Insights Into the Origin of N2O in Lake Vida Brine

University of Nevada, Reno

Insights into the origin of N2O in Lake Vida brine

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in

Environmental Science and Health

Gareth Trubl

Dr. Alison Murray/Thesis Advisor

August 2013

We recommend that the thesis prepared under our supervision by

Gareth Trubl

entitled

Insights into the origin of N2O in Lake Vida brine

be accepted in partial fulfillment of the

requirements for the degree of

Master of Science

Alison Murray, Ph.D., Advisor

Glenn Miller, Ph.D., Committee Member

Jerry Qualls, Ph.D., Committee Member

Nathanial Ostrom, Ph.D., Committee Member

Christian Fritsen, Ph.D., Graduate School Representative

Marsha H. Read, Ph.D., Dean, Graduate school

August 2013

I. Abstract

Lake Vida is a briny, frigid, inhospitable anoxic lake, and sealed beneath

16m of ice, there is an 18 to 21% NaCl based brine with a high concentration of

N2O. Biological and abiotic denitrification pathways of N2O production were evaluated to determine the origin(s) of the high concentration of N2O in Lake

Vida brine. Sixteen bacterial isolates from Lake Vida brine spanning six genera

(Marinobacter, Psychrobacter, Exiguobacterium, Arthrobacter, Microbacterium, and Kocuria) were tested for the capability of incomplete or complete denitrification. Four of the sixteen isolates evaluated showed evidence of N2 production. The isolate from each genus that produced the most N2O was further tested to see if total protein production correlated with N2O production.

Marinobacter sp. strain LV411 had the highest N2O production rate of 60 ± 6µM day−1 and a generation time of 30 ± 0 hours grown at 10°C. Isolate LV411 produced nearly 8-fold more N2O, than the next highest N2O producer

(Exiguobacterium sp. strain LV05br1). Lake Vida bacterial isolates demonstrated the ability to denitrify, in which four of them showed the potential for complete denitrification. The potential, therefore exists that these isolates, Marinobacter sp., in particular, could influence the Lake Vida brine nitrogen biogeochemistry.

The total N2O produced by Marinobacter sp. strain LV411 and Psychrobacter sp. strain LV181 each had a SP value around 0‰ that is similar to that reported previously for production by denitrifying bacteria. The net isotope effect values for

δ15N, δ18O, and SP for LV411 and LV181 were different from each other, ii suggesting different fractionation factors along the denitrification pathway. The potential for abiotic denitrification in Lake Vida brine was also demonstrated. The

2+ − − addition of ferrous iron (Fe ), nitrite (NO2 ), or nitrate (NO3 ) stimulated N2O

15 production in mercuric chloride (HgCl2) amended Lake Vida brine. A N tracer experiment, using Lake Vida brine, revealed a decrease in N2O concentration in several amendments with a decrease in δ15N bulk isotope values, confirming

N2O production and suggesting further reduction of N2O to N2. Lake Vida brine treated with 10mM FeSO4 and 10mM NaNO2 had a N2O production rate of 10µM day−1, a final SP value of −1.5 ± 0.5‰, a final δ15N value of 24.9 ± 0.4‰, and a final δ18O value of −14.5 ± 0.1‰. The SP value is within the range of SP values

2+ − − reported from chemodenitrification. The addition of Fe , NO2 , or NO3 also stimulated CO2 production in HgCl2 amended Lake Vida brine, likely from acidification of the large reservoir of DIC in Lake Vida brine. The N2O SP values from biological and abiotic sources in Lake Vida were not significantly different and therefore cannot be used to discern a difference between the two sources

(Marinobacter sp. strain LV411 was 1.2 ± 0.7‰, Psychrobacter sp. strain LV181 was −1.9 ± 0.8‰ and abiotic treatment with 10mM FeSO4 + 10mM NaNO2 was

−1.5 ± 0.5‰). A comparison of N2O SP values, along with evidence of possible biological and abiotic N2O production, suggests the source of N2O in Lake Vida brine could be the result of either process or a combination of both biological and abiotic sources.

iii

Dedication

I dedicate this thesis to my wife Brittany Trubl for inspiring and dealing with me in all my years of study. You came into my life and totally changed my world. You are my best friend and I am deeply in love with you.

Acknowledgments

Funding for this research was provided by the NSF ANT-0739681 to A.E.M. and

C.H.F. G.T. was also supported by the NRES department at UNR and the DEES at DRI. Thanks are extended to Emanuele Kuhn, Kathryn Bywaters, KC King,

Paula Matheus-Carnevali, Dr. Glenn Miller, Dr. Christian H. Fritsen, Dr. Jerry

Qualls, Dr. Nathaniel E. Ostrom, Dr. Fabien Kenig, Dr. Hasand Gandhi, Vivan

Peng, Olivia Rassuchine, Dr. Jeremy Dodsworth, and Dr. Alison E. Murray for advice and technical assistance. v

TABLE OF CONTENTS

SECTION PAGE

Abstract ...... i

Dedication ...... iii

Acknowledgments ...... iv

Table of Contents ...... v

List of Tables ...... vi

List of Figures ...... vii

Introduction ...... 1

Experimental Procedures………………………………………………………………5

Results ...... 14

Discussion ...... 23

Conclusion ...... 32

References ...... 34

Supplemental Tables and Figures ...... 41

Appendices ...... 46

Appendix A: Detection of genes in the denitrification pathway ...... 46

Appendix B: Salinity and temperature tolerance of Lake Vida isolates .... 50 vi

LIST OF TABLES

Table 1. N2O production rates and generation times for Lake Vida bacterial

isolates ...... 17

Table 2. N2O concentration and stable isotopic data ...... 18

Table 3. Abiotic N2O and CO2 production rates ...... 20

15 Table 4. Enrichment of N in abiotic N2O production ...... 21

Table 5. Isotopic composition of N and O in N2O ...... 22

vii

LIST OF FIGURES

Figure 1. Biological N2O production with and without C2H2 ...... 15

Figure 2. N2O production correlated to total protein production for Lake Vida

bacterial isolates ...... 16

I. Introduction

High concentrations of nitrous oxide (N2O) are a feature of several aquatic systems in Antarctica’s McMurdo Dry Valleys (Priscu et al., 1993; Priscu, 1996).

Lakes Vanda, Bonney, Joyce, Fryxell, Hoare, Miers (Priscu, 1997), Vida (Murray et al., 2012), and Don Juan Pond (sediment; Samarkin et al., 2010) all contain a

N2O concentration exceeding air saturation (~320ppb; Forster et al., 2007), by at least 500%. The N2O in the McMurdo Dry Valleys is thought to be a result of active, biologically mediated processes. N2O in Lake Bonney occurs in both oxic and anoxic environments. N2O was detected in the East (41µM; Priscu et al.,

1996) and West lobes (1µM; Priscu, 1997) of Lake Bonney. The high concentration of N2O in the East lobe is thought to result from nitrifier- denitrification due to the high potential for nitrification in aerobic zones and low rates of denitrification in anaerobic zones (Priscu 1997). The N2O detected in the

West lobe appears to be from biological denitrification due to the appearance of

N2O in anoxic water below 13m (Priscu et al., 1993).

High concentrations of N2O were recently detected in sediment surrounding Don Juan Pond that ranged from 1.16 to 2.03µM (Samarkin et al.,

2010). N2O was not detected in Don Juan pond brine, presumably because the pond is extremely shallow, ice free, and undergoes wetting and drying cycles where the produced N2O could degas. Experimentation regarding N2O production in Don Juan Pond led to a realization that N2O in these sediments

− − occurred immediately when NO3 or NO2 and an iron source are added to Don 2

Juan Pond brine (Samarkin et al., 2010). Abiotic denitrification has since been proposed as the process producing the N2O via an abiotic serpentinization-like reaction(s) (Evans, 2008; Samarkin et al., 2010):

− 2+ + 2NO2 + 6Fe + 5H2O  N2O + 2Fe3O4 + 10H

− 2+ + 2NO3 + 12Fe + 11H2O  N2O + 4Fe3O4 + 22H

Serpentinization is a metamorphic process that has been most frequently implicated in environments with moderate to high heat and moderate to high pressure (Müntener, 2010; Kelley et al., 2005; Sleep et al., 2004). Heat and

2+ pressure cause anaerobic oxidation of Fe by H2O, leading to the formation of

H2 and magnetite (Fe3O4). Serpentinization reactions producing H2 from alteration of ultramafic rocks has been shown to occur in environments that are anoxic and low temperature (Abrajano et al., 1990; Sano et al.,1993; Charlou et al., 2002) and has also been proposed to occur on the icy Jovian moon Europa

(Zolotov and Shock, 2004):

1.5Fe2SiO4 + H2O  Fe3O4 + 1.5 SiO2 +H2

3FeSiO3 + H2O  Fe3O4 + 3SiO2 + H2

FeS +4H2O  Fe3O4 + 3H2S + H2 3

Lake Vida, which lies in Victoria Valley, is an ice-sealed aquatic system with brine below 16m of ice. The cold (−13°C) anoxic brine has unusual

− + geochemical properties (~1,100µM NO3 , ~3,600µM ammonium (NH4 ), and

~260µM Fe), has extremely high concentrations of N2O (~87µM), and bacteria that are metabolically active at low levels (Murray et al., 2012). Victoria Valley geology is dominated by pyroxene rich rock and sediment (Clarkson, 1981;

Murray et al., 2012) and is similar to iron rich rock and sediment in Don Juan

Pond (Samarkin et al., 2010). The measured N2O from Lake Vida places this environment as one of the highest reported N2O concentrations in any natural ecosystem. The origin of the N2O in Lake Vida brine is the subject of the current study in hopes of determining if isotopic analysis may help in determining between a biological and abiotic origin(s).

Dissimilatory denitrification is a widespread, energy yielding respiratory

− pathway used by microbes living in oxygen-limited environments in which NO3 is reduced to yield N2O and nitrogen (N2) (Payne, 1973). The complete respiratory

− denitrification pathway involves four key genes: nap or nar (NO3 reduction to

− − NO2 ), nir (NO2 reduction to nitric oxide (NO)), nor (NO reduction to N2O), and nos (N2O reduction to N2) (Thauer et al., 1977; Ye et al., 1993; Zumft, 1997).

Several microbes do not complete this entire pathway and produce N2O as the final product. This shorter version of denitrification accounts for a large portion of global N2O production (Ronen et al., 1988; Bouwman et al., 2002).

A potentially high source of uncharacterized N2O production from natural

− 2+ sources occurs when NO2 reacts chemically with Fe in a process called 4

chemodenitrification (Postma, 1990; Rakshit et al., 2008; Picardal, 2012). N2O is produced abiotically from anthropogenic (automobiles and industry; Toyoda et al., 2008) and natural sources (sediment and aquatic systems; Tai and Dempsey,

2009; Thorn and Mikita, 2000). The primary source of N2O in the atmosphere is from agriculture practices that have increased denitrification and nitrifier- denitrification rates (Galloway et al., 2004; Ronen et al., 1988; Bouwman et al.,

2002). Over the past century N2O concentrations have been steadily increasing due to anthropogenic emissions (Rasmussen and Khalil, 1986; Prinn et al., 1990;

Machida et al., 1995). N2O is a key component of climate science and is of great interest to many astrobiologists, because N2O is considered to be a biosignature for Earth-like atmospheres (Segura et al., 2005; Rauer et al., 2011) and is present in many extreme environments (Priscu, 1997).

Sources of biological N2O production can be determined based on site preference (SP) isotope values. SP refers to the position (α or β) in the N2O compound, in which 15N is preferentially occupied; the central N atom is referred to as the α-position and the terminal N atom is referred to as the β-position

(Yoshida and Toyoda, 2000). N2O SP can be used to discern a difference between N2O production from nitrification and denitrification (Toyoda et al., 2005;

Sutka et al, 2006; Ostrom and Ostrom, 2011) as distinct SP values have been reported for these processes; ~33‰ for nitrification and ~0‰ for denitrification

(Sutka et al, 2006). Abiotic N2O, however, has been reported to have a wide range of SP values from −45‰ to 50‰ (Toyoda et al., 2008; Samarkin et al.,

2010). SP analysis of N2O from Lake Vida brine (−3.6 ± 0.3‰), suggests the 5

origin could be from biological or abiotic sources or a combination of both

biological and abiotic sources (Murray et al., 2012).

Bacterial isolates recovered from Lake Vida brine and Lake Vida brine

amended with several chemical and mineral amendments were tested for

denitrification potential and N2O production was analyzed for isotopic

fractionation and SP values. These experiments were conducted to help address

questions regarding: (1) the origin of N2O measured in Lake Vida brine; (2)

denitrifying capability of Lake Vida brine isolates; (3) abiotic denitrification in Lake

Vida brine; and (4) N2O SP values and application to discerning differences

between biological and abiotic origins in Lake Vida brine.

II. Experimental Procedures

Denitrification Capabilities of Lake Vida Brine Isolates

Bacterial cultivation efforts using brine collected in 2005 and 2010 have

resulted in a collection of 59 bacterial isolates (Mondino et al. (2009), Christian

Fritsen (pers comm.), and Kuhn et al. (in prep)), 16 of which were non-redundant

at the 16S rRNA level (over 1295 base pairs). The 16 distinct isolates spans six

genera (Marinobacter, Psychrobacter, Exiguobacterium, Arthrobacter,

Microbacterium, and Kocuria) and were used in the biological experiments.

Supplemental Table 1 lists all isolates and associated information. The Lake Vida

brine isolates were characterized for their ability to denitrify in three different

experiments (i) complete denitrification, aimed to test the difference in N2O vs. N2 6

production, (ii) N2O production and growth rate determination, aimed to test if

N2O production was coupled to protein production and (iii) net isotope effect for biological N2O, aimed at obtaining isotopic fractionation data, specifically SP values. Similar growth conditions were used in all experiments which included the isolates were grown in triplicate in R2A + 5% NaCl, with a 10% acetylene

(C2H2; Oremland and Taylor, 1975; Teissier and Torre 2002) and 90% N2 headspace, with 10mM KNO3, inverted, and incubated at 10°C in borosilicate serum bottles capped with geomicrobial rubber stoppers. The negative controls for each experiment were sterile medium without inoculum. The C2H2 used in each biological experiment was generated by adding three grams of calcium carbide granules (Thermo Fisher Scientific, Waltham, MA) to a 120mL serum bottle that was capped with a butyl rubber stopper (Geomicrobiological

Technologies, Inc). Using a syringe, 3mL of distilled water was injected into the serum bottle. The bottle was vigorously mixed and some headspace was released using the syringe to remove water vapor (Bixler and Coberly, 1953).

i. Complete Denitrification. The capacity of the non-redundant suite of

Lake Vida brine isolates to produce both N2O and N2 was determined. Sixteen isolates and negative controls were grown in conditions previously stated, along with a paired culture or negative control with no C2H2. The same amount of medium (5mL) and inoculum (100µL; grow under aerobic conditions) was used for each isolate and serum bottle. After five days of incubation, 100µL of headspace from each replicate was measured via gas chromatography with an electron capture detector (GC-ECD; Lovelock, 1958). The GC-ECD was a 7

Shimadzu Gas chromatograph equipped with a 63Nickel electron capture detector. The column was a 80/100 Hayesep Q resin in a 2 m length 1/8 in diameter column. The carrier gas was 95% Argon with 5% methane and the flow rate was 45mL x minute-1. GC-ECD parameters followed from Mosier and

Klemedtsson (1994). N2O was quantified using a 7 point standard curve ranging from 0.1-2000 ppmv (Scotty gases, Air Liquide America Specialty Gases,

Plumsteadville, PA). A calibration curve was generated by plotting peak area vs. ppmv. The equation from the linear regression analysis of the calibration curve was used to calculate the ppmv of the Lake Vida samples. The ppmv amount was then converted to µM by using the ideal gas law PV=nRT. Note that the acetylene block technique is not completely effective (Yu et al., 2010) and the effectivness has not been tested on these isolates.

ii. N2O production and growth rate determination. One isolate from each genus (Marinobacter sp. strain LV411, Psychrobacter sp. strain LV181,

Exiguobacterium sp. strain LV05br1, Arthrobacter sp. strain LV10R520-3,

Microbacterium sp. strain LV10fR510-14, and Kocuria sp. strain LV10fR520-8) was selected (based on ability to denitrify from the complete denitrification experiment). Experiments ran (under previously stated conditions) for 12 days and 100µL of headspace was measured via GC-ECD (with parameters previously listed) and on day(s) 0, 4, 6, 8, 10, and 12, 1mL of media was subsampled for total protein analysis. Total protein analysis was carried out according to microplate microassay instructions via Quick Start Bradford Protein

Assay (Bio-Rad Laboratories, Inc., Hercules, CA). Microplates were measured 8 with a Spectromax spectrophotometer (Molecular Devices, Sunnyvale, CA) using a preloaded Braford protein assay program. A generation time was established for each isolate tested by using the measured protein concentrations and the equation g=(log10 Nt − log10 N0)/log102 (Powell 1956). N2O production rates were determined by measuring N2O concentration and dividing by the number of days grown.

iii. Net isotope effect for biological N2O production. Isolates Marinobacter sp. strain LV411 and Psychrobacter sp. strain LV181 were chosen for a final experiment to obtain isotopic data on N2O production. These two isolates were chosen, because they represented the majority of bacteria in the brine based in

16S rRNA gene sequencing (Murray et al. 2012). The isolates were grown (in conditions previously stated) for 12 days. Fifteen replicates from each isolate were grown and the headspace (three replicates from each isolate) of their respective serum bottle was measured via mass spectroscopy at the end of the

12 day time series. After 4, 6, 8, 10, and 12 days, a set of replicates from each isolate were inactivated by the addition of a final concentration of 0.5 mM HgCl2 and inverted and incubated at −20°C. A multi-collector GV Instruments Isoprime mass spectrometer interfaced with a continuous Trace Gas inlet system was used to obtain N2O isotopic fractionation and SP values. The MS was prepared to quantify ion signals at masses 30, 31, 44, 45, and 46. A cryogenic trap was setup to remove CO2 and volatiles. Air samples were used to calibrate the MS.

The headspace of each replicate was measured first via GC-ECD. Using a gas- tight syringe, 500 µL of headspace was extracted and injected into an evacuated 9 serum bottle. The sample incubated in the serum bottle for 15 minutes before being measured with the MS. The MS collected data on δ15N, δ18O, and SP

15 15 14 values. The δ N is defined as (Rsample/Rstandard −1)* 1000, where R is N/ N.

18 18 16 The δ O is defined as (Rsample/Rstandard −1)* 1000, where R is O/ O. Final isotopic ratios in per mil notation were corrected for mass overlap and reported

15 18 relative to atmospheric N2 (for δ N) or Standard Mean Ocean Water (δ O) according to Toyoda and Yoshida (1999). N2O production from biological denitrification is a multi-step process, with fractionation present at each step. The fractionation occurs from preferential cleavage of 14N by each enzyme in the four step process. Fractionation takes place over several steps, therefore the total fractionation is presented as the net isotope effect. The net isotope effect is

15 15 defined as δ NN2O − δ Ninorganic-N. The Rayleigh distillation equation (Mariotti et al., 1981) was used to determine net isotope effect. A second GC-ECD was implemented with the same parameters, but with helium as the carrier gas at a

−1 flow rate of 25mL minute . N2O concentrations were calculated via calculations previously stated for GC-ECD and via MS. The MS N2O concentration in ppmv was calculated by plotting peak height versus N2O standards. The conversion of ppmv to µM was calculated by using the ideal gas law PV=nRT.

Abiotic Denitrification in Lake Vida Brine

Lake Vida brine used in these experiments was collected from the same sample time and depth (18.5m) collected Nov. 30, 2010 (Murray et al., 2012).

The brine was stored in a −10°C walk in freezer under anoxic conditions for ~2.5 10 years at the Desert Research Institute prior to use in these experiments. Lake

Vida brine was amended with several different treatments in three experiments (i)

N2O production via chemodenitrification, aimed to determine abiotic N2O production and brine reactive agents; (ii) N2O tracer, aimed to confirm N2O production; (iii) Natural isotope abundance as a function of chemodenitrification, aimed to obtain isotopic fractionation data and determine SP values. Each experiment was conducted by anaerobically transferring Lake Vida brine from

500mL serum bottles to evacuated 15mL serum bottles in a −10°C walk in freezer. Tubing and T-locks were used, along with N2 to transfer Lake Vida brine from the 500mL serum bottles to the 15mL serum bottles. Lake Vida brine with no amendments and the Lake Vida brine treated with HgCl2 (final concentration of 0.5mM HgCl2) to eliminate biological activity were used in these experiments as controls.

i. N2O production via chemodenitrification. The N2O production via chemodenitrification experiment used 5mL of Lake Vida brine while for the other two experiments, the serum bottles were completely filled (15 mL each).

Experiments included Lake Vida brine amended with chemical and mineral treatments to determine chemodenitrification potential. Lake Vida brine was amended (in triplicate) with 8 different treatments in 15mL serum bottles, capped with butyl rubber stoppers and crimped, inverted, and incubated at −13.4°C and at room temperature (~23°C) for 9 days. Amendements included: brine, HgCl2 amended brine, finely crushed pyroxene rich rock, sediment, FeSO4, KNO3, 11

NaNO2, FeSO4 + KNO3, FeSO4 + NaNO2, FeSO4 + KNO3 + H2O, FeSO4 +

NaNO2. The FeSO4, NaNO2, and KNO3 treatment concentrations were estimated to be 10% of in situ Lake Vida brine conditions reported in Murray et al (2012);

(26µM FeSO4, 2.73µM NaNO2, and 112µM KNO3).

A 27-m ice core with intercalated sediment layers was recovered from

Lake Vida in 2010 (Murray et al. 2012). The depth horizon 2528cm to 2548cm below lake surface was used here, and was homogenized for this experiment. In addition finely crushed pyroxene rich rock was collected during the 2010 field expedition. It is a riverine deposit collected in the inside of second meander of

Kyte Stream, the stream located on the west side of Lake Vida. The rock is made up of sand and many grains are well rounded and the concentration of pyroxene

(54%, 416 grains counted, 35% clinopyroxene, 19 % orthopyroxene) is more elevated than for other sands we collected in river streams, most probably as a result of increased density of pyroxene of other mineral constituents that were favorably deposited on the inside of the meander. The rock used was also made up of 12.3% of dolerite fragments and it is thought the high pyroxene content is derived from weathered dolerite. This rock was used in these experiments, because it was rich in pyroxene (contained iron) and it was likely to be somewhat close to what is present in the sediment core.

All chemical amendments were prepared in 15 mL serum bottles, capped with butyl rubber stoppers, and crimped, then degassed under a N2 stream for 10 minutes. The chemical amendments were anaerobically tranferred to the 15mL serum bottles with 5mL of brine via a syringe flushed with N2. Three grams of 12 crushed pyroxene rich rock and sediment were added to 15 mL serum bottles that were capped, crimped, then autoclaved. Five milliters of Lake Vida brine was anaerobically transfered (condtitions previously stated) to each bottle. The treatments FeSO4 + KNO3 + H2O and FeSO4 + NaNO2 + H2O were made up of

5mL anoxic distilled water, 18% NaCl, and 10% in situ concentrations: ~100µM

− − for NO3 , ~3µM for NO2 , and ~30 µM for Fe (Murray et al., 2012). A Neslab RTE

110 Recirculating Chiller and Heater was filled with a mixture of 50% distilled water and 50% Molecular Biology-Grade glycerol (Thermo Fisher Scientific,

Waltham, MA) and housed the −13.4°C samples. The chiller was prepared next to the GC-ECD with tubing from the chiller lining the inside of a cooler. The 15mL serum bottles were incubated within the chiller, which produced a continouse flow of −13.4°C glycerol and water mixture. Each day for 9 days, 100µL of headspace was measured via GC-ECD. The GC-ECD paramteres were the same as prevuiously noted and the standards and calculations were the same as well. Significance was determined by running a t test and obtaining a p value at the 95% confidence level.

15 ii. N2O tracer. N-N2O enrichments were conducted to confirm abiotic N2O production. Lake Vida brine was transferred anaerobically (as previously noted) into 15mL serum bottles. The serum bottles were inoculated, in triplicate, with one of the following treatments (followed by them being inverted and incubated in

15 15 15 a −14°C freezer): N-NaNO2, N-NaNO2 + 1g crushed pyroxene rich rock, N-

15 15 NaNO2 + 1g sediment, N-NaNO3, N-NaNO3 + 1g crushed pyroxene rich rock,

15 15 15 and N-NaNO3 + 1g sediment. The N-NaNO2 and N-NaNO3 were ordered 13 from Cambridge Isotope laboratories Inc, Andover, MA (anoxic chemical

15 15 amendments were prepared as previously described). The N-NaNO2 and N-

NaNO3 treatment concentrations were 10% of in situ conditions previously stated

(Murray et al., 2012). Approximately every two weeks (in triplicate) ~500µL of brine was removed via a gas tight syringe and transferred into an evacuated

15mL serum bottle and let degas at room temperature for 15 minutes and then measured via MS (conditions and controls were the same as previously described).

iii. Natural isotope abundance as a function of chemodenitrification. An

15 experiment was conducted to determine δ N-N2O isotopic enrichments factors and N2O SP values. Lake Vida brine was transferred anaerobically (as previously noted) into 15mL serum bottles. The serum bottles were amended with a final concentration of 10mM FeSO4 and 10mM NaNO2 (in triplicate), inverted, and incubated in a freezer set to −14°C. About every two weeks (in triplicate) about

500µL of brine was removed with a gas tight syringe and transferred into an evacuated 15mL serum bottle to degas at room temperature for 15 minutes and then measured via MS (conditions, controls, and calculations were the same as noted before). There is no preferential cleavage of 14N in abiotic production of

N2O, therefore fractionation occurs in one step and is reported as a fractionation factor. The Rayleigh distillation equation (Mariotti et al., 1981) was used to determine the fractionation factor. To account for isotopic inflation, each replicate was plotted separately and linear regressions were calculated. Scott et al. (2004) 14

showed the potential for the linear regression of the Rayleigh distillation equation

to be inflated, but provides the most accurate estimate of fractionation factors.

III. Results

Denitrification Capabilities of Lake Vida Brine Isolates

i. Complete Denitrification. Lake Vida bacterial isolates were incubated

with and without the presence of C2H2 to evaluate their ability to produce N2O

and N2. All isolates tested had the capability to produce N2O rates ranging from

−1 0.2 to 20µM day N2O. Exiguobacterium sp. strain LV05br1 had the highest N2O

−1 −1 production rate of 20µM day and 6µM day with or without C2H2. Marinobacter

−1 sp. strain LV10S had the lowest N2O production rate of 0.2µM day in the

presence of C2H2 and Marinobacter sp. strain LV10R510-11A had the lowest

N2O production rate of 0.1µM without C2H2. Four of the sixteen isolates tested

produced significantly (p< 0.05; t test) more N2O in the presence of C2H2

suggesting the capability to reduce N2O to N2. These included Marinobacter sp.

strain LV411, LV10R510-11A, Psychrobacter sp. strain LV181, and

Exiguobacterium sp. strain LV05br1 (Fig. 1). Twelve of the sixteen isolates tested

did not produce more N2O in the presence of C2H2. The average N2O production

rate for Marinobacter and Psychrobacter were the same (4 ± 3µM day−1 and 4 ±

−1 −1 −1 2µM day in the presence of C2H2 and 2 ± 1µM day and 2 ± 1µM day without

C2H2). An average production rate for the other genera was not applicable as

there were less than three isolates in those genera. 15

Figure 1. Biological N2O production with and without C2H2. Isolates are grouped by genus (Ex is Exiguobacterium, Ar is Arthrobacter, Ko is Kocuria, and Microbacterium is Mi. Error bars represent one standard deviation from the mean for three replicates. Asterisk denotes isolates that produced more N2O in the presence of C2H2 (p < 0.05; t test).

ii. N2O production and growth rate determination. One isolate from each genus was selected (based on capacity to produce N2O) and was further tested to determine if N2O production was correlated to protein production. All isolates

−1 had a N2O production rate ranging from 3 to 58 µM day . Total protein

−1 −1 production ranged from 0.8 to 7µg mL day . N2O production relative to total protein production (Fig. 2) was positively correlated (p< 0.05; t test) for each isolate. Marinobacter sp. strain LV411 had the highest N2O production rate of 60

± 6µM day−1 and a generation of 30 ± 0 hours at 10°C (Table 1). Marinobacter sp. strain LV411 produced nearly 8-fold more N2O, than the next highest producer 16

(Exiguobacterium sp. strain LV05br1, which had a N2O production rate of 9 ± 1

µM day−1 and a generation of 30 ± 0 hours at 10°C). The other Lake Vida isolates

−1 had N2O production rates ranging from 3 to 7µM day and generation times ranging from 40 to 50 hours at 10°C (Table 1).

Figure 2. N2O production correlated to total protein production for Lake Vida bacterial isolates. (A) Marinobacter sp. strain LV411, (B) Psychrobacter sp. strain LV181, (C) Exiguobacterium sp. strain LV05br1, (D) Arthrobacter sp. strain LV10R520-3, (E) Microbacterium sp. strain LV10fR510-14, and (F) Kocuria sp. strain LV10fR520-8. The x-axis depicts N2O in µM. The y-axis depicts total protein in µg x mL−1. Error bars represent one standard deviation from the mean for three replicates. The correlation coefficent (r) value is in the upper right corner of each box. The correlation between N2O production and total protein production was significant for each isolate (p < 0.05; t test).

Table 1. N2O production rates and generation times for Lake Vida bacterial isolates. Isolates are listed from highest N O production rate to lowest N O 2 2 production rate. The fastest generation time and fastest N2O production rate are in bold. Generation times are also listed and in hours. Error represents one standard deviation from the mean for three replicates.

Generation time N2O production rate Isolate n (hours) (µM day−1) Marinobacter sp. strain LV411 30 ± 0 60 ± 6 3 Exiguobacterium sp. strain LV05br1 30 ± 0 9 ± 1 3 Psychrobacter sp. strain LV181 40 ± 1 7 ± 1 3 Arthrobacter sp. strain LV10R520-3 40 ± 0 4 ± 0 3 Microbacterium sp. strain LV10fR510-14 50 ± 2 3 ± 0 3 Kocuria sp. strain LV10fR520-8 50 ± 0 3 ± 0 3

iii. Net isotope effect for Biological N2O. Marinobacter and Psychrobacter were the most abundant genera detected in Lake Vida brine (Murray et al., 2012) and therefore Marinobacter sp. strain LV411 and Psychrobacter sp. strain LV181 were utilized to quantify and isotopically analyze N2O. Isotopic analysis of N2O produced by Marinobacter sp. strain LV411 and Psychrobacter sp. strain LV181

15 18 revealed non-zero net isotope effects for δ N-N2O, δ O-N2O, and SP (Table 2).

The net isotope effect value from Marinobacter sp. strain LV411 was different than the net isotope effect for Psychrobacter sp. strain LV181. Marinobacter sp.

15 18 strain LV411 final day δ N-N2O value was −11.0 ± 1.7‰, δ O-N2O value was

28.4 ± 0.7‰, and SP value was 1.2 ± 0.7‰ (Table 2). Psychrobacter sp. strain

15 18 LV181 final day δ N-N2O value was −15.5 ± 2.0‰, δ O-N2O value was 27.6 ±

0.8‰, and SP value was −1.9 ± 0.8‰ (Table 2). N2O production rates were 70 ±

30µM for LV411 and 20 ± 7µM for LV181. 18

Table 2. N2O concentration and stable isotopic data. Gas production and associated stable isotope results for Marinobacter sp. strain LV411 and Psychrobacter sp. strain LV181 were determined over 12 days. Psychrobacter sp. strain LV181 measured at day 4 were below the limit of detection (N/D); day 8 only had two replicates, therefore both values are shown. Error represents one standard deviation from the mean. Asterisk (*) denotes the row of net isotope effect values. A polynomial relationship shows a stronger correlation (r2 = 0.9) for Marinobacter sp. strain LV411’s SP net isotope effect, therefore it is given.

15 15 α 15 β 18 Isolate Time (days) N2O (µM) δ N-N2O δ N -N2O δ N -N2O δ O-N2O SP n 4 6 ± 3 −19.9 ± 0.8 −24.4 ± 1.3 −15.4 ± 0.8 21.0 ± 2.1 −9.0 ± 1.4 3 6 30 ± 50 −18.9 ± 3.4 −22.2 ± 2.6 −15.6 ± 4.2 23.9 ± 2.2 −6.6 ± 1.6 3 8 200 ± 200 −19.2 ± 0.6 −21.1 ± 1.6 −17.3 ± 0.9 26.7 ± 1.1 −3.8 ± 2.4 3 Marinobacter sp. strain LV411 10 500 ± 400 −17.1 ± 1.8 −18.5 ± 4.9 −15.7 ± 2.0 26.6 ± 0.6 −2.9 ± 6.6 3 12 800 ± 400 −11.0 ± 1.7 −10.4 ± 2.0 −11.6 ± 1.5 28.4 ± 0.7 1.2 ± 0.7 3 * −65.6 −110.1 −21.1 −38.7 −1317.1x2 + 2402.1x − 1093.2 15 4 *N/D 6 9 ± 10 −24.1 ± 1.2 −29.5 ± 0.90 −18.7 ± 1.4 23.1 ± 0.6 −10.8 ± 0.6 3 8 8; 80 −23.4; -23.4 −28.3;-28.8 −18.6, -18.1 24.0; 22.9 −9.7; -10.7 2 Psychrobacter sp. strain LV181 10 100 ± 40 −17.9 ± 0.5 −19.3 ± 0.6 −16.4 ± 0.4 27.4 ± 0.6 −2.9 ± 0.5 3 12 200 ± 90 −15.5 ± 2.0 −16.5 ± 2.3 −14.6 ± 1.7 27.6 ± 0.8 −1.9 ± 0.8 3 * −352.9 −528.9 −176.9 −185.1 −352.0 11

Abiotic Denitrification in Lake Vida Brine

i. N2O production via chemodenitrification. Lake Vida brine was amended with potentially reactive substrates to determine the potential for abiotic N2O production. HgCl2 was added to Lake Vida brine to eliminate biological activity and was determined to catalyze N2O production (p < 0.05; t test); therefore each treatment was compared to the HgCl2 amended brine treatment (Table 3). The results revealed significant (p < 0.05; t test) abiotic N2O production in treatments brine + KNO3, brine + FeSO4, and brine + FeSO4 + NaNO2 at −13.4°C and in all treatments at 23°C (Table 3). The treatment with brine + FeSO4 + NaNO2 had the

−1 greatest N2O production rate of 1.51 ± 0.47µM day at −13.4°C and 3.25 ±

0.70µM day−1 at 23°C. There was a temperature effect observed (p < 0.05; t test) for treatments brine, brine + crushed pyroxene rich rock, and brine + FeSO4 + 19

NaNO2. Each treatment had a greater rate of N2O production, when incubated at room temperature vs. −13.4°C. Treatment FeSO4 + NaNO2 + H2O had a greater

N2O production rate at room temperature vs. −13.4°C (this treatment had no

Lake Vida brine). The formation of iron oxyhydroxides was observed in treatments brine + KNO3, brine + NaNO2, brine + FeSO4, brine + FeSO4 + KNO3, and brine + FeSO4 + NaNO2 at both temperatures.

HgCl2 also catalyzed CO2 production (p < 0.05; t-test) for −13.4°C and therefore each treatment was compared to the HgCl2-amended brine (Table 3).

For 23°C treatments were compared to brine without HgCl2. Significant (p < 0.05; t test) CO2 production occurred in treatments brine + FeSO4 + KNO3, brine +

FeSO4 + NaNO2, and brine + FeSO4 at −13.4°C and no amendments at 23°C

(Table 3). The brine + FeSO4 + NaNO2 amendment had the highest CO2 production rate of 0.05 ± 0.01µM at −13.4°C. A temperature effect (p < 0.05; t test) was observed with all the treatments having a greater CO2 production rate when incubated at −13.4°C vs. room temperature, except brine + FeSO4 +

KNO3, which had a greater CO2 production rate at room temperature.

Treatments FeSO4 + KNO3 + H2O and FeSO4 + NaNO2 + H2O had a greater N2O production rate at −13.4°C vs. room temperature (these treatments had no Lake

Vida brine).

Table 3. Abiotic N2O and CO2 production rates. N2O and CO2 production rates at −13.4ºC and 23ºC for brine and experimental additions. Values in bold and grey shading had a significant production rate (p < 0.05; t test) compared to HgCl2 amended brine. Asterisk (*) denotes the treatments without brine. Error represents one standard deviation from the mean.

−13.4°C N2O 23°C N2O −13.4°C CO2 23°C CO2 Treatment production rate production rate production rate production rate (µM day−1) (µM day−1) (µM day−1) (µM day−1) Brine 0.02 ± 0.02 0.13 ± 0.03 0.00 ± 0.00 0.02 ± 0.00

HgCl2-amended brine 0.84 ± 0.08 0.73 ± 0.03 0.03 ± 0.00 0.01 ± 0.00 Brine + sediment 0.74 ± 0.21 0.70 ± 0.04 0.02 ± 0.01 0.01 ± 0.01 Brine + crushed pyroxene rich rock 0.70 ± 0.21 1.07 ± 0.15 0.04 ± 0.01 0.02 ± 0.00

Brine + KNO3 1.11 ± 0.10 0.95 ± 0.03 0.04 ± 0.02 0.02 ± 0.00

Brine + NaNO2 1.19 ± 0.23 1.18 ± 0.12 0.04 ± 0.02 0.02 ± 0.00

Brine + FeSO4 1.42 ± 0.28 1.11 ± 0.01 0.02 ± 0.01 0.01 ± 0.00

Brine + FeSO4 + KNO3 1.22 ± 0.41 1.49 ± 0.12 0.00 ± 0.02 0.02 ± 0.00

Brine + FeSO4 + NaNO2 1.51 ± 0.47 3.25 ± 0.70 0.05 ± 0.01 0.02 ± 0.00

*FeSO4 + KNO3 + H2O 0.01 ± 0.00 0.02 ± 0.00 2.34 ± 0.00 2.03 ± 0.01 *FeSO4 + NaNO2 + H2O 0.33 ± 0.77 1.24 ± 0.07 2.39 ± 0.00 2.03 ± 0.01

15 15 15 ii. N2O tracer. An experiment with N enrichments ( N-NaNO3 and N-

NaNO2) was performed to accurately determine the rate of N2O production as concentration measurements alone do not always provide accurate rates. N2O

15 concentrations declined over the incubation period for N-NaNO3 + crushed

15 pyroxene rich rock from 46.4 ± 5.9µM to 11.7 ±13.8µM and N-NaNO3 + sediment from 46.4 ± 5.9µM to 13.6 ± 15.3µM (Table 4). There was no change in

15 N2O concentrations for all other treatments. The δ N values decreased for all

15 15 treatments amended with N-NaNO2 and no change for the N-NaNO3 (Table

18 15 4). No chnage was observed in δ O value, except in N-NaNO2 + crushed pyroxene rich rock, which decreased from 5.8 ± 0.7 to 3.0 ± 0.2µM.

15 Table 4. Enrichment of N in abiotic N2O production. N2O concentrations, δ15N, and δ18O values are listed for brine treatments enriched in 15N. Error represents one standard deviation from the mean for three replicates.

Time Treatment N O (µM) δ15N δ18O n (days) 2 15 N-NaNO2 1 40 ± 10 −20.5 ± 0.3 4.8 ± 0.3 3 15 N-NaNO2 2 40 ± 10 −19.4 ± 0.7 6.2 ± 1.0 3 15 N-NaNO2 11 50 ± 5 −17.7 ± 0.9 5.6 ± 0.1 3 15 N-NaNO2 29 40 ± 10 −15.8 ± 1.9 5.2 ± 1.0 3 15 N-NaNO2 51 40 ± 2 −13.2 ± 1.4 7.1 ± 2.4 3 15 N-NaNO2 64 40 ± 3 −11.8 ± 2.2 5.5 ± 0.4 3 15 N-NaNO2 79 40 ± 6 −12 ± 1.7 4.1 ± 2.5 3 15N-NaNO + crushed 2 1 50 ± 5 −20.7 ± 0.1 4.4 ± 0.1 3 pyroxene rich rock 15N-NaNO + crushed 2 2 50 ± 9 −19.7 ± 0.2 5.5 ± 0.1 3 pyroxene rich rock 15N-NaNO + crushed 2 11 50 ± 2 −17.3 ± 0.5 5.5 ± 0.4 3 pyroxene rich rock 15N-NaNO + crushed 2 29 50 ± 1 −14.5 ± 1.2 5.1 ± 0.3 3 pyroxene rich rock 15N-NaNO + crushed 2 51 30 ± 9 −9.4 ± 0.2 20.2 ± 7.3 3 pyroxene rich rock 15N-NaNO + crushed 2 64 40 ± 10 −9.5 ± 1.8 7.0 ± 0.4 3 pyroxene rich rock 15N-NaNO + crushed 2 79 40 ± 3 −7.5 ± 2.7 3.0 ± 0.2 3 pyroxene rich rock 15 N-NaNO2 + sediment 18 60 ± 2 −16.3 ± 1.3 5.7 ± 0.9 3 15 N-NaNO2 + sediment 40 30 ± 10 −10.5 ± 4.2 16.8 ± 4.3 3 15 N-NaNO2 + sediment 53 40 ± 2 −10.3 ± 1.6 6.8 ± 0.5 3 15 N-NaNO2 + sediment 68 40 ± 9 −7.8 ± 2.0 5.8 ± 1.4 3 15 N-NaNO3 19 30 ± 20 −18.7 ± 0.8 9.7 ± 5.5 3 15 N-NaNO3 41 40 ± 6 −20.1 ± 0.2 10.4 ± 3.3 3 15 N-NaNO3 54 40 ± 20 −20.3 ± 0.4 4.7 ± 0.1 3 15 N-NaNO3 73 30 ± 20 −20.0 ± 0.3 5.5 ± 0.0 3 15N-NaNO + crushed 3 19 40 ± 4 −20.0 ± 0.1 3.7 ± 1.1 3 pyroxene rich rock 15N-NaNO + crushed 3 41 30 ± 5 −20.1 ± 0.2 6.0 ± 2.3 3 pyroxene rich rock 15N-NaNO + crushed 3 54 30 ± 20 −20.0 ± 0.2 4.8 ± 0.1 3 pyroxene rich rock 15N-NaNO + crushed 3 73 10 ± 10 −19.8 ± 0.4 5.0 ± 0.1 3 pyroxene rich rock 15 N-NaNO3 + sediment 20 30 ± 10 −19.5 ± 0.3 6.1 ± 0.2 3 15 N-NaNO3 + sediment 41 20 ± 8 −20.0 ± 0.2 4.8 ± 0.2 3 15 N-NaNO3 + sediment 54 20 ± 10 −19.7 ± 0.7 5.0 ± 0.4 3 15 N-NaNO3 + sediment 73 10 ± 20 −19.8 ± 0.6 5.5 ± 0.4 3

iii. Natural isotope abundance as a function of chemodenitrification. The results from N2O production via chemodenitrification experiment (i) revealed abiotic N2O production, but the results from N2O tracer experiment (ii) revealed a decrease in N2O concentration for several treatments. Therefore another abiotic experiment was conducted, adding a surplus of reactants (10mM FeSO4 and

10mM NaNO2) to determine a N2O production rate, along with obtaining isotope values for N2O. Lake Vida brine treated with 10mM FeSO4 and 10mM NaNO2

−1 had a N2O production rate of 10µM day (Supplemental Fig. 3), a N2O final day

SP value of −1.5 ± 0.5‰ (Table 5), and a fractionation factor of −37.8‰ for δ15N-

18 N2O, −3.6‰ for δ O-N2O, and 3.3‰ for SP (Table 5). The pH of the brine decreased from 6.20 to 5.71 over the 88 day incubation period. N2O in Lake Vida brine with no amendments had a SP value of −12.6 ± 0.8‰, a δ15N bulk isotope

15 15 value for N2 of −0.3‰, a δ N bulk isotope value for KNO3 of 3 ± 0‰, and a δ N bulk isotope value for NaNO2 of −50 ± 0‰.

Table 5. Isotopic composition of N and O in N2O. Isotopic data reflects the average (± standard deviation) for n measurements of the final time point. The pH listed for LV brine 2010 is the initial pH for the treatment 10mM FeSO4 + 10mM NaNO2. The pH listed for 10mM FeSO4 + 10mM NaNO2 is the final pH after the 88 day incubation period. Asterisk (*) denotes row with fractionation values.

15 15 α 15 β 18 Sample N2O pH δ N δ N δ N δ O SP n LV Brine 2010 50 ± 6 6.2 −19.9 ± 0.3 −26.1 ± 0.1 −13.6 ± 0.7 5.8 ± 0.7 −12.6 ± 0.8 3 900 ± 30 5.7 −24.9 ± 0.4 −25.6 ± 0.3 −24.1 ± 0.5 −14.5 ± 0.1 −1.5 ± 0.5 5 10mM FeSO4 + NaNO2 * −37.8 −36.2 −39.4 −3.6 3.3 15

IV. Discussion

High concentrations of N2O is a feature of many aquatic systems in the

McMurdo Dry Valleys and has been intensely researched (Priscu, 1997). Both

biological and abiotic N2O production have been observed in different aquatic

envrionments within the McMurdo Dry Valleys (Priscu 1997, Samarkin et al.,

2010, Murray et al., 2012). Microbial denitrification has been reported in all three

domains of life and also occurs abiotically. Biological denitrification is reported to

account for 24.1 x 1012 mol year −1 of nitrogen compounds released to the

atmosphere, 71% of which is derived from aquatic environments (Canfield et al.,

2010). In spite of this knowledge, anthropogenically impacted sites (Bouwman,

1990) and terrestrial sites have been studied more than aquatic sites (Priscu,

1997). Advances in isotope techniques over the past decade have revealed new

ways to decipher the orgin of N2O sources from ecosystems based upon mass

spectrometric analysis (Toyoda and Yoshida, 1999; Yoshida and Toyoda, 2000;

Sutka et al., 2006). Here biological and abiotic experiments were conducted to

provide insight into the origin of N2O in Lake Vida brine and to determine if SP

values provide a basis to distinguish between biological and abiotc processes.

Biological experiments to determine the potential for N2O production

A wide range of microorganisms have the potential to release N2O and N2

into the environment via biological denitrification (Zumft, 1997). It is important to

identify which denitrifiers produce N2O or N2 as their end product, because it will 24 provide a basis to improve models of the nitrogen cycle to become more accurate (Galloway et al., 2004; Canfield et al., 2010). There is a need for more research on biological denitrification in cold environments as very few studies have examined this and to better predict the effect of cold temperatures on denitrification (Canion et al., 2013). The complete denitrification experiment (i) revealed all isolates produced N2O, though the organisms isolated here were all isolated under aerobic conditions. This revealed all the Lake Vida isolates are facultative anaerobes. Likewise, a majority (75%) of Lake Vida brine isolates only had the capacity for incomplete denitrification, revealing the potential for the isolates to produce N2O. Exiguobacterium sp. strain LV05br1 produced the most

N2O in this experiment, potentially as a result of quick up regulation of the cellular machinery to adapt to anoxic conditions. Exiguobacterium sp. are not known denitrifiers, though are known to be facultative anaerobes (Borsodi et al., 2010).

A search for genes in the denitrification pathway (FunGene, http://fungene.cme.msu.edu/index.spr) revealed most Exiguobacterium sp. do not contain the genes required for denitrification, but the following sp. do contain at least one gene in the pathway: Exiguobacterium sp. AT1b, nosZ, accession number ACQ71262; Exiguobacterium sibiricum 255-15, nosZ, accession number

ACB60078.

Nitrifying denitrification and denitrification are both important microbial processes which can yield N2O. Bacterial cultures isolated from Lake Vida brine

− have the capacity to produce N2O. The loss of NO3 from denitrification can occur through several different processes: assimilatory denitrification (Bazylinski and 25

Blakemore, 1983), dissimilatory denitrification (respiratory denitrification), and

− 2+ NO3 dependent Fe oxidation (Picardal, 2012); therefore two necessary criteria must be met to establish if denitrification is respiratory: (1) N2 or N2O is produced

− − − − from NO3 or NO2 and (2) the reduction of NO3 or NO2 is correlated to a growth yield increase (Mahne and Tiedje, 1995). The N2O production and growth rate determination experiment (ii) revealed the ability of each isolate to produce

N2O. N2O production was correlated to total protein production, suggesting potential for these isolates to perform respiratory. Further studies to optimize growth conditions and demonstrate electron acceptor consumption are needed to confirm respiratory denitrification however.

The Lake Vida isolates exhibited two growth relationships. Isolates LV411,

LV05br1, LV10fR510-14, and LV10fR520-8 had an excellent correlation between total protein production and N2O production (r > 0.93). Isolates LV181 and

LV10R520-3 showed a positive correlation (r > 0.83), but other processes could be occurring as well. Other proposed pathways that may be occurring include chemosynthesis (Murray et al., 2012) and fermentation. Each isolate exhibited

N2O production in a fashion similar to a typical microbial growth curve. This is not unusual as N2O production is linked to protein production, which would increase in log phase and level off in stationary phase. These results revealed

Marinobacter as the most robust genus, cultured from the brine, in terms of ability to denitrify.

Marinobacter sp. were the most abundant genus detected in Lake Vida brine (25% of rRNA gene sequences; Murray et al., 2012). Marinobacter sp. have 26 been isolated all over the world and have exhibited a high capacity to grow below freezing temperatures and at high salinities (Takai et al., 2005; Fernandez-

Linares et al., 1996), along with the capability to denitrify (Joye et al., 1996). A search for genes in the denitrification pathway (FunGene, http://fungene.cme.msu.edu/index.spr) revealed several Marinobacter sp. do contain the genes required for denitrification. Marinobacter sp. strain LV411’s generation time of 30 ± 0 hours was slow compared to other Marinobacter sp. grown at 10°C or room temperature over a range of salinities (~3.5 hours; Takai et al., 2005 and 2 to15 hours; Fernandez-Linares et al., 1996). The bacteria were grown at 10°C, because it seemed to be an optimal temperature (based on preliminary experiments) for the isolates. The temperature used here does not reflect Lake Vida’s temperature (10°C used in experiments and −13.4°C for Lake

Vida brine) and temperature is known to have an effect on denitrification rates

(Joye et al., 1996). Denitrification rates and generation times in Lake Vida brine are likely to be much lower.

Psychrobacter sp. were the second most abundant genus detected in

Lake Vida brine (13%; Murray et al., 2012). Most Psychrobacter sp. are strictly aerobic, but they have a high capacity to grow below freezing temperatures and at high salinities (Bozal 2003; Yumoto et al., 2003; Bakermans et al., 2006). A search for genes in the denitrification pathway (FunGene, http://fungene.cme.msu.edu/index.spr) revealed most Psychrobacter sp. do not contain the genes required for denitrification, but the following sp. do contain at least one gene in the pathway: Psychrobacter sp. 1501, nirK and norB, 27 accession number EGK14836; Psychrobacter sp. PRwf-1, norB, accession number ABQ94463; Psychrobacter cryohalentis K5, norB, accession number

ABE74567; Psychrobacter arcticus 273-4, norB, accession number AAZ18631.

The closest relatives to Psychrobacter sp. strain LV181 (LV40 and LV414) do not have the ability to perform denitrification (Mondino et al., 2009), suggesting the ability for denitrification could have been acquired through horizontal gene transfer. Heylen et al., (2006) revealed that for common denitrifiers (i.e.

Pseudomonas), nir gene diversity is not (always) congruent with 16S rRNA gene phylogeny and there is a high potential for environmental influences and possible horizontal gene transfer. Psychrobacter sp. have previously been shown to contain the ability for denitrification, but this is the first study to our knowledge to provide a N2O production rate and show Psychrobacter sp. as facultative anaerobic denitrifiers (Shwu-Ling et al., 1999; Zheng et al., 2011).

Marinobacter and Psychrobacter have related strains all over the world

(especially in marine environments), with several species able to survive very cold and saline environments, and several species in these genera are known to have the capability of denitrification (Bonin et al., 2002; Shwu-Ling et al., 1999;

Zheng et al., 2011). The other Lake Vida isolates did not have comparable data to evaluate the novelty of denitrification, which warrants further study.

Marinobacter sp. and Psychrobacter sp. were the most abundant bacteria in Lake Vida brine and therefore isolates LV411 and LV181 were chosen for a final biological experiment to obtain isotope data on N2O production. Isotopic analysis of N2O produced by Marinobacter sp. strain LV411 and Psychrobacter 28 sp. strain LV181 revealed a non-zero net isotope effect during the 12 day incubation period (Table 2). The net isotope effect for denitrification is the result

− of fractionation from the three or four sequential reactions, where NO3 is reduced to N2O or N2, in which isotopic fractionation occurs at each reaction step

(Sutka et al., 2008). The final day SP value of N2O was consistent with denitrification SP values previously reported (Sutka et al., 2006). The large net isotope effect observed could be due to several different outcomes. A potential cause of the fractionation observed in both isolates may be a result of NO and

N2O leaking out of the bacterial cell. This scenario best describes the difference in net isotope effect values measured for LV411 and LV181. Both sp. have a nirS and nosZ gene (Appendix A, Table 1), therefore each sp. do not have a nirK gene and may use different forms of the nor gene (used to reduce NO to N2O) or one sp. may be better at utilizing NO. Another potential cause of the high fractionation could result from the fact that in most cases less than 10% of the

− initial NO3 was consumed, therefore creating a high chance of error as a result of inflated fractionation values (Scott et al., 2004). Lastly, the results could be due to the nature of the organisms studied here; the bacteria used are psychrotrophs and grew relatively slow compared to other isolates used to obtain a net isotope effect (Sutka et al., 2006; Sutka et al., 2008). This study demonstrated nitrogen physiology and potential net isotope effect of N2O produced by psychotropic microorganisms by providing N2O production rates,

N2O isotopic values, and generation times of Lake Vida bacteria. Culture 29 independent studies are needed to assess whether or not Marinobacter and

Psychrobacter perform denitrification in situ.

− − Abiotic N2O production via the addition of Fe, NO3 , or NO2

Abiotic denitrification is not well understood, in which N2O or N2 are produced from geochemical reactions (Samarkin et al., 2010; Picardal 2012).

Abiotic denitrification has been shown to occur predominately when Fe2+ reacts

− with NO2 (Picardal 2012), but has recently been proposed to occur via a serpentinization-like reaction (Evans, 2008; Samarkin et al., 2010). Studies quantifying abiotic N2O production via a serpentinization in cryoecosystems in particular, are few, yet understanding the kinetics and natural situations that lead to such geochemistry are important to both terrestrial and planetary studies.

The results from the N2O production via chemodenitrification experiments conducted here revealed the potential for chemodenitrification in Lake Vida brine.

N2O production catalyzed by HgCl2 was suspected as many metals are shown to increase reaction rates (Buresh and Moraghan, 1976; Ottley et al., 1997) and was observed in a preliminary experiment with Lake Vida brine (Ostrom, pers

Comm.). Significant N2O production (p < 0.05; t test) occurred in treatments with

− − NO3 , NO2 , and FeSO4 added to Lake Vida brine. The treatment with FeSO4 +

NaNO2 had the greatest N2O production rate at each temperature, suggesting

2+ − Fe and NO2 are key reactants in Lake Vida brine abiotic N2O production. This

+ − is not surprising as Fe2 and NO2 are key reactants in chemodenitrification taking place in other aquatic and terrestrial environments (Picardal, 2012). Lake 30

Vida abiotic N2O production could still be occurring in the Lake Vida subsurface,

− − given that iron (256µM) and NO2 (273µM) and NO3 (1120µM) concentrations remain abundant (Murray et al., 2012). Iron oxyhydroxides could visually be observed in each treatment and more so in the treatments with higher N2O

2+ − production. The use of Fe and NO2 to produce N2O and iron oxyhydroxides is consistent with the proposed serpentinization-like reaction between Don Juan

Pond brine and iron rich rock proposed by Samarkin et al. (2010).

Significant CO2 production occurred in several treatments. There was

2+ greater CO2 production in treatments with Fe . In most cases N2O production occurred, CO2 production occurred as well, suggesting the C and N cycles in

Lake Vida brine are linked. The CO2 production may have been stimulated by the acidification of dissolved inorganic carbon (DIC) from hydrogen ions created by

2+ − − the addition of Fe , NO2 , or NO3 to Lake Vida brine (Murray et al., 2012,

Marion et al., 2013).

Our knowledge of the nitrogen cycle is continually evolving and new pathways are still being described (Galloway et al., 2004; Canfield et al., 2010).

15N enrichments are widley used in nitrogen experiments to determine production rates of various N cycling pathways. The N2O tracer experiment revealed that in

15 the N-NaNO2 treatments N2O was generated and due to no change in N2O concentration, N2O was reduced to N2 at the same rate N2O was generated. Also

15 15 the N-NaNO3 + crushed pyroxene rich rock and N-NaNO3 + sediment

15 18 treatments exhibited a decrease in N2O with no change in δ N or δ O,

15 suggesting N2O was reduced to N2. No change in N2O concentration for N- 31

− NaNO3, suggests the addition of NO3 did not stimulate N2O production, but the addition of iron through crushed pyroxene rich rock and sediment stimulated a reduction of N2O to N2 (Yu et al., 2010).

The natural abundance isotopic results revealed a N2O SP value that is consistant with abitoic N2O SP values reported in Don Juan Pond brine ranging

15 from −45‰ to 21‰ (Samarkin et al., 2010). The N2O δ N (−24.9 ± 0.4‰) and

SP (−1.5 ± 0.5‰ ) values from Lake Vida brine amended with 10mM FeSO4 +

15 10mM NaNO2 best resemble the N2O δ N Soil pore 2007 measurments of −34.5

± 0.2‰ and −35.4 ± 0‰ and SP values for 2008 flux of 1.3 ± 2.1‰ and 2008 soil pore measurments of 1.2‰ and 1.9‰ in Samarkin et al (2010). The δ18O value of −14.5 ± 0.1‰ from Lake Vida brine amended with 10mM FeSO4 + 10mM

NaNO2 is not close to any reported values (50 to 119‰) in Samarkin et al (2010).

Little fractionation was observed for δ18O and this result is reasonable given that exchange of elemental oxygen with water during N2O production is expected

(Casciotti et al., 2007), thus the exchange negates any isotope fractionation effect. The small fractionation value of 3.3‰ measured in the brine amended with

10mM FeSO4 + 10mM NaNO2 over the 88 day incubation period, demonstrates that SP is a good tracer of abiotic N2O production. The N2O SP value for Lake

Vida brine reported in this study of −12.6 ± 0.8 is different than the previously reported Lake Vida brine N2O SP value of −3.6 ± 0.3‰ (Murray et al., 2012). It is proposed the difference in SP value is due to biological and abiotic processes that took place during the two year period in which the brine was stored prior to the experiments in this study. A precipitate was apparent in the archived brine 32

(also observed right away in the experiments) used in the experiment, in which we suspect that iron oxyhydroxides may have precipitated, shifting the speciation balance and reactivity potential of the iron in the brine. N2O has been considered an excellent biosignature (Segura et al., 2005; Rauer et al., 2011), but the experimental data here reveals N2O is not a good biosignature due to the potential for abiotic N2O production and abiotic N2O SP values being within the range of biological values. The results from the abiotic experiments expands the current understanding of abiotic N2O production and SP values from abiotic sources, along with demonstrating the potential for abiotic N2O production in

Lake Vida brine.

V. Conclusion

N2O SP values can be used to distinguish between biological sources of

N2O (i.e. nitrification and denitrification), but no literature is present describing the potential for N2O SP values to distinguish between biological and abiotic processes. Lake Vida bacterial isolates demonstrated the ability to denitrify, in which four of them showed the potential for complete denitrification. The potential, therefore exists that these isolates, Marinobacter sp., in particular, could influence the Lake Vida brine nitrogen biogeochemistry. Abiotic N2O production is not well studied and this study provides experimental data demonstrating abiotic N2O production in Lake Vida brine at in situ and room

2+ − − temperatures (using Fe , NO3 , and NO2 as reactants). The N2O SP values 33 from biological and abiotic sources in Lake Vida were not significantly different and therefore cannot be used to discern a difference between the two potential sources (Marinobacter sp. strain LV411 was 1.2 ± 0.7‰, Psychrobacter sp. strain

LV181 was −1.9 ± 0.8‰ and abiotic treatment with 10mM FeSO4 + 10mM

NaNO2 was −1.5 ± 0.5‰). A comparison of N2O SP values, along with evidence of possible biological and abiotic N2O production, suggests the source of N2O in

Lake Vida brine could be the result of either process or a combination of both biological and abiotic sources.

VI. References

Abrajano, T. A., Sturchio, N. C., Kennedy, B. M., Lyon, G. L., Muehlenbachs, K., and Böhlke, K. J. (1990), Geochemistry of reduced gas related to serpentinization of the Zambales Ophiolite, Philippines, Appl. Geochem., 5, 625–630.

Bakermans, C., Ayala-del-Rıo, H. L., Ponder M. A., Vishnivetskaya, T., Gilichinsky, D., Thomashow, M. F., and Tiedje, J. M. (2006). Psychrobacter cryohalolentis sp. nov. and Psychrobacter arcticus sp. nov., isolated from Siberian permafrost. Int J Syst Evol Micr, 56, 1285–1291. doi:10.1099/ijs.0.64043-0.

Bazylinski, D. A., and Blakemore, R. P. (1983). Denitrification and assimilatory nitrate reduction in Aquaspirillum magnetotacticum. Appl Environ Microb, 46(5), 1118–1124.

Bixler, H. and Coberly, C. W. (1953). Wulff process acetylene. Ind Eng Chem, 45 (12), 2596–2606. DOI: 10.1021/ie50528a018.

Bonin, P., Tamburini, C., and Michotey, V. (2002). Determination of the bacterial processes which are sources of nitrous oxide production in marine samples. Water Res, 36, 722–732. Borsodi, A. K., Kiss, R. I., Cech, G., Vajna, B., Tôth, E. M., and Márialigeti, K. (2010). Diversity and activity of cultivable aerobic planktonic bacteria of a saline lake located in Sovata, Romania. Folia Microbiol, 55(5), 461–466. 10.1007/s12223-010-0077-7. Bouwman, A. E. (ed.) (1990). Soils and the greenhouse effect. John Wiley, Chichester.

Bouwman, A. F., Boumans, L. J. M., & Batjes, N. H. (2002). Emissions of N2O and NO from fertilized fields: Summary of available measurement data. Global Biogeochemcy, 16(4), 1058-1071.

Bozal, N. Montes, J. M., Tudela E., and Guinea, J. (2003). Characterization of several Psychrobacter sp. isolated from Antarctic environments and description of Psychrobacter luti sp. nov. and Psychrobacter fozii sp. nov. Int J Syst Evol Micr, 53(4), 93-1100. doi:10.1099/ijs.0.02457-0.

Buresh, R. J. and Moraghan, J. T. (1976). Chemical reduction of nitrate by ferrous iron. J Environ Qual, 5, 320–325. 35

Canfield, D. E.,Glazer, A. N., Falkowski, P. G. (2010). The Evolution and Future of Earth's Nitrogen Cycle. Science, 330,192. DOI: 10.1126/science.1186120.

Canion, A., Prakash, O., Green, S. J., Jahnke, L., Kuypers, M. M. M., and Kostka, J. E. (2013). Isolation and physiological characterization of psychrophilic denitrifying bacteria from permanently cold Arctic fjord sediments (Svalbard, Norway). Environ Microbiol,15(5), 1606–1618.

Casciotti, K. L., Böhlke, J. K., McIlvin, M. R., Mroczkowski, J., and Hannon, J. E. (2007). Oxygen isotopes in nitrite: analysis, calibration, and equilibration. Anal. Chem, 79, 2427-2436.

Charlou, J. L., Donval, J. P., Fouquet, Y., Jean-Baptiste, P., and Holm, N. (2002), Geochemistry of high H2 and CH4 vent fluids issuing from ultramafic rocks at the Rainbow hydrothermal field (36°140’N, MAR), Chem. Geol., 191, 345–359.

Clarkson, P. D. (1981). Geology of the Shackleton Range IV. The Dolerite Dykes. Brit Antarct Surv B, 53, 210-212.

Evans, B. W. (2008). Control of the Products of Serpentinization by the 2+ Fe Mg−1 Exchange Potential of Olivine and Orthopyroxene. J Petrol, 49, 1873-1887.

Fernandez-Linares, L., Acquaviva, M., Bertrand, J-C., Gauthier (1996). Effect of sodium chloride concentration on growth and degradation of eicosane by the marine halotolerant bacterium Marinobacter hydrocarbonoclasticus. Syst Appl Microbiol, 19(1),113–121. Galloway, J. N., Dentener, F. J., Capone, D. G., Boyer, E. W., Howarth, R. W., Seitzinger, S. P., Asner, G. P., Cleveland, C. C., Green, P. A., Holland, E. A., Karl, D. M., Michaels, A. F., Porter, J. H., Townsend, A. R., and Vöosmarty, C. J. (2004) Nitrogen cycles: past, present, and future. Biogeochemistry, 70(2), 153-226.

Joye, S. B., Smith S. V., Hollibaugh, J. T., Paerl, H. W. (1996). Estimating denitrification rates in estuarine sediments: a comparison of stoichiometric and acetylene based methods. Biogeochemistry, 33(3), 197-215.

Lovelock, J. E. (1958). A sensitive detector for gas chromatography. J Chromatogr A, 1, 35-46. 36

Machida, T., Nakazawa, T., Fujii, Y., Aoki, S., & Watanabe, O. (1995). Increase in the atmospheric nitrous oxide concentration during the last 250 years. Geophys Res Lett, 22(21), 2921-2924.

Mahne, I., & Tiedje, J. M. (1995). Criteria and methodology for identifying respiratory denitrifiers. Appl Environ Microb, 61, 1110-1115.

Marion, G. M., Murray, A. E., Wagner, B., Fritsen, C. H., Kenig, F., & Doran, P. T. (2013). Carbon sequestration and release from antarctic lakes: Lake Vida and West Lake Bonney (McMurdo Dry Valleys). Aquat Geochem, 19, 135- 145.

Mariotti, A., Germon, J. C., Hubert, P., Kaiser, P., Letolle, R., Tardieux, A., and Tardieux, P. (1981). Experimental determination of nitrogen kinetic isotope fractionation: some principals; Illustration for the denitrification and nitrification processes. Plant Soil, 62, 413-430.

Mondino, L. J., Asao, M., and Madigan, M. T. (2009). Cold-active halophilic bacteria from the ice-sealed Lake Vida, Antarctica. Arch Microbiol, 191(10), 785-790. doi: 10.1007/s00203-009-0503-x.

Mosier, A.R. and Klemedtsson, L. (1994). Measuring denitrification in the field. In: Weaver, R. W. & Soil Science Society of America. (Eds.) Methods of soil analysis. part 2, microbiological and biochemical properties. Madison, Wis: SSSA.

Müntener, O. (2010). Serpentine and serpentinization: A link between planet formation and life. Geology, 38, 959-960. doi: 10.1130/focus102010.1.

Murray, A. E., Kenig, F., Fritsen, C. H., McKay, C. P., Cawley, K. M., Edwards, R., Doran, P. T. (2012). Microbial life at −13°C in the brine of an ice-sealed Antarctic lake. P Natl Acad Sci USA, 109(50), 20626-20631. doi: 10.1073/pnas.1208607109.

Oremland, R. S., & Taylor, B. F. (1975). Inhibition of methanogenesis in marine sediments by acetylene and ethylene: validity of the acetylene reduction assay for anaerobic microcosms. Appl Microbiol, 30, 707-709.

Ostrom, N. E., & Ostrom, P. H. (2011). The isotopomers of nitrous oxide: analytical considerations and application to resolution of microbial production pathways. Handbook of Environmental Isotope Geochemistry (pp. 453-476), Amsterdam: Elsevier Scientific Pub. Co.

Ostrom, N. E., Pitt, A., Sutka, R., Ostrom, P. H., Grandy, A. S., Huizinga, K. M., & Robertson, G. P. (2007). Isotopologue effects during N2O reduction in 37

soils and in pure cultures of denitrifiers. J Geophys Res-Biogeo, 112(G2), 2005-2012.

Ottley, C. J., Davison, W., and Edmunds, W. M. (1997). Chemical catalysis of nitrate reduction by iron (II). Geochim Cosmochim Acta, 61, 1819–1828.

Payne, W. J. (1973). Reduction of nitrogenous oxides by microorganisms. Bacteriol Rev, 37(4), 409-452.

Pérez, T., Garcia-Montiel, D., Trumbore, S., Tyler, S., Camargo, P., Moreira, M., Piccolo, M., Cerri, C. (2006). Nitrous oxide nitrification and denitrification 15N Enrichment factors from Amazon forest soils. Ecol Appl, 16(6), 2153– 2167.

Picardal, F. (2012). Abiotic and microbial interactions during anaerobic − transformations of Fe(II) and NOx . Front. Microbio. 3, 112. doi: 10.3389/fmicb.2012.00112.

Postma, D. (1990). Kinetics of nitrate reduction by detrital Fe(II)-silicates. Geochim Cosmochim Acta, 54, 903–908.

Powell, E. O. (1956). Growth rate and generation time of bacteria, with special reference to continuous culture. J Gen Microbial, 15, 492-511.

Prinn, R., Cunnold, D., Rasmussen, R., Simmonds, P., Alyea, F., Crawford, A., Rosen, R. (1990). Atmospheric emissions and trends of nitrous oxide deduced from 10 years of ALE–GAGE data. J Geophys Res, 95, 18369– 18385.

Priscu, J. C. (1997). The biogeochemistry of nitrous oxide in permanently ice- covered lakes of the McMurdo Dry Valleys, Antarctica. Glob Change Biol, 3, 301-315.

Priscu, J. C., Downes, M. T., & McKay, C. P. (1996). Extreme supersaturation of nitrous oxide in a poorly ventilated Antarctic lake. Limnol Oceanogr, 41(7), 1544-1551.

Priscu, J. C., Ward, B. B., & Downes, M. T. (1993). Water column transformations of nitrogen in Lake Bonney, a perennially ice-covered antarctic lake. Antarct J US, 28, 237-239.

Rakshit, S., Matocha, C. J., & Coyne, M. S. (2008). Nitrite reduction by siderite. Soil Sci Soc AM J, 72, 1070-1077. 38

Rasmussen, R. A., & Khalil, M. A. (1986). Atmospheric trace gases: trends and distributions over the last decade. Science, 232(4758), 1623-1624. doi: 10.1126/science.232.4758.1623.

Rauer, H., Gebauer, S., Paris, P. V., Cabrera, J., Godolt, M., Grenfell, J. L., Schreier, F. (2011). Potential biosignatures in super-earth atmospheres. Astron Astrophys, 529, 1-25.

Ronen, D., Magaritz, M., & Almon, E. (1988). Contaminated aquifers are a forgotten component of the global N2O budget. Nature, 335, 57-59.

Samarkin, V., Madigan, M., Bowles, M., Casciotti, K., Priscu, J., McKay, C., & Joye, S. (2010). Abiotic nitrous oxide emission from the hypersaline Don Juan Pond in Antarctica. Nat Geosci, 3, 341-344.

Sano, Y., Urabe, A., Wakita, H., and Wushiki, H. (1993), Origin of hydrogen- nitrogen gas seeps, Oman, Appl. Geochem., 8, 1–8.

Schmidt, H. L., Werner, R. A., Yoshida, N., & Well, R. (2004). Is the isotopic composition of nitrous oxide an indicator for its origin from nitrification or denitrification? A theoretical approach from referred data and microbiological and enzyme kinetic aspects. Rapid Commun Mass Sp, 18(18), 2036-2040. doi: 10.1002/rcm.1586.

Scott, K., M., Lu, X., Cavanaugh, C. M., and Liu, J. S. (2004). Optimal methods for estimating kinetic isotope effects from different forms of the Rayleigh distillation equation. Geochim Cosmochim Acta, 68(3), 433–442. doi:10.1016/S0016-7037(03)00459-9. Segura, A., Kasting, J. F., Meadows, V., Cohen, M., Scalo, J., Crisp, D., Tinetti, G. (2005). Biosignatures from Earth-like planets around M dwarfs. Astrobiology, 5(6), 706-725. doi: 10.1089/ast.2005.5.706. Shwu-Ling, P., Nyuk-Min, C., Chei-Hsiang, C. (1999). Potential applications of aerobic denitrifying bacteria as bioagents in wastewater treatment. Bioresource Techmol, 68(2), 179-185.

Sutka, R. L., Adams, G. C., Ostrom, N. E., & Ostrom, P. H. (2008). Isotopologue fractionation during N2O production by fungal denitrification. Rapid Commun Mass Sp, 22(24), 3989-3996.

Sutka, R. L., Ostrom, N. E., Ostrom, P.H., Breznak, J. A., Gandhi, H., Pitt, A. J., Li, F. (2006). Distinguishing nitrous oxide production from nitrification and denitrification on the basis of isotopomer abundances. Appl Environ Microb, 72(1), 638-644. doi: 10.1128/AEM.72.1.638-644.2006. 39

Tai, Y. L., & Dempsey, B. A. (2009). Nitrite reduction with hydrous ferric oxide and Fe(II): stoichiometry, rate, and mechanism. Water Res, 43(2), 546- 552. doi: 10.1016/j.watres.2008.10.055.

Takai, K., Moyer, C. L., Miyazaki, M., Nogi, Y., Hirayama, H., Nealson, K. H., Horikoshi, K. (2011). Marinobacter alkaliphilus sp. nov., a novel alkaliphilic bacterium isolated from subseafloor alkaline serpentine mud from Ocean Drilling Program Site 1200 at South Chamorro Seamount, Mariana Forearc. Extremophiles, 9, 17–27. DOI 10.1007/s00792-004-0416-1. Taylor, R. E. (1987). Radiocarbon dating: An Archeological perspective. Academic Press, New York. Teissier, S., & Torre, M. (2002). Simultaneous assessment of nitrification and denitrification on freshwater epilithic biofilms by acetylene block method. Water Res, 36(15), 3803-3811.

Thauer, R. K., Jungermann, K., & Decker, K. (1977). Energy conservation in chemotrophic anaerobic bacteria. Bacteriol Rev, 41(1), 100-180.

Thorn, K. A., & Mikita, M. A. (2000). Nitrite fixation by humic substances: Nitrogen-15 nuclear magnetic resonance evidence for potential intermediates in chemodenitrification. Soil Sci Soc Am J, 64, 568-582.

Toyoda, S., Mutobe, H., Yamagishi, H., Yoshida, N., & Tanji, Y. (2005). Fractionation of N2O isotopomers during production by denitrifier. Soil Biol Biochem, 37, 1535–1545.

Toyoda, S., Yamamoto, S., Arai, S., Nara, H., Yoshida, N., Kashiwakura, K., & Akiyama, K. (2008). Isotopomeric characterization of N2O produced, consumed, and emitted by automobiles. Rapid Commun Mass Sp, 22(5), 603-612. doi: 10.1002/rcm.3400.

Toyoda, S., & Yoshida, N. (1999). Determination of nitrogen isotopomers of nitrous oxide on a modified isotope ratio mass spectrometer. Anal Chem, 71, 4711–4718.

Ward, B. B., and Priscu, J. C. (1997). Detection and characterization of denitrifying bacteria from a permanently ice-covered Antarctic Lake. Hydrobiologia, 347, 57-68.

Ye, R. W., Fries, M. R., Bezborodnikov, S. G., Averill, B. A., & Tiedje, J. M. (1993). Characterization of the structural gene encoding a copper- containing nitrite reductase and homology of this gene to DNA of other denitrifiers. Appl Environ Microbiol, 59(1), 250-254. 40

Yoshida, N., & Toyoda, S. (2000). Constraining the atmospheric N2O budget from intramolecular site preference in N2O isotopomers. Nature, 405(6784), 330-334. doi: 10.1038/35012558.

Yu, K., Seo, D.-C., & DeLaune, R. D. (2010). Incomplete acetylene inhibition of nitrous oxide reduction in potential denitrification assay as revealed by using 15N-nitrate tracer. Commun Soil Sci Plan, 41(18), 2201-2210. doi: 10.1080/00103624.2010.504800.

Yumoto, I., Hirota, K., Sogabe, Y., Nodasaka, Y., Yokota, Y., Hoshino, T. (2003). Psychrobacter okhotskensis sp. nov., a lipase-producing facultative psychrophile isolated from the coast of the Okhotsk Sea. Int J Syst Evol Micr, 53(6), 1985-1989. doi: 10.1099/ijs.0.02686-0.

Zheng, H. Y., Liu, Y., Sun, G. D., Gao, X. Y., Zhang, Q. L., Liu, Z. P., (2011). Denitriﬁcation characteristics of a marine origin psychrophilic aerobic denitrifying bacterium. J Environ Sci, 23(11): 1888–1893. doi: 10.1016/S1001-0742(10)60615-8.

Zolotov, M.Y. and Shock, E.L. (2004) A model for low temperature biogeochemistry of sulfur, carbon, and iron on Europa. J. Geophys. Res. 109, 2003JE002194.

Zumft, W. G. (1997). Cell biology and molecular basis of denitrification. Microbiol Mol Biol Rev, 61(4), 533-616.

VII. Supplemental Figures and Tables

Figure S1. Net Isotope Effect for Marinobacter sp. strain LV411 and Psychrobacter sp. strain LV181. The x-axis (−(f*lnf)/(1−f)) represents the − proportion of NO3 not consumed. The y-axis for each graph is the isotopic data measured via MS. The slopes represent the net isotope effect. A polynomial regression line best fit the data from biological isolate Marinobacter sp. strain LV411, there the regression equation is present instead of slope.

Figure S2. Fractionation factors for abiotic N2O produced from Lake Vida 15 15 α brine treated with 10mM FeSO4 + 10mM NaNO2. (A) δ N-N2O, (B) δ N -N2O, 15 β 18 (C) δ N -N2O, (D) δ O-N2O, and (E) Site Preference (SP).The x-axis − (−(f*lnf)/(1−f)) represents the proportion of NO3 not consumed. The y-axis for each graph is the isotopic data measured via MS. The slopes represent the fractionation factor.

Figure S3. Abiotic N2O production rate from natural isotope abundance as a function of chemodenitrification experiment. N2O was measured at five time points over an 88 day incubation period. The y-axis represents N2O in µM. The x- axis represents time in days.

y = -65.6x + 46.2 -(f*lnf)/(1-f) A 0 R² = 0.6 -5 -10 δ15N 411

‰) -15 δ15N 181 -20 N ( N -25 15 y = -352.9x + 329.5 δ -30 R² = 0.9 0.85 0.9 0.95 1 -(f*lnf)/(1-f) y = -110.1x + 87.1 B 0 R² = 0.7 -5

-10 δ15α 411 (‰) -15 α -20 N δ15α 181

15 -25

δ -30 -35 y = -528.9x + 500.8 R² = 0.8 0.85 0.9 0.95 1 C -(f*lnf)/(1-f) y = -21.1x + 5.3 0 R² = 0.1 -5 δ15β 411

-10 (‰)

β -15 δ15β 181 N

15 -20

δ -25 y = -176.9x + 158.3 0.85 0.9 0.95 1 R² = 0.8

-(f*lnf)/(1-f) y = -38.7x + 62.8 D 35 R² = 0.3 30 25 20 d18O 411 (‰) 15

O O 10 d18O 181 18

δ 5 y = -185.1x + 208.9 0 R² = 0.7 0.85 0.9 0.95 1 -(f*lnf)/(1-f) y = -1317.1x2 + 2402.1x - 1093.2 E 5 R² = 0.9 0

-5 SP 411 SP 181 -10 SP (‰) SP y = -352.0x + 342.5 -15 R² = 0.7 0.85 0.9 0.95 1

y = -37.8x + 9.1 A -24 -25 R² = 0.9 -26 -27

N (‰) N -28

-29 δ 0.9 0.92 0.94 0.96 0.98 1 −(f*lnf)/(1−f) y = -3.6x - 12.7 B -14 R² = 0.0

-15

-16

-17 O (‰) O

18 δ -18 0.9 0.92 0.94 0.96 0.98 1 −(f*lnf)/(1−f) -25 y = -36.2x + 7.1 C -26 R² = 0.7

-27

-28 (‰)

α -29 N -30 15 δ 0.9 0.92 0.94 0.96 0.98 1

−(f*lnf)/(1−f) D y = -39.4x + 11.1 -24 R² = 0.9 -25

-26

-27

(‰)

β -28 N -29

15 δ 0.9 0.92 0.94 0.96 0.98 1 −(f*lnf)/(1−f)

E y = 3.3x - 4.0 2 R² = 0.0 1 0 -1

-2 SP (‰) SP -3 0.9 0.92 0.94 0.96 0.98 1 −(f*lnf)/(1−f)

y = 10.4x - 31.7 10mM FeSO4 + 10mM NaNO2 R² = 1.0 1000 900 800 700 600

500

O O (µM) 2

N 400 300 200 100 0 0 10 20 30 40 50 60 70 80 90 Time (days)

Table S1. Lake Vida bacterial isolates. Isolates were collected from three labs and used in this study. The accession number, genus, redundancy, lab the isolate came from, and nearest neighbor in GeneBank are listed.

VIII. Appendices

Appendix A: Detection of genes in the denitrification pathway.

Respiratory denitrification is a complex process involving four separate − − sets of genes including (1) nas, nar, or nap for reduction of NO3 to NO2 , (2) nir − for reduction of NO2 to NO, (3) nor for reduction of NO to N2O, and (4) nos for reduction of N2O to N2 (Zumft, 1997). The culture collection used in this study included: Marinobacter sp. strain LV10S, Marinobacter sp. strain LV411, Marinobacter sp. strain LV10MA510-1, Marinobacter sp. strain LV10R510-2, Marinobacter sp. strain LV10R510-5, Marinobacter sp. strain LV10R510-8, Marinobacter sp. strain LV10R510-11A, Psychrobacter sp. strain LV181, Psychrobacter sp. strain LV40, Psychrobacter sp. strain LV414, Psychrobacter sp. strain LV10R520-6, Exiguobacterium sp. strain LV05br1, Arthrobacter sp. strain LV10R520-3, Kocuria sp. strain LV10fR520-8, Microbacterium sp. strain LV10fR510-14, and Microbacterium sp. strain LV10fR510-15. The culture collection was screened using specific primers as follow: NirS1F and NirS6R for nirS gene (Braker et al., 1998), NirK1F and NirK5R for nirK gene (Braker et al., 1998), Nos661F and Nos1773R for typical nosZ gene (Scala and Kerkhof, 1998), and DGF and DGR for the atypical nosZ gene (Sanford et al., 2012). PCR Conditions for each primer employed followed the respective article. After amplification reactions, the amplicons were run on a 1.2% agarose gel by electrophoresis and bands of approximately 890 bp (nirS), 514 bp (nirK), 880 bp (typical or atypical nosZ) were extracted from the gel, purified, and sequenced. The QIAquick Gel Extraction Kit (Qiagen, Valencia, CA) was used for purification of the bands extracted and sequences were read on a Prism 3730 DNA Analyzer (Applied Biosystems, Carlsbad, CA, USA). The nir and nosZ denitrification genes were targeted as these genes have the most literature available and yield the most information. If the nir gene is detected, then the isolate is considered to have the potential for denitrification and if the nosZ gene is detected, then the isolate is considered to have the potential for complete denitrification. The nirS gene was detected in seven of the sixteen isolates (Appendix A, Table 1): LV10S, LV411, LV10R510-8, LV181, LV40, LV414, and LV10fR510-15. The nosZ gene was confirmed in eight of the sixteen isolates (Appendix A, Table 1): LV10S, LV411, LV10MA510-1, LV10R510-5, LV10R510-8, LV10R510-11A, LV181, and LV40. Five out of the sixteen isolates had a complete set of denitrification genes (nir and nosZ): LV10S, LV411, LV10R510-8, LV181, and LV40.

Appendix A, Table 1. Lake Vida bacterial isolates and identified genes in the denitrification pathway. The Lake Vida bacterial isolate list is a collection from three labs (Mondino et al., 2009; Kuhn et al., in prep; Fritsen pers comm.). The collection spans six genera. Complete denitrifier status was determined based of the results from the Complete Denitrification experiment (i).

Genes in the denitrification pathway Isolate Genus Complete denitrifier nirK nirS nosZ LV10S Marinobacter X X LV411 Marinobacter X X X LV10MA510-1 Marinobacter X LV10R510-2 Marinobacter LV10R510-5 Marinobacter X LV10R510-8 Marinobacter X X LV10R510-11A Marinobacter X X LV181 Psychrobacter X X X LV40 Psychrobacter X X LV414 Psychrobacter X LV10R520-6 Psychrobacter LV05br1 Exiguobacterium X LV10R520-3 Arthrobacter LV10fR520-8 Kocuria LV10fR510-14 Microbacterium LV10fR510-15 Microbacterium X

A comparison of denitrification genes to the results from the complete denitrification experiment (i) revealed nine of the sixteen isolates that produced N2O, did not harbor a nir gene. The seven isolates that a nir gene was detected in had the nirS gene and none had the nirK gene. This is not surprising as only Marinobacter ELB-17 has been identified to contain a nirK, with most other Marinobacter sp. containing a nirS (Jones et al., 2008). Three of the seven Marinobacter sp. strains (LV10S, LV411, and LV10R510-8) contained nirS genes, while four were negative and of those four, three (LV10MA510-1, LV10R510-2, and LV10R510-11A) are closely related to Marinobacter ELB-17. Five isolates with a confirmed nosZ gene did not produce more N2O in the presence of C2H2, suggesting that gene was not expressed. Exiguobacterium sp. strain LV05br1 was the only complete denitrifier that did not harbor a nosZ gene. Atypical primers (listed above) were employed to try and detect alternative nosZ forms. Atypical primers were designed to be degenerative and detect nosZ genes that are not typically detected (Sanford et al., 2012). These primers are more degenerative and therefore would target more sequence variations in the nosZ gene (Sanford et al., 2012). The atypical primers employed did not detect any additional nosZ genes. The lack of nosZ gene detection in these isolates further shows a possible variation in the nosZ gene, in which the typical and atypical primers employed did not target the gene. Further research is needed to design more nosZ primers to account for the nosZ variability. The information presented 48 on functional genes in the respiratory denitrification pathway is limited to confirmation by PCR and sequencing of PCR products. These experimental results revealed four isolates that had the capacity for complete denitrification, but molecular gene surveys only detected a complete set of denitrification genes in two isolates.

References

Braker, G., Fesefeldt, A., & Witzel, K. P. (1988). Development of PCR Primer Systems for Amplification of Nitrite Reductase Genes (nirK and nirS) To Detect Denitrifying Bacteria in Environment Samples. APPL ENVIRON MICROB, 64(10), 3769-3775.

Jones, C. M., Stres, B., Rosenquist, M., and Hallin, S. (2008). Phylogenetic Analysis of Nitrite, Nitric Oxide, and Nitrous Oxide Respiratory Enzymes Reveal a Complex Evolutionary History for Denitriﬁcation. Mol Biol Evol, 25(9), 1955-1966. doi: 10.1093/molbev/msn146.

Mondino, L. J., Asao, M., and Madigan, M. T. (2009). Cold-active Halophilic bacteria from the ice-sealed Lake Vida, Antarctica. Arch Microbiol, 191(10), 785-790. doi: 10.1007/s00203-009-0503-x.

Sanford, R. A., Wagner, D. D., Wu, Q., Chee-Sanford, J. C., Thomas, S. H., Cruz-García, C., Rodríguez, G., Massol-Deyá, A., Krishnani, K. K., Ritalahti, K. M., Nissen, S., Konstantinidis, K. T., and Löffler, F. E. (2012). Unexpected nondenitrifier nitrous oxide reductase gene diversity and abundance in soils. P Natl Acad Sci USA, 109(48), 19709-19714. doi: 10.1073/pnas.1211238109.

Scala, D. J., & Kerkhof, L. J. (1998). Nitrous oxide reductase (nosZ) gene- specific PCR primers for detection of denitrifiers and three nosZ genes from marine sediments. FEMS Microbiol Lett, 162(1), 61-68.

Zumft, W. G. (1997). Cell biology and molecular basis of denitrification. Microbiol Mol Biol Rev, 61(4), 533-616.

Appendix B: Salinity and temperature tolerance of Lake Vida isolates. Lake Vida is considered one of the most extreme environments on earth, with an unusual geochemistry (Murray et al., 2012). Bacteria were revitalized from Lake Vida brine samples brought back from the 2005 expedition (Mondino et al., 2009). The enrichment cultures revealed aerobic, heterotrophic bacteria from the genera Marinobacter and Psychrobacter (Mondino et al., 2009). Analysis of these bacteria under several growth conditions and their morphology provided insight into how microbes survive in hypersaline cryoecosystems. The Marinobacter and Psychrobacter sp. were halotolerant or halophilic, psychrotolerant or psycrophilic, and grew in a range of temperatures from −8 to 25°C. Marinobacter and Psychrobacter are genera known to be able to survive in hypersaline and very cold environments (Bakermans et al., 2006; Zhang et al., 2008). In 2010 another field expedition went to Lake Vida and collected brine samples (Murray et al., 2012). Culture efforts of Kuhn et al. (in prep) revealed six genera of bacteria: Marinobacter, Psychrobacter, Exiguobacterium, Arthrobacter, Kocuria, and Microbacterium. The aim of this work was to characterize each Lake Vida bacterial isolate in terms of growth response to temperature and salinity gradients. Lake Vida cultures were collected from Mondino et al. (2009), Christian Fritsen (pers comm.), and Kuhn et al. (in prep), which together constituted the Lake Vida isolate collection. The 16 distinct isolates were grown under several salinities to determine tolerance to salinity. The collection was grown on R2A + 5% NaCl aerobically and incubated at 10°C. Several generations of each isolate was grown to allow acclimation of the microbes to the growth conditions and provide backup cultures. Cultures were also persevered in 20% glycerol and stored in cryotubes in a −80°C freezer. The salinity treatments were made by adding NaCl to premade R2A broth (Teknova Hollister, CA). When the medium was turbid (determined visually) for each isolate, 100µL of inoculum was added to 5mL of R2A and R2A + 10% NaCl and incubated at 10°C. The isolates were checked every three days for up to 90 days. Once turbidity was visually apparent, the isolate was considered to be able to growth at that salinity and then 100µL of inoculum was added to R2A + 15% NaCl. The isolates were stepped up gradually in percentage of salinity: 5, 10, 13, 16, 20 and 24%. Salinity was confirmed using a LW Scientific Refractometer. Each isolate from the Lake Vida culture collection was incubated under several temperatures to determine levels of tolerance to temperature. All cultures were grown in R2A + 5% NaCl aerobically and incubated at 10°C. Once several generations were grown, 100µL was added to 5mL of R2A + 5% NaCl and incubated at room temperature (~22°C) and 0°C. The isolates incubated at room temperature where placed in a plastic container with a beaker of water to prevent the cultures from drying out. The isolates were checked every three days for up to 90 days. Once turbidity was visually apparent, the isolate was considered to be able to growth at that temperature and then 100µL of inoculum was added to R2A + 10% NaCl and incubated at −9.9°C and 100µL of inoculum was added to 51

R2A + 5% NaCl and incubated at 26°C (same conditions to prevent drying as before). The cultures were then incubated at 43°C (same conditions as for 26°C incubation). Salinity experiment All isolates grew in a salinity range of 5% to 20%, six showed growth at 24% (LV411, LV10S, LV181, LV414, LV40, and LV10fR510-15) and three (LV411, LV181, and LV414) were unable to grow without salt (Appendix B, Table 1). Isolates LV411, Psychrobacter sp. strain LV181, and Psychrobacter sp. strain LV414 are considered halophiles based on their ability to grown in high salt concentration and inability to grow without salt; every other isolate was labeled as halotolerant, because although they could grow at salinities up to seven times saltier than the ocean, they could grow without salt. These results here varied from the salinities results reported in Mondino et al. (2009), in that LV40 could grow in the absence of salt and each isolate could grow in salinities greater than 15%. The salinity optimum for this study was estimated to be 5%, based off visual observations during incubations. Many of the isolates from Lake Vida brine revealed novel salinity limits. Kocuria is a rare species that has only recently been cultured from a marine environment and exhibits a high tolerance to salinity up to 15% (Kim et al., 2004). Microbacterium is a genus of bacteria that are not known to tolerate even moderate levels of salt (Kageyama, et al., 2007). Recently Microbacterium sediminis was described to grow in up to 8% salinity (Yu et al., 2013). Arthrobacter arilaitensis, Microbacterium foliorum, Exiguobacterium sp., Marinobacter hydrocarbonoclasticus, and Psychrobacter aquaticus were shown to grow between 5 to 15% salinity and were isolated from saline Lake Red in Sovata, Romania (Borsodi et al., 2010). These results further define Marinobacter and Psychrobacter sp. as bacteria with the capability to survive in extreme hypersaline environments and expand upon the abilities of Exiguobacterium, Kocuria, Arthrobacter, and Microbacterium to survive in extreme environments.

Appendix B Table 1. Salinity growth range of Lake Vida bacterial isolates. An “X” marks each salinity the Lake Vida bacterial isolates were able to grow in. Growth was determined turbidimetrically by visual inspection.

Salinity (%) Isolate ID Genus 0 5 10 13 16 20 24 LV10R520-3 Arthrobacter X X X X X X LV05br1 Exiguobacterium X X X X X X LV10fR520-8 Kocuria X X X X X X LV10R510-5 Marinobacter X X X X X X LV10R510-8 Marinobacter X X X X X X LV10R510-11A Marinobacter X X X X X X LV10MA510-1 Marinobacter X X X X X X LV10R510-2 Marinobacter X X X X X X LV411 Marinobacter X X X X X X LV10S Marinobacter X X X X X X X LV10fR510-15 Microbacterium X X X X X X LV10fR510-14 Microbacterium X X X X X X X LV10R520-6 Psychrobacter X X X X X X LV414 Psychrobacter X X X X X X LV181 Psychrobacter X X X X X X LV40 Psychrobacter X X X X X X

Temperature experiment All isolates grew from −9.9°C to 22°C, and only four (LV10R520-3, LV10MA510-1, LV10R510-8, and LV10fR510-14) showed growth at 43°C (Appendix B, Table 2). Microbacterium sediminis has been shown to grow from 4 to 50°C (Yu et al., 2013). Exiguobacterium sibiricum has been shown to grow from −2.5 to 40°C (Rodrigues et al., 2006). Arthrobacter agilis has been shown to grow from −5 to 40°C (Steven et al., 2007; Fong et al., 2001). Kocuria sp. isolated from the Antarctic have been shown to grow in media from 5 to 30°C (Reddy et al., 2003). Each Lake Vida isolate was able to grow in a colder condition than previously reported. These results show the high resilience and unique properties of Lake Vida bacterial isolates by exhibiting growth over a 30 to 50°C temperature gradient.

Appendix B, Table 2. Temperature growth range of Lake Vida bacterial isolates. An “X” marks each temperature the Lake Vida bacterial isolates were able to grow in. Growth was determined turbidimetrically by visual inspection.

Temperature (°C) Isolate ID Genus −9.9 5 0 10 22 26 43 LV10R520-3 Arthrobacter X X X X X X X LV05br1 Exiguobacterium X X X X X LV10fR520-8 Kocuria X X X X X LV10R510-5 Marinobacter X X X X X X X LV10R510-8 Marinobacter X X X X X X X LV10R510-11A Marinobacter X X X X X LV10MA510-1 Marinobacter X X X X X LV10R510-2 Marinobacter X X X X X LV411 Marinobacter X X X X X LV10S Marinobacter X X X X X LV10fR510-15 Microbacterium X X X X X LV10fR510-14 Microbacterium X X X X X X X LV10R520-6 Psychrobacter X X X X X LV414 Psychrobacter X X X X LV181 Psychrobacter X X X X LV40 Psychrobacter X X X X X

References Bakermans, C., Ayala-del-Rıo, H. L., Ponder M. A., Vishnivetskaya, T., Gilichinsky, D., Thomashow, M. F., and Tiedje, J. M. (2006). Psychrobacter cryohalolentis sp. nov. and Psychrobacter arcticus sp. nov., isolated from Siberian permafrost. Int J Syst Evol Micr 56, 1285–1291. doi:10.1099/ijs.0.64043-0. Borsodi, A. K., Kiss, R. I., Cech, G., Vajna, B., Tôth, E. M., and Márialigeti, K. (2010). Diversity and activity of cultivable aerobic planktonic bacteria of a saline lake located in Sovata, Romania. Folia Microbiol, 55(5), 461–466. 10.1007/s12223-010-0077-7. Fong, N., Burgess, M., Barrow, K., and Glenn, D. (2001). Carotenoid accumulation in the psychrotrophic bacterium Arthrobacter agilis in response to thermal and salt stress. Appl Microbiol Biot, 56(5-6), 750-756. 10.1007/s002530100739. Kageyama, A., Takahashi, Y., Matsuo, Y., Adachi, K., Kasai, H., Shizuri, Y., and Ōmura, S. (2007). Microbacterium flavum sp. nov. and Microbacterium lacus sp. nov., isolated from marine environments. Actinomycetologica, 21, 53-58. Kim, S. B., Nedashkovskaya, O. I., Mikhailov, V.V., Han, S. K., Kim, K-O., Rhee, M-S., and Bae, K. S. (2004). Kocuria marina sp. nov., a novel actinobacterium islated from marine sediment. Int J Syst Evol Micr, 54, 1617-1620.

Mondino, L. J., Asao, M., and Madigan, M. T. (2009). Cold-active halophilic bacteria from the ice-sealed Lake Vida, Antarctica. Arch Microbiol, 191(10), 785-790. doi: 10.1007/s00203-009-0503-x.

Reddy, G. S. N., Prakash, J. S. S., Prabahar, V., Matsumoto, G. I., Stackebrandt, E., and Shivaji, S. (2003). Kocuria polaris sp. nov., an orange-pigmented psychrophilic bacterium isolated from an Antarctic cyanobacterial mat sample. Int J Syst Evol Micr, 53(1), 183–187. 10.1099/ijs.0.02336-0. Rodrigues, D. F., Goris, J., Vishnivetskaya, T., Gilichinsky, D., Thomashow, M. F., and Tiedje, J. M. (2006). Characterization of Exiguobacterium isolates from the Siberian permafrost. Description of Exiguobacterium sibiricum sp. nov. Extremophiles,10(4), 285-294. 10.1007/s00792-005-0497-5. Steven, B., Briggs, G., McKay, C. P., Pollard, W. H., Greer, C. W., Whyte, L. G. (2007). Characterization of the microbial diversity in a permafrost sample from the Canadian high Arctic using culture-dependent and culture- independent methods. FEMS Microbiol Ecol, 59(2), 513–523. DOI:10.1111/j.1574-6941.2006.00247.x. 55

Yu, L., Lai, Q., Yi, Z., Zhang, L., Huang, Y., Gu, L., and Tang, X. (2013).. Microbacterium sediminis sp. nov., a psychrotolerant, thermotolerant, halotolerant and alkalitolerant actinomycete isolated from deep-sea sediment. Int J Syst Evol Micr, 63(1), 25-30. 10.1099/ijs.0.029652-0. Zhang, D. C., Li, H. R., Xin, Y. H., Chi, Z. M., Zhou, P. J., Yu, Y. 2008). Marinobacter psychrophilius sp. nov., a psychrophilic bacterium isolated from the arctic. Int J Syst Evol Micr, 58, 1463-1470.