Microbial Community Succession in an Unvegetated,
Recently Degiaclated Soli
Diana R. Nemergut^’^, Suzanne P. Anderson^’^, Cory C. Cleveland^, Andrew P. Martin^,
Amy E. Miller^’^, Anton Seimon^ and Steven K. Schmidt^
(1)
(2)
(3)
(4)
(5)
(6)
INSTAAR, An Earth and E nvironm ental Systems Institute, University o f Colorado, Boulder, CO 80309, USA
E nvironm ental Studies Program , U niversity o f Colorado, Boulder, CO 80309, USA
D epartm ent o f Geography, University o f Colorado, Boulder, CO 80309, USA
Ecology and Evolutionary Biology, University o f Colorado, Boulder, CO 80309, USA
N ational Park Service, Anchorage, AK 99501, USA
International Research Institute for Clim ate Prediction, C olum bia University, Palisades, NY 10964, USA
Received: 4 M ay 2006 / Accepted: 30 M ay 2006
Abstract
P rim ary succession is a fundam ental process in m acro ­
ecosystem s; how ever, if an d h o w soil d ev elo p m en t
influences m icro b ial c o m m u n ity stru c tu re is p o orly
u nderstood. Thus, we investigated changes in the bacterial
com m unity along a chronosequence o f three unvegetated,
early successional soils (~ 2 0 -y ear age gradient) from a
receding glacier in southeastern P eru using m olecular
phylogenetic techniques. W e fo u n d th a t evenness, phylo­
genetic diversity, and the n u m b er o f phylotypes were
lowest in the youngest soils, increased in the interm ediate
aged soils, an d plateaued in the oldest soils. This increase
in diversity was com m ensurate w ith an increase in the
n u m b er o f sequences related to co m m o n soil bacteria in
the old er soils, in clu d in g m em bers o f the divisions
A cidobacteria, B acteroidetes, a n d V errucom icrohia.
Sequences related to the C om am onadaceae clade o f the
B etaproteobacteria were d o m in an t in the youngest soil,
decreased in abundance in the interm ediate age soil, and
were n o t detected in the oldest soil. These sequences are
closely related to culturable hetero tro p h s from rock and
ice environm ents, suggesting th a t they originated from
organism s living w ithin or below the glacier. Sequences
related to a variety o f nitrogen (N )-fixing clades w ithin
the Cyanobacteria were ab u n d an t along the chronose­
quence, com prising 6 -40% o f phylotypes along the age
gradient. A lthough there was no obvious change in the
overall abundance o f cyanobacterial sequences along
the chronosequence, there was a d ram atic shift in the
abundance o f specific cyanobacterial phylotypes, w ith the
Correspondence to: D iana R. N em erg u t; E-m ail: nem ergut@ colorado.
ed u
interm ediate aged soils containing the greatest diversity
o f these sequences. M ost soil biogeochem ical character­
istics show ed little change along this ~20-year soil age
gradient; however, soil N pools significantly increased
w ith soil age, perhaps as a result o f the activity o f the
N -fixing Cyanobacteria. O u r results suggest that, like
m acrobial com m unities, soil m icrobial com m unities are
structured by substrate age, and th at they, too, undergo
predictable changes through time.
Introduction
P rim ary succession, the establishm ent and gradual change
in plant com m unities on newly inhabited substrate, is
a fundam ental ecological process [9, 11, 16] th at rem ains
an active focus o f b o th co m m u n ity and ecosystem
ecology research (e.g., [7, 8, 13, 55, 57]). P rim ary
succession occurs on new ly exposed o r d ep o sited
substrates, including lava flows, sand dunes, landslides,
and glacial till. In systems characterized by continuous
substrate inputs, such as active lava flows and receding
glaciers, chronosequences result, w ith older, m o re
w eathered soils farther from the substrate source. Here,
distance can serve as a proxy for tim e, and ecosystem
developm ent can be investigated along a spatial gradient
w ith o u t the need for long-term observations o f the
same site.
C hronosequence studies have revealed correlations
betw een plant com m unity com position and substrate
age. Integrating these observations w ith knowledge o f
p lant biology has been key in developing m odels that
describe changes in ecosystem n u trien t dynam ics and soil
DOI: 10.1007/s00248-006-9144-7 • © Springer Sdence+Business Media, Inc. 2006
D .R . N
developm ent th ro u g h tim e. For example, it has long been
recognized th a t lichens o r plants th a t form sym biotic
associations w ith nitrogen (N )-fixing m icrobes, including
Dryas an d Alnus, are am ong the first colonizers o f fresh
substrate (reviewed in [30]). This observation led to a
paradigm th a t describes nitrogen dynam ics in prim ary
successional com m unities: because N is absent from the
new m ineral substrate, organism s th a t fix nitrogen are
com petitive (and hence abundant) in early stages o f
succession. Furtherm ore, the presence o f N fixers is
tightly linked to the accum ulation o f soil N [5, 12, 30],
and increases in soil nitrogen availability m ay facilitate
colonization by other, later successional species [8, 56].
Because o f their ability to fix nitrogen, m icro organ­
isms are clearly central to early ecosystem developm ent,
and som e studies even suggest th a t m icrobial activity
prom otes p lant establishm ent [8, 56] an d grow th [6].
However, only a few studies have exam ined the m icrobial
com m unities an d n u trie n t dynam ics in the youngest,
unvegetated sections o f chronosequences. The investiga­
tions th a t do exist su p p o rt increases in m icrobial biom ass
[47, 48] an d m icrobial activity [54] w ith soil age. O ther
studies dem onstrate decreases in substrate-induced res­
piratio n (SIR) p er u n it biom ass w ith soil age, suggesting
th at older com m unities store m o re carbon th a n they
respire [34, 41]. Finally, small su b u n it ribosom al RNA
(SSU rRNA) gene fingerprinting suggests b o th a decrease
in diversity and an increase in the m icrobial co m m unity
evenness w ith soil age [48].
D espite the evidence for active, dynam ic m icrobial
com m unities in these barren, unvegetated soils, little data
exist on the types o f m icroorganism s inhabiting these
environm ents, or how shifts in these m icrobial co m m u ­
nities correlate w ith soil age. This is likely due to the
difficulties inh eren t in describing m icrobial co m m unity
com position, as an estim ated less th a n 1% o f m ic ro ­
organism s can be cultured using traditional techniques
[2, 35]. However, the recent application o f m olecular
analyses, in p articular the use o f SSU rRNA gene clone
libraries, has allowed for m icrobial com m unity surveys
w ith o u t th e biases associated w ith cu ltivation-based
techniques. SSU rRNA clone library techniques have
been used to exam ine fungal [24] an d archaeal [33]
diversity along recently deglaciated chronosequences,
but, to o u r knowledge, n o studies have provided detailed,
phylogenetic data o n the bacterial com m unities in these
developing soils. Thus, we set forth to exam ine patterns
in bacterial com m unity com position along the soil age
gradient o f a glacial foreland. W e sam pled unvegetated
sites w ith n o visible organic m atter along the soil age
gradient o f the Puca glacier in the Peruvian Andes to
address som e basic questions about bacterial succession.
First, we were interested in determ ining if m icrobial
com m unities, like p lan t com m unities, undergo p red ict­
able changes in species com position as soil develops. In
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
sim pler, less com plex m icro b ial system s, in clu d in g
developing biofilm s (e.g., [29]) and com posting systems
(e.g., [43]), m icrobial com m unities exhibit successional
patterns, and populations are replaced in som ew hat
consistent and predictable ways. However, the ability to
recognize such patterns in soils m ay be lim ited by the
extrem e m icroscale diversity o f soils. Next, we investigated
how shifts in the m icrobial com m unity com position
correlate w ith n u trien t dynam ics along the chronose­
quence o f unvegetated soils. A lthough linking populations
and processes has been a m ajor focus o f macroecology,
there is debate about w hether or n o t in form ation on the
structure o f m icrobial com m unities is inform ative at the
ecosystem scale (e.g., [3, 14, 25]). However, it is possible
th at in unvegetated soil relationships betw een n u trie n t
dynam ics and m icrobial populations m ight be m ore
evident. In this region o f the Andes, pollen deposition
rates on the nearby Quelccaya icecap are am ong the
highest ever recorded [38]. Thus we hypothesized th at
this allochthonous carbon (C) in p u t w ould select for
heterotrophs in these young soils. In addition, because
the C/N ratio o f pollen is higher th a n m icrobial biom ass,
and N is absent from new substrate, we hypothesized
th at heterotrophic nitrogen fixers w ould be abu n d an t in
these soils.
Materials and Methods
S tu d y S ites a n d Sam ple Collection.
In A ugust 2003,
we collected soil sam ples from the recently exposed
forelands o f the P uca glacier. This glacier lies in the
Laguna Sibinacocha basin, in the C ordillera V ilcanota
range in the Peruvian Andes (elevation, ~5000 m) near
the Quelccaya ice cap. The m oraines in this region are
com posed o f rock w ith high quartz and calcite content
(A nderson and Blum, unpublished data). The forelands
o f the Puca glacier are largely unvegetated at distances
o f up to 500 m, w ith plants covering less th an ~ 10% of
the ground. The first vascular colonizers o f this terrain
are predom inately grasses o f the genera Ephedra and
Calam agrostis. P re c ip ita tio n in th is area is highly
seasonal; o f the approxim ately 1000 m m o f annual
precipitation, about 60% falls betw een D ecem ber and
M arch [15, 52]. M onthly m ean tem peratures vary by
less th a n 3°C, b u t diurnal variation changes markedly,
ranging from about 12°C in the w et season to about 22°C
in the w inter dry season. D uring the dry season, surface
soil tem peratures at o u r sites oscillate between -5 °C and
25°C [32].
W e established three transects parallel to the fore­
fro n t o f the Puca Glacier. The transects were chosen to
represent regions o f progressively older soils, an d thus
w ere at distances o f 0 m (youngest soil), 100 m
(interm ediate age), and 500 m (oldest soil) from the
glacial forefront. They are hereafter referred to as the 0-,
D .R . N
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
100-, and 500-m transects. F our replicate surface soil
samples (0 -5 cm) were th en collected along each transect
at ra n d o m ly ch o sen p o sitio n s. Each rep licate was
obtained by m arking a 50 x 50 cm area aro u n d the
collection point, an d th en bulking three individual soil
samples from w ithin this area. These replicate samples
are hereafter referred to as Om-1, Om-2, Om-3, Om-4,
lOOm-1, an d so on. Soils were scooped in sterile conical
tubes, im m ediately placed o n ice, and kept in the dark
at 0°C for tran sp o rt to B oulder, CO, USA (approxi­
m ately 1 week). Soil samples were th en divided into two
aliquots: one was placed at -8 0 °C for DNA extraction
an d the o th er was stored at 4°C for biogeochem ical
analyses. Along each o f the three transects (0, 100, and
500 m ), we m easured biogeochem ical param eters on all
fou r replicate soil samples, an d we random ly selected
two o f these replicates for m olecular analyses. Replicate
soil samples chosen for m olecular analyses were all at
distances o f at least 25 m (along the transects) from one
another. Because these are n o t soils in the traditional
sense, an d contain an array o f fine particles to fairly
large rocks, we did n o t sieve soils, b u t we did use soil
particles <2 m m for analyses.
Soil Dating.
In August 2003, the U TM coordinates
o f the glacier term inus were 276193 E, 8476734 S (P ro ­
visional South A m erican 1956). A sequence o f isochrones
o f glacier term in u s location derived from air photo,
satellite an d g ro und-based GPS m easurem ents for the
perio d 1931-2003 suggest th a t the 0 m soils were less th an
1 year old, the 100 m soils were IM years old, and the 500 m
soils were approximately 20 years old (Seimon, unpublished
data).
DNA Extraction, 163 rRNA G ene Clone Libraries, an d
S e q u e n c e Analysis.
D NA was extracted from soil
samples using a m odification o f the protocol described
by M ore et al. [31]. Soil (0.5 g) was added to 0.3 g o f
each 0.1-, 0.5-, and 1-m M -sized glass beads w ith 1 mL
phosphate extraction buffer, an d agitated for 2 m in
w ith a bead beater. Samples were centrifuged at 14,000
rp m for 10 m in an d the supern atan t was rem oved to a
fresh tube. Next, 200 pL o f 7.5 m M am m o n iu m acetate
was added an d samples were incubated o n ice for 5 m in.
Samples were centrifuged at 14,000 rp m for 3 m in and the
supern atan t was then rem oved to a fresh tube. Samples
were extracted w ith an equal volum e o f phenol/chloroform /isoam yl alcohol (25:24:1). D NA was precipitated
w ith 200 pL isopropanol overnight at -20°C . Samples
were th en centrifuged at 14,000 rp m for 20 m in and the
supern atan t was rem oved an d discarded. D NA pellets
were w ashed w ith 1 mL o f 70% ethanol. Eight separate
D N A e x tra c tio n s w ere p o o le d a n d p u rifie d o v er
Sepharose 4B (Sigma, St. Louis, M O , USA) packed
colum ns as described by Jackson et al. [23].
A pproxim ately 30 ng o f DNA was am plified w ith the
prim ers 515f and 1492r [27]. The reaction conditions
consisted o f 400 uM each prim er, 200 pM each dNTP,
and 1.25 U o f Taq DNA polym erase (Prom ega, M adison,
W l, USA) in Taq DNA polym erase buffer containing
2.5 m M M gCl 2 (Prom ega). After an initial denaturation
step at 94°C for 1 m in, 30 cycles o f 94°C for 1 m in, 58 +
5°C for 30 s, and 72°C for 2.5 m in w ith a term inal 10-m in
extension at 72°C were perform ed. Polym erase chain
reaction (PGR) products from six reactions w ith differ­
ent annealing tem peratures were pooled and gel purified
by using Q lA quick gel purification colum ns (Qiagen,
Valencia, CA, USA). Purified products were ligated into
the vector pC R 2.1 (Invitrogen, Carlsbad, CA, USA) and
transform ed into Escherichia coli following the m a n ­
ufactu rer’s instructions.
For each cloning reaction, 96 colonies were in o cu ­
lated into a 96-well deep-dish plate containing 1.5 mL
Luria b ro th (1% NaCl, 1% tryptone, 0.5% yeast extract)
w ith 50 pg/m L am picillin. C ultures were agitated at
200 rp m for 14 h at 37°C. Plasm id DNA was extracted by
using a m odified protocol for 96-well plates based the
procedure described by Sam brook and Russell [40]. DNA
pellets were air-dried for 20 m in and resuspended in
50 pL T ris-H C l (pH 8.5).
Inserted 16S rRNA genes were PCR-am plified from
the plasm ids by using the prim ers M13F and M U R
(Invitrogen). The reaction conditions consisted o f 400 nM
each prim er, 200 pM each dNTP, and 1.25 U o f Taq
DNA polym erase in Taq DNA polym erase buffer con­
taining 2.5 m M M gCl 2 . After an initial d en atu ratio n step
at 94°C for 1 m in, 30 cycles o f 94°C for 1 m in, 58°C for
30 s, and 72°C for 2.5 m in w ith a term inal 10-m in ex­
tension at 72°C were perform ed. Excess prim ers and n u ­
cleotides were rem oved by adding 0.5 pL Exonuclease 1
and 0.5 pL Shrim p Alkaline Phosphatase to 20 pL o f PCR
reactions and incubating at 37°C for 15 m in (New
England Biolabs, Beverly, MA, USA). These enzymes
were th en deactivated w ith a 15-m in incubation at 85°C.
The 16S rRN A genes were sequenced w ith the T7
p ro m o ter prim er and the M 13-9 prim er by using the
BigDye T erm inator Cycle Sequencing kit v. 3.0 (PE
Biosystems, Foster City, CA, USA) following the m a n ­
ufactu rer’s directions. Sequencing products were ana­
lyzed at the Iowa State U niversity D NA Sequencing
Facility.
Sequences were edited in Sequencher 4.1 (Gene Codes
Co., A nn A rbor, M l, USA) and subjected to BLAST
searches [1]. 16S rRNA gene sequences were th en sub­
jected to chim era check in Ribosom al D atabase Project
(RDP) [10] as well as w ith Bellerophon [19] and aligned
in Dr. Phil H u g e n h o ltz’s 16S rRN A ARB database
(h ttp ://rd p 8 .c m e .m su .e d u /h tm l/a lig n m e n ts.h tm l). A p ­
proxim ately six chim eras were detected in each soil
sam ple and excluded from fu rth er analyses. Closely
D .R . N
related sequences from the ARB database an d from
BLAST searches were used as reference taxa for phyloge­
netic analyses. A lignm ents were subjected to Bayesian
phylogenetic analysis im plem ented in MRBAYES [20]
using the GTR -I- gam m a m odel o f evolution w ith
1,000,000 generations, sam pling trees every 100 gener­
ations. B urnin values were determ ined by plotting the
likelihood scores against generation n u m b er an d retain ­
ing trees for w hich stationarity was evident. T axonom ic
affiliations were assigned based on b o th BLAST m atches
and Bayesian phylogenetic analysis. O TU abundance was
calculated by generating a distance m atrix in ARB and
im portin g it in to D O T U R [42]. Phylogenetic trees for the
Cyanobacteria and B etaproteobacteria were constructed
in PA U P 4.0b [51] u sin g b o th th e d ista n c e an d
parsim ony optim ality criterion an d boo tstrap analysis
(1000 replicates). The S hannon index o f diversity was
calculated using the equation
H ' = ~ ^ P i Inpi
w here p; is th e p ro p o rtio n o f cover c o n trib u te d by th e
ith species [45]. P h ylogenetic diversity was calculated by
tran sfo rm in g th e se q u en ce d a ta in to pairw ise distances
using th e T a m u ra -N e i m o d e l o f evolution w ith a =
0.5, a n d calculating th e average pairwise d ifferen ce (in
% seq u e n c e differen ce) usin g A R LEQ U IN [44]. A d o p t­
in g d iffe re n t m o d e ls o f se q u e n c e ev o lu tio n w hen
estim atin g pairwise se q u e n c e d ifferen ce yield ed q u a n ­
titative differences, b u t d id n o t affect th e p a tte rn across
sites.
Soil Parameters.
For total carbon an d nitrogen
analyses, soils were g ro u n d to a fine pow der, and a su b ­
sam ple o f each soil was a c id -tre ate d fo r carb onate
rem oval. Briefly, 30 mL o f 0.5 N H C l were added to 2 g
soil. A fter an h o u r at ro o m tem perature, soil slurries were
centrifuged at 5000 rp m an d the acid was rem oved. Acidtreated soils were th en w ashed twice w ith d iH 2 0 . All soils
(treated an d u n treated) were dried at 70°C for 18 h.
A pproxim ately 30 m g o f dried soil was packaged in to tin
capsules an d % C and N was m easured by using a Carlo
E rba c o m b u s tio n -re d u c tio n elem ental analyzer (CE
Elantech, Lakewood, NJ, USA).
Labile (resin-extractable) phosp h o ru s content was
determ ined o n soil samples by using a m odification o f
the protocol described by Tiessen an d M oir [53]. First,
soil samples were dried at 40° C for 10 h, an d 1 g o f dried
soil was placed in a 50-m L conical tube w ith 30 mL
deionized H 2 O and a charged anion-exchange m em brane
strip. After shaking at 200 rp m for 16 h, resin strips were
rem oved and placed in a 50-m L conical tube, then
ex tracted w ith 20 mL o f 0.5 N H C l. P h o sp h o ru s
c o n ce n tra tio n s w ere d e te rm in e d colo rim etrically by
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
using the m odified version o f the M urphy-R iley m ethod
described by Tiessen and M oir [53].
Soil pH was determ ined in a 1:2 soil/deionized w ater
slurry, by using a digital pH m eter (Fisher Scientific,
Pittsburgh, PA, USA). G ravim etric soil m oisture content
was determ ined by weight loss after 24 h at 60°C.
Statistics.
W e used one-w ay ANOVAs (SYSTAT,
V ersion 10) w ith distance from the glacier as the cate­
gorical variable to test for differences in soil C, N, P, pH ,
m oisture, and n u trien t ratios along the chronosequence.
16S rRNA gene libraries were com pared by using b o th the
FST and P test as described by M artin [28]. In addition,
we introduce a new test statistic th at quantifies the
differences in the phylogenetic com position o f m icrobial
com m unities based on com paring the likelihood scores
o f individual and com bined com m unities. To do so, we
generated neighbor-joining trees for each com m unity
after transform ing the sequence data into a pairwise
distance m atrix using the HKY + gam m a + invariant
sites m odel o f sequence evolution. The gam m a rate value
and the p ro p o rtio n o f invariant sites were estim ated to
be 0.58 and 0.30, respectively, based on a likelihood
analysis o f a large n u m b er o f sequences. U sing the same
m odel o f sequence evolution, we determ ined the -In
likelihood for the data using PAUP 4.0b [51]. The same
analysis was applied to all pairs o f soil 16S rRNA gene
libraries. The statistic is
D ab = [~ In T a b /( ~ In L-a H— In Tb)] ~ 0.5,
w here -InL is th e negative log lik elih o o d score fo r the
d a ta fro m a given co m m u n ity a n d A a n d B are th e two
d a ta sets fo r w hich we can calculate th e lik elih o o d o f
th e d a ta given th e m o d e l o f evolution a n d th e topology.
If th e two co m m u n ities are identical, nam ely, th a t
-InfiA = -InLe, th e n -lnLAB= -InfiA = -InL e a n d D ab = 0.
Phylogenetic sim ilarity o f co m m u n ities was th e n visual­
ized using n eig h b o r-jo in in g cluster analysis.
A c c e ssio n N um bers.
All sequences fro m this
study were deposited in Genbank. The sequences from
the OMl library were nam ed w ith the prefix 0M 1_ and
were assigned to the accession num bers D Q 513833D Q 5 13909; sequences from the 0M2 library were nam ed
w ith the prefix 0M 2_ and were assigned to the accession
nu m b ers D Q 513910-D Q 513971; sequences from the
lOOMl library were nam ed w ith the prefix 100M1_ and
were assigned to the accession num bers D Q 513972DQ514041; sequences from the 100M2 library were
nam ed w ith the prefix 100M2_ and were assigned to
the accession num bers D Q 514042-D Q 514115; sequences
from the 500M1 library were nam ed w ith the prefix
500M 1_ and were assigned to the accession num bers
D .R . N
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
phylotypes in the 0 m sites are very closely related. In
contrast, lineage per tim e plots for the older soils are
m ore gradual, fu rth er supporting the idea th at these sites
h arbored greater phylogenetic diversity th a n the 0 m soils
(Fig. 2; Table 1). Interestingly, although there are no
obvious differences in the diversity sam pled betw een the
100 an d 500 m soils, the C h a o l-d e riv e d p redicted
n u m b er o f lineages for the 100 m soils is 600, twice as
high as the average for the 500 m soils (Fig. 2).
It is im p o rta n t to n o te here th a t we have n o t
exhaustively sam pled the com m unities in these soils,
w hich could com plicate o u r interpretations o f changes in
m icrobial com m unity diversity over the soil age gradient.
In particular, rarefaction analysis reveals th at diversity at
O TU definitions o f <5% sequence difference are u n d e r­
sam pled (Fig. 3). However, this analysis also suggests th at
at an O TU definition o f <10% these soils were bettersam pled. This tren d is also revealed by com paring lineage
p er tim e plots (i.e., plotting the n u m b e r o f unique
lineages present in a collection o f sequences over a range
o f given phylogenetic distances) w ith plots o f the
D Q 5 1 4 1 1 6 -D Q 5 1 4 1 8 4 ; sequences fro m th e 500M 2
library were nam ed with the prefix 500M2_ and were
assigned to the accession num bers DQ514185-DQ514243.
Results
16S rRNA G ene Libraries.
Across all six sites sampled,
we obtained 411 sequences and detected a total o f 181
distinct 168 rRNA phylotypes (>3% sequence difference
from all o th er sequences) representing 15 m ajo r bacterial
and 2 plastid lineages (Fig. 1). N otably, overall patterns
o f m icrobial diversity varied along the chronosequence.
For example, phylotype and division-level richness as
well as overall phylogenetic diversity o f the sequences was
highest in the older soils (100 an d 500 m ) an d lowest in
the youngest soils (Fig. 1, Table 1). Likewise, the Shannon
diversity index im plied th a t evenness increased from the
0 to the 100 m soils; evenness appeared to be sim ilar in
the 100 and 500 m soils (Table 1). Lineage p er tim e plots
(lower line in Fig. 2) feature a steep initial slope in the
youngest soils, im plying th a t m any o f the individual
p G e m m a tim o n a d etes
■A cidobacteria t— B e ta p ro te o b a c te ria
I^andidate D iv is io n W Y O
■B etaproteobacteria
R eplicate
- V erru co m icro b ia
C ya n o b a c teria
B a c tero id etes
I—C ya n o b a c teria
S tra m e n o p ile p la stid
~ G am m a p ro te o b a c te ria
p G am m a p ro te o b a c te ria
~ B a ctero id etes
-r> .
i.
■I
p B e ta p ro te o b a c te ria
B e ta p ro te o b a c te ria
A c id o b a cteria
C ya n o b a c teria
^ C y a n o b a c te r ia
“ S tra m e n o p ile p la stid
Om
A lp h a p ro te o b a c te ria
C ya n o b a c teria
B a c te ro id e te s
- S tra m e n o p ile p la stid
- A lp h a p ro te o b a c te ria
100 m
*—A c id o b a c te ria
- V erru co m icro b ia
- C an d id ate D iv is io n W Y O
—G em m a tim o n a d etes
C ya n o b a c teria
—B a c te ro id e te s
- A lp h a p ro te o b a c te ria
^ A c id o b a c te r ia
5 0 0 111
Figure 1. R e la tiv e a b u n d a n c e o f SS U rD N A p h y lo ty p e s in c lo n e lib ra rie s f ro m d iffe re n t soils. C olors r e p re s e n t d iv isio n - o r su b d iv is io n level clad es a n d black lines re p re s e n t in d iv id u a l p h y lo ty p e s. N u m b e r o f se q u e n c e s = 77 (0 m - 1 ), 62 (0 m - 2 ), 70 (1 0 0 m - 1 ), 74 (1 0 0 m - 2 ),
69 (5 0 0 m - 1 ), 59 (5 0 0 m - 2 ). Q A lp h a p ro te o b a c te ria , ^ Cyanobacteria, ^ B e ta p ro te o b a c te ria , Q Acidobacteria, Q Bacteroidetes, □ S tra­
m e n o p ile p l a s t i d , ^ Planctomycetes, ^ c a n d id a te division O PIO , Q G a m m a p ro te o b a c te ria , ^ ChloroflexiJ"^ Chlorophyta p lastid , ^ D eltap ro te o b a c te ria , ^ Verrucomicrobia, □ Deinococcus, Q Actinohacteria, □ Gemmatimonadetes, □ c a n d id a te division WYO.
D.R.
N
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
Table 1. Diversity indices for the six clone libraries
D iv isio n s ( o r p r o te o b a c te r ia l s u b d iv is io n s)
P h y lo ty p e s (> 3 % se q u e n c e d iffe re n c e )
S h a n n o n in d e x
P h y lo g e n e tic d iv e rsity
0 m -1
0 m -2
100 m -1
100 m -2
5 0 0 m -1
5 0 0 m -2
5
16
0.8
18.8
9
25
3.9
18.9
12
42
4.9
31.2
11
47
4.8
28.2
12
39
5.1
32.3
12
38
4.6
32.9
n onparam etric diversity estim ator C h a d over a range o f
genetic distances (Fig. 2). These com parisons show th at
although is clear th a t we have n o t sam pled the com plete
diversity at distances o f less th a n 0.1 (10% sequence
difference), the lines do converge at a distance o f about
0.1 for each soil. This im plies th at at this phylogenetic
i"
600
600
500
500
400
400
300
^
distance— som ew here betw een “genus” and “division”
level differences— we have sam pled enough diversity to
m ake fairly accurate claims about the diversity in these
soils.
The frequencies o f sequences related to specific clades
varied m arkedly along the chronosequence (Fig. 1). M ost
300
H
O
200
200
100
100
0
0
0
0.3
0.2
0.1
0.1
genetic distance
1400
1400
1200
1200
1000
1000
800
(/)
3p
(/)
3
H
o
600
o
200
200
0.3
0.2
0.2
0.3
600
400
0.1
0.3
800
400
0
0.2
genetic distance
0.4
genetic distance
0
0.1
genetic distance
e
500
450
400
350
300
D
1—
( )
Figure 2. L in e ag e p e r tim e p lo t
250
200
(low er line) a n d e s tim a te d d iv e rsity
150
p lo ts (u p p e r lin e w ith e r r o r b a rs) fo r
100
th e six sa m p le s, (a) O m - 1 , (b ) 0 m -
50
2, (c) 100 m - 1 , (d ) 100 m -2 , (e) 500
0
m -1 , (f) 500 m -2 . O T U s a n d th e
0.1
0.2
genetic distance
0.3
0.1
0.2
genetic distance
C h a o l d iv e rsity e s tim a te w e re c a l­
c u la te d in D O T U R [42].
D .R . N
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
address this, we applied the FST statistic and the P test,
tw o statistical tools developed to com pare clone library
data [28]. These tests o f w hether pairs o f com m unities
are com posed o f significantly different pools o f phyloge­
netic diversity revealed th at the tw o samples from the
oldest soils were statistically indistinguishable using both
hierarchical analysis o f variance (FST statistic) and
phylogenetic character m a p p in g (P test) (d ata n o t
show n). Sam ples fro m the youngest soils h a rb o red
sim ilar phylogenetic diversity (a nonsignificant FST test
w hen corrected for m ultiple com parisons) b u t contained
different sets o f clades (significant P test), a p attern th at
reflects the presence o f different, b u t very closely related,
clades in the two samples (data n o t show n). All other
dram atically, sequences related to the Betaproteobacteria
d o m in a te d in th e y o u n g est soil (6 3 -6 6 % o f to tal
sequences), decreased in abundance in the 100 m soil
(6-19% ), an d were rare in the oldest soil (1-3% ). The
in te rm e d ia te -a g e d a n d o ld est soils h a rb o re d m o re
sequences related to com m o n soil bacteria, including
Acidobacteria, Verrucomicrohia, an d Bacteroidetes. These
older soils also appeared to contain m ore o f the m ost
divergent lineages sam pled in this study, including
sequences related to the candidate division WYO.
A lthough it appeared th a t som e p atterns in the data
were consistent betw een replicates (Figs. 1 and 2; Table 1),
o u r sam pling schem e d id n o t allow us to test for
significant differences using trad itio n al statistics. To
30
40
35
25
30
20
25
15
Z)
I----
■.■■■■
10
o
.................
20
15
10
5
5
0
0
0
10
20
30
40
50
60
70
80
0
10
sequences sampled
30
20
40
50
60
70
sequences sampled
60
70
50
60
50
40
40
30
3 30
I—
o
20
20
10
10
0
0
0
10
20
30
40
50
60
70
0
10
sequences sampled
20
30
40
50
60
70
80
sequences sampled
50
45
40
35
3
I—
o
30
Figure 3. R a re fa c tio n c u rv e s c o n ­
25
s tr u c te d f ro m v a rio u s O T U d e f in i­
20
t io n s (se q u e n c e d ifferen ces: b o ld
15
lin e = 1 % ; s q u a re s = 3% ; tria n g le s =
5% ; h a tc h m a rk s = 10% ) f o r th e six
10
sa m p le s, (a) O m - 1 , (b ) 0 m -2 , (c)
5
0
100 m - 1 , (d ) 100 m -2 , (e) 5 00 m -1 ,
0
10
20
30
40
50
sequences sampled
60
70
0
10
20
30
40
sequences sampled
50
60
(f) 500 m - 2 . O T U s w e re c a lc u la te d
i n D O T U R [42].
D .R . N
pairwise com parisons for b o th tests were significant,
im plying the o th er soils sam pled h arb o red different
com m unities o f bacteria (data n o t show n).
N ext, we u sed a lik e lih o o d -b a se d m e th o d fo r
estim ating the phylogenetic distance betw een co m m u n i­
ties, an d we sum m arized the sim ilarities using cluster
analysis (Fig. 4). The results m irro red the FST and P
tests, and the 0 m soils grouped together to the exclusion
o f all o th er samples, suggesting th a t these replicate
sam ples fro m th e y o u n g est soils h a rb o re d sim ilar
lineages. Likewise, the 500 m soils grouped together to
the exclusion o f all o th er libraries, sup p o rtin g the fact
th at the replicate samples from the oldest soils contained
sim ilar organism s. The tw o replicates from the 100 m
sites, however, were about as different from each o th er as
either was from the 0 o r 500 m soils.
Phylogenetic A nalyses.
W e applied a m ore de­
tailed phylogenetic analysis to the betaproteobacterial
sequences, a bacterial lineage th a t appeared to d ram a ti­
cally decrease in abundance along the soil chronosequence
(Fig. 1). N early all o f the betaproteobacterial sequences
from the 0 m soils clustered w ithin the C om am onadaceae
group (Fig. 5). The bootstrap su p p o rt for the exact ph y ­
logenetic position o f these sequences w ithin this group
was weak, likely due to the lack o f difference in the 16S
rRNA genes o f very closely related organism s. However,
these sequences appeared to be specifically related to the
genera Variovorax, Polaromonas, an d Rhodoferax. In
addition, these sequences were closely related to two
genes obtained from basal ice o f the John Evans Glacier,
N unavut, C anada [49]. A lthough the 100 an d 500 m soils
also h arb o red a few betaproteobacterial lineages, these
soils contain lineages th a t are phylogenetically distinct
from those in the 0 m soil (Fig. 5). This suggests th a t n o t
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
only does the abundance o f the B etaproteobacteria shift
considerably along the soil age gradient, b u t also th a t the
types o f B etap ro teo b acteria in the oldest soils are
different from the younger soils (Fig. 5).
W ith the exception o f the 0 m -2 library, sequences
related to the C yanobacteria com prised at least 20% of
the genes in each clone library (Fig. 1); therefore, we
applied detailed phylogenetic analyses to these sequences
as well (Fig. 6). There were different types o f cyanobacte­
rial sequences in the clone libraries from these sites, and
the diversity appeared to shift notably along the c h ro ­
nosequence. In the 0 m soils, the m ajority o f sequences
were related to Chamaesiphon subglobosus w ithin the
Chroococcales, and Leptolyngbya sp. SV l-M K -52 w ithin
the Oscillatoriales. The diversity o f cyanobacterial sequen­
ces increased in the 100 m soils, and included sequences
related to b o th the Oscillatoriales and the Nostocaceae.
T he 500 m soils co n ta in e d a less diverse suite o f
cyanobacterial genes, including an a b u n d an t sequence of
uncertain phylogenetic position, w hich appears to be
distantly related to a 16S rRNA gene from a seafloor
sedim ent sample (Fig. 6).
Soil Nutrient a n d Physical Parameters.
N utrient
pools in these soils were low, approxim ately 1 m g g^^
organic carbon, 0.1 m g g^^ nitrogen and 2 pg g^^ resinextractable P, whereas pH and inorganic C concentrations
were relatively high in all soils (Table 2). There was a
sig n ific a n t in c rea se in soil N c o n te n t a lo n g th e
chronosequence. In addition, there was a decreasing,
b u t nonsignificant tren d in resin-extractable P along the
soil age gradient. W ater content in all soils was low, and
the 500 m soils were significantly drier th a n the other
soils at the tim e o f sam pling. Soil C /N ratios did n o t
significantly vary along the chronosequence (Table 2).
Discussion
100 m - 1
500 m- 2
100 m- 2
500 m - 1
Om- 2
Figure 4. N e ig h b o r -jo in in g to p o lo g y s u m m a r iz in g th e p a irw ise
d iffe re n c e s in th e p h y lo g e n y o f e a c h o f th e c o m m u n itie s b a s e d o n
th e lik e lih o o d te s t sta tistic . B ra n c h le n g th s a re p r o p o r tio n a l to
th e p h y lo g e n e tic d iffe re n c e b e tw e e n c o m m u n itie s .
A lth o u g h m o st soil b io g e o ch em ic al c h a rac teristic s
show ed little change along the ~ 20-year soil age gradient
(Table 2), o u r m olecular phylogenetic analyses revealed
dram atic shifts in the associated m icrobial com m unity
com position along the chronosequence (Figs. 1, 2, 3, 4, 5
and 6; Table 1). O u r results suggested b o th significant
changes in m icrobial com m unity com position along the
chronosequence, as well as an overlap in the replicate
com m unities o f b o th the youngest and the oldest soils
(Fig. 4). In addition, o u r results suggest th at autotrophic
N fixers are abu n d an t in these soils (Figs. 1 and 6), and
m ay be im p o rta n t in the biogeochem istry o f newly
deglaciated, unvegetated soils. Below, we fu rther elabo­
ra te o n h o w specific in fo rm a tio n a b o u t m icro b ial
com m unity com position can advance o u r understanding
about general successional patterns in this system.
D .R . N
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
A lc a lig en es faecalis
. P ig m e n tip tia g a ktiHae
I A l c a i i g e n a c e a e
_ B u rk h o ia eria c e p a c ia
I B u i K h o t d e r i a c e a e
' C u p r ia v /d u s n e c a io r
D Ltganeila z o o g lp e o iP e s
J a p th ip o b a a eriu n n lividijm
□
_ T ^
O x a b b a cteo c e a e
H e rb a sp iriltijm fr /s in g e n s e
fibodofera^ fermentani
m
G l a c i e r b a s a l i c e J E G . c 3
I
I
G l a
J * C .^
4Z
c i e r b a s a l ic e J E G e l
□
Poiai'O'POPas ^acuolafa
a
—
Comamoredaceae
v a rio i/o ra x p a r a d o x u s
I
C o m a m o r r a s r e s fo s ie r o n ^
' ----------------H y d r o g e n o p h a g s d e tlu v ii
B r a c iiy m o n a s d e n irriflc a n s
■L a rrtp ro p e d ia b y a lia a
X e n o p b ila s a z o v o f a n s
#?am/fi)acfe'' irenc/irrerjsJS
A cid o vo ra x a v e n a e cilrulii
— D eiftia I s u r u b a te n s is
P o b r iv iv a i go/ar/poSOS
t,epfoiPr/x cb o /o d n ii
r F
Sp/>earornus p a ra p s
ED
B u r k h o l d e r i a l e s G e n e r a
i n c e r t a e s e d i s
IVfffsuarja crtiiDsafi/rabba
T P io m op as rP erm o su /^ ta
ScP /ege/erra in e r n io d ep o /y p r er a n s
□
— O.DOS substitutions/sile
Figure 5. N e ig h b o r -jo in in g p h y lo g e n e tic tre e s h o w in g r e p re s e n ta tiv e se q u e n c e s f ro m th e clo n e s re la te d to th e B e ta p ro te o b a c te ria .
R e ctan g les re p re s e n t se q u e n c e s f ro m th is stu d y . T h e re la tiv e size o f th e re c ta n g le sy m b o liz e s th e n u m b e r o f se q u e n c e s in th e 0 m (w h ite ),
100 m ( d o tte d ), a n d 500 m (b lack ) soils. T re e is r o o te d w ith R h o d o b a cter sp . H T C C 5 1 5 (A Y 584573) a n d S p h in g o m o n a s oligophenolica
(A B 0 1 8 4 3 9 ). *: A p a r s im o n y a n d d is ta n c e b o o ts tr a p v a lu e o f 80 o r h ig h e r. A c c e ssio n n u m b e rs : A lcaligenes fa eca lis (A Y 667065),
P ig m en tip h a g a k u lla e (A F 2 8 2 9 1 6 ), B u rkh o ld eria cepacia (A Y 099314), C u p ria v id u s necator (A F 1 9 1 7 3 7 ), D u g a n ella zoogloeoides (D 1 4 2 5 6 ),
J a n th in o h a c te riu m liv id u m (A Y 581141), H erh a sp irillu m frisin g en se (A Y 043372), R h o d o fe ra x fe r m e n ta n s (D 1 6 2 1 2 ), G lacier b a s a l ice JE G .C 3
(D Q 2 2 8 3 9 5 ), G lacier b a s a l ice JE G .e l (D Q 2 2 8 4 0 3 ), P o la ro m o n a s va cu o la ta (U 1 4 5 8 5 ), V a rio vo ra x p a ra d o x u s (A F 5 3 2 8 6 8 ), C o m a m o n a s
testosteroni (A J6 0 6 3 3 6 ), H yd rogenophaga d e flu v ii (A J585993), B ra ch ym o n a s d en itrifica n s (D 1 4 3 2 0 ), L a m p ro p e d ia h ya lin a (A Y 291121),
X e n o p h ilu s a zo vo ra n s (A F 2 8 5 4 1 4 ), R a m lih a c te r henchirensis (A F 4 3 9 4 0 0 ), A c id o v o ra x a ven a e c itru lli (A F 1 3 7 5 0 6 ), D elftia tsu ru h a ten sis
(A J6 0 6 3 3 7 ), R u h r iv iv a x g elatinosus (A M 0 8 6 2 4 2 ), L e p to th rix discophora (X 9 7 0 7 0 ), S ph ea ro tilu s n a ta n s (Z 1 8 5 3 4 ), M its u a ria ch ito sa n ita h id a
(A Y 8 5 6 8 4 1 ), T h io m o n a s th erm o su lfa ta (U 2 7 8 3 9 ), Schlegelella th erm o d ep o lym era n s (A Y 538709).
G en era l P a tte rn s in M icrobial D iversity.
The
n u m b er o f uniq u e sequences represented in the clone
libraries and the S hannon index o f diversity doubled
betw een the 0 an d 100 m soils, an d th en plateaued in the
500 m soils (Fig. 1, Table 1). Similarly, the n u m b er o f
bacterial divisions and phylogenetic diversity increased
by approxim ately 60% betw een the 0 and 100 m and
leveled off in the 500 m soils (Table 1). This initial
increase in diversity (from 0 to 100 m ) corroborates
o ther estimates o f functional diversity based on m icrobial
D .R . N
processes an d enzyme activity in recently deglaciated soils
[54]. In addition, it is sim ilar to the w ell-docum ented
increases in p lan t com m u n ity diversity observed in suc­
cessional soils [30], and a recent study th a t d em onstrated
an increase in archaeal diversity along a successional
gradient [33]. O u r results are n o t consistent w ith a
ribosom al fingerprinting study, which suggest a decrease
in m icrobial diversity w ith soil age in early successional
soils [48]. However, clone library techniques are m ore
sensitive to sequence variation in the 16S rRNA molecule,
and therefore m ay be a m ore accurate reflection o f the
com m unity dynamics. This increase in the diversity o f the
m icrobial com m unity m ay be attributable to an increase
in the diversity and am o u n t o f energy and carbon inputs
LB3-63 Antaictic lake ice
•tolyngbya sp. SVl-M K-52
LB3-T6 A ntarctic lake ice
Fr121 Antarctic microbial m at
P sauPanabaena tremuia
UIZ36 siliceous rock
BSC- T6- 2 gypsum rock
Leptolyngbya sp
L B Y rSA ntarctK lake ice
Jtp —- Cham aesiption su bglobosu s
'---------- LB3-47 Antarctic lake ice
-
Oscillalorra sp PCC 7112
M icrocoleus vaginatus PCC 9802
Tyctionema bourrellyt
AKIW114 2 urban aero so l
Sym ploca sp. V P377
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
into the system, both in the forms o f available light and
organic m aterial from deposition, perm itting growth o f a
broader suite o f microorganisms. In addition, changes in
microecological interactions (e.g., competition) could result
in a m ore diverse microbial community.
Interestingly, Ju m pponen [24] show ed a decrease in
fungal diversity along an early successional, unvegetated
gradient o f Lym an glacier in the N o rth Cascade M o u n ­
tains (W ashington, USA). H e attrib u ted this decrease in
diversity to the presence o f a diverse, b u t dorm ant, spore
b an k in younger soils th at was subsequently displaced by
a less diverse, b u t active fungal com m unity. Additionally,
Ju m pponen [24] observed high spatial heterogeneity in
the fungal com m unities in recently deglaciated soils, as
the m ajority o f the 37 phylotypes th at he sequenced were
fo u n d in only 1 o f the 16 soil samples th a t he exam ined.
This m ay suggest th a t fungal com m unity assembly is
m ore stochastic o r variable th an bacterial succession in
early successional soils.
Microbial Dispersal.
Sequences related to the Beta­
proteobacteria were m ost d o m in an t in the youngest soils,
and th eir abundance decreased m arkedly w ith soil age.
Phylogenetic analysis placed the m ajority o f the sequences
from the 0 and 100 m soils in the C om am onadaceae
cluster o f the B etaproteobacteria (Fig. 5). O rganism s from
this clade have been cultured from several “frozen”
en v iro n m en ts, in clu d in g glacial ice cores [46] and
p erm anent A ntarctic lake ice [17]. In addition, related
sequences have been fo u n d in 168 rRNA gene libraries
from subglacial ice above Lake V ostok, A ntarctica [37],
as well as from u n d erneath the Bench Glacier in Alaska
Figure 6. N e ig h b o r-jo in in g p h y lo g e n e tic tree sh o w in g re p re se n ta tiv e
se q u en ces f ro m th e clo n es re la te d to th e C y a n o b a c te ria . R ectangles
re p re se n t se q u en ces f ro m th is stu d y . T h e relative size o f th e rectan g le
O D P 1230B 22.23 su b seafio o r sedim ent
D
sym bolizes th e n u m b e r o f se q u en ces in th e 0 m (w h ite), 100 m
( d o tte d ), a n d 500 m (b lack ) soils. T re e is r o o te d w ith R hodobacter sp.
H T C C 5 1 5 (A Y 584573) a n d S p h ingom onas oligophenolica
(A B 018439). *: A p a r s im o n y a n d d ista n c e b o o ts tr a p v a lu e o f 80 o r
h ig h er. A ccession n u m b e rs: LB 3-53 (A F 076159), L eptolyngbya sp.
Mosloc s p 9229
Nostob caicicoia VI
N ostocsp. 6936
Hostoceaapt<icum
^ N ephrom a par lie cy an o b io rt
N ostoc sp.
Nadutaria spbaeracarpa
P seudanabaena trem u la (A F 218371), M 1Z 36 (A B 179527), B S C -16-2
A n abaen a augstum alis
Foonema oceiiaium
frichormus doliolum
9802 (A F 284803), T ychonem a bourrellyi (A B 045897), A K IW 1142
S V l-M K -5 2 (A Y 239604), LB 3-76 (A F 076158), F rl2 1 (A Y 151728),
(A Y 422697), L eptolyngbya sp. (X 84809), LB 3-75 (A F 076164),
C h a m aesiphon subglobosus (A Y 170472), LB 3-47 (A F 076163),
O scillatoria sp. P C C 7112 (A B 074509), M icrocoleus va g in a tu s P C C
□Lepiolyngbya sp. OB32 4 8 0 4
(D Q 1 2 9 6 4 5 ), Sym ploca sp . V P 3 7 7 (A F 306497), 66863799
Leptolyngbya sp. 0BB 19S12
Leptolyngbya sp. 0BB 30S02
(A J630448), N ostoc sp. 8938 (A Y 742454), N ostoc edaphicum
(A B 177177), N ostoc sp. P C C 9229 (A Y 742451), N ostoc calcicola V I
(A J630449), N e p h ro m a p a rile c y a n o b io n t (A F 506257), N osto c sp.
(A F 027655), N o d u h r ia sphaerocarpa (A J781149), A n a b a en a
au g stu m a lis (A J630458), S tig o n em a ocellatum (A J544082), Trichorm u s d o lio lu m (A J630455), L eptolyngbya sp . 0B B 24S04 (A J639893),
L eptolyngbya sp. 0BB 19S12 (A J639895), L eptolyngbya sp. 0BB30S02
0.01 substitution S/Site
(A J639892).
D .R . N
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
Table 2. Biogeochemical parameters for the three soils
S a m p le (m )
0
100
500
A p p ro x. age (years)
O rganic C
N
P
pH
% H 2O
C /N
< 1
1 -4
20
1.29 (0 .3 4 )“
1.14 (0 .1 8 )“
1.59 (0 .6 3 )“
0.08 (0 .0 0 )“
0.09 (0 .0 0 )“
0.11 ( O .O lf
2.52 (0 .8 5 )“
1.71 (0 .1 7 )“
1.53 (0 .3 6 )“
7.5 (0 .2 )“
7.6 (0 .1 )“
7.5 (0 .1 )“
9 .1 6 (1 .2 3 )“
9.81 (0 .9 0 )“
2 .5 4 (2.26)*”
15.3 (3 .2 0 )“
12.9 (2 .2 3 )“
15.3 (7 .5 0 )“
O rganic C and N are expressed in mg-g , P is the resin-extractable fraction expressed in pg-g . N um bers in parentheses are standard deviations, values
followed by the same letter are n ot significantly different from one another (P < 0.05). For each m easurem ent, n - 4.
and John Evans Glacier in N unavut, Canada [49]. O ther
sequences from the Com am onadaceae cluster have been
obtained from inside a deep sea rock [21], as well as from
the vadose zone o f a California Mollisol [26].
The presence o f close relatives in rock and ice
environm ents suggests th a t these bacteria m ay be living
in o r u n d e r the glacier, an d th at m elting glacial ice or
exposure o f subglacial rock m ay seed the youngest soils
w ith organism s from the C om am onadaceae. If so, this
m ay explain the fact th a t the youngest soils appeared to
h arb o r different, b u t very closely related clades in the two
samples (Figs. 1 an d 4). If these organism s originated
from w ithin or beneath the glacier, they could have been
physically an d genetically iso lated for h u n d re d s to
thousands o f years, allowing for local speciation events.
This hypothesis is in agreem ent w ith other studies th at
dem onstrated b o th functional an d phylogenetic differen­
tiatio n o f bacteria from geographically isolated frozen
environm ents [50].
O n average, the phylotypes from the youngest soils
were m uch m ore closely related to oth er sequences in
G enbank th a n th e phylotypes fro m th e older soils
(average BLAST score ~ I7 0 0 bits in the 0 m soils;
~ I4 0 0 bits in all older soils). This p a tte rn is prim arily
driven by the dom inance o f the B etaproteobacteria in the
youngest soils; if these phylotypes are rem oved from the
analyses the average BLAST score averages ~ I4 0 0 bits in
the 0 m soils as well. Some o f the very closely related
betaproteobacterial sequences in G enbank are from other
frozen environm ents aro u n d the w orld (Fig. 5), suggest­
ing th a t geographically separated ice environm ents m ay
select for sim ilar suites o f organism s. However, m any o f
the closely related betaproteobacterial sequences are from
o th er types o f environm ents, ranging from drinking water
to co ntam inated soils. This m ay indicate th at early m icro ­
bial colonizers are m ore “ co sm opolitan”, and are replaced
by m ore endem ic species along the soil age gradient.
Nutrient Dynam ics.
The C om am onadaceae is a
phenotypically diverse group o f organism s (Fig. 5), w hich
makes functional predictions for this ab u n d an t group
difficult. Some o f the cultured relatives o f these organisms
are autotrophs, and can fix carbon photosynthetically [39]
or using H 2 as an electron source [22]. However, m ost
other isolates from the Com am onadaceae are hetero­
tro p h s [58], a n d are capable o f m etabolizing b o th
recalcitrant [36] and m ore labile [46] carbon sources.
The high rates o f pollen deposition in this region [38]
m ay provide an im p o rta n t carbon source for these
organism s in an otherwise C-lim ited environm ent. A ddi­
tionally, closely related organisms th at have been isolated
from sim ilar environm ents have been cultured on organic
carbon sources [17, 46], suggesting th at these bacteria live
a heterotrophic lifestyle. This is sim ilar to m acrobial
successional patterns, where heterotrophs (insects) are
often am ong the first organism s to arrive in recently
deglaciated soils [18].
Photosynthetic au to tro p h s appeared to be com m on
in all o f the soil samples th a t we analyzed. For example,
sequences related to the C yanobacteria were abu n d an t in
m ost o f the soils (Fig. I). In addition, sequences sim ilar
to stram enopile plastids were p ro m in en t in the 100 m
libraries, com prising 16% and 8% o f genes from these
soils (Fig. I). The im portance o f photosynthesis in early
succesional soils is obvious; in the absence o f a large pool
o f organic carbon, C-fixing organism s w ould clearly have
an advantage over heterotrophs. The significance of
photosynthesis in early successional com m unities is also
su pported by w ork in aquatic systems [4] th at show
increasing chlorophyll a abundance w ith distance from a
glacier. Interestingly, although there was an increasing
trend, there was no significant difference in the organic C
pools along the chronosequence (Table 2). This suggests
th at although these carbon-fixing organism s appear to
m ake up a large p o rtio n o f the m icrobial com m unity,
they are n o t adding detectable am ounts o f fixed carbon
to these soils. However, m any C yanobacteria can live both
a u to tro p h ic ally an d m ixotrophically, an d th e actual
biogeochem ical roles o f the organism s represented by
these sequences in these soils are unknow n. The lack o f a
significant increase in organic C pools also suggests that
although rates o f deposition are high in this region,
cum ulative pollen inputs do n o t result in significant
increases in soil organic m a tte r in this tim e frame.
O u r data also contain som e interesting parallels to N
cycling in plant successional com m unities. The signifi­
cance o f N fixation in the ecosystem dynam ics o f young
plant com m unities is well established, and o u r w ork
dem onstrates th at free-living N fixation m ay be im p o r­
ta n t p rio r to plant establishm ent. For example, phyloge­
netic analysis (Fig. 6) show ed th a t the overwhelm ing
m ajority o f C yanobacteria in these soils are closely
D .R . N
related to heterocystous (e.g., Anabaena, Nostoc) and
nonheterocystous (e.g., Oscillatoria, Microcoleus) n itro ­
gen fixers. If these organism s have the ability to fix
nitrogen, this could explain the significant increase in N
in the older soils (Table 2). Indeed, recent w ork from our
laboratory (Reed an d Schm idt, unpub lish ed data) shows
th at N fixation (acetylene reduction) rates in the 100 m
soils are m u ch higher th a n in the 0 m soils. However, it
should be em phasized th a t the organism s responsible for
this activity are still unknow n, and th at N fixation m ay
also be driven by h etero tro p h ic organism s in this system.
M any o f the 168 rRNA genes th a t we sequenced from
these soils are only distantly related to cultured organ­
isms, and m ay originate from organism s th at are capable
o f nitrogen fixation.
Finally, the 168 rRNA gene library data from this
study m ay explain results from oth er studies reporting a
m icrobial shift fro m r to K selection (decrease in
culturability) [47] along the successional gradient. Cya­
nobacteria w ould n o t have been cultured using the
culture conditions an d m edia used in the 8igler et al.
[47] study. In addition, o u r later successional soils
(Fig. 1) show an increase in other, difficult-to-culture
groups including Acidobacteria, candidate division WYO,
G em m a tim o n a d etes, a n d V errucom icrobia, p o ssib ly
explaining this observed switch from r- to K- selected
m icroorganism s w ith increasing soil age. Bacteria from
these groups are com m o n soil bacteria, b u t are difficult
to isolate using traditional culturing techniques. This
m ay suggest th a t these organism s utilize com plex carbon
substrates th a t are m ore likely to be present in older,
m ore developed soils, an d th a t are n o t traditionally used
in culturing efforts. Alternatively, this m ay su p p o rt th at
the grow th o f these bacteria requires interactions w ith
o ther types o f bacteria, w hich w ould in tu rn m ake them
very difficult to isolate in p u re cultures.
M acroscale controls over m icrobial processes have
long been a focus o f ecology; however, the relationship
betw een m icrobial p opulations and processes is poorly
understood. This w ork dem onstrates th a t there m ay be
environm ental-level influences o n bacterial co m m unity
succession, and th a t these p atterns o f com m unity replace­
m en t m ay parallel plan t com m unity developm ent. For
example, o u r study suggests that, like m acrobial h etero ­
trophs, m icrobial hetero tro p h s are likely to be am ong the
first colonizers o f new soils. In addition, biogeochem ical
and sequence analyses suggest th a t nitrogen fixation is
im p o rta n t in these very young, b arren soils, m uch as it is
in the early stages o f p lan t com m u n ity developm ent.
Finally, the initial increases in m icrobial co m m unity
diversity observed in this study are sim ilar to patterns
seen in p lan t com m unities. Thus, despite the small-scale
heterogeneity o f soil environm ental conditions, m icrobial
populations are likely controlled, to som e extent, by
factors w hich also regulate m acrobial com m unities.
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
Acknowledgments
The authors th a n k Allen Meyer, Preston 8owell, and Julia
Rosen for field assistance, D an Liptzin for help w ith the
biogeochem ical analysis, Alex Blum for help w ith the
XRD analysis, and Alan T ow nsend for valuable discus­
sions. W e also acknowledge the helpful suggestions of
several thoughtful reviewers. Funding was provided from
the N8F M icrobial Observatories Program , grant MCB0084223, and the D epartm ent o f Ecology and E volution­
ary Biology at the U niversity o f C olorado, Boulder,
th ro u g h the M arion and G ordon A lexander M em orial
8cholarship for research in m o n tan e biology.
References
1. A ltschul, SF, G ish, W , M iller, W , M yers, EW , L ipm an, DJ (1990)
Basic local alig n m en t search tool. J M ol Biol 215: 403M rl0
2. A m a n n , RI, L udw ig, W , S chleifer, K H (1995) P h y lo g en etic
id en tifica tio n a n d in-situ d etectio n o f in d iv id u al m icrob ial cells
w ith o u t cultivation. M icro b io l Rev 59: 143-169
3. A n d ren , O , B russaard, L, C larholm , M (1999) Soil o rg an ism
influence o n ecosystem -level processes bypassing th e ecological
hierarchy? A ppl Soil Ecol 11: 177-188
4. B attin, TJ, W ille, A, Saltier, B, P senner, R (2001) P hylogenetic an d
fu n ctio n al hetero g en eity o f se d im e n t biofilm s along en v iro n m en tal
g rad ien ts in a glacial stream . A ppl E n v iro n M icro b io l 67: 7 9 9 -807
5. B o rm an n , BT, Sidle, R C (1990) C hanges in p ro d u c tiv ity an d
d istrib u tio n o f n u trie n ts in a ch ro n o seq u e n ce a t G lacier Bay
N atio n al P ark, Alaska. J Ecol 78: 561-578
6. G arrillo-G astaneda, G, M u n o z, JJ, P eralta-V idea, JR, G om ez, E,
T ie m a n n b , KJ, D uarte-G ard e a, M , G ardea-T orresdey, JL (2002)
Alfalfa gro w th p ro m o tio n b y b acteria grow n u n d e r iro n lim itin g
co n d itio n s. A dv E nviron Res 6: 391-399
7. C hadw ick, OA, D erry, LA, V itousek, P M , E luebert, BJ, Eledin, LO
(1999) C hanging sources o f n u trie n ts d u rin g fo u r m illio n years o f
ecosystem developm ent. N a tu re 397: 491Mr97
8. C h ap in , FS, W alker, LR, Fastie, GL, S h arm an , LG (1994) M ech a­
n ism s o f p rim a ry succession follow ing d eglaciation at G lacier Bay,
Alaska. Ecol M o n o g r 64: 149-175
9. C lem ents, EE (1928) P la n t Succession a n d Indicators: A D efinitive
E d itio n o f P la n t Succession an d P lan t In d icato rs. FIW W ilson,
N ew Y ork
10. Cole, JR, Ghai, B, M arsh, TL, Farris, RJ, W ang, Q , K ulam , SA,
C h a n d ra, S, M cG arrell, D M , S chm idt, T M , G arrity, C M , Tiedje,
JM (2003) T he R ibosom al D atabase P ro ject (R D P -II): preview ing
a new a u to a lig n e r th a t allow s reg u lar u p d a te s a n d th e new
p ro k ary o tic taxonom y. N ucleic A cids Res 31: 442-443
11. C ooper, W S (1923) T he recen t ecological h isto ry o f G lacier Bay,
Alaska: II. T he p re se n t vegetation cycle. E cology 4: 2 23-24 6
12. C rocker, RL, M ajor, J (1955) Soil d ev elo p m en t in relatio n to
v egetation a n d surface age a t G lacier Bay, Alaska. J Ecol 43: 427M,48
13. Fastie, GL (1995) C auses a n d ecosystem consequences o f m u ltip le
pathw ays o f p rim a ry succession a t G lacier Bay, Alaska. E cology 76:
1899-1916
14. Finlay, BJ, M aberly, SG, C o o p er, JI (1997) M icrobial diversity an d
ecosystem fu n ctio n . O ikos 80: 209-213
15. G arreaud, R D , M u n o z , R (2004) T he d iu rn a l cycle in circu latio n
an d cloudiness o v er th e su b tro p ical so u th e ast Pacific: a m o d elin g
study. J C lim ate 17: 1699-1710
16. G leason, H A (1927) F u rth e r views o n the succession co ncept.
E cology 8: 2 99-326
17. G o rd o n , DA, P riscu, J, G io v an n o n i, S (2000) O rigin a n d phylogeny
D .R . N
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
e m e r g u t e t a l .:
S o il M
ic r o b ia l
C
o m m u n it y
S u c c e s s io n
o f m icro b es living in p e rm a n e n t A n tarctic lake ice. M icro b Ecol 39:
197-202
H o d k in so n , ID , C oulson, SJ, H a rriso n , J, W ebb, N R (2001) W h a t
a w o n d erfu l w eb th ey weave: spiders, n u trie n t c a p tu re a n d early
ecosystem d ev elo p m en t in th e high A rctic— som e c o u n te r-in tu i­
tive ideas o n c o m m u n ity assem bly. O ikos 95: 349-352
H u b er, T, F au lk n er, G, H u g en h o ltz, P (2004) B ellerophon: a
p ro g ra m to d etect chim eric sequences in m u ltip le sequence align­
m en ts. B io in fo rm atics 20:2317-2319
H uelsenbeck, JP, R o n q u ist, F (2001) MRBAYES: bayesian in fer­
ence o f phylo g en etic trees. B io in fo rm atics 17: 754-755
In ag ak i, F, T ak ai, K, K o m atsu , T, S ak ih am a, Y, In o u e , A,
H o rik o sh i, K (2002) P rofile o f m icrobial c o m m u n ity stru ctu re
an d p resen ce o f en d o lith ic m icro o rg an ism s inside a deep-sea rock.
G eo m icro b io l J 19: 535-552
Irgens, RL, G osink, JJ, Staley, JT (1996) Polaromonas vacuolata gen
no v , sp. nov, a psychrophilic, m arin e, gas v acuolate b acteriu m
fro m A ntarctica. In t J Syst B acteriol 46: 822-826
Jackson, GR, H a rp e r, JP, W illoughby, D, R oden, EE, G hurchill, PE
(1997) A sim ple, efficient m e th o d fo r th e se p ara tio n o f h u m ic
sub stan ces a n d D N A fro m en v iro n m en tal sam ples. A ppl E nviron
M icro b io l 63: 4993-4995
Ju m p p o n e n , A (2003) Soil fungal c o m m u n ity assem bly in a
p rim a ry successional glacier fo refro n t ecosystem as in ferred fro m
rD N A sequence analyses. N ew P h y to l 158: 569-578
K em p, PE, Aller, JY (2004) B acterial diversity in aq u atic a n d o th e r
en v iro n m en ts: w h a t 16S rD N A libraries can tell us. FEM S M ic ro ­
b io l Ecol 47: 161-177
L aM o n tag n e, M G , Schim el, JP, H o ld en , PA (2003) G om parison o f
su b su rface a n d surface soil bacterial c o m m u n itie s in G alifom ia
grasslan d as assessed b y te rm in a l re stric tio n frag m en t length
p o ly m o rp h ism s o f P G R -am plified 16S rR N A genes. M icro b Ecol
46: 2 1 6 -2 2 7
Lane, DJ (1991) 16S/23S rR N A Sequencing. In: S tackebrandt, E,
G oodfello w , M (Eds.) N ucleic A cid T ech n iq u es in B acterial
System atics. W iley, G hichester, p p 115-175
M a rtin , AP (2002) P hylogenetic ap p ro ach es fo r d escribing a n d
co m p arin g th e diversity o f m icrobial co m m u n ities. A ppl E nviron
M icro b io l 68: 3673-3682
M artin y , AG, Jorgensen, T M , A lbrechtsen, H J, A rvin, E, M o lin , S
(2003) L o n g -term succession o f stru c tu re a n d diversity o f a biofilm
fo rm e d in a m o d el d rin ld n g w ater d istrib u tio n system . A ppl
E n v iro n M icro b io l 69: 6899-6907
M atth ew s, JA (1992) T he E cology o f R ecently-D eglaciated T errain:
A G eoecological A p p ro ach to G lacier F orelands an d P rim a ry
Succession. G am bridge U niversity Press, G am bridge
M o re, M I, H errick, JB, Silva, M G , G hiorse, W G , M adsen, EL
(1994) Q u an titativ e cell-lysis o f in d ig en o u s m icro o rg an ism s a n d
ra p id e x tractio n o f m icro b ial D N A fro m sed im en t. A ppl E nviron
M icro b io l 60: 1572-1580
N em erg u t, D R (2004) E vo lu tio n a n d ecology o f high a ltitu d e soil
m icro b ia l co m m u n ities. P h.D . d issertatio n . U niversity o f G olorad o , B oulder, GO
N icol, G W , T scherko, D, Em bley, T M , P rosser, JI (2005) P rim a ry
succession o f soil G renarchaeota across a receding glacier foreland.
E n v iro n M icro b io l 7: 337-347
O h to n e n , R, Fritze, H , P en n an en , T, Ju m p p o n e n , A, T rappe, J
(1999) E cosystem p ro p erties an d m icrobial c o m m u n ity changes
in p rim a ry succession o n a glacier forefront. O ecologia 119:
23 9 -2 4 6
Pace, N R (1997) A m o lecu lar view o f m icrobial diversity a n d the
b io sp h ere. Science 276: 734-740
P ark , IH , Ka, JO (2003) Iso latio n a n d characterizatio n o f 4-(2,4d ich lo ro p h en o x y )b u ty ric acid -d eg rad in g b acteria fro m agricultural
soils. J M icro b io l B iotechnol 13: 2 43-250
37. P riscu, JG, A dam s, EE, Lyons, W B, V oytek, M A, M ogk, D W ,
Brow n, RL, M cKay, GP, Takacs, GD, W elch, KA, W olf, GF,
K irshtein, JD, Avci, R (1999) G eom icrobiology o f subglacial ice
above Lake V ostok, A ntarctica. Science 286: 2141-2144
38. Reese, GA, Liu, KB (2003) P ollen dispersal a n d d ep o sitio n o n the
Q uelccaya Ice G ap, P eru. Phys G eogr 23: 4 4 -5 8
39. R oberts, G P, L udden, P W (1992) N itro g en fixation b y p h o to sy n ­
thetic bacteria. In: Stacey, G, B urris, R H , Evans, H J (Eds.) Biological
N itro g en F ixation. G h ap m an & H all, N ew Y ork, p p 135-165
40. S am brook, J, Russell, D W (2001) M o lecu lar G loning: A L ab o rato ry
M anual. G old S pring H a rb o r L ab o rato ry Press, G old S p rin g H a rb o r
41. S chipper, LA, D egens, BP, Sparling, GP, D u n can , LG (2001)
Ghanges in m icrobial h e te ro tro p h ic diversity along five p la n t
successional sequences. Soil Biol B iochem 33: 2093-2103
42. Schloss, P D , H an d elsm an , J (2005) In tro d u c in g D O T U R , a
c o m p u te r p ro g ra m for defining o p eratio n al tax o n o m ic u n its an d
estim atin g species richness. A ppl E nviron M icro b io l 71: 1 5011506
43. Schloss, PD , H ay, AG, W ilson, DB, G ossett, JM , W alker, LP (2005)
Q uantifying bacterial p o p u la tio n d ynam ics in c o m p o st u sin g 16S
rR N A gene probes. A ppl M icro b io l B iotechnol 66: 45 7 -463
44. S chneider, S, Roessli, D , Excoffier, L. 2000. A rlequin: A softw are
for p o p u la tio n genetics d a ta analysis. V er 2.000. G enetics an d
B iom etry Lab, D ept, o f A n th ro p o lo g y , U niversity o f G eneva
45. S h an n o n , DE, W eaver, W (1949) T he M a th em atic al T h eo ry o f
G o m m u n icatio n . U niversity o f Illinois Press, U rb an a
46. S heridan, P P , M iteva, VI, Brenchley, JE (2003) P hylogenetic analysis
o f an aero b ic psy ch ro p h ilic e n ric h m e n t cu ltu res o b ta in e d fro m a
G reenland glacier ice core. A ppl E nviron M icro b io l 69: 2 1 5 3 -2160
47. Sigler, W V , Grivii, S, Zeyer, J (2002) B acterial succession in glacial
forefield soils characterized b y c o m m u n ity stru c tu re , activ ity an d
o p p o rtu n is tic gro w th dynam ics. M icro b Ecol 44: 306-316
48. Sigler, W V , Zeyer, J (2002) M icrobial diversity an d activ ity along
the forefields o f tw o receding glaciers. M icro b Ecol 43: 3 9 7 ^ 0 7
49. S lddm ore, M , A n d erso n , SP, S harp, M , F oght, J, L anoil, BD (2005)
G o m p ariso n o f m icro b ia l c o m m u n ity co m p o sitio n s o f tw o su b g la­
cial en v iro n m en ts reveals a possible role for m icrobes in chem ical
w eath erin g processes. A ppl E nviron M icro b io l 71: 698 6 -6997
50. Staley, JT, G o sin k , JJ (1999) P oles a p a rt: b io d iv e rsity a n d
b io g eo g rap h y o f sea ice bacteria. A n n u Rev M icro b io l 53: 189-215
51. Sw offord, DL (2001) P hylogenetic Analysis U sing P a rsim o n y (* an d
O th e r M e th o d s), 4 th edn. S in au er A ssociates, S u n d erlan d , M A
52. T h o m p so n , LG, M o sley-T hom pson, E, A m ao , BM (1984) E l-N in o
so u th e rn oscillation events reco rd ed in th e stratig rap h y o f the
tro p ical Q uelccaya ice cap, P eru. Science 226: 50-53
53. Tiessen, H , M oir, JO (1993) G haracterization o f available P b y
sequential extraction. In: G arter, M R (Ed.) Soil S am pling an d
M e th o d s o f Analysis. Lewis P ublishers, Boca R ato n , p p 7 5 -8 6
54. T scherko, D , R ustem eier, J, R ichter, A, W anek, W , K andeler, E
(2003) F u n ctio n al d iversity o f th e soil m icro flo ra in p rim a ry
succession across tw o glacier forelands in th e G entral Alps. E u ro J
o f Soil Sci 54: 6 85-696
55. V itousek, P M (2004) N u trie n t Gycling a n d L im itation : H a w a i’i as
a M odel System . P rin c eto n U niversity Press, P rin c eto n
56. W alker, LR, G larkson, BD, Silvester, W B, G larkson, BR (2003)
G olonization dynam ics a n d facilitative im p acts o f a n itro g en -fix in g
sh ru b in p rim a ry succession. J Veg Sci 14: 2 77-290
57. W ardle, DA, W alker, LR, B ardgett, RD (2004) E cosystem p ro p e r­
ties a n d forest decline in co n trastin g lo n g -te rm chron o seq u en ces.
Science 305: 509-513
58. W en, AM , Fegan, M , H ayw ard, G, G halcraborty, S, Sly, LI (1999)
P hylogenetic relatio n sh ip s am o n g m em b ers o f th e G o m am o n ad aceae, a n d d escrip tio n o f D elftia acidovorans (den D o o ren d e Jong
1926 a n d T am ao k a et al. 1987) gen. nov., com b. nov. In t J Syst
B acteriol 49: 567-576