in Bradford 2013), both of which influence decomposition rates. In contrast, effect traits

(hereafter “functional” traits as in Srivastava et al. 2012) are characteristics that influence ecosystem properties or processes such as nutrient cycling and trace gas emissions.

Together, knowledge about the response and functional traits of all species in a community may help predict how composition and functioning will shift in response to environmental change (Garnier et al. 2004, Allison and Martiny 2008, Suding et al. 2008, Williams et al.

2010, Diaz et al. 2013).

The phylogenetic relatedness of community members may also provide information about ecosystem functioning (reviewed in Cavender-Bares et al. 2009, Philippot et al.

2010). This is based on the idea that species more closely related should, due to sharing a recent common ancestor, possess a similar suite of traits. Studies on microbes (Langille et al. 2013) and plants (Cadotte et al. 2008) have shown the phylogenetic structure of a community is often correlated with functioning. In addition, a recent synthesis paper examining the phylogenetic conservatism of functional traits in bacteria and archea found that the vast majority of carbon-use traits are non-randomly distributed, although the depth at which these traits were conserved varied (Martiny et al. 2013). The sensitivity of microbes to environmental change could also be patterned by phylogeny, such that perturbed communities consist of more closely related species than would be expected by chance

(Evans and Wallenstein 2014). If this is the case for leaf litter microbial communities, the phylogenetic composition of a community might help estimate the ecosystem’s decomposition rate in a suite of environmental conditions.

Even if traits are not phylogenetically conserved, correlations between functional and response traits may also help predict future process rates. Some evidence suggests that traits often differ in their contribution to a functional process versus an environmental response, and thus are not likely correlated (Reich et al. 2001). Alternatively, certain traits like body size or traits related to niche breadth may be strongly related to both. For example, Larsen et al. (2005) found that female body size of bees influenced pollination efficiencies (function) and the probability of extinction due to agricultural intensification

(response), leading to a strong decline in pollination function due to agricultural intensification. Additionally, Lennon et al. (2012) observed that functional traits like niche breadth and optimum water potential contributed to respiration rates and also accurately predicted a bacteria’s response to decreased moisture.

In this study, we isolated 16 bacteria from a complex grassland leaf litter community.

To characterize the functional and response traits of the bacteria, we performed physiological and metabolic assays along a realistic temperature gradient ranging from just above the ecosystem’s mean to its summer highs. These assays yielded distributions of the functional traits at different temperatures (Figure 3.1). From these data, we also estimated the temperature response (Q10; response trait) of each functional trait for each strain.

We then tested if these traits, as measured by a variety of metrics, were phylogenetically conserved. We hypothesized that the temperature response of leaf litter bacteria would be more strongly phylogenetically conserved than that the functional, substrate use traits. The optimal temperature of a bacterium is presumably determined by a variety of genetic factors. Thus, we would expect this complex trait to be relatively deeply

conserved (Jain et al. 1999, Martiny et al. 2013), similar to moisture responses of soil bacteria (Lennon et al. 2012, Placella et al. 2012, Evans and Wallenstein 2014). In contrast, the use of particular carbon substrates is a simple trait that can be enabled by the presence of just one or two genes. Thus, the ability to use these substrates have likely evolved many times, either by lateral transfer or point mutation, and is therefore usually only finely conserved amongst bacteria. Finally, we examined whether the functional and response traits were correlated with one another. A correlation between these traits would suggest a tradeoff between substrate use and temperature sensitivity. Although we had no a priori reason to predict such a trade off, a correlation between these traits would aid in constraining predictions about how temperature change might alter bacterial decomposition rates.

Methods

Strain isolation and identification

The bacterial strains used in this study were isolated from leaf litter collected from the

Loma Ridge research site (Potts et al. 2012) located 5 km north of Irvine, California, USA

(33 44’N, 117 42’W, 365 m elevation) in January, April, July, and September of 2011.

During this period, the site was dominated by the annual grass genera Avena, Bromus, and

Lolium the annual forb genera Erodium and Lupinus and the native perennial grass

Nassella pulchra. The temperatures in this system are moderate, with an average winter

(November through April) high of 21.0°C and low of 6.2°C and an average summer (May through October) high of 27.3°C and low of 13.1°C; the mean annual temperature is 17°C.

Once collected, the litter was ground and then washed with sterilized deionized (DI) water. Litter fragments (106-212 µm) were suspended in sterile DI water and the resulting solution was passed through a 100 µm filter (EMD Millipore, Billerica MA), and 20 ml of the

filtered litter solution was mixed with 20 ml of 0.6% carboxymethylcellulose solution (an emulsifier). This slurry was then plated onto Luria Broth (LB) agar plates or litter agar plates and allowed to incubate at 21°C for up to two weeks (as described in Matulich and Martiny

2015). Individual colonies were transferred to fresh LB agar plates at least three times to purify the cultures. Prior to the experiments, the cultures were stored at -80°C in a glycerol solution (52% glycerol, 2% MgSO4, and 2% Tris-Cl; pH=7.6).

We identified each bacterial isolate by PCR amplification and Sanger sequencing of the 16S rRNA gene. DNA extract from each isolate was added to a PCR cocktail containing

1.5 units of HotMasterTaq polymerase (5PRIME, Gaithersburg MD) 1x PreMixF (FailSafe),

0.3 µM of each primer (pA: 5'- AGAGTTTGATCCTGGCTCAG-3' and pH: 5'-

AAGGAGGTGATCCAGCCGCA-3'; Edwards et al. 1989), and H2O to a final volume of 31

µl. Following an initial denaturation step at 95°C for 4 min, PCR was cycled 30 times at

95°C for 40 sec, 55.5°C for 30 sec, 72°C for 2 min, and a final extension at 72°C for 3 min

30 sec. All PCRs here and below were performed on a PTC-100 thermocycler (Bio-Rad,

Hercules CA). The taxonomic identity of each isolate was determined by matching our sequences to the most closely related cultured representative using the BLAST tool in the

GenBank database (Figure 3.6). We aligned the 16S rRNA sequences using the SINA aligner (www.arb- silva.de; Pruesse et al. 2012). A maximum likelihood tree was estimated with 100 bootstrap replications using a transition/transversion ratio = 2, a constant base rate variation among sites, and empirical base frequencies using PHYLIP v. 3.68 (Felsenstein

2006).

Simple carbon use assay

Approximately 10 days before each assay, glycerol stocks of 16 bacterial strains were removed from -80°C and streaked onto LB agar plates. Plates were incubated at room

temperatures (roughly 23°C) for approximately 6 days or until single colonies appeared.

Single colonies were picked from each culture and inoculated into 25 ml of LB media. To allow for temperature acclimation, liquid cultures were grown with agitation (100 rpm) at each experimental temperature (18°C, 22°C, or 26°C) for 5 days and transferred at least once to new media. On the day of inoculation, each culture was transferred to a 50 ml conical centrifuge tube and centrifuged at 4500 rpm for 10 minutes. The supernatant was discarded and the pellet was re-suspended in sterile saline solution (0.9% NaCl). This washing process was repeated a total of four times to remove residual LB media.

After washing, cultures were diluted in saline solution to achieve a cell density with an optical density at 600 nm (OD600) of 0.100 ± 0.05, and 100 µl of the diluted cell suspension was inoculated into each well of an Ecoplate (Biolog, Hayward CA). Ecoplates contain 31 unique carbon substrates in triplicate plus 3 carbon-free controls controls in a 96 well-plate format. Through substrate consumption and consequent cellular respiration, an indicator dye is reduced and a purple pigment is produced. After inoculation, plates were incubated at the same temperature used during acclimation (18°C, 22°C, or 26°C) for ten days. Substrate utilization (OD590) and bacterial growth (turbidity; OD750) were assessed at time 0, and 1, 2, 3, 5, 7, and 10 days using an automated Synergy 4 (BioTek, Winooski VT) microplate reader (Figure 3.2a). A substrate was considered utilized if the OD590 (mean based on three replicates) after 10 days was ≥0.30, which exceeded the maximum OD590 of the substrate-free controls. Sanguibacter sp. was removed from all Ecoplate analyses due to the development of pigment in negative control wells.

To estimate growth parameters, we fit the substrate utilization data to a modified

Gompertz equation:

* $ µmaxe ' - y = Aexp+ "exp& (# " t) +1) . + y0 , % A ( /

65 !

where y is cell density (or in this case mean OD590 of a particular substrate), A is the carrying capacity, µmax is maximum growth rate, and λ is the lag time between inoculation and the initiation of exponential growth, and y0 is the initial inoculation density (Zwietering et al. 1990).

The response bacterial growth to temperature was determined for every utilized substrate by calculating the Q10, the factor by which a biological process increases in rate for every 10°C change in temperature, based on µmax values:

# 10 & % ( # µ &$ 26°C"18°C' Q 18°C 10"µ max = % ( $ µ26°C ' where µ18°C and µ26°C are the estimated µmax for the coldest (18°C) and warmest (26°C) incubation temperatures.! To determine the sensitivity of substrate yield, Q10-OD590 values were calculated using the mean final OD590 value after 10 days.

To test if the similarity of substrate use activity and temperature sensitivity was correlated with phylogenetic distance, we performed Mantel tests based on 999 permutations. Specifically, we compared the bacterial phylogenetic distance matrix with euclidean distance matrices of substrate usage (substrate use richness, mean OD590, or mean µmax) or temperature sensitivity (Q10-µmax or Q10-OD590). The mean OD590 and mean µmax represented the mean value across all three temperature treatments.

We assessed substrate use efficiency of each strain by comparing the residuals of a linear correlation between microbial turbidity (OD750) and substrate use (OD590), where positive or negative residuals denote a relative increase or decrease in substrate use efficiency. To try to remove any effect of initial inoculum density, we used data collected on the final sampling day. To determine if substrate use efficiency varied by temperature, we performed an ANCOVA.

Complex litter use experiment

Leaf litter (comprised of roughly 12% lignin, 48% cellulose, 33% hemmicellulose, 4% crude protein, and 3% ethanol soluble carbohydrates; (Baker and Allison 2015)) from the

Loma Ridge site was collected in the summer of 2010, ground in a Wiley mill to pass a #20 mesh screen, and sterilized with roughly 23 kGy of gamma irradiation and autoclaving for

30 minutes. Sterilized litter (50 mg) was then added to each microcosm, an autoclaved glass vial (40 ml) with a sampling septum containing 4 g sterile sand. One culture (100 µl of the diluted cell suspension with 700 µl of saline solution) was slowly pipetted into each microcosm. Microcosms were placed in large plastic containers to keep humidity near 100% and stored at either 18°C, 22°C, or 26°C. Each strain (n=16) and temperature (n=3) combination contained three replicate microcosms for a total of 144 microcosms.

CO2 concentrations in the microcosm headspace were measured on days 1, 2, 3, 6,

8, and 10 to estimate cumulative CO2 production, our metric for litter decomposition (Figure

3.2b). To keep the microcosms at atmospheric CO2 concentrations, the vial caps were kept loose except for 24 hours before each sampling. For each measurement, an 8 ml subsample of headspace gas was withdrawn by syringe and injected into an infrared gas analyzer (PP-Systems EGM-4). To determine the temperature responses of leaf litter decomposition, we compared the time required for each strain to produce 5000 ppm of CO2 at 18°C, 22°C, and 26°C. Following an approach by Conant et al. (2008), we estimated the

Q10 of litter decomposition by dividing the time taken to respire 5000 ppm of CO2 at the coldest temperature (t18°C) by time taken at the warmest temperature (t26°C) and correcting for the actual incubation temperature differential (26°C-18°C):

" 10 % $ ' " %# 26°C(18°C& t18°C Q10 = $ ' # t 26°C &

! 67

Additional statistical analyses

Because the main effect of temperature on substrate usage and litter decomposition was non-linear, we used a mixed model ANOVA (“lme4” package) to analyze the effect of strain, temperature, their interaction, and substrate (when applicable) on respiration, substrate use richness, OD590, and µmax. Strain and substrate were treated as random factors and temperature as a fixed, categorical factor. We assessed the phylogenetic signal of functional and response traits among bacterial strains using 999 permutations of

Bloomberg’s K statistic as encoded in the R package ‘picante’. The input tree for all phylogenetic analyses was a maximum likelihood tree estimated from original alignment using the majority consensus tree of 100 bootstrap runs as a topological guide.

Results

Substrate use traits

Grassland litter bacteria varied greatly in their functional traits, as measured by their use of simple carbon substrates and complex leaf litter (Figure 3.3). We first compared the ability of each isolate to use the 31 substrates on the Ecoplates (Table 3.1). All but two substrates, 2-hydroxy benzonic acid and phenylethylamine, were used by at least one strain. At the intermediate temperature (22°C), the strains used an average of 11.6 substrates. The breadth of this use ranged greatly, between 0 and 24 substrates depending on the strain (Figure 3.S1). Two Actinobacteria strains, Schumannella luteola and a

Frigoribacterium sp., did not use any of the substrates (Figure 3.S1). The most commonly utilized substrates were the carboxylic acids pyruvic acid methyl ester, D-galacturonic acid, and glycyl-L-glutamic acid (Table 3.2).

Beyond binary use, we assayed the degree to which the strains could use each substrate by estimating growth rate (µmax) and comparing a metric of relative yield (OD590 on

day 10). Across all substrates, the strains differed greatly in both their growth rates and relative yields (p<0.001; Table 3.1; Figure 3.3a). For instance, the ability of Pseudomonas sp. 1 to utilize pyruvic acid methyl ester at 26°C was 10x faster than Pseudomonas synxntha (µmax = 4.67 and 0.467, respectively).

We further quantified substrate use efficiency (SUE), by comparing substrate utilization versus cell abundance. In contrast to the other trait measurements, SUE did not significantly vary across the litter bacteria, although it did vary by substrate (p<0.001; Table

3.1; Figure 3.3b).

Similar to the simple substrates, the litter bacteria differed greatly in their abilities to decompose complex leaf litter (Table 3.1). Decomposition, measured as cumulative CO2 production, varied significantly across bacterial taxa; strains ranged from producing ~5,500-

23,500 ppm CO2 during the 10 period (Figures 3.3c, 3.4).

Response to temperature

Generally, substrate use by the isolates depended on the temperature of the assay.

The median number of simple substrates used by each strain was highest at the intermediate temperature 22°C (n=13) and declined to 10 substrates at 18°C and 9 substrates at 26°C (Figure 3.S2). Similarly, mean yield (OD590) and growth rate (µmax) on the

Ecoplates peaked at 22°C. In contrast, SUE significantly increased across the three temperatures (Figure 3.5). Finally, decomposition rate was highest at 26°C, but the mean rate at 22°C tended to be lower than at 18°C (Figure 3.4).

In addition to these broad trends in temperature response, the litter bacteria varied in their responses to temperature, whether growing on simple or complex substrates. Indeed, strain and temperature significantly interacted to affect the mean yield (OD590) and SUE

(mixed-model ANOVA; Table 3.1) on the Ecoplates. Similarly, we observed a significant strain and temperature interaction on CO2 mineralization rates (Table 3.1; Figure 3.S3).

Phylogenetic signal of substrate use

The overall substrate use profile, defined as the suite of simple substrates used by each strain, was correlated with phylogenetic distance (Mantel Test: rM=0.23, p=0.32). This was also the case when we examined the µmax values for the suite of substrates used by each strain (Mantel Test: rM=0.20, p=0.74; Table 2). In addition, the average µmax across all utilized substrates was also strongly conserved among strains (K-statistic: p=0.002), but substrate richness was not (Figure 3.6). The ability to use particular substrates also displayed a phylogenetic pattern, and their patterns were consistent if we examined each temperature treatment individually (data not shown) or took the average of all three temperatures (Table 3.2). For example, mannitol was metabolized by strains in all three phyla but the maximum potential growth rate (umax) was generally higher in strains belonging to the Proteobacteria phyla (Figure 3.S4). Alternatively, α-cyclodextrin was only utilized by two stains, Pedobacter borealis and Chryseobacterium sp., both of which belonged to the phylum Bacteriodetes.

Average substrate use efficiency (SUE) of a strain (averaged across temperatures and substrates) was also phylogenetically conserved (K-statistic: p=0.02; Figure 3.6).

Interestingly, this pattern disappeared when we examined each temperature individually.

The SUE of two individual substrates, D-galacturonic acid andα-Cyclodextrin, also displayed a displayed a significant phylogenetic pattern (Table 3.2). For example, the D- galacturonic acid was utilized most efficiently by the two Duganella zoogloeoides strains

(Figure 3.S4).

The ability to utilize the complex leaf litter also displayed a significant phylogenetic pattern (Figure 3.6), and this signal remained whether we examined each temperature individually (data not shown) or averaged data together (K-statistic: p=0.001). On average,

Proteobacteria produced more CO2 on the leaf litter than Bacteriodetes and Actinobacteria.

Phylogenetic signal of temperature response

Unlike functional traits, we found little evidence of a phylogenetic pattern in the sensitivity of taxa to changes in temperature (Figure 3.6, Table 3.2). This was the case when we examined the Q10 of mineralization rates (K-statistic: p=0.80; Figure 3.6), and the average µmax or OD590 across all utilized substrates (K-statistic: p=0.52 and 0.74, respectively). However, a few individual substrates did have Q10 values (based on both µmax and OD590) that were phylogenetically conserved (Table 3.2). These significant correlations did not appear to be related to substrate type (polymer, amino acid, etc.).

Correlations between substrate use & temperature response

The functional and response traits of the litter bacteria were generally not correlated.

A taxon’s average µmax tended to be negatively related with Q10-µmax, but this pattern was not statistically significant (R2=0.16, p=0.09; Figure 3.7a). In addition, we found no correlation between a strain’s average mineralization rate and it’s sensitivity to temperature (Q10;

Figure 3.7b).

Discussion

The collection of co-occurring leaf litter bacteria varied substantially in their abilities to use a variety of carbon substrates. While this variation has been previously documented

(as seen in Romani et al. 2006, Allison et al. 2009, Lebauer 2010), there are surprisingly few studies that assay such parameters for strains isolated from the same community.

Although isolation most certainly provides a biased picture, these data reveal an estimate of the distribution of functional traits in a particular community and the breadth of occupied niche space (Lennon et al. 2012). Such data can also be used to parameterize trait-based models that predict how microbial communities, and their ecosystem functions, will respond

to changing environmental conditions (e.g., Follows and Dutkiewicz 2011, Allison 2012,

Bouskill et al. 2012, Scheiter et al. 2013).

The simple substrate assays revealed a large range in potential substrate use among leaf litter bacteria, with some taxa appearing to specialize on a small number of substrates, while others metabolize a higher number. The use of the Ecoplate system to analyze microbial traits is not without its weaknesses. Indeed, the simple-carbon substrates do not cover the whole diversity of potential substrates that can be found in leaf litter, and microbial activity in the Ecoplate may not be identical to that in nature. Despite these weaknesses, we observed that strains that utilized more Ecoplate substrates also had

2 higher rates of CO2 production on complex leaf litter (R =0.08, p=0.05).

Changes in temperature also influenced the bacteria’s ability to utilize simple and complex carbon substrates as previously observed (reviewed in Bradford (2013). For example, umax and substrate use richness of the Chryseobacterium sp. strain (family

Flavobacteriaceae) increased with temperature. Indeed, earlier studies demonstrated that many strains of Chryseobacterium can grow at temperatures as high as 37°C (Vandamme et al. 1994). On the other hand, the SUE and substrate use richness of both Curtobacterium strains (family Microbacteriaceae) decreased with temperature.

While warmer temperatures generally increase decomposition rates through a combination of increased enzymatic activity, metabolism, and growth (increased abundance) (Ratkowsky et al. 1982, Lloyd and Taylor 1994, Sheik et al. 2011), we observed mixed responses to increasing temperatures. In general, the usage of simple carbon substrates (via Ecoplates) decreased with temperature, although the Ecoplate substrate use efficiency and the decomposition of complex leaf litter mildly increased with temperature.

Unlike many other studies (Ratkowsky et al. 1982, Fierer et al. 2005, Bradford et al. 2010), we measured temperature response using a relatively narrow temperatures near the

average high of the natural environment (Kimball et al. 2014). Given that much of the decomposition that occurs in this system take place during the cool, rainy season when temperatures are low and moisture/relative humidity is at its highest (Allison et al. 2013), it is therefore possible the temperatures used in this study exceed the temperature optimum of these strains. If this is the case, determining the range of temperatures that are biologically relevant to microbial communities, and their functions, will be essential to assessing ecosystem responses to environmental change.

The ability of the leaf litter taxa to utilize carbon substrates was a least partially phylogenetically conserved. This was the case whether we analyzed simple carbon substrates (via Ecoplates) or natural, more complex, leaf litter. For example, strains belonging to the phyla Proteobacteria, on average, utilized more substrates and had higher decomposition rates that strains belonging to the phyla Actinobacteria and Bacteriodetes.

These Proteobacteria (class γ-Protebacteria and β-Proteobacteria) are often characterized as fast growers that target lower molecular weight (more labile) compounds (Elsas et al.

2007, Treseder et al. 2011). As labile substrates are the first targets of decomposition, it's likely that the C mineralized in our assay originated from these low molecular weight compounds. Temperature response did not appear to be phylogenetically conserved. This is in contrast to previous studies, which have observed environmental conditions like light and moisture availability to result in the phylogenetic, clustering of microbial communities

(e.g., Moore et al. 1998, Evans and Wallenstein 2014). In this system in particular, we previously observed significantly conserved responses to drought and nitrogen addition in the field, at least among closely related taxa (<0.4% similar in 16S sequence)(Amend et al, in revision). However, this example also raises the possibility that the temperature response is similarly conserved at a fine genetic scale. In this case, our selection of 16 strains across three phyla reduced our ability to detect a phylogenetic signal at finer genetic scales (within

phyla), as well as limits our ability to pinpoint the depth of conservation (e.g., using metric such as consenTRAIT; Martiny et al. 2013).

We did not detect a correlation between the functional traits and temperature response of these leaf litter bacteria. While many of an organism’s traits are expected to be correlated (for instance, traits describing a plant’s leaves; Wright et al. 2004), the mechanism driving their linkage varies. For example, trait correlations could be due to biological or physical tradeoffs (e.g., size and nutrient uptake in phytoplankton; Litchman and Klausmeier 2008). In addition, coupling of genetic mechanisms could also be produced when the same genes influence multiple traits (i.e. pleiotropy).

Although accompanied by recognized biases, culture-based studies offer a path for characterizing the distribution of traits in a microbial community and its response to changing environmental conditions. These microbial trait distributions can also be incorporated into models that predict how climate change will influence ecosystem functioning (Todd-Brown et al. 2012, Wieder et al. 2013, Hararuk et al. 2014). As this area of research grows, future studies should determine whether the patterns observed here

(e.g., is substrate use phylogenetically correlated) apply to other microbial communities

(residing on leaf litter or other environments). Further work should also evaluate if the response of particular taxa to environmental change varies when present in different ecosystems. Overall, these types of studies will improve our understanding of how the functional and response traits are distributed in a microbial community, and may help predict how composition, and microbial-controlled ecosystem functions, will shift in response to future environmental change.

Acknowledgements

This work was funded by the NSF Major Research Instrumentation program

(1126749); the US Department of Education GAANN program; and the U.S. Department of

Energy, Office of Science, Office of Biological and Environmental Research (BER), under

Award Number DE-PS02-09ER09-25 (A.C.M., S.D.A, M.L.G., K.K.T., and J.B.H.M). We thank Steve Allison for providing valuable feedback and support, Steven Nguyen for assistance in developing the Ecoplate protocols, and Claudia Weihe for aiding in data collection.

Tables and Figures

Table 3.1: ANOVA results for functional traits. Overall mixed-model ANOVA results for factors affecting substrate yield (OD590), growth rate (µmax), substrate use efficiency (SUE), and CO2 production rates. Substrate and Strain terms are treated as random factors, while the Temperature term is treated as a fixed factor.

OD590 µmax SUE CO2 production

Source of variation df Chisq P df Chisq P df Chisq P df Chisq P Substrate 1 162.2 <0.001 1 85.6 <0.001 1 110.8 <0.001 NA NA NA Strain 1 33.6 <0.001 1 21.1 <0.001 1 1.3 0.3 1 15.3 <0.001 Temperature 2 12.7 0.002 2 13.6 0.001 2 7.9 0.02 1 10.1 0.006 Strain*Temperature 1 7.9 0.005 1 0.4 0.6 1 24.3 <0.001 1 7.1 0.008

Table 3.2: Results for each Biolog Ecoplate substrate, including the number of taxa that utilized each substrate. Also reported are the p-values from Bloomberg's K-statistic testing for a phylogenetic signal of average Ecoplate yield (OD590), growth rate (µmax), substrate use efficiency (SUE), and temperature sensitivities of growth rate (Q10-µmax) and yield (Q10-OD590).

Taxa Biolog Ecoplate Substrate Richness SUE OD590 µmax Q10-µmax Q10-OD590 Amine Phenylethylamine 0 NA NA NA NA NA Putrescine 6 0.28 0.19 0.28 0.83 0.84 Amine Acid L-Arginine 8 0.13 0.17 0.09 0.57 0.60 L-Asparagine 9 0.23 0.10 0.01 0.84 0.84 L-Phenylalanine 1 0.10 0.14 0.09 0.10 0.11 L-Serine 9 0.49 0.01 0.03 0.79 0.11 L-Threonine 7 0.09 0.09 0.36 0.79 0.14

77 Carbohydrate D-Cellobiose 11 0.32 0.29 0.08 0.66 0.66 D-Galactonic Acid γ-Lactone 8 0.22 0.07 0.35 0.83 0.76 D-Mannitol 9 0.42 0.00 0.01 0.78 0.76 D-Xylose 11 0.23 0.37 0.47 0.75 0.31 D,L-α-Glycerol Phosphate 8 0.13 0.02 0.03 0.69 0.74 Glucose-1-Phosphate 9 0.27 0.21 0.24 0.86 0.85 i-Erythritol 5 0.44 0.36 0.41 0.28 0.21 N-Acetyl-D-Glucosamine 8 0.29 0.05 0.25 0.80 0.78 α-D-Lactose 7 0.46 0.82 0.13 0.85 0.13 β-Methyl-D-Glucoside 6 0.62 0.21 0.11 0.04 0.41 Carboxylic Acid D-Galacturonic Acid 12 0.01 0.36 0.02 0.84 0.77 D-Glucosaminic Acid 6 0.03 0.11 0.31 0.07 0.31 D-Malic Acid 8 0.32 0.42 0.44 0.90 0.90 Glycyl-L-Glutamic Acid 12 0.45 0.02 0.08 0.81 0.80

Itaconic Acid 7 0.56 0.49 0.52 0.50 0.65 Pyruvic Acid Methyl Ester 12 0.07 0.29 0.09 0.73 0.73 α-Ketobutyric Acid 1 0.10 0.03 0.10 0.10 0.11 γ-Hydroxybutyric Acid 4 0.21 0.33 0.49 0.52 0.47 Phenolic Compound 2-Hydroxy Benzonic Acid 0 NA NA NA NA NA 4-Hydroxy Benzonic Acid 5 0.95 0.29 0.31 0.89 0.88 Polymer Glycogen 5 0.98 0.79 0.76 0.24 0.04 Tween 40 11 0.16 0.30 0.09 0.60 0.77 Tween 80 8 0.28 0.25 0.62 0.86 0.89 α-Cyclodextrin 2 0.05 0.00 0.01 0.04 0.03

!$/.&0/$%;*")#%1@AB2

+)/#*)95#)&0%&' 350,#)&0":%;*")#/ @ABA

@AB< 3*$45$0,6%17%8#*")0/2

+AC>? +<=>? !"#$%&'%()##$*%+$,&-.&/)#)&0%1+2

Figure 3.1: Theoretical distribution of functional and response traits. Theoretical distribution of one functional trait, litter decomposition rate (D), and one response trait, temperature sensitivity (Q10), for 16 bacterial taxa. Also illustrated is a correlation between functional and response traits, such that better litter degraders at 18°C are, on average, less sensitive to changes in temperature.

(b) (a)

590 6000 ) -1

2.0 4000

Temperature 1.0 2000 26oC 22oC production (ppm day 18oC 2 N-Acetyl-D-Glucosamine OD CO 0 0 1 2 3 5 7 10 0 1 2 3 6 8 Day Day

Figure 3.2: Sample Ecoplate and CO2 production data. The effect of temperature on Erwinia billingiae’s (a) ability to utilize the Biolog substrate N-acetyl-D-glucosamine and (b) mineralize leaf litter C. Each value plotted is mean (± SEM) based on three replicates.

3#4 3&4 304

1 :)*;!*102+3<+%()#/1'4

í í í 56

! %!&'()#(*+,'*+-../0/*102+3%,-4 78 3URGXFWLRQ SSP[94 "#$

Figure 3.3: Distribution of functional traits. The distribution of three functional traits for all bacterial taxa: (a) maximum potential growth rate (µmax) on Ecoplate substrates, (b) substrate use efficiency (SUE) on Ecoplate substrates, and (c) CO2 production from the mineralization of leaf litter C. All functional traits are based on the mean value across all temperatures and, when applicable, Ecoplate substrates.

18OC

0 22OC

0 Frequency (# Strains) 26OC

5 10 15 20 25

3 CO2 Production (ppm x 10 )

Figure 3.4: Distribution of CO2 production trait. The distribution of one functional trait, CO2 mineralization rate, across each temperature treatment. Each box represents the cumulative CO2 production rate for 1 bacterial strain (16 total). Stains are shaded according to their mineralization rate at 18°C, where light and dark shaded boxes are relatively low and high producers at 18°C, respectively. Vertical, dashed line indicate the mean CO2 production each temperature treatment.

0.9

0.6 urbidity) T ( 750 OD Temperature 0.3 18OC

22OC

26OC

1 2 3

OD590 (Substrate Usage)

Figure 3.5: Substrate yield and bacterial biomass correlation. Correlation of substrate yield (OD590) and bacterial biomass (well turbidity; OD750) for each of the three environmental treatments after 10 days of incubation. Each point represents the average substrate yield (OD590) and well turbidity (OD750) of a utilized substrate (final OD590>0.30) based on three replicates. More positive slopes indicate higher substrate use efficiency. The usage of a particular substrate was highly correlated with corresponding changes in well turbidity (R2=0.84, p<0.001). In general, substrate use efficiency increased with temperature (ANCOVA: p=0.002).

6>1234!6789:;<:4!"JH023<:4( /012345!6789:;<:4!"=492>;<:>0?(

6789:;<:4 6IJ 71<5 !G*,N71<5 !AB4;HF?499 =492>;<:>0?

!"#$%&'&()"*+,$&-#".#("!"#$%&&&&'(

!"#$%&'&()"*"/(0(12)*"#$%&&&)'( ,.,&

!"#$%&'&()"*"34!*!"#$%&&&)+(

!"#$%&'&()"*"34!)!!"#$%&&&&,(

5-67(7)*87,,7(97)#!"#$%&&&)-(

:$9)(#,,)*;&&9,&#&7%#"!*!"#$%&&&&+(

:$9)(#,,)*;&&9,&#&7%#"!)!"#$%&&&&&(

<)(9$78).1#-*"34!"#$%&&&*'(

=-79&-78).1#-7$'*"34!"#$%&&&*)(

<.2$')((#,,)*,$1#&,)!"#$%&&&**(

>7.-&8).1#-7$'*"34!"#$%&&&*+(

?$-1&8).1#-7$'*"34!"#$%&&&*-(

?$-1&8).1#-7$'*+,)..$'+).7#("!"#$%&&&,'(

!#%&8).1#-*8&-#),7"*AB=CDDDEDF

=,)@&8).1#-7$'*"34*AB=CDDDEEF

?2-/"#&8).1#-7$'*"34*AB=CDDDEGF

AB4;B4!=492>;<:>0?!"221( @,,, '&&& *,@@@ *&,,, *-&&& *%@@@ ),,,,

D<5>171!20:4?:><3!C;0E:F!;<:4!"71<5( , ,.&& ,.@@ *., *.&& *.@@ )., 6789:;<:4!=>HF?499 , M ' *) *@ ), )M

6789:;<:4!I94!JKK>H>4?HL!"6IJ( N,.,@ N,.,M N,.,) , ,.,) ,.,M ,.,@

=42>;<:>0?!G*, , ,.&& ,.@@ *., *.&& *.@@ ).,

O>030C!G*,N71<5 , ,.&& ,.@@ *., *% && -,

Figure 3.6: Phylogenetic tree of bacterial strains with functional & response traits. Maximum likelihood pylogenetic tree of bacterial strains based on the majority consensus of 100 bootstrap runs using Phylip. Plotted next to each taxa is the summary of functional and response traits for simple and complex substrates. All functional traits are based on the mean value across all temperatures and, when applicable, Ecoplate substrates.

4/7 4>7 "&C

A3.,9.>/+,93-/ %! =+,-*.>/+,93-/ ?/+,93-.@9,9: "&B ) 7 10

"! $! "&%

"&#

# "! 8 F&!"BG1;F!&!H Response Trait (Q 89:;.*:9123/-,14<

"&!

!&' "&! "&' "!!!! "'!!! #!!!!

()*+,-.*/0123/-,14)5/67 ()*+,-.*/0123/-,14D)5)0/,-E9189:;-3/,-.*7

Figure 3.7: Functional and response trait correlation. Correlation between the functional traits (a) µmax and (b) cumulative CO2 production and each trait’s sensitivity to temperature (Q10). Symbols correspond that to that found in Figure 3.6.

!"#$%&'&()"*"+,!" !"#$%&'&()"*-.$&/#"0#(" 4$5)(#..)*6&&5.&#&3%#"!" !"#$%&'&()"*"=(>(2;)* 20 10

!"#$%&'&()"*"+,!# 4$5)(#..)*6&&5.&#&3%#"!# 7/83(3)*13..3(53)#

:0;$')((#..)*.$2#&.) 9$/2&1)02#/3$'*"+, :)(5$31)02#/*"+, @30/&1)02#/3$'*"+,* 20 10 NA

9,*-.)00$'-)03#(" 9;/="#&1)02#/3$'*"+, <.)?&1)02#/3$'*"+, !#%&1)02#/*1&/#).3"

Subsrtrate Use Richness 20 10

18 22 26 18 22 26 18 22 26 18 22 26 Temperature (oC)

Figure 3.S1:8Substrate use richness for each strain. The number of substrates utilized by the bacterial taxa across the three temperatures.

Substrate Richness 5

0 18 22 26

Temperature (oC)

Figure 3.S2:9Substrate use richness at each temperature. Substrate use richness for all bacterial taxa for each temperature treatment. N=15 for each temperature treatment. For each boxplot, the line within the box shows the median substrate use richness, endpoints of the box show the 25% and 75% quartile range, and whiskers show the 0% and 100% quartile range (lower and upper limits of the data, respectively).

!"#$%&'&()"*"+,*! !"#$%&'&()"*9/$&5#"3#(" -$.)(#//)*0&&./&#&1%#""! !"#$%&'&()"*"6(7(48)* 25000 15000 5000 !"#$%&'&()"*"+,*# -$.)(#//)*0&&./&#&1%#"*# <5=1(1)*21//1(.1)# ;51.&512)34#51$'*"+, 25000 (ppm)

2 15000 5000 >38$')((#//)*/$4#&/) :$54&2)34#51$'*"+, >)(.$12)34#5*"+, ?135&2)34#51$'*"+,* 25000 15000 5000

Cumulative CO :,*9/)33$'9)31#(" :856"#&2)34#51$'*"+, ;/)@&2)34#51$'*"+, !#%&2)34#5*2&5#)/1" 25000 15000 5000 18 22 26 18 22 26 18 22 26 18 22 26

Temperature (oC)

Figure 3.S3:10Mineralization rates for each bacterial strain. Cumulative CO2 production rates for each bacterial strain across three temperature treatments. N=3 for each box.

/01232456789:4!24:; B29:5583

<=> !6?2@ !A*,06?2@

!"#$%&'&()"*+,$&-#".#("!"#$%&&&&'(

!"#$%&'&()"*"/(0(12)*"#$%&&&)'(

!"#$%&'&()"*"34!*!"#$%&&&)+(

!"#$%&'&()"*"34!)!!"#$%&&&&,(

5-67(7)*87,,7(97)#!"#$%&&&)-(

:$9)(#,,)*;&&9,&#&7%#"!*!"#$%&&&&+(

:$9)(#,,)*;&&9,&#&7%#"!)!"#$%&&&&&(

<)(9$78).1#-*"34!"#$%&&&*'(

=-79&-78).1#-7$'*"34!"#$%&&&*)(

<.2$')((#,,)*,$1#&,)!"#$%&&&**(

>7.-&8).1#-7$'*"34!"#$%&&&*+(

?$-1&8).1#-7$'*"34!"#$%&&&*-(

?$-1&8).1#-7$'*+,)..$'+).7#("!"#$%&&&,'(

!#%&8).1#-*8&-#),7"*AB=CDDDEDF

=,)@&8).1#-7$'*"34*AB=CDDDEEF

?2-/"#&8).1#-7$'*"34*AB=CDDDEGF

,.,&

B2@:?6?!D85E95:23!178F5G!725E!"6?2@( , ,.&& ,.CC *., *.&& *.CC ).,

/01232456789:4!24:;!<=> 0,.*, 0,.,% 0,.,& , ,.,& ,.,% ,.*,

B299:583!A*,06?2@ , ,.&& ,.CC *., *.&& *.CC ).,

Figure 3.S4:11Phylogenetic tree of bacterial strains with Ecoplate substrate. Maximum likelihood pylogenetic tree of bacterial strains based on the majority consensus of 100 bootstrap runs using Phylip. Plotted next to each taxa are some functional and response traits based on individual Ecoplate substrates. All functional traits are based on the mean value across all temperatures.

REFERENCES

Allison, S. D. 2012. A trait-based approach for modelling microbial litter decomposition. Ecology Letters 15:1058-1070. Allison, S. D. and J. D. Jastrow. 2006. Activities of extracellular enzymes in physically isolated fractions of restored grassland soils. Soil Biology & Biochemistry 38:3245-3256. Allison, S. D., D. S. LeBauer, M. R. Ofrecio, R. Reyes, A. M. Ta, and T. M. Tran. 2009. Low levels of nitrogen addition stimulate decomposition by boreal forest fungi. Soil Biology & Biochemistry 41:293-302. Allison, S. D., Y. Lu, C. Weihe, M. L. Goulden, A. C. Martiny, K. K. Treseder, and J. B. Martiny. 2013. Microbial abundance and composition influence litter decomposition response to environmental change. Ecology 94:714-725. Allison, S. D. and J. B. H. Martiny. 2008. Resistance, resilience, and redundancy in microbial communities. Proceedings of the National Academy of Sciences of the United States of America 105:11512-11519. Alster, C. J., D. P. German, Y. Lu, and S. D. Allison. 2013. Microbial enzymatic responses to drought and to nitrogen addition in a southern California grassland. Soil Biology & Biochemistry 64:68-79. Anderson, M. J., R. N. Gorley, and K. R. Clarke. 2008. PERMANOVA+ for PRIMER: Guide to Software and Statistical Methods, Plymouth, UK. Arguez, A., I. Durre, S. Applequist, R. S. Vose, M. F. Squires, X. G. Yin, R. R. Heim, and T. W. Owen. 2012. NOAA's 1981-2010 U.S. Climate Normals an Overview. Bulletin of the American Meteorological Society 93:1687-1697. Baker, N. R. and S. D. Allison. 2015. Ultraviolet photodegradation facilitates microbial litter decomposition in a Mediterranean climate. Ecology. Balser, T. C. and M. K. Firestone. 2005. Linking microbial community composition and soil processes in a California annual grassland and mixed-conifer forest. Biogeochemistry 73:395-415. Bardgett, R. D., R. D. Lovell, P. J. Hobbs, and S. C. Jarvis. 1999. Seasonal changes in soil microbial communities along a fertility gradient of temperate grasslands. Soil Biology & Biochemistry 31:1021-1030. Barnard, R. L., C. A. Osborne, and M. K. Firestone. 2013. Responses of soil bacterial and fungal communities to extreme desiccation and rewetting. Isme Journal 7:2229-2241. Bell, C. W., D. T. Tissue, M. E. Loik, M. D. Wallenstein, V. Acosta-Martinez, R. A. Erickson, and J. C. Zak. 2014. Soil microbial and nutrient responses to 7 years of seasonally altered precipitation in a Chihuahuan Desert grassland. Glob Chang Biol 20:1657-1673. Bell, T., J. A. Newman, B. W. Silverman, S. L. Turner, and A. K. Lilley. 2005. The contribution of species richness and composition to bacterial services. Nature 436:1157-1160. Berlemont, R., S. D. Allison, C. Weihe, Y. Lu, E. L. Brodie, J. B. H. Martiny, and A. C. Martiny. 2014. Cellulolytic potential under environmental changes in microbial communities from grassland litter. Front Microbiol 5. Bouskill, N. J., J. Tang, W. J. Riley, and E. L. Brodie. 2012. Trait-based representation of biological nitrification: model development, testing, and predicted community composition. Front Microbiol 3:364. Bradford, M. A. 2013. Thermal adaptation of decomposer communities in warming soils. Front Microbiol 4. Bradford, M. A., B. W. Watts, and C. A. Davies. 2010. Thermal adaptation of heterotrophic soil respiration in laboratory microcosms. Global Change Biology 16:1576-1588.

Bragazza, L., A. Buttler, J. Habermacher, L. Brancaleoni, R. Gerdol, H. Fritze, P. Hanajík, R. Laiho, and D. Johnson. 2012. High nitrogen deposition alters the decomposition of bog plant litter and reduces carbon accumulation. Global Change Biology 18:1163-1172. Braid, M. D., L. M. Daniels, and C. L. Kitts. 2003. Removal of PCR inhibitors from soil DNA by chemical flocculation. Journal of Microbiological Methods 52:389-393. Bray, J. R. and J. T. Curtis. 1957. An Ordination of the Upland Forest Communities of Southern Wisconsin. Ecological Monographs 27:326-349. Cadotte, M. W., B. J. Cardinale, and T. H. Oakley. 2008. Evolutionary history and the effect of biodiversity on plant productivity. Proceedings of the National Academy of Sciences of the United States of America 105:17012-17017. Caporaso, J. G., J. Kuczynski, J. Stombaugh, K. Bittinger, F. D. Bushman, E. K. Costello, N. Fierer, A. G. Pena, J. K. Goodrich, J. I. Gordon, G. A. Huttley, S. T. Kelley, D. Knights, J. E. Koenig, R. E. Ley, C. A. Lozupone, D. McDonald, B. D. Muegge, M. Pirrung, J. Reeder, J. R. Sevinsky, P. J. Tumbaugh, W. A. Walters, J. Widmann, T. Yatsunenko, J. Zaneveld, and R. Knight. 2010. QIIME allows analysis of high-throughput community sequencing data. Nature Methods 7:335-336. Caporaso, J. G., C. L. Lauber, W. A. Walters, D. Berg-Lyons, J. Huntley, N. Fierer, S. M. Owens, J. Betley, L. Fraser, M. Bauer, N. Gormley, J. A. Gilbert, G. Smith, and R. Knight. 2012. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. Isme Journal 6:1621-1624. Cardinale, B. J., K. L. Matulich, D. U. Hooper, J. E. Byrnes, E. Duffy, L. Gamfeldt, P. Balvanera, M. I. O'Connor, and A. Gonzalez. 2011. The functional role of producer diversity in ecosystems. Am J Bot 98:572-592. Carpenter, S. R. 1996. Microcosm experiments have limited relevance for community and ecosystem ecology. Ecology 77:677-680. Castro, H. F., A. T. Classen, E. E. Austin, R. J. Norby, and C. W. Schadt. 2010. Soil Microbial Community Responses to Multiple Experimental Climate Change Drivers. Applied and Environmental Microbiology 76:999-1007. Cavender-Bares, J., K. H. Kozak, P. V. Fine, and S. W. Kembel. 2009. The merging of community ecology and phylogenetic biology. Ecol Lett 12:693-715. CEC. 2003. Global Climate Change and California: Potential Implications for Ecosystems, Health, and the Economy - Consultant Report. Publication# 500-03-058CF, California Energy Commission. Clarke, K. R. and R. M. Warwick. 2001. Change in marine communities: an approach to statistical analysis and interpretation. 2nd Edition edition, Plymouth, UK. Cleveland, C. C., S. C. Reed, A. B. Keller, D. R. Nemergut, S. P. O'Neill, R. Ostertag, and P. M. Vitousek. 2014. Litter quality versus soil microbial community controls over decomposition: a quantitative analysis. Oecologia 174:283-294. Conant, R. T., R. A. Drijber, M. L. Haddix, W. J. Parton, E. A. Paul, A. F. Plante, J. Six, and J. M. Steinweg. 2008. Sensitivity of organic matter decomposition to warming varies with its quality. Global Change Biology 14:868-877. Cotrufo, M. F., P. Ineson, and A. P. Rowland. 1994. Decomposition of Tree Leaf Litters Grown under Elevated Co2 - Effect of Litter Quality. Plant and Soil 163:121-130. Cregger, M. A., C. W. Schadt, N. G. McDowell, W. T. Pockman, and A. T. Classen. 2012. Response of the soil microbial community to changes in precipitation in a semiarid ecosystem. Appl Environ Microbiol 78:8587-8594. Cruz-Martinez, K., K. B. Suttle, E. L. Brodie, M. E. Power, G. L. Andersen, and J. F. Banfield. 2009. Despite strong seasonal responses, soil microbial consortia are more resilient to long-term changes in rainfall than overlying grassland. Isme Journal 3:738-744. de Vries, F. T., P. Manning, J. R. Tallowin, S. R. Mortimer, E. S. Pilgrim, K. A. Harrison, P. J. Hobbs, H. Quirk, B. Shipley, J. H. Cornelissen, J. Kattge, R. D. Bardgett, and N.

Johnson. 2012. Abiotic drivers and plant traits explain landscape-scale patterns in soil microbial communities. Ecol Lett. DeAngelis, K. M., E. L. Brodie, T. Z. DeSantis, G. L. Andersen, S. E. Lindow, and M. K. Firestone. 2009. Selective progressive response of soil microbial community to wild oat roots. Isme Journal 3:168-178. DeSantis, T. Z., P. Hugenholtz, N. Larsen, M. Rojas, E. L. Brodie, K. Keller, T. Huber, D. Dalevi, P. Hu, and G. L. Andersen. 2006. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Applied and Environmental Microbiology 72:5069-5072. Diaz, S., A. Purvis, J. H. C. Cornelissen, G. M. Mace, M. J. Donoghue, R. M. Ewers, P. Jordano, and W. D. Pearse. 2013. Functional traits, the phylogeny of function, and ecosystem service vulnerability. Ecol Evol 3:2958-2975. Edgar, R. C. 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26:2460-2461. Edwards, U., T. Rogall, H. Blocker, M. Emde, and E. C. Bottger. 1989. Isolation and Direct Complete Nucleotide Determination of Entire Genes - Characterization of a Gene Coding for 16s-Ribosomal Rna. Nucleic Acids Research 17:7843-7853. Elsas, J. D. v., J. K. Jansson, and J. T. Trevors. 2007. Modern soil microbiology. 2nd edition. CRC Press/Taylor & Francis, Boca Raton, FL. Evans, S. E. and M. D. Wallenstein. 2014. Climate change alters ecological strategies of soil bacteria. Ecology Letters 17:155-164. Felsenstein, J. 2006. PHYLIP (Phylogeny Inference Package). Fenn, M. E., J. S. Baron, E. B. Allen, H. M. Rueth, K. R. Nydick, L. Geiser, W. D. Bowman, J. O. Sickman, T. Meixner, D. W. Johnson, and P. Neitlich. 2003a. Ecological effects of nitrogen deposition in the western United States. Bioscience 53:404-420. Fenn, M. E., M. A. Poth, A. Bytnerowicz, J. O. Sickman, and B. K. Takemoto. 2003b. Effects of ozone, nitrogen deposition, and other stressors on montane ecosystems in the Sierra Nevada. Pages 111-155 in M. J. A. R. A. Andrzej Bytnerowitcz, editor. Developments in Environmental Science. Elsevier. Fierer, N., A. Barberan, and D. C. Laughlin. 2014. Seeing the forest for the genes: using rnetagenomics to infer the aggregated traits of microbial communities. Front Microbiol 5. Fierer, N., J. M. Craine, K. McLauchlan, and J. P. Schimel. 2005. Litter quality and the temperature sensitivity of decomposition. Ecology 86:320-326. Fierer, N., M. S. Strickland, D. Liptzin, M. A. Bradford, and C. C. Cleveland. 2009. Global patterns in belowground communities. Ecology Letters 12:1238-1249. Follows, M. J. and S. Dutkiewicz. 2011. Modeling diverse communities of marine microbes. Ann Rev Mar Sci 3:427-451. Gardes, M. and T. D. Bruns. 1993. ITS primers with enhanced specificity for basidiomycetes-- application to the identification of mycorrhizae and rusts. Mol Ecol 2:113-118. Garnier, E., J. Cortez, G. Billes, M. L. Navas, C. Roumet, M. Debussche, G. Laurent, A. Blanchard, D. Aubry, A. Bellmann, C. Neill, and J. P. Toussaint. 2004. Plant functional markers capture ecosystem properties during secondary succession. Ecology 85:2630- 2637. German, D. P., M. N. Weintraub, A. S. Grandy, C. L. Lauber, Z. L. Rinkes, and S. D. Allison. 2011. Optimization of hydrolytic and oxidative enzyme methods for ecosystem studies. Soil Biology & Biochemistry 43:1387-1397. Green, J. L., B. J. M. Bohannan, and R. J. Whitaker. 2008. Microbial biogeography: From taxonomy to traits. Science 320:1039-1043. Griffiths, B. S., K. Ritz, R. Wheatley, H. L. Kuan, B. Boag, S. Christensen, F. Ekelund, S. J. Sorensen, S. Muller, and J. Bloem. 2001. An examination of the biodiversity-ecosystem

function relationship in arable soil microbial communities. Soil Biology & Biochemistry 33:1713-1722. Gutknecht, J. L. M., C. B. Field, and T. C. Balser. 2012. Microbial communities and their responses to simulated global change fluctuate greatly over multiple years. Global Change Biology 18:2256-2269. Habekost, M., N. Eisenhauer, S. Scheu, S. Steinbeiss, A. Weigelt, and G. Gleixner. 2008. Seasonal changes in the soil microbial community in a grassland plant diversity gradient four years after establishment. Soil Biology & Biochemistry 40:2588-2595. Hararuk, O., M. J. Smith, and Y. Luo. 2014. Microbial models with data-driven parameters predict stronger soil carbon responses to climate change. Glob Chang Biol. Harris, R. F. 1981. Effect of water potential on microbial growth and activity. Pages 23-95 in J. F. Parr, W. R. Gardner, and L. F. Elliott, editors. Water Potential Relations in Soil Microbiology. ASA, Madison, WI. Hattenschwiler, S., A. V. Tiunov, and S. Scheu. 2005. Biodiversity and litter decomposition interrestrial ecosystems. Annual Review of Ecology Evolution and Systematics 36:191- 218. Hawksworth, D. L. 2001. The magnitude of fungal diversity: the 1.5 million species estimate revisited. Mycological Research 105:1422-1432. Hobbie, S. E. 1996. Temperature and plant species control over litter decomposition in Alaskan tundra. Ecological Monographs 66:503-522. IPCC. 2013. The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovern- mental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA. Isbell, F., P. B. Reich, D. Tilman, S. E. Hobbie, S. Polasky, and S. Binder. 2013. Nutrient enrichment, biodiversity loss, and consequent declines in ecosystem productivity. Proc Natl Acad Sci U S A 110:11911-11916. Jain, R., M. C. Rivera, and J. A. Lake. 1999. Horizontal gene transfer among genomes: The complexity hypothesis. Proceedings of the National Academy of Sciences of the United States of America 96:3801-3806. Kennedy, N., E. Brodie, J. Connolly, and N. Clipson. 2006. Seasonal influences on fungal community structure in unimproved and improved upland grassland soils. Canadian Journal of Microbiology 52:689-694. Khanduri, V. P., C. M. Sharma, and S. P. Singh. 2008. The effects of climate change on plant phenology. Environmentalist 28:143-147. Kim, M., W. S. Kim, B. M. Tripathi, and J. Adams. 2014. Distinct bacterial communities dominate tropical and temperate zone leaf litter. Microb Ecol 67:837-848. Kimball, S., M. L. Goulden, K. N. Suding, and S. Parker. 2014. Altered water and nitrogen input shifts succession in a southern California coastal sage community. Ecological Applications 24:1390-1404. Kominoski, J. S., T. J. Hoellein, J. J. Kelly, and C. M. Pringle. 2009. Does mixing litter of different qualities alter stream microbial diversity and functioning on individual litter species? Oikos 118:457-463. Lage, M. D., H. E. Reed, C. Weihe, C. M. Crain, and J. B. Martiny. 2010. Nitrogen and phosphorus enrichment alter the composition of ammonia-oxidizing bacteria in salt marsh sediments. ISME J 4:933-944. Lane, D. J. 1991. 16S/23S rRNA sequencing. Pages 115-175 in E. Stackebrandt and M. Goodfellow, editors. Nucleic acid techniques in bacterial systematics. John Wiley and Sons, Ney York, USA. Langille, M. G., J. Zaneveld, J. G. Caporaso, D. McDonald, D. Knights, J. A. Reyes, J. C. Clemente, D. E. Burkepile, R. L. Vega Thurber, R. Knight, R. G. Beiko, and C.

Huttenhower. 2013. Predictive functional profiling of microbial communities using 16S rRNA marker gene sequences. Nat Biotechnol 31:814-821. Larcher, W. 2000. Temperature stress and survival ability of Mediterranean sclerophyllous plants. Plant Biosystems 134:279-295. Larsen, T. H., N. M. Williams, and C. Kremen. 2005. Extinction order and altered community structure rapidly disrupt ecosystem functioning. Ecology Letters 8:538-547. Lavorel, S. and E. Garnier. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology 16:545-556. Lebauer, D. S. 2010. Litter degradation rate and beta-glucosidase activity increase with fungal diversity. Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 40:1076-1085. Lennon, J. T., Z. T. Aanderud, B. K. Lehmkuhl, and D. R. Schoolmaster. 2012. Mapping the niche space of soil microorganisms using taxonomy and traits. Ecology 93:1867-1879. Litchman, E. and C. A. Klausmeier. 2008. Trait-Based Community Ecology of Phytoplankton. Annual Review of Ecology Evolution and Systematics 39:615-639. Lloyd, J. and J. A. Taylor. 1994. On the Temperature-Dependence of Soil Respiration. Functional Ecology 8:315-323. Loranger-Merciris, G., L. Barthes, A. Gastine, and P. Leadley. 2006. Rapid effects of plant species diversity and identity on soil microbial communities in experimental grassland ecosystems. Soil Biology & Biochemistry 38:2336-2343. Martiny, A. C., K. Treseder, and G. Pusch. 2013. Phylogenetic conservatism of functional traits in microorganisms. ISME J 7:830-838. Martiny, J. B. H., B. J. M. Bohannan, J. H. Brown, R. K. Colwell, J. A. Fuhrman, J. L. Green, M. C. Horner-Devine, M. Kane, J. A. Krumins, C. R. Kuske, P. J. Morin, S. Naeem, L. Ovreas, A. L. Reysenbach, V. H. Smith, and J. T. Staley. 2006. Microbial biogeography: putting microorganisms on the map. Nature Reviews Microbiology 4:102-112. Matulich, K. L. and J. B. H. Martiny. 2015. Microbial composition alters the response of litter decomposition to environmental change. Ecology 96:154-163. Matulich, K. L., C. Weihe, S. D. Allison, A. S. Amend, R. Berlemont, M. L. Goulden, S. Kimball, A. C. Martiny, and J. B. H. Martiny. 2015. Temporal variation overshadows the response of leaf litter microbial communities to simulated global change. ISME J (in press). Mcclaugherty, C. A., J. Pastor, J. D. Aber, and J. M. Melillo. 1985. Forest Litter Decomposition in Relation to Soil-Nitrogen Dynamics and Litter Quality. Ecology 66:266-275. McGrady-Steed, J., P. M. Harris, and P. J. Morin. 1997. Biodiversity regulates ecosystem predictability. Nature 390:162-165. Mcguire, K. L. and K. K. Treseder. 2010. Microbial communities and their relevance for ecosystem models: Decomposition as a case study. Soil Biology & Biochemistry 42:529- 535. Menzel, A., T. H. Sparks, N. Estrella, E. Koch, A. Aasa, R. Ahas, K. Alm-Kubler, P. Bissolli, O. Braslavska, A. Briede, F. M. Chmielewski, Z. Crepinsek, Y. Curnel, A. Dahl, C. Defila, A. Donnelly, Y. Filella, K. Jatcza, F. Mage, A. Mestre, O. Nordli, J. Penuelas, P. Pirinen, V. Remisova, H. Scheifinger, M. Striz, A. Susnik, A. J. H. Van Vliet, F. E. Wielgolaski, S. Zach, and A. Zust. 2006. European phenological response to climate change matches the warming pattern. Global Change Biology 12:1969-1976. Mille-Lindblom, C. and L. J. Tranvik. 2003. Antagonism between bacteria and fungi on decomposing aquatic plant litter. Microbial Ecology 45:173-182. Moore, L. R., G. Rocap, and S. W. Chisholm. 1998. Physiology and molecular phylogeny of coexisting Prochlorococcus ecotypes. Nature 393:464-467. Niklaus, P. A., D. Alphei, D. Ebersberger, C. Kampichler, E. Kandeler, and D. Tscherko. 2003. Six years of in situ CO(2) enrichment evoke changes in soil structure and soil biota of nutrient-poor grassland. Global Change Biology 9:585-600.

Padgett, P. E. and E. B. Allen. 1999. Differential responses to nitrogen fertilization in native shrubs and exotic annuals common to Mediterranean coastal sage scrub of California. Plant Ecology 144:93-101. Philippot, L., S. G. Andersson, T. J. Battin, J. I. Prosser, J. P. Schimel, W. B. Whitman, and S. Hallin. 2010. The ecological coherence of high bacterial taxonomic ranks. Nat Rev Microbiol 8:523-529. Placella, S. A., E. L. Brodie, and M. K. Firestone. 2012. Rainfall-induced carbon dioxide pulses result from sequential resuscitation of phylogenetically clustered microbial groups. Proceedings of the National Academy of Sciences of the United States of America 109:10931-10936. Potts, D. L., K. N. Suding, G. C. Winston, A. V. Rocha, and M. L. Goulden. 2012. Ecological effects of experimental drought and prescribed fire in a southern California coastal grassland. Journal of Arid Environments 81:59-66. Prosser, J. I., B. J. M. Bohannan, T. P. Curtis, R. J. Ellis, M. K. Firestone, R. P. Freckleton, J. L. Green, L. E. Green, K. Killham, J. J. Lennon, A. M. Osborn, M. Solan, C. J. van der Gast, and J. P. W. Young. 2007. Essay - The role of ecological theory in microbial ecology. Nature Reviews Microbiology 5:384-392. Pruesse, E., J. Peplies, and F. O. Glockner. 2012. SINA: Accurate high-throughput multiple sequence alignment of ribosomal RNA genes. Bioinformatics 28:1823-1829. Ramirez, K. S., J. M. Craine, and N. Fierer. 2012. Consistent effects of nitrogen amendments on soil microbial communities and processes across biomes. Global Change Biology 18:1918-1927. Ratkowsky, D. A., J. Olley, T. A. Mcmeekin, and A. Ball. 1982. Relationship between Temperature and Growth-Rate of Bacterial Cultures. Journal of Bacteriology 149:1-5. Reed, H. E. and J. B. Martiny. 2013. Microbial composition affects the functioning of estuarine sediments. ISME J 7:868-879. Reeder, J. and R. Knight. 2010. Rapidly denoising pyrosequencing amplicon reads by exploiting rank-abundance distributions. Nature Methods 7:668-669. Reich, P. B., J. Knops, D. Tilman, J. Craine, D. Ellsworth, M. Tjoelker, T. Lee, D. Wedin, S. Naeem, D. Bahauddin, G. Hendrey, S. Jose, K. Wrage, J. Goth, and W. Bengston. 2001. Plant diversity enhances ecosystem responses to elevated CO2 and nitrogen deposition. Nature 410:809-812. Romani, A. M., H. Fischer, C. Mille-Lindblom, and L. J. Tranvik. 2006. Interactions of bacteria and fungi on decomposing litter: Differential extracellular enzyme activities. Ecology 87:2559-2569. Rossman, A. Y., R. E. R. E. Tulloss, T. E. O'Dell, and R. G. Thorn. 1998. Protocols for an all taxa biodiversity inventory of fungi in a Costa Rican conservation area. Parkway Publishers, Boone, N.C. Scheiter, S., L. Langan, and S. I. Higgins. 2013. Next-generation dynamic global vegetation models: learning from community ecology. New Phytol 198:957-969. Schimel, J. P. 1995. Ecosystem consequences of microbial diversity and community structure. Pages 239-254 in F. S. Chapin and C. Korner, editors. Arctic and Alpine Biodiversity: patterns, causes, and ecosystem consequences. Springer-Verlag, Berlin. Schimel, J. P. 2001. Biogeochemical Models: Implicit versus Explicit Microbiology. Pages 177- 183 in E. D. Schulze, M. Heimann, S. Harrison, E. Holland, J. Lloyd, I. C. Prentice, and D. Schimel, editors. Global Biogeochem Cycles in the Climate System. Academic Press, New York. Schimel, J. P., J. Bennett, and N. Fierer. 2005. Microbial community composition and soil nitrogen cycling: is there really a connection? . Pages 171-188 in R. D. Bardgett, M. Usher, and D. Hopkins, editors. Biological Diversity and Function in Soils. Cambridge University Press, New York, USA.

Schimel, J. P. and J. Gulledge. 1998. Microbial community structure and global trace gases. Global Change Biology 4:745-758. Schimel, J. P. and S. M. Schaeffer. 2012. Microbial control over carbon cycling in soil. Front Microbiol 3:348. Schlesinger, W. H. and J. A. Andrews. 2000. Soil respiration and the global carbon cycle. Biogeochemistry 48:7-20. Seifert, J. 1961. The influence of moisture and temperature on the number of bacteria in the soil. Folia Microbiologica 6:268-272. Sheik, C. S., W. H. Beasley, M. S. Elshahed, X. H. Zhou, Y. Q. Luo, and L. R. Krumholz. 2011. Effect of warming and drought on grassland microbial communities. Isme Journal 5:1692-1700. Singh, B. K., R. D. Bardgett, P. Smith, and D. S. Reay. 2010. Microorganisms and climate change: terrestrial feedbacks and mitigation options. Nature Reviews Microbiology 8:779-790. Specht, R. L., E. J. Moll, F. Pressinger, and J. Sommerville. 1983. Moisture regime and nutrient control of seasonal growth in Mediterranean ecosystems. Pages 120–132 in F. J. Kruger, D. T. Mitchell, and J. U. M. Jarvis, editors. Mediterranean-type Ecosystems: The Role of Nutrients. Springer, Berlin. Srivastava, D. S., M. W. Cadotte, A. A. M. MacDonald, R. G. Marushia, and N. Mirotchnick. 2012. Phylogenetic diversity and the functioning of ecosystems. Ecology Letters 15:637- 648. Strickland, M. S., C. Lauber, N. Fierer, and M. A. Bradford. 2009. Testing the functional significance of microbial community composition. Ecology 90:441-451. Suding, K. N., S. Lavorel, F. S. Chapin, J. H. C. Cornelissen, S. Diaz, E. Garnier, D. Goldberg, D. U. Hooper, S. T. Jackson, and M. L. Navas. 2008. Scaling environmental change through the community-level: a trait-based response-and-effect framework for plants. Global Change Biology 14:1125-1140. Swift, M. J., O. W. Heal, and J. M. Anderson. 1979. Decomposition in terrestrial ecosystems. Blackwell Scientific Publications, Oxford. Taylor, B. R., D. Parkinson, and W. F. J. Parsons. 1989. Nitrogen and Lignin Content as Predictors of Litter Decay-Rates - a Microcosm Test. Ecology 70:97-104. Tedersoo, L., T. Jairus, B. M. Horton, K. Abarenkov, T. Suvi, I. Saar, and U. Koljalg. 2008. Strong host preference of ectomycorrhizal fungi in a Tasmanian wet sclerophyll forest as revealed by DNA barcoding and taxon-specific primers. New Phytol 180:479-490. Tiunov, A. V. and S. Scheu. 2005. Facilitative interactions rather than resource partitioning drive diversity-functioning relationships in laboratory fungal communities. Ecology Letters 8:618-625. Todd-Brown, K. E. O., F. M. Hopkins, S. N. Kivlin, J. M. Talbot, and S. D. Allison. 2012. A framework for representing microbial decomposition in coupled climate models. Biogeochemistry 109:19-33. Treseder, K. K., S. N. Kivlin, and C. V. Hawkes. 2011. Evolutionary trade-offs among decomposers determine responses to nitrogen enrichment. Ecology Letters 14:933-938. Treves, D. S., B. Xia, J. Zhou, and J. M. Tiedje. 2003. A two-species test of the hypothesis that spatial isolation influences microbial diversity in soil. Microb Ecol 45:20-28. Truper, H. G. 1992. Prokaryotes - an Overview with Respect to Biodiversity and Environmental Importance. Biodiversity and Conservation 1:227-236. Turner, S., K. M. Pryer, V. P. W. Miao, and J. D. Palmer. 1999. Investigating deep phylogenetic relationships among cyanobacteria and plastids by small submit rRNA sequence analysis. Journal of Eukaryotic Microbiology 46:327-338. Vandamme, P., J. F. Bernardet, P. Segers, K. Kersters, and B. Holmes. 1994. New Perspectives in the Classification of the Flavobacteria - Description of Chryseobacterium

Gen-Nov, Bergeyella Gen-Nov, and Empedobacter Nom Rev. International Journal of Systematic Bacteriology 44:827-831. Voriskova, J. and P. Baldrian. 2013. Fungal community on decomposing leaf litter undergoes rapid successional changes. ISME J 7:477-486. Wallenstein, M. D. and E. K. Hall. 2012. A trait-based framework for predicting when and where microbial adaptation to climate change will affect ecosystem functioning. Biogeochemistry 109:35-47. Wang, Q., G. M. Garrity, J. M. Tiedje, and J. R. Cole. 2007. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Applied and Environmental Microbiology 73:5261-5267. Weber, C. F., R. Vilgalys, and C. R. Kuske. 2013. Changes in Fungal Community Composition in Response to Elevated Atmospheric CO2 and Nitrogen Fertilization Varies with Soil Horizon. Front Microbiol 4:78. White, T. J., T. D. Bruns, S. Lee, and J. Taylor. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pages 315–322 in M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White, editors. PCR Protocols: a guide to methods and applications. Academic Press, Ney York, USA. Wieder, W. R., G. B. Bonan, and S. D. Allison. 2013. Global soil carbon projections are improved by modelling microbial processes. Nature Climate Change 3:909-912. Williams, N. M., E. E. Crone, T. H. Roulston, R. L. Minckley, L. Packer, and S. G. Potts. 2010. Ecological and life-history traits predict bee species responses to environmental disturbances. Biological Conservation 143:2280-2291. Wolkovich, E. M., B. I. Cook, J. M. Allen, T. M. Crimmins, J. L. Betancourt, S. E. Travers, S. Pau, J. Regetz, T. J. Davies, N. J. Kraft, T. R. Ault, K. Bolmgren, S. J. Mazer, G. J. McCabe, B. J. McGill, C. Parmesan, N. Salamin, M. D. Schwartz, and E. E. Cleland. 2012. Warming experiments underpredict plant phenological responses to climate change. Nature 485:494-497. Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J. Cavender-Bares, T. Chapin, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Garnier, P. K. Groom, J. Gulias, K. Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J. Midgley, M. L. Navas, U. Niinemets, J. Oleksyn, N. Osada, H. Poorter, P. Poot, L. Prior, V. I. Pyankov, C. Roumet, S. C. Thomas, M. G. Tjoelker, E. J. Veneklaas, and R. Villar. 2004. The worldwide leaf economics spectrum. Nature 428:821-827. Yuste, J. C., J. Penuelas, M. Estiarte, J. Garcia-Mas, S. Mattana, R. Ogaya, M. Pujol, and J. Sardans. 2011. Drought-resistant fungi control soil organic matter decomposition and its response to temperature. Global Change Biology 17:1475-1486. Zak, D. R., W. E. Holmes, D. C. White, A. D. Peacock, and D. Tilman. 2003. Plant diversity, soil microbial communities, and ecosystem function: Are there any links? Ecology 84:2042- 2050. Zavaleta, E. S., M. R. Shaw, N. R. Chiariello, B. D. Thomas, E. E. Cleland, C. B. Field, and H. A. Mooney. 2003. Grassland responses to three years of elevated temperature, CO(2), precipitation, and N deposition. Ecological Monographs 73:585-604. Zwietering, M. H., I. Jongenburger, F. M. Rombouts, and K. Vantriet. 1990. Modeling of the Bacterial-Growth Curve. Applied and Environmental Microbiology 56:1875-1881.