Metabolism and Biodegradation of Spacecraft Cleaning Reagents by Strains of Spacecraft-Associated Acinetobacter

Abstract Spacecraft assembly facilities are oligotrophic and low-humidity environments, which are routinely cleaned using alcohol wipes for benchtops and spacecraft materials, and alkaline detergents for floors. Despite these cleaning protocols, spacecraft assembly facilities possess a persistent, diverse, dynamic, and low abundant core microbiome, where the Acinetobacter are among the dominant members of the community. In this report, we show that several spacecraft-associated Acinetobacter metabolize or biodegrade the spacecraft cleaning reagents of ethanol (ethyl alcohol), 2-propanol (isopropyl alcohol), and Kleenol 30 (floor detergent) under ultraminimal conditions. Using cultivation and stable isotope labeling studies, we show that ethanol is a sole carbon source when cultivating in 0.2 × M9 minimal medium containing 26 μM Fe(NH4)2(SO4)2. Although cultures expectedly did not grow solely on 2-propanol, cultivations on mixtures of ethanol and 2-propanol exhibited enhanced plate counts at mole ratios of ≤0.50. In support, enzymology experiments on cellular extracts were consistent with oxidation of ethanol and 2-propanol by a membrane-bound alcohol dehydrogenase. In the presence of Kleenol 30, untargeted metabolite profiling on ultraminimal cultures of Acinetobacter radioresistens 50v1 indicated (1) biodegradation of Kleenol 30 into products including ethylene glycols, (2) the potential metabolism of decanoate (formed during incubation of Kleenol 30 in 0.2 × M9), and (3) decreases in the abundances of several hydroxy- and ketoacids in the extracellular metabolome. In ultraminimal medium (when using ethanol as a sole carbon source), A. radioresistens 50v1 also exhibits a remarkable survival against hydrogen peroxide (∼1.5-log loss, ∼108 colony forming units (cfu)/mL, 10 mM H2O2), indicating a considerable tolerance toward oxidative stress under nutrient-restricted conditions. Together, these results suggest that the spacecraft cleaning reagents may (1) serve as nutrient sources under oligotrophic conditions and (2) sustain extremotolerances against the oxidative stresses associated with low-humidity environments. In perspective, this study provides a plausible biochemical rationale to the observed microbial ecology dynamics of spacecraft-associated environments.


Introduction
S pacecraft assembly, test, launch, and operational procedures that minimize the biological contamination of explored environments are critical to ensuring the integrity of future life-detection missions, and in mitigating irreversible impacts to any native biochemical states (Space Studies Board, 2000, 2006. To assist in reducing the probability of contamination, NASA planetary protection requirements for Mars include the assembly of spacecraft in clean room facilities, trajectory biasing for outgoing spacecraft, low impact probabilities for orbiting spacecraft, and partial sterilization of lander and rover spacecraft (at the sub-or full-system levels) (NASA, 2011;Frick et al., 2014).
Among the more diverse members of this core microbiome are the Acinetobacter, a Gram-negative, nonspore forming, and strictly aerobic genus commonly found in soil and water environments, and increasingly associated with multiantibiotic resistance outbreaks in varying clinical settings (Bergogne-Bérézin et al., 2008). In the context of spacecraft microbiology, nonpathogenic strains of Acinetobacter have been isolated and detected in diverse spacecraft-associated environments, including the surface of the preflight Mars Odyssey orbiter (La Duc et al., 2003), floors in the assembly facility for the Mars Phoenix lander (Ghosh et al., 2010), surfaces in the assembly facilities for the Herschel spacecraft (Moissl-Eichinger et al., 2013), surfaces of the International Space Station (ISS) (Castro et al., 2004), and in the drinking water of the ISS (La Duc et al., 2004). Molecular community analyses further suggest that the diversity (and/or abundance) of Acinetobacter increases during spacecraft assembly, as was observed for the Mars Phoenix lander, where the relative abundance of Acinetobacter among all genera (in operational taxonomic units) increased *10-fold upon commencement of assembly and enforcement of the surface and floor cleaning protocols .
In fact, for the Mars Phoenix lander, phylogenetic changes across several genera were quantitated over the course of the spacecraft assembly process, which included sampling before, during, and after assembly . Together, these net phylogenetic changes (before vs. after assembly) showed (1) relative increases in abundance for the Acinetobacter, Ralstonia, and Brevundimonas (all Gram-negative), (2) relative decreases for Mycoplana, Sphingomonas, and Pseudomonas, and (3) <1% or no net changes for most Gram-positive genera such as Streptococcus and the spore-forming Bacillus.
In further detail, and as a comparative example, the Acinetobacter and Streptococcus, respectively, increased *10 and *100-fold during the assembly process (from *4% to *38% and from *0.4% to *55%, respectively), likely due to contamination arising from increased personnel and activities within the facilities. However, after assembly, and despite the routine cleaning procedures, the Acinetobacter further increased to *48% (or a net *10-fold increase) to ultimately represent the numerically dominant genus within the postassembly community. In contrast, the Streptococcus reduced *180-fold in abundance and reduced to 0.3%.
These observations of a dynamic and persistent spacecraft microbial community support the hypothesis that the core microbiome is composed of members that harbor a biochemical potential to tolerate the cleaning procedures, and survive the oligotrophic and low-humidity environments of the assembly facilities (La Moissl-Eichinger et al., 2013). Accordingly, the objective of this work was to measure the ability of spacecraft-associated Acinetobacter to metabolize and biodegrade spacecraft cleaning reagents, and survive under extreme conditions, when cultivated under nutrient-restricted conditions.
Cultivations were performed using capped and parafilmed culture vessels (threaded screw caps for tubes; sterile rubber plugs or aluminum foil for Erlenmeyer flasks). In comparison, cultivations performed in loosely capped culture vessels yielded reliable growth, but irreproducible phenotypic measures, most likely due to evaporation of ethanol. Control experiments included cultures containing all components of the medium except the bacterial inoculum (inoculumnegative) or ethanol carbon source (ethanol-negative); in all cases, the controls yielded no growth.

Growth rates and plate counts
Growth rates were determined by nonlinear regression (KaleidaGraph, Synergy Software; and Microsoft Excel) using a modified version of the Gompertz equation [Eq. (1)] (Begot et al., 1996). Regression parameters included the time of measurement (t), growth rate (k), lag time (L), and the maximum log DOD DOD min value (A). For the regressions, the final OD values for the inoculum-negative controls were negligible, thereby indicating no measurable contamination or formation of nonbiological particulate matter after cultivation.
For all plate count measurements, aliquots were removed from the cultures, decimally diluted in 0.2 · M9 (by 10-100,000-fold), and spread in 20 mL aliquots onto LB agar plates. After incubation for 15-24 h at 32°C, plates bearing clearly isolated colonies, but no more than *300, were enumerated from at least three biological replicates, using at least two technical replicates per trial. For this study, LB agar plates were used for reproducibility purposes, as plating onto M9/ethanol agar plates provided unreliable cell counts (due to inconsistent adsorption of ethanol onto the agar plates across trials, and loss of ethanol and cracking of the agar during incubation). Negligible colony counts were obtained from the inoculum-negative and ethanol-negative controls.

Alcohol dehydrogenase kinetics
The kinetics of alcohol dehydrogenase were measured by absorbance spectroscopy (Beckman Coulter DU640). Midlog phase cultures (80 mL) of A. radioresistens 50v1 were prepared in 0.2 · M9 (with 26 mM Fe 2+ ) containing 16 mM ethanol. Cells were harvested by centrifugation at 6000 g at 4°C for 10 min (Beckman Coulter AllegraÔ 21R), the supernatants were discarded, and cell pellets thoroughly washed by resuspending in 20 mL 1· PBS, harvesting, and repeating the procedure another two times; at the penultimate step, samples were aliquoted and the final cell pellets stored at -80°C.
Protein extracts were prepared from 0.1-0.2 g of the cell pellets (wet mass), which were thawed on ice, resuspended in 10 mL 50 mM HEPES buffer (pH 7.5), and ultrasonicated at 30 s intervals (with 30 s intervals on ice) for a total of 2 min of ultrasonication at a power of 5 (Virtis Virsonic Sonicator). The suspensions were then centrifuged at 6000 g for 20 min at 4°C, the supernatants and pellets saved, and independently analyzed for alcohol dehydrogenase activity. Pelleted samples were resuspended in 500 mL of 50 mM HEPES (pH 7.5) containing 1% Triton X-100 (v/v). All samples were stored on ice and immediately analyzed.
As determined in control experiments, reproducible rates (n ‡ 3) were only obtained when simultaneously using NAD + and an exogenous electron acceptor, such as XTT or DCIP (2,6-dichloroindophenol), with XTT providing lower overall standard deviations. All downstream kinetic studies were performed using 2.5 mM NAD + and 5.0 mM XTT in 50 mM HEPES buffer (pH 7.5), and substrate concentrations of 0.25-10 mM for ethanol or 2-propanol. Reactions were initiated by the addition of 100-200 mL sample and followed by monitoring the change in absorbance at 470 nm every 2 s for 200-800 s.
All reactions were thoroughly but gently mixed, 1 mL in final volume, and performed in 1.0 mL Plastibrand Ò disposable UV cuvettes at 22°C. Reaction rates (n ‡ 3) were determined by linear regression over a minimum of 200 s using an R 2 of ‡0.95, and converted to relevant units using the molar extinction coefficient for XTT (3.70 · 10 4 M -1 cm -1 ). For control purposes, rates of background reduction of XTT by the cellular extracts, in the absence of substrate, were also measured.
Specific activities were expressed as pkat/mg protein (1 pkat = 1 · 10 -9 katals = conversion of 1 pmole of substrate per second), where protein concentrations of the cell lysates were measured using Bio-Rad standard and DC protein assays following the manufacturer's instructions. Michaelis-Menten pseudoparameters were calculated through nonlinear least-squares fitting of the rate data (IC 50 Toolkit; ic50.tk), which provided the parameters of maximum specific activity of catalysis (pkat/mg) and apparent K M (mM); fits were obtained using aggregated data from several trials (n = 7-9, ethanol; n = 3-7, 2-propanol), and standard deviation was calculated from the standard error of the regression.

Stable isotope labeling and metabolite profiling
Stationary phase cultures (2 mL) of A. radioresistens 50v1 (OD *0.9, *9 · 10 8 cfu/mL) were prepared as described using ethanol (natural abundance) or 13 C 1,2 -labeled ethanol under ultraminimal conditions (0.2 · M9, 26 mM Fe 2+ ). All cultures were harvested by centrifugation at 3500 g for 15 min at 4°C (Beckman Coulter Allegra 21R), the supernatants were discarded, and resulting cell pellets were washed by resuspending in 2-6 mL 1· PBS. The suspensions were centrifuged again, the supernatant was discarded, and washing procedure was repeated once more. The final cell pellets were partly dried by centrifugal evaporation (*12 h; Cen-tiVap Console), stored at -80°C, and ultimately analyzed by untargeted analyses of the primary metabolites by gas chromatography-mass spectrometry (GC-MS).
Cell pellets were extracted by thawing the samples on ice, resuspending in a 50/50 mixture of acetonitrile and water (*1.5 mL/cell pellet), and vortexing for 3-5 min at a setting of 3000 (Vortex Genie Vortexer 2). Next, samples were clarified by centrifugation (12,000 g, 4 min, 4°C), and the supernatants carefully transferred to 2 mL centrifuge tubes and immediately dried to *10 mL by centrifugal evaporation ( ‡8 h at medium heat; DNA 110 Savant DNA SpeedVac). The concentrated extracts were diluted *100-fold using 1 mL of 50/50 acetonitrile/H 2 O, vortexed again, and transferred in 10 mL aliquots to microvolume glass inserts (American Chromatography), which had been inserted into 2 mL centrifuge tubes. Samples were again dried by centrifugal evaporation (*10 min, medium heat) and stored at -80°C or immediately prepared for GC-MS analysis.
Untargeted analyses were also performed on the extracellular fractions of cultures grown in the absence and presence of 0.1% and 1.0% (v/v) Kleenol 30. For these samples, late-log cultures (2 mL) were separated by centrifugation (as described), and the supernatants were removed and separated into 500 mL aliquots, dried by centrifugal evaporation, and stored at -80°C. Upon analysis, the dried samples were directly derivatized using 90 mL MSTFA with 1% trimethylchlorosilane (with FAMEs standards), followed by heating at 37°C for 30 min. All samples were transferred to wide mouth crimp top vials (American Chromatography Supplies, New Jersey) and sealed with an 11 mm crimp cap. Samples were analyzed by GC-MS within 24 h of derivatization.
All samples were analyzed on an Agilent Technologies 6890N Network GC System connected to an Agilent Technologies 5973 Inert Mass Selective Detector and outfitted with a 7683B Series Injector. Separations were performed using an Agilent J&WDB-5ms GC-capillary column (30 m · 0.25 mm i.d.) and an integrated 10 m guard column. Samples (1 mL) were introduced by splitless injection using a constant helium gas (99.999% purity) flow of 1 mL/min. Injection port temperature was maintained at 250°C, and the separation was performed using an initial oven temperature of 60°C (for 1 min) and a temperature ramp of 10°C/min to 320°C (hold for 5 min). Mass spectra were obtained by electron impact ionization at 70 eV, using an ion source temperature of 230°C and collected over a mass range of 60-600 m/z (2.71 scans/s).
Deconvolution and identification were performed on the raw data using the Automated Mass Spectral Deconvolution and Identification System software (AMDIS, National Institute of Standards and Technology) using a retention index of -2% and a corrected match factor of at least 700. The data were integrated using Agilent Mass Hunter Quantitative Analysis B.07.00M after translation with the GC MSD Translator (Agilent). Analyte abundances were compared using parametric t-tests, and multiple hypothesis testing was corrected for using a Benjamini-Hochberg false discovery rate of 0.10 (Microsoft Excel).

Cultivations under ultraminimal conditions
Growth rates and viability were measured for multiple spacecraft-associated Acinetobacter strains under ultraminimal conditions (0.2 · M9, 16 mM ethanol, 26 mM Fe 2+ ). The tested strains included A. radioresistens 50v1, which was isolated from the surface of the preflight Mars Odyssey orbiter, and A. proteolyticus 2P01AA, A. johnsonii 2P08AA, A. johnsonii 2P07AA, A. oryzare 2P08MC, A. guillouiae 2P07PB, and A. guillouiae 2P07PC, which were isolated from the floor of the assembly facility for the Mars Phoenix lander. As shown in Figure 1A and Supplementary Fig. S1 (see Supplementary Data at https://www.liebertpub.com/suppl/doi/10.1089/ast .2017.1814), except for the 2P01AA strain, all spacecraftassociated Acinetobacter strains grew on ethanol as the sole carbon source, with growth rates ranging from *0.36 to 0.53 h -1 (in 16 mM or 0.1% v/v ethanol), where the fastest rates were exhibited by A. johnsonii 2P08AA (0.53 -0.03 h -1 ).
At 10-fold higher concentrations of ethanol (160 mM or 1.0% v/v), rates for the 2P08AA and 2P07AA strains were *1.5-fold lower ( p < 0.05), whereas rates for the 50v1, 2P07PB, and 2P07PC strains were, respectively, and statistically equivalent across 16 and 160 mM ethanol. As displayed in Figure 1B, viable cultures were confirmed by plate counts on mid-log phase cultures, which provided cell densities ranging from *4 · 10 8 to 1 · 10 9 cfu/mL. Most strains displayed similar cell densities when grown on 16 or 160 mM ethanol; however, plate counts for the 2P07PB strain were *1.8-fold higher (1.1 · 10 9 -1.5 · 10 8 cfu/mL) in 16 mM ethanol.
As a species-level comparison, growth rates were also measured across a range of ethanol concentrations (2-650 mM) for A. radioresistens 50v1 and A. radioresistens 43998 T , which served as a nonspacecraft-associated control for this study ( Fig. 2A). For the 50v1 strain, rates were fastest at 8 mM ethanol (0.46 -0.02 h -1 ) and gradually trended downward between 40 and 650 mM (from 0.43 -0.04 to 0.11 -0.02 h -1 ). For the type strain, growth rates were essentially equivalent across the range of tested ethanol concentrations; overall, the rates (*0.22 h -1 ) between *2 and 40 mM ethanol were *2-fold lower than those of the 50v1 strain. For both strains, growth at concentrations <2 mM provided irreproducible results.
Together, these trends indicated that cell viabilities at midlog phase were maximally enhanced in the presence of low mole ratios of 2-propanol, and that a concentration of 30 mM 2propanol was not inhibitory, as the growth rate was unaffected. At a higher mole ratio of 0.85, however, longer cultivation times were required to reach mid-log phase, as indicated by the slower growth rates (in 170 mM 2-propanol), and cell viabilities at mid-log phase were not enhanced. Unsurprisingly, no growth was observed under these conditions when using 2propanol as a sole carbon source (mole ratio of 1.0).

Stable isotope labeling
Cultivation on ethanol as a sole carbon source was confirmed by untargeted metabolomics analysis and stable isotope profiling using 13 C-labeled ethanol ([1,2-13 C 2 ]ethanol). Cultures of A. radioresistens 50v1 were prepared in ultraminimal medium (0.2 · M9, 26 mM Fe 2+ ) using 16 mM ethanol or 13 C-labeled ethanol, and the cell extracts were analyzed by GC-MS ( Supplementary Fig. S2A, B).
Confirmation of ethanol incorporation was obtained by comparison of these metabolites with those extracted from cells cultivated on 13 C-labeled ethanol. As representative examples, mass spectra for oleic acid [(E)-9-octadecenoic acid] and trehalose (a-d-glucopyranosyl-(141)-a-d-glucopyranose) are provided in Figure 3. In Figure 3A (highlighted by the arrow), the molecular ion for the MSTFA-derivatized version of oleic acid (trimethylsilyl ester of oleic acid, 354.6 g/mol) is observed at a mass-to-charge ratio (m/z) of 354. In comparison, mass spectra for oleic acid obtained from cultures grown on 13 Clabeled ethanol yielded a molecular ion at 372 m/z, representing a gain of 18 mass units, consistent with 13 C incorporation at each of the 18 carbons of oleic acid (18:1 D9 ). For trehalose, in Figure 3B, the fragment ion at 361 m/z represented a six-carbon product resulting from scission at the glycosidic bond to yield two identical fragment ions (Füzfai et al., 2008). In comparison, mass spectra of trehalose obtained from cultures grown on 13 C-labeled ethanol yielded a fragment ion peak at 367 m/z, representing a gain of 6 mass units, consistent with 13 C incorporation at each of the carbons in the fragment ions.

Kinetics of ethanol and 2-propanol oxidation
The alcohol dehydrogenase activities in cellular extracts of A. radioresistens 50v1 were measured against the substrates of ethanol and 2-propanol (Fig. 4). Cultures were prepared under ultraminimal conditions (0.2 · M9, 26 mM Fe 2+ ) using 16 mM ethanol as the sole carbon source. Extracts were prepared by ultrasonication, and the membrane (resuspended pellet) and soluble (supernatant) fractions were separately tested for alcohol oxidation activities. Comparisons of the protein extracts revealed approximately fivefold higher specific activities in the membrane fractions. Kinetic studies on the membrane fractions using the substrates of ethanol or 2-propanol (Fig. 4A, B) revealed standard Michaelis-Mententype behavior, with nonlinear regressions providing maximum specific activities of 23 -3 pkat/mg and 1.4 -0.4 pkat/ mg (and apparent K M values of *0.3 and *0.4 mM), respectively. As shown in Figure 4C, the calculated maximum specific activities for ethanol were *16-fold higher than that of 2-propanol.
As summarized in Figure 5B, impacts to the extracellular metabolome were supported by decreases in abundances for several hydroxy-and ketoacids, including 2-ketoglutarate, a-hydroxyglutarate, 2-ketoisocaproate, citrate, and 4-hydroxyphenyllactate. Furthermore, control experiments with Kleenol 30 (Fig. 5A) showed that incubations in 0.2 · M9 (containing 26 mM Fe 2+ and 16 mM ethanol), in the absence of A. radioresistens 50v1, yielded substantial increases in the abundances of decanoate (10:0). However, in the presence of A. radioresistens 50v1, decanoate was not detected, suggesting either biodegradation or metabolism of this product.

Discussion
In this study, we measured the ability of several strains of spacecraft-associated Acinetobacter to metabolize or biodegrade the reagents used to clean spacecraft, and surfaces and floors of spacecraft assembly facilities. Cultivations in the presence of the cleaning reagents were conducted under ultraminimal conditions to roughly approximate the lownutrient and low-particulate (inorganics and organics) conditions of the assembly facilities. Cultivations were performed in 0.2 · M9 minimal medium supplemented with 26 mM Fe 2+ , where alkali, halogen, and main group elements (i.e., N, S, and P, excluding oxygen) were <15 mM in concentration, alkaline earth metals were £0.4 mM, and heavy metals contaminants were collectively £0.001% of the medium constituents. In comparison with prior work, early cultivations of the Acinetobacter on ethanol relied upon inorganic-rich medium (Abbott et al., 1973) containing a variety of supplemented metalloids and transition metals (e.g., 0.1-100 mM concentrations of Fe, Mn, Cu, Co, Zn, Mo, and B), with other media containing organic components such as citrate (Du Preez et al., 1981), and yeast autolysate and pantothenate (Pirog and Kuz'minskaya, 2003).
Under more extreme nutrient-restricted conditions (0.2 · M9, 26 mM Fe 2+ ), most of the tested spacecraft-associated strains grew on ethanol as a sole carbon source (six out of the seven tested strains). Cultivations provided high cell counts at midlog phase (10 8 -10 9 cfu/mL), where generation times (or doubling times) ranged from 0.56 to 0.80 h in 16 mM (0.1% v/v) ethanol, and from 0.85 to 1.1 h in 160 mM (1.0% v/v) ethanol. All viable spacecraft-associated strains grew optimally at lower ethanol concentrations, with A. radioresistens 50v1 having an optimal substrate concentration range with an upper limit of 2-40 mM. In contrast, the A. radioresistens type strain did not exhibit an optimal substrate concentration, and generation times (at £16 mM ethanol) were approximately twofold slower than all tested strains. These comparisons support a trend between lower optimal substrate concentrations and spacecraft association, which would be a necessary correlation for any postulated metabolism under the oligotrophic conditions of the assembly facilities.
For this study, all downstream analyses focused on A. radioresistens 50v1, which is the best characterized strain among the spacecraft-associated Acinetobacter; in addition, at the species level, A. radioresistens has been detected on Mars-bound spacecraft and in the ISS (surfaces and drinking water) (La Duc et al., 2003;Castro et al., 2004;McCoy et al., 2012;Moissl-Eichinger et al., 2013;Schuerger et al., 2013;Derecho et al., 2014). For A. radioresistens 50v1, use of ethanol as a sole carbon source was confirmed through stable isotope labeling (using 13 C-labeled ethanol) and untargeted analyses, which supported full enrichment of 13 C in the primary metabolome. Consistent with the known microbiology of A. radioresistens (Nishimura et al., 1988), 2-propanol (isopropyl alcohol, isopropanol) did not serve as a sole carbon source; however, cultivations on mixtures of ethanol and 2propanol (70/30 and 50/50 mixtures) showed enhanced growth, as indicated by statistically significant increases in plate counts (cfu/mL) at mole ratios of £0.50.
Biochemical interrogations supported the oxidation of both ethanol and 2-propanol by extracts of A. radioresistens 50v1, likely due to a membrane-associated alcohol dehydrogenase (Singer and Finnerty, 1985). Michaelis-Menten kinetics provided a low K M value for ethanol (*0.3 mM), consistent with the faster generation times obtained at the lower ethanol concentrations in ultraminimal medium. Kinetic studies also indicated a requisite need for both NAD + and an exogenous electron acceptor (XTT or DCIP), thereby supporting the formation of NADH (although transiently in cellular extracts) from either alcohol substrate, which in vivo (and in the presence of intact membranes) would directly support respiration.
Molecular and cultivation studies show that the floor of the assembly facilities also harbor strains of Acinetobacter (Ghosh et al., 2010;La Duc et al., 2012), with human-based activities possibly assisting in physical transport of these microbes across the facilities. Molecular experiments with A. radioresistens 50v1 show that Kleenol 30, an alkaline floor detergent, is biodegraded under ultraminimal conditions into lower molecular weight ethylene glycols.
Furthermore, untargeted analysis of the extracellular metabolome showed decreased abundances for several hydroxyand ketoacids in the presence of Kleenol 30. Interestingly, these hydroxy-and ketoacids share iron-binding properties (Drechsel et al., 1993;Schofield and Zhang, 1999;Yue et al., 2003) and, hence, reveal a potential impact to the transport/ metabolism of extracellular iron. Our studies also indicate that decanoate (formed during incubation in Kleenol 30 in 0.2 · M9/Fe/ethanol) is biodegraded by A. radioresistens 50v1 (Fig. 5), with the results (Fig. 5B) suggesting concomitant increases in abundances of suberic acid (octanedioic acid), which along with acetyl-CoA would potentially be a metabolic product of o-oxidation of decanoate (Donoghue and Trudgill, 1975;Kunz and Weimer, 1983;Van Bogaert et al., 2011).
Under ultraminimal conditions, we also show that A. radioresistens 50v1 exhibits a remarkable oxidative extremotolerance when cultivated on ethanol as a sole carbon source (*1.5-log reduction, *10 8 cfu/mL, 10 mM H 2 O 2 ). This assessment is based on comparisons with other nonspore forming radiation and oxidation-resistant bacteria. For instance, the survivability of A. radioresistens 50v1 in nutrient-poor medium is comparable with that of Deinococcus radiodurans R1 (*1-log reduction in 33 mM H 2 O 2 ) and that of Vibrio rumoiensis S-1T (*1.5-log reduction in 0.4 mM H 2 O 2 ) when cultivated in nutrient-rich medium (e.g., LB, tryptic soy broth, and peptone yeast extract starch) (Arrage et al., 1993;Ichise et al., 1999). Hence, by extension, these results support the potential for oxidative extremotolerance under oligotrophic conditions, which is significant, as desiccating environments (such as the assembly facilities) are thought to promote oxidative stress through the formation of reactive oxygen species (Billi and Potts, 2002;Franca et al., 2007).
In the context of survival in the assembly facilities, therefore, these combined results support the potential for ethanol, 2-propanol, and perhaps Kleenol 30 to (1) serve as carbon or energy sources under oligotrophic conditions and (2) sustain extremotolerances against the oxidative stresses associated with low-humidity environments. For recent Mars missions, however, the surface cleaning procedures were predominantly performed using isopropyl alcohol wipes (rather than ethanol) (La Benardini et al., 2014). Accordingly, and as a potential survival mechanism, members of the core microbiome likely remained on the surface after wiping, wetting by the residual cleaning reagents likely initiated a basal metabolic activity, and the resulting activity was likely attenuated upon evaporation of the cleaning reagents (analogous to our experiments using loosely capped culture tubes).
In turn, the cycles of wetting/drying, resulting from highfrequency cleaning, likely imposed certain stresses (such as oxidative stress) on the microorganisms, with the residual evaporates forming a baseline vapor abundance in the facilities. In support, recent measurements show that the vapor abundances of 2-propanol are *0.1 ppm in facilities maintained at ISO classes 8 and 7 standards (Dworkin et al., 2018); hence, these observations support the potential for 2-propanol to serve as a perpetual nutrient source, as the acquisition of volatile organics is a known survival tactic for soil bacteria (Hanzel et al., 2011;Modrzyński et al., 2016). In this combined perspective, and after substrate acquisition, the Acinetobacter could have slowly metabolized 2-propanol to yield minimal but potentially sufficient amounts of NADH to support survival. Moreover, when accounting for surface communities within the assembly facilities, our work opens the possibility that metabolic contributions may arise from other members of the core spacecraft microbiome.

Conclusion
In sum, this study provides a plausible biochemical rationale to the observed microbial ecology dynamics of spacecraft assembly facilities, as spacecraft-associated microorganisms (such as the Acinetobacter) may metabolize/ biodegrade spacecraft cleaning reagents and exhibit extreme oxidative tolerances under the oligotrophic and lowhumidity conditions. This work also adds to the range of known survival features for the spacecraft-associated Acinetobacter, which include extreme tolerances toward aqueous hydrogen peroxide, under nutrient-poor (this study) and nutrient-rich conditions (Derecho et al., 2014), desiccation (McCoy et al., 2012, sequential exposures to oxidative and radiative stressors (McCoy et al., 2012), heat treatments (80°C for 15 min) (Moissl-Eichinger et al., 2013), and exposures to martian atmospheric and pressure conditions (Schuerger et al., 2013).
In the framework of planetary protection, therefore, the Gram-negative and nonspore forming Acinetobacter may tolerate partial sterilizations with vaporous hydrogen peroxide, and survive the heat treatments associated with the NASA Standard Assay, which could possibly impact treatments and measurements for missions requiring very low bioburden values (e.g., life detection and Special Regions missions). Furthermore, our work lends support toward the use of differing (and rotating) spacecraft-compatible cleaning reagents as a means of controlling the core spacecraft microbiome.