Microbiomes of the dust particles collected from the International Space Station and Spacecraft ...

18 Pages·2015·1.84 MB·English

by  Aleksandra Checinska

Checinskaetal.Microbiome (2015) 3:50
DOI10.1186/s40168-015-0116-3
RESEARCH Open Access
Microbiomes of the dust particles collected
from the International Space Station and
Spacecraft Assembly Facilities
Aleksandra Checinska1, Alexander J. Probst2, Parag Vaishampayan1, James R. White3, Deepika Kumar4,
Victor G. Stepanov4, George E. Fox4, Henrik R. Nilsson5, Duane L. Pierson6, Jay Perry7 and Kasthuri Venkateswaran1*
Abstract
Background: The International Space Station (ISS) is a unique built environment due to theeffectsof microgravity,
space radiation, elevated carbon dioxide levels, and especially continuous humanhabitation.Understandingthe
composition of the ISSmicrobialcommunitywill facilitatefurther development of safety and maintenance
practices. The primary goal of this study was to characterize theviable microbiome of theISS-builtenvironment.
A second objective was to determine if the built environments ofEarth-based cleanrooms associated with space
exploration are an appropriatemodel ofthe ISSenvironment.
Results: Samples collectedfrom the ISSand two cleanrooms atthe JetPropulsion Laboratory (JPL, Pasadena, CA)
were analyzedby traditional cultivation, adenosine triphosphate (ATP), and propidium monoazide–quantitative
polymerase chain reaction(PMA-qPCR) assays to estimate viable microbialpopulations. The 16S rRNA gene Illumina
iTag sequencing was used to elucidatemicrobial diversity and explore differences between ISS and cleanroom
microbiomes. Statisticalanalyses showed that members ofthe phyla Actinobacteria, Firmicutes, and Proteobacteria
were dominant inthe samples examined but varied inabundance. Actinobacteria were predominant in theISS
samples whereas Proteobacteria, least abundant inthe ISS, dominated inthecleanroom samples. The viable
bacterial populations seen by PMA treatment were greatly decreased.However, the treatment did not appear to
have an effect onthe bacterial composition (diversity) associated witheach sampling site.
Conclusions: The results of this study providestrong evidence that specific human skin-associated microorganisms
make a substantial contribution to theISS microbiome, which is not the case in Earth-based cleanrooms. For
example, Corynebacterium and Propionibacterium (Actinobacteria) but not Staphylococcus (Firmicutes) species are
dominant on theISS interms ofviable and total bacterial community composition. The results obtained will
facilitate future studies to determine how stablethe ISSenvironment is over time. The presentresults also
demonstrate the value of measuring viable cell diversity and population size atany sampling site. This information
can be used to identifysitesthat can be targetedfor more stringent cleaning. Finally, the results will allow
comparisons with other built sitesand facilitatefuture improvements onthe ISSthatwill ensure astronaut health.
Keywords: International Space Station, Air,Surface,Microbiome,Closed habitat,Cleanroom, Propidium monoazide
*Correspondence:[email protected]
1JetPropulsionLaboratory,BiotechnologyandPlanetaryProtectionGroup,
CaliforniaInstituteofTechnology,M/S89-24800OakGroveDr.,Pasadena,
CA91109,USA
Fulllistofauthorinformationisavailableattheendofthearticle
©2015Checinskaetal.OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.0
InternationalLicense(http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinkto
theCreativeCommonslicense,andindicateifchangesweremade.TheCreativeCommonsPublicDomainDedicationwaiver
(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated.
Checinskaetal.Microbiome (2015) 3:50 Page2of18
Background are controlled for humidity, temperature, and circula-
The microbial characterization of the International tion. Furthermore, NASA quality assurance engineers
Space Station (ISS) has been mostly limited to trad- perform periodic audits to ensure that certified-facility
itional culture-based microbiology and selective molecu- cleanliness levels conform to the requirements delineated
lar biology methods, such as Sanger sequencing, for for both the ISS and Earth cleanrooms. The ISS environ-
supporting tasks such as water remediation, food safety, ment is zero gravity with exposure to space radiation and
and crewmember health [1–5]. Since built environments elevated carbon dioxide levels. By necessity the air is re-
are known to have specific microbiomes [6], it is of the circulatedwhereasonEarththecleanroomsareconstantly
highest interest to the National Aeronautics and Space replenished with fresh air. An important difference is
Administration (NASA) scientific community to explore human habitation, which does not occur in the Earth
the environmental microbiome of the ISS as a closed cleanrooms. However, the human traffic in Earth clean-
environment. Moreover, the National Research Council rooms is actually rather high (at least 50+ people in a
(NRC) specifically recommended that NASA study givenworkingday)whencomparedtotheISSwhereonly
changesinmicrobialpopulations inresponse to selective 6astronautsareallowedatasingletime.
pressure associated with microgravity, which character- This study is the first to analyze samples from the ISS
izes life aboard the ISS [7]. Previous studies show that air and surface using the traditional (colony counts [i.e.,
permanent changes have occurred within the microbial cultivable bacteria]) and state-of-the-art molecular tech-
species during experiments aboard the ISS [8, 9]. Al- niques (adenosine triphosphate [ATP] and quantitative
though next-generation sequencing (NGS) analyses are polymerase chain reaction [qPCR] assays) to measure
now broadly implemented in many microbiology-related the abundance of microorganisms (i.e., live and dead
scientific fields, especially in microbial ecology [10, 11] cells). Furthermore, the abundance and diversity of
and human microbiome projects [12, 13], use of these viable bacterial community were assessed using molecu-
techniques for closed habitats has just begun [14] and lar methods (PMA treatment followed by qPCR [i.e.,
warrantsmore research. bacterial burden] or Illumina-based 16S rRNA sequen-
“Deep” sequencing of ISS samples can answer questions cing [i.e., microbial diversity]). Additionally, the micro-
on abundance and diversity of the microorganisms. How- bial diversity of the ISS was compared with samples
ever, differentiating viable and yet-to-be-cultivable micro- from Jet Propulsion Laboratory (JPL) spacecraft assem-
bial populations requires an appropriate sample processing bly facility (SAF) cleanrooms, which also represent
technology[15].Theuseofthereagentpropidiummonoa- closed andenvironmentally controlled built ecosystems.
zide (PMA) before DNA extraction eliminates cells with a
compromised membrane. The PMA-based assay thus Results
allows for a more accurate approximation of the vi- Microbialburden
able microbial community in terms of richness as well The samples analyzed during this study were the follow-
as abundance [15]. Due to the technically rigorous ing: ISS HEPA filter particulates, vacuum cleaner bag
methods required for culturing many microorganisms, components of ISS (ISS Debris), JPL Class 10 K clean-
characterization of human-associated microbial popu- room (JPL-SAF Debris), and JPL Class 1 K cleanroom
lations in the ISS environment remains a significant (JPL-103 Debris) (Table 1). Microbial burdens estimated
challenge.However,itisimportanttomonitorthepresence via traditional cultivation-dependent and cultivation-
of any opportunistic pathogenic microorganisms. As long- independent methods are given in Table 2. The bacteria
duration human missions are planned in the future, detec- capable of growth in nutrient-rich media were in the
tion of human pathogens and possible mitigation practices range of 105 CFU per gram in the ISS HEPA filter
must be developed. In addition, understanding of the ISS sample and an order of magnitude greater (106 CFU per
microbiome could facilitate the necessary maintenance of gram)fortheISSDebris.Relatively,thebacterialpopula-
thisclosed habitatandtherebyassistinpreventing degrad- tions were larger in the ISS Debris sample than in the
ationofitscomponentsbysomemicroorganisms[4]. JPL cleanroom debris samples. When compared to the
Both ISS environment and cleanrooms on Earth are ISS Debris, the JPL-SAF Debris harbored an order of
maintained by high-efficiency particulate arrestance magnitude smaller bacterial population (105 CFU per
(HEPA) filters. Importantly, considerable prior know- gram) and the JPL-103 Debris sample was two orders of
ledge of the microbial communities associated with magnitude fewer (1.4×104 CFU per gram). In general,
these cleanrooms exists from prior studies. The ISS has cultivable fungi were less represented than bacteria in all
been exposed to these communities indirectly by the samples tested. The cultivable fungal population was
fact that some of the cargos sent to the ISS were pack- two logs less for the ISS samples when compared to
aged in these specific NASA cleanrooms. As in the case their bacterial counts, whereas the JPL cleanroom
of the ISS closed habitat, the Earth-based cleanrooms samples exhibited one-log difference. The ISS Debris
C
h
e
c
in
s
k
a
e
Table1CharacteristicsofsamplescollectedfromISScabinlocationsandEarth-basedcleanroomswherespacecraftcomponentsareassembled ta
l.
Samplename Replicates Location Cleanroomclassification Source Air/surface Specifications Usagetime Model Missionactivities M
ic
ISSHEPA 1 ISS None Filterelementparticles Air HEPArated,retains 40months Partno.SV810010-1, Returnedaboard ro
b
99.97%particles0.3μm; Serialno.0049;HEPA STS-134/ULF6in io
m
20-meshinletscreenhas mediasuppliedby May2011 e
841μmsieveopenings FlandersFilters,Inc.; (2
Nomexinletscreen 01
5
ISSDebris 2 ISS None vacuumcleaner Surface Vacuumbagretains 1day ISSvacuumcleaner Expedition31; ) 3
bagdust particles>6μm; returnedaboard :50

HEPA-ratedfilter Soyuzflight29S
retainsparticles inJuly2012
>0.3μm
JPL-SAFDebris 2 JPL-SAF 10K Vacuumbagdust Surface Retains99.7% 70days NilfiskGM80,81620000 Nomajormission
particles>3μm
JPL-103Debris 1 JPL-103 1K Vacuumbagdust Surface Retains99.7% >180days NilfiskGM80,81620000 Nomajormission
particles>3μm
Table2TotalandviablemicrobiologicalcharacteristicsofparticlesaccumulatedinISSandotherEarth-basedcleanrooms
Sample Cutivablebacterialpopulation qPCR-basedbacterialpopulation Viablebacterialpopulation ATP-basedmicrobialpopulation Viablemicrobialpopulation
(CFU/g)thatare: (16SrRNAcopies/g) (B/A×100) (RLU/g): (D/C×100)
Bacteria Fungi Untreated(A) PMA-treated(B) TotalATP(C) IntracellularATP(D)
ISSHEPA 8.17×105 2.34×103 1.7×109 2.91×107 1.7 2.06×107 3.54×105 1.7
ISSDebris 1.34×106 5.02×104 4.5×108 1.21×107 2.7 3.50×107 2.65×107 75.7
JPL-SAFDebris 1.28×105 5.05×104 3.1×108 2.07×108 66.8 1.91×107 2.10×106 11.0
JPL-103Debris 1.40×104 3.30×103 4.3×108 1.90×107 4.5 2.60×106 1.20×106 46.2
P
a
g
e
3
o
f
1
8
Checinskaetal.Microbiome (2015) 3:50 Page4of18
possessed a similar fungal burden as that of the JPL-SAF Among 19 fungal strains sequenced, 13 strains were
Debris sample, which is higher than the more stringent identified to six different species. Out of 19 fungal
JPL-103 area. The ISS HEPA harbored less fungi when strains, 16 were isolated from the ISS samples and the
comparedtotheISSDebris. other three strains were from JPL samples. All the culti-
The qPCR-based assay, which measured DNA from vated fungal strains affiliated phylogenetically to the
both dead and live cells, estimated that the ISS HEPA members of the Ascomycota phylum. The most common
sample had the highest bacterial density (1.7×109 16S fungal isolate in the ISS samples was Aspergillus niger,
rRNA copies per grams) compared to all other samples, although other diverse species of Aspergillus were also
which were at least one-log less abundant in total identified, along with Penicillium as the second most
bacterial burden. This trend was not noticed in the sam- dominant genus (Additional file 1: Table S1). Based on
ples that were treated with PMA (viable bacterial popu- these results, most of the cultivated bacteria or fungi
lation), where ~107 16S rRNA gene copies per gram from the ISS locations were not noticed in the Earth-
accounted for all samples except in the JPL-SAF, which analog cleanrooms.
exhibited ~108 16S rRNA gene copies. The percentage
of the viable bacterial population (PMA-treated samples) Pyrosequencing-derivedbacterialdiversity
was ~1.7 % in the ISS HEPA, ~2.7 % in ISS Debris, and The number of pyrosequences belonging to various
~4.5 % in the JPL-103 samples. However, approximately bacterialphylaandoperationaltaxonomicunits(OTUs)is
67%ofthebacteria intheJPL-SAFDebris were viable. presented in Additional file 1: Table S2. Approximately,
The total ATP contents derived from both dead and 100,000 reads of bacterial pyrosequences of >500 bp in
alive cells were in the range of 107 relative light unit length were generated from eight samples during this
(RLU) per gram, exceptfor the sample collected from the study. When the software mothur [16] was employed
Class 1 K JPL-103 cleanroom, which showed ~2.6×106 for the bioinformatics analysis, ~71 % of the se-
RLU per gram. The microbial estimation carried out quences were deemed to be of good quality and used
via an intracellular ATP assay (viable cells only) indi- for further analysis, resulting in 70,669 sequences
cated thatonly1.7%ofthetotalmicroorganismswerevi- (Additional file 1: Table S2).
ableintheISSHEPAsample.InthecaseoftheISSDebris Among PMA-untreated samples, the ISS Debris
sample,thispercentagewasmuchhigher(75%);similarly, yielded the highest number of sequences (30,111 reads)
JPL-SAFandJPL-103constitutedapproximately11%and whereas the lowest numberof sequences was seen in the
46%viablemicroorganisms,respectively. ISS HEPA (1,720). Likewise, more sequences were re-
trieved from the JPL-SAF (8,482 reads) as compared
with the JPL-103 (5,062) sample. Contrary to the pyrose-
Culture-basedmicrobialdiversity quence abundance, higher numbers of OTUs were
The 16S rRNA sequencing-based identification and present in the JPL-SAF sample (1,448 OTUs) than in the
phylogenetic affiliation of the bacterial and fungal strains ISS Debris (452 OTUs) sample. However, as noticed in
isolated during this study are given in Additional file 1: the pyrosequence abundance, the number of OTUs was
Table S1. Among 41 bacterial strains isolated and identi- also higher in the JPL-103 sample (301 OTUs) as
fied, 36 belonged to 29 different species, five strains were comparedtotheISSHEPA(133 OTUs)sample.
onlyidentifiedtothegenuslevel.Themajorityoftheiden- The PMA-untreated samples constituted ~65 % of the
tified isolates belong to Firmicutes and only four strains high-quality sequences,whileonly 35%ofthesequences
wereaffiliatedtothemembersoftheProteobacteriagroup. were retrieved from PMA-treated samples (25,294
ThehighestnumberofisolateswasrepresentedbyBacillus sequences). Among the PMA-treated (viable) portions,
and Staphylococcus genera. While Bacillus was dominant ISS HEPA and ISS Debris samples possessed ~38 % and
in both ISS samples, Staphylococcus was only present in 60%ofsequences,respectively.SimilartotheISSDebris
ISS Debris. Both genera were either absent or underrepre- sample, a higher percent of viable sequences was
sented in the JPL-SAF and JPL-103 samples. Single repre- retrieved from the JPL-SAF Debris (77 %) but surpris-
sentatives of other spore-forming lineages were also ingly, only ~3 % of the viable sequences were from the
identified (Paenibacillus, Brevibacillus, and Solibacillus). PMA-treatedJPL-103Debrissample.
Notably, five strains belonging to the B. anthracis-cereus- Almost all of the viable bacterial pyrosequences
thuringiensisgroupwereisolatedfromtheHEPAfilter,and retrieved from the ISS HEPA filter were members of
moredetailedphylogenetic,pathogenic,andwholegenome Corynebacterium. This organism was also dominant in
sequence characterizations are underway to confirm the theISSDebrissample(Additionalfile1:TableS2).Genera
functional characteristics of these isolates. The pathogenic such as Propionibacterium and Staphylococcus were only
propertiesspecifictoB.anthraciswerenotpresentinthese present in the ISS Debris sample. While the ISS-HEPA
fivestrains(datanotshown). sample contained fewer sequence reads representing only
Checinskaetal.Microbiome (2015) 3:50 Page5of18
five genera, the ISS Debris sample was dominated by low sequences associated with the order Actinomycetales
abundance of more diverse genera (Stenotrophomonas, and the genus Staphylococcus. In contrast, the JPL-SAF
Sphingomonas, Pseudomonas, Delftia). Bacillus species Debris sample possessed larger numbers of Bacillus and
were predominant when traditional cultivation methods unclassified Proteobacteria genera sequences, whereas
were employed (Additional file 1: Table S1), but Bacillus JPL-103 Debris mimicked ISS locations and members of
pyrosequences were not retrieved. When PMA-treated Actinobacteriaweremoredominant.
samples from the ISS samples were compared to the In order to detect differentially abundant taxonomic
Earth-analog cleanroom samples, the dominant phyla groups between PMA-treated and untreated samples at
were found to be very similar (Actinobacteria, Proteo- the phylum through genus levels, several statistical
bacteria, Firmicutes). However, on the genus level, the analyses were conducted (see “Methods” section). Hier-
JPL-SAF yielded many more viable sequences repre- archical clustering of samples using genus taxonomic
senting diverse genera (data not shown). profiles resulted in clustering of samples by collection
site but not strong clustering of paired PMA or no PMA
Illumina-derivedbacterialdiversity samples (Fig. 2). Both the negative binomial test and
The Illumina sequencing generated ~6.8 million high- Fisher’s exact test provided valuable information and
quality reads, which is ~100 times more coverage than supported the hierarchical clustering scenario. Searching
was obtained with pyrosequencing yield Additional file 2: for differentially abundant genus-level groups with sig-
Figure S1. Since Illumina sequences were short (~150 bp nificant P values using the negative binomial test, we
each),itwasnotpossibletoreliablyresolvetaxonomicaf- saw that the vast majority of differentially abundant taxa
filiation beyond the genus level. We characterized high- were relatively depleted in the PMA-treated group. The
qualityreadsusingtheRibosomalDatabaseProject(RDP) only well-represented genera that were relatively
classifier[17]andsummarizedcommunitycompositionat enriched were Clostridium (0.34 % vs. 2.46 %, P=0.009)
multiple taxonomic levels. Among the ISS-associated and Rheinheimera (0.005 % vs. 0.472 %, P=7E-06). In
samples, profiles were dominated by Actinobacteria, Ba- contrast, genera such as Streptococcus, Veillonella,
cilli,andClostridia,whilesamplesfromtheJPL-associated Lactobacillus, Bifidobacterium, and Neisseria were all
sites maintained higher levelsofAlphaproteobacteriaand relatively enriched in the untreated samples. Alpha di-
Gammaproteobacteria (Fig. 1). At the family level, the dif- versity estimators also indicated a significant drop in di-
ferences between ISS and JPL profiles were substantial, versity associated with PMA treatment. On average,
with complete distinction between the two sites using PMA samples had 48 % fewer OTUs than their un-
actinobacterial members alone—the ISS samples were treatedpairs(pairedT-testP=0.004).Thisconsistentre-
dominated by Corynebacterium, while the actinobacterial duction in the PMA samples was also true for Faith’s
groups among JPL samples were predominantly diversity,Chao1,andShannondiversitymeasures(Fig.3)
Geodermatophilaceae. and is consistent with the depletion of many taxa identi-
Actinobacteria, Firmicutes, and Proteobacteria phyla to- fiedbydifferential abundanceanalysis.
gether constituted 80 % or more in all cases. When To better characterize the shared community compos-
compared on the family level, sequences associated ition among samples, beta-diversity metrics (e.g., Bray-
with any family did not exceed more than 10 % of Curtis distances) were computed and evaluated using
the total sequences. Characteristics of bacterial phyla principal coordinate analysis (PCoA). Figure 4 displays
that are dominant during this study are depicted in PCoA plots associated with the R1 dataset. The samples
Table 3. A clear-cut trend was noticed on the pres- from the same site are highly similar in composition re-
ence of bacterial phylum dominance on the ISS as gardless of PMA treatment. Using Procrustes analysis,
well as in the Earth-analog cleanroom samples. we compared PCoA plots generated for both the R1 and
Viable sequences arising from Actinobacteria were R2 datasets. In this case, Procrustes analysis was
the most abundant in the ISS HEPA and ISS Debris employed to transform the R2-based principal coordin-
samples (~95 and ~66 %, respectively), while ate set (by rotation, scaling, and translation) to minimize
Proteobacteria was present in the least numbers the distances between corresponding points in the R1
(0.41 and ~3 %, respectively). Among these three principal coordinate space. We found that virtually iden-
dominant phyla, the lowest percentage was found for tical distances were derived between the R1 and R2
Firmicutes and differed significantly between clean- PCoA plots, further supporting the clustering of PMA
room samples as well (JPL-SAF: 11.05 % and anduntreated paired samples.Thishigh-resolution com-
JPL-103: 0.98 %). In-depth analysis showed that the parison stood in contrast to broader composition com-
ISS HEPA sample accumulated more sequences of parisons, such as the hierarchical clustering above. At
the genus Corynebacterium than ISS Debris sample. the OTU level, the bacterial composition was strongly
In addition to the latter, the ISS Debris had associated with the sampling site, and the reduced
Checinskaetal.Microbiome (2015) 3:50 Page6of18
overall diversity of PMA treatment did not dramatically microbiomes. These differences are displayed in ordin-
impactthe composition relative tosampleorigin. ation analyses (Additional file 3: Figure S2) and are
supported by Permutational Multivariate Analysis of
Variance (PERMANOVA) and Multi-Response Permu-
Pyrosequencing-derivedarchaealdiversity
tation Procedure (MRPP) analyses (Additional file 1:
When the samples were subjected to the archaeal
Table S4). Irrespective of the sample origin, the viable
characterization, neither qPCR nor PCR for next-
community profile was highly similar to the total
generation sequencing generated archaeal amplification
community profile, revealing no significant differences in
products (data not shown). It is likely that the presence
PERMANOVAorMRPPindices.However,thedifferences
of archaea in these samples was at a very low concentra-
between the viableand total communityprofile wereout-
tion (<100 gene copies per PCR reaction), below the
weighed by the substantial differences of the ISS and
detection limits of the technology [14]. Alternatively,
Earth-analog cleanroom microbiome, which ranged from
materials associated with the dust samples might have
0.1 to 0.5 concerning the chance-corrected within-group
inhibited the archaeal DNA amplification. However, this
agreement (Additionalfile1:Table S4).Bacterialand fun-
is unlikely because the dust samples spiked with a puri-
gal taxa that showed a significant difference between ISS
fied archaeal DNA did exhibit an appropriate band after
and Earth-analog cleanroom samples are presented in
PCR amplification (data not shown).
Additionalfile3:FigureS2.
Pyrosequencing-derivedfungaldiversity
In total, ~35,000 fungal pyrosequences were retrieved
Discussion
from the eight samples. The detailed breakdown of the
The safety and health of spaceflight crewmembers are of
number of sequences and OTUs is given in Table 4. The
the highest importance for current and future missions.
comparisonismadeontheclasslevel,asfungaltaxonomy
Individuals living and/or working in built environments
is not as well described on the genus level as bacterial
are often susceptible to health issues associated with mi-
taxonomy when it comes to sequencing of environmental
croorganisms [18, 19]. Moreover, the microbial ecology
substrates. The fungal OTUs represented 153 distinct
of ISS remains largely unknown, as study efforts have
taxa. Most OTUs were obtained for three classes:
been mostly focused on microbiological surveillance
Dothideomycetes, Eurotiomycetes, and Tremellomycetes.
using cultivation procedures. The NRC recommended
As noticed for the bacterial abundance, fungal species
use of state-of-the-art molecular biology techniques to
richness was also location specific. The dry cleanroom
develop better microbial monitoring of future closed
surfaces were found to harbor different fungal species
habitat(s) and response system(s) against potential bio-
(mostly Dothideomycetes and Eurotiomycetes) whereas
hazards originating from microbiological sources, using
ISSsamplespossessedmoreEurotiomycetes,Saccharomy-
the ISS as a test bed [1, 4]. Exploration of the microbial
cetes,andExobasidiomycetesmembers.
diversity associated with unusual built-in environments
When treated with PMA, no viable fungal sequences
such as the ISS would further contribute to the indoor
were retrieved from the ISS HEPA sample. In contrast,
microbiome research. This will benefit the development
viable fungal sequences for Dothideomycetes and
of spaceflight applications as well as basic and applied
Tremellomycetes were common in the ISS Debris
researchonEarth.
samples. Viable fungal diversity was far higher in the
The main goal of this study was to determine ISS air
JPL-SAF Debris sample than in JPL-103 Debris. The
and surface viable microbiomes and unveil the differ-
JPL-SAF Debris sample revealed the highest diversity
ences of viable microbiomes on the ISS and Earth-based
that are viable and the highest number of sequences.
cleanrooms. The ISS is a very unique environment due
The great number of Eurotiomycetes sequences was con-
to microgravity, and a suitable Earth analog of this
firmed by cultivation of Aspergillus and Penicillium
ecosystem is not available. The cleanrooms used for
isolatesfrom theISSHEPAandISSdebris.
comparison in this study are environmentally controlled
and represent oligotrophic conditions in an Earth-based
Significantdifferencesofmicrobialcommunitiesbetween setting. Comprehensive characterization and cataloguing
ISSandEarthcleanroomsamples of microbial species in NASA cleanrooms has been
The differences between the ISS and Earth-analog taking place since the 1970s. This extensive data set
cleanroom microbiomes were assessed using multi- renders these facilities the best-characterized closed
variate statistics. Independent analyses of bacterial environments with limited human traffic that the ISS
sequences derived from pyrosequencing and Illumina samplescanbecomparedwith[20–24].Alowerincidence
sequencing revealed a significant difference in the of cultivable microorganisms in Earth cleanrooms than
community profile of ISS and Earth-analog cleanroom ISS surfaces suggests that regular thorough maintenance
Checinskaetal.Microbiome (2015) 3:50 Page7of18
Fig.1TaxonomicprofilesofR1samplesattheclasslevel
Table3CharacteristicsofdominantbacterialphylaasmeasuredbyIlluminaiTagmethodduringthisstudy
ISSHEPA ISSHEPA ISSDebris ISSDebris JPL-SAF JPL-SAF JPL-103 JPL-103
total viable total viable total viable total viable
Totalnumberofsequences 553,176 587,569 1,148,047 1,116,419 1,029,984 1,472,777 489,047 397,607
Percentageofsequencesthat 90.92 99.65 92.35 98.26 80.25 81.84 87.96 90.46
belongtothreedominantphyla
Actinobacteria
Percentageofsequences 63.28 95.28 40.52 66.54 28.76 25.25 23.79 21.46
Numberofgenera 78 55 62 38 122 116 103 71
Numberofdominantgenera(>100sequences) 16 7 28 16 71 76 28 24
Firmicutes
Percentageofsequences 24.83 3.97 45.67 28.48 7.70 11.05 6.60 0.98
Numberofgenera 118 67 100 31 152 150 101 53
Numberofdominantgenera(>100sequences) 50 17 65 18 53 69 15 7
Proteobacteria
Percentageofsequences 2.81 0.41 6.16 3.24 43.80 45.55 57.57 68.02
Numberofgenera 95 65 89 30 189 191 143 92
Numberofdominantgenera(>100sequences) 22 7 49 10 96 104 56 29
Checinskaetal.Microbiome (2015) 3:50 Page8of18
Fig.2Hierarchicalclusteringofsamplesusingtaxonomicprofilesatthegenuslevel(R1dataset).Thetaxonomicprofilesclusteredbasedon
samplinglocation.Thecolorscalereflectslog-normalizedproportionalvalues(e.g.−1~10%,−2~1%,−3~0.1%).JPL-SAFsamplesclustering
togetherandshowingseveraluniquelowabundancememberswerenotfoundintheJPL-103DebrisandISSsamples(e.g.,Mollicutes,Nitros-
pira,andmembersofChloroflexi).RowsandcolumnsareclusteredindependentlyusingthefurthestneighboralgorithmwithaEuclideandis-
tancemetric.oorder,ffamily,ggenus
and cleaning practices are required to reduce microbial qPCR [25–27], next-generation sequencing [15, 28, 29],
burdeninclosedhabitats. or PhyloChipG3 [15]. The utilization of PMA combined
Multiple studies have used PMA for the estimation of with Illumina MiSeq technology for exploring viable
viable bioburden, but they were mostly interpreted by microbiome of ISS-related environmental surfaces is not
Checinskaetal.Microbiome (2015) 3:50 Page9of18
Fig.3PMAtreatmentisassociatedwithareductioninalphadiversity.Alphadiversityvaluesarenormalizedasapercentageoftheirmeanvalue
acrossalleightsamplesinthechart
yet reported. The NGS technology resulting in informa- the dead but intact spores [32]. Despite this, none of the
tion on millions of base pairs might have included other current methodologies can precisely estimate viable
genetic information from dead cells including contamin- populations; PMA treatment has the highest potential
ant DNA associated with the sample processing reagents to lead to the substantial differentiation between dead
[30, 31]. Thus, PMA treatment would eliminate these and live cells [15]. The statistical analysis of the viable
contaminants, making it possible to elucidate viable bacterial community in this study revealed the signifi-
microbiomes. However, it was documented that PMA cant variance in viable microbiome (Fig. 4; Additional
might not work for all kinds of microbial populations, file 1: Table S3 and S4), between the ISS and the
especiallysporeswherethedyewasnotabletopenetrate cleanroom samples.
Fig.4Principalcoordinateanalysisplot(R1data)basedonBray-Curtisdistances.Percentageofvarianceexplainedbyeachprincipalcoordinate
axisisshowninparentheses
Checinskaetal.Microbiome (2015) 3:50 Page10of18
Table4Pyrosequencing-basedfungalphylapresentinISSandEarth-basedcleanroomsamples
Phylum/class Numberofpyrosequencesthatareretrievedfrom:
(numberofgenera)
Number ISSHEPA ISSHEPA ISSDebris ISSDebris JPL-SAF JPL-SAF JPL-103 JPL-103
ofOTU total viable total viable Debristotal Debrisviable Debristotal Debrisviable
Ascomycota
Dothideomycetes(25) 46 116 24 2681 2997 2089 1184 2362
Eurotiomycetes(7) 24 508 6278 1537 265 269 7 63
Leotiomycetes(2) 2 16
Arthoniomycetes 8 51 29 1
Norank 1 3
Saccharomycetes(5) 6 2536 2859 46 5
Sordariomycetes(11) 12 233 66 34
Basidiomycota
Agaricomycetes(3) 4 148 8
Cystobasidiomycetes(3) 7 18 43 424 1020
Exobasidiomycetes 6 498 555 3
Microbotryomycetes(3) 8 108 401 280 111 8
Pucciniomycetes 3 2 18
Tremellomycetes(5) 23 11 4110 204 175 90 186
Chytridiomycetes 2 4 1
Earlydivergingfungallineages 1 5 3
The ISS HEPA filter was in service for ~40 months non-spore-forming bacteria were considered a potential
whereas debris collected from the ISS surface was only contamination risk of spacecraft due to their resistance
1 day old (Table 1). The HEPA filter system employed in to desiccation and ultraviolet radiation [34]. In this
ISS was designed to revitalize the air by passing through study, the viable population of the Actinobacteria did
a heat-exchange process. Hence, particulates associated not constitute high population of the Earth cleanrooms,
with ISS HEPA filter might have been exposed to drier although this can be explained by the different gowning
conditions than the crew living quarter debris (relative procedures required before entering these facilities un-
humidity >50 %) and therefore the microorganisms that like practiced in the ISS. In this study, both the ISS
could withstand dry conditions only survived in the HEPA and ISS Debris samples contained high number
HEPA filter. Any microbial species that could withstand of viable Corynebacterium sequences whereas Propioni-
a prolonged period of time (~40 months during this bacterium sequences were more in the ISS Debris
study) requires survival capabilities against harsh condi- sample. Since these microorganisms require special
tions, and members of the Actinobacteria phylum are growth conditions, they were not isolated when trad-
known to withstand desiccation, dry conditions, and itional cultivation methods were employed during this
high pH [33, 34], which explains their abundance in ISS study. Although the Propionibacterium species repre-
HEPA filters (~95.28 %; Table 3). Comparatively, se- sent natural skin commensals, P. acnes is considered an
quences of viable Actinobacteria were retrieved in more opportunistic pathogen that leads to various infections
numbers in ISS surface debris (~66 %) than in Earth [36, 37]. Similar concerns refer to Corynebacterium,
cleanroomdebris(~25%)samples. which has received attention for being a genus contain-
The members of Actinobacteria phylum represent ing several opportunistic pathogens [38]. Even though
human skin commensals and also soil-borne microorgan- viable sequences of these opportunistic pathogens were
isms[35].ThedominanceofActinobacteriamightbedue retrieved from both ISS locations, their virulence char-
to the presence of astronauts who live and perform work acteristics are to be studied before correlating their
activities on board the ISS. However, characterization of influence on human health in a closed habitat. The risk
astronauts’ skin microbiome is warranted and such of acquiring infection from opportunistic bacterial and
studies will add more insights into the source of these fungal pathogens might pose a threat to crewmembers’
bacteria associated with the ISS environmental surfaces health and needs to be studied in the future.
and atmosphere. Actinobacteria have been previously In contrast, both of the Earth-based cleanrooms con-
reported in spacecraft assembly cleanrooms, and these tained more Proteobacterial members when compared