Ecology, 93(7), 2012, pp. 1550–1559
Ó 2012 by the Ecological Society of America
Biotic contexts alter metal sequestration and AMF effects
on plant growth in soils polluted with heavy metals
SYDNEY I. GLASSMAN1
AND
BRENDA B. CASPER
Department of Biology, University of Pennsylvania, Philadelphia, Pennsylvania 19104 USA
Abstract. Investigating how arbuscular mycorrhizal fungi (AMF)–plant interactions vary
with edaphic conditions provides an opportunity to test the context-dependency of
interspecific interactions. The relationship between AMF and their host plants in the context
of other soil microbes was studied along a gradient of heavy metal contamination originating
at the site of zinc smelters that operated for a century. The site is currently under restoration.
Native C3 grasses have reestablished, and C4 grasses native to the region but not the site were
introduced. Interactions involving the native mycorrhizal fungi, non-mycorrhizal soil
microbes, soil, one C3 grass (Deschampsia flexuosa), and one C4 grass (Sorghastrum nutans)
were investigated using soils from the two extremes of the contamination gradient in a full
factorial greenhouse experiment. After 12 weeks, plant biomass and root colonization by
AMF and non-mycorrhizal microbes were measured. Plants from both species grew much
larger in soil from low-contaminated (LC) origin than high-contaminated (HC) origin. For S.
nutans, the addition of a non-AMF soil microbial wash of either origin increased the efficacy
of AMF from LC soils but decreased the efficacy of AMF from HC soils in promoting plant
growth. Furthermore, there was high mortality of S. nutans in HC soil, where plants with
AMF from HC died sooner. For D. flexuosa, plant biomass did not vary with AMF source or
the microbial wash treatment or their interaction. While AMF origin did not affect root
colonization of D. flexuosa by AMF, the presence and origin of AMF did affect the number of
non-mycorrhizal (NMF) morphotypes and NMF root colonization. Adding non-AMF soil
biota reduced Zn concentrations in shoots of D. flexuosa. Thus the non-AMF biotic context
affected heavy metal sequestration and associated NMF in D. flexuosa, and it interacted with
AMF to affect plant biomass in S. nutans. Our results should be useful for improving our basic
ecological understanding of the context-dependency of plant–soil interactions and are
potentially important in restoration of heavy-metal-contaminated sites.
Key words: AMF; arbuscular mycorrhizal fungi; C3 grasses; C4 grasses; context-dependency;
Deschampsia flexuosa; heavy metal pollution; microbial wash; non-mycorrhizal soil microbes; Palmerton;
Sorghastrum nutans; symbioses.
INTRODUCTION
The ability to generalize findings of ecological studies
is often hampered by the failure to understand how
biotic and abiotic contexts shape the species interactions under investigation (Agrawal et al. 2007). One
area that needs explicit investigation is how the context
in which mutualisms operate affects their magnitude
and direction (Agrawal et al. 2007). As the most
widespread and abundant mutualism in the terrestrial
biosphere (Fitter and Moyersoen 1996), the symbiosis
between plants and arbuscular mycorrhizal fungi
(AMF) provides an excellent model for studying how
contextual frameworks affect symbioses, because both
biotic and abiotic contexts are known to influence how
Manuscript received 5 November 2010; revised 12 December
2011; accepted 16 December 2011; final version received 3
February 2012. Corresponding Editor: G. S. Gilbert.
1 Present address: Department of Environmental Science
Policy Management, University of California, Berkeley,
California 94720 USA. E-mail: [email protected]
AMF affect host plant performance (Hoeksema et al.
2010).
In exchange for plant-derived carbon, AMF provide
their hosts with increased surface area and absorption of
nutrients (Smith and Read 2008). AMF also help plants
deal with a variety of environmental stresses such as
drought (Auge 2001) and soil fungal pathogens (AzconAguilar and Barea 1996), and they improve soil health
(Jeffries et al. 2003) and soil structure (Wright and
Upadhyaya 1998). AMF can also help plants cope with
heavy metal stress, in some cases selectively sequestering
the metals away from plant tissues (Hildebrandt et al.
2007). As heavy metal pollutants cause severe environmental damage worldwide (Dudka and Adriano 1997),
understanding the context-dependence of AMF in heavy
metal stress may be critically important to restoration
efforts.
We studied AMF and host plant interactions across a
heavy metal contamination gradient. We expected that
both the soil environment and co-occurring soil microbes might strongly affect the strength of the
1550
July 2012
CONTEXT-DEPENDENCY OF SOIL MICROBES
mycorrhizal mutualism in terms of root colonization
and effect on host plant performance. Heavy metals can
be toxic to plants as well as fungi (Kelly et al. 2003,
Andrade et al. 2009), although metal-tolerant ecotypes
are known in both organisms (Bradshaw 1952, Del Val
et al. 1999). While the diversity of AMF species may be
reduced in heavy metal soils (Pawlowska et al. 1996,
Doherty 2009), species present may be particularly metal
tolerant and effective in promoting plant growth in
those soils (Galli et al. 1994, Gamalero et al. 2009).
Evidence from naturally metalliferous serpentine soils
corroborates the existence of edaphic AMF ecotypes
(Schechter and Bruns 2008, Ji et al. 2010).
Biotic contexts might play an important role in
determining AMF function along any environmental
gradient. There is increasing evidence that non-mycorrhizal soil microbes have significant effects on the
formation and outcome of the mycorrhizal symbiosis
(Frey-Klett et al. 2007). Although studies have examined how bacteria or AMF might improve phytoremediation efficiency or plant growth in heavy metal soils,
few studies exist in which a combination of microorganisms or even multiple species of AMF or bacteria are
used (Gamalero et al. 2009, Ji et al. 2010, Wehner et al.
2010). Because a diverse microbial community exists on
or near plant roots, studies using single species of AMF
or pathogens are not realistic (Wehner et al. 2010) nor
do they take into account that many communities
include guilds of interacting species (Stanton 2003).
In order to gain a better understanding of how
changing abiotic (soil) and biotic (non-mycorrhizal soil
microbes) contexts affect the function of native AMF
communities along a gradient of heavy metal pollution,
we performed a fully factorial greenhouse experiment
with two plant species in which we manipulated soil
type, non-mycorrhizal soil microbes, and AMF communities. Specifically, we asked two questions. (1) Does
the effect of AMF on plant growth depend on the
context of either the soil background or the soil
microbial background (or interactions between the
two)? (2) How does the native C3 grass vs. the
introduced C4 grass respond to the same soil, AMF,
and other microbial treatments?
MATERIALS
AND
METHODS
Site description.—One of the most extensive heavymetal-contaminated sites in the United States is located
on Blue Mountain, near the town of Palmerton in Carbon
County, Pennsylvania (Oyler 1993; see Plate 1). Over 2000
acres on the north-facing slope were contaminated with
zinc, cadmium, copper, and lead from emissions of two
zinc smelters operated by the New Jersey Zinc Company
from 1898 until 1980 (Roberts et al. 2002). In the 1970s,
concentrations of metals within 1 km of the smelter were
reported to reach levels of 135 000 ppm Zn, 1750 ppm Cd,
2000 ppm Cu, and 2000 ppm Pb (Buchauer 1973). The
pollution severely impeded growth and reestablishment
of trees (Jordan 1975), reduced density and diversity of
1551
lichens (Nash 1975), decreased soil microbial diversity
(Jordan and Lechevalier 1975), and negatively impacted
soil fauna (Strojan 1978) and terrestrial vertebrates
(Storm et al. 1994). Revegetation efforts, begun in 2002
(Kunkle 2006), included aerial deposition of seed,
fertilizer, and lime. Native C3 grasses, especially Deschampsia flexuosa and Danthonia spicata (see Plate 1),
reestablished naturally; scattered individuals of C4
grasses native to Pennsylvania but not to the site (Latham
et al. 2007) were introduced: Andropogon gerardii,
Schizachyrium scoparium, and Sorghastrum nutans (see
Latham et al. [2007], section 3.2, for a complete
description). Our site has particular conservation value
because it is one of the largest expanses of native
grasslands left in the state (Latham et al. 2007) and is
traversed by the heavily used Appalachian Trail.
A distinct gradient of heavy metal contamination with
metals decreasing in concentration with distance from
the smelter has been documented several times (Buchauer 1973, Johnson and Richter 2010); here we present
data from a 2009 study, comparing bioavailable heavy
metals (Appendix A) in freshly collected field soils
beneath D. flexuosa and Da. spicata at the high and low
end of the gradient and for soils collected in similar
locations but analyzed after autoclaving (Doherty 2009).
The soils we used were also taken from each end of the
contamination gradient but not from the exact locations
of those previously (J. Doherty, personal communication). Our low-contaminated (LC) soils were collected
within a mixed grassland of Da. Spicata and D. flexuosa
and then homogenized. Our high-contaminated (HC)
soils were from a sparsely vegetated area closer to the
site of the abandoned smelter than HC soils collected in
2006.
Study species.—We selected D. flexuosa and S. nutans
for study. S. nutans, a C4 grass, is an obligate mycotroph
(Hartnett and Wilson 1999, Castelli and Casper 2003),
while D. flexuosa, a C3 grass, is less dependent on
mycorrhizae (J. Doherty, personal communication).
Based on spore morphospecies, arbuscular mycorrhizal
fungi (AMF) communities at the site are more diverse
under C4 grasses than under C3 grasses (Glassman and
Thompson, unpublished data).
Arbuscular mycorrhizal fungal communities.—A full
description of the AMF morphospecies found at
Palmerton are listed in Appendix B, but the most
common are Acaulospora mellea and Glomus rubiforme.
While we did not quantify spore abundance in this
study, observations corroborate findings from a previous study that AMF spores were more abundant and
diverse in LC than in HC soils at the site (Doherty 2009).
Trap cultures.—Soil collections used to establish
AMF trap cultures, which would serve as sources of
inoculum for the greenhouse experiment, were established in September 2009. Approximately 1 L of soil was
taken beneath each of 15 D. flexuosa individuals in LC
soil and 15 in HC soil and used within days to set up
trap cultures with three different host plants (Danthonia
1552
SYDNEY I. GLASSMAN AND BRENDA B. CASPER
spicata and D. flexuosa from the site and Sorghum
Special Effort (Sorghum bicolor 3 S. sudanese Hybrid 3;
American Seed, Spring Grove, Pennsylvania, USA) in
order to increase spore diversity and numbers. Each
culture contained 600 mL of field soil from one
collection point mixed with 900 mL autoclaved sand.
Sorghum seeds were rinsed for 20 min to remove
powdery fungicide and then sterilized in 50% EtOH
for 30 min, rinsed in water for 20 min, and dried for a
day. Seeds of Danthonia spicata and D. flexuosa used for
trap cultures were not sterilized. Trap cultures were
watered twice daily for four months and harvested in
late January 2010. Soils were placed in plastic bags and
put to rest in the cold room for almost a month to
enhance infectivity of spores (R. Koide, personal
communication).
Greenhouse experiment.—Seeds of D. flexuosa were
collected from the site in September 2009. Seeds of S.
nutans were collected at Anderson Prairie, Emmet
County, Iowa, USA, in 2005. (C4 grass seeds used in
revegetating Blue Mountain were also obtained from a
commercial prairie seed source.) Soils were collected
from several points within the high and low ends of the
metal contamination gradient in November 2009, steam
sterilized in an autoclave for 1 h at 1008C at 1 atm
(101.317 kPa) then again after 24 h of rest. Repeated
autoclaving at low temperature minimizes solubilization
of heavy metals and changes in P availability.
A full factorial experiment was set up with each
species in February 2010. For D. flexuosa, the experiment included two soil contamination levels (HC, LC),
three AMF treatments (originating in HC soil, LC soil,
or no AMF), and three non-mycorrhizal soil microbe
(MIC) treatments (HC, LC, or none) 310 replicates,
yielding 180 pots. For S. nutans, treatment combinations
lacking AMF were omitted since its growth is minimal
without AMF, yielding 120 pots.
AMF spores to be used as inoculum were extracted
from 11 trap cultures of each soil type (HC and LC)
using the modified wet-sieve method (McKenney and
Lindsey 1987), pooled by soil type and surface sterilized
in an antibiotic solution containing 500 mg/L streptomycin and 500 mg/L ampicillin for at least 24 h
(modified from Abbott et al. 1992). Immediately before
planting, spores were rinsed of antibiotic, concentrated
on a 38-lm sieve, and then placed in a beaker containing
300 mL of deionized water. Spores were shaken to
suspend them in water while 2.5-mL aliquots were
removed and transferred directly onto roots of monthold seedlings that had been grown from seed sterilized
by soaking in 70% EtOH for 3 min, rinsed with water for
10–15 min, placed immediately onto trays of autoclaved
yellow bar sand, and watered several times a day
without fertilizer. Based on three samples for each soil
type, approximate numbers of spores (mean 6 SD) per
2-mL aliquot were as follows for the two most abundant
AMF species: 254 6 58 (A. mellea) and 91 6 32 (G.
rubiforme) derived from LC soils, and 148 6 35 (A.
Ecology, Vol. 93, No. 7
mellea) and 80 6 26 (G. rubiforme) derived from HC
soils. No attempt was made to equalize the number of
spores present in the different inocula as we wanted
them to represent natural differences in spore communities in HC and LC soils.
Seedlings were then immediately planted into 500-mL
pots containing three parts autoclaved soil of either HC
or LC origin and one part autoclaved sand. After
planting, pots receiving a MIC treatment were given 10
mL of a microbial wash of either HC or LC origin,
prepared by creating a slurry of 1500 mL fresh bulk soil
mixed with 2500 mL water. This soil slurry was then
passed through a series of sieves, the smallest being 20
lm (Klironomos 2002). The microbial wash was
expected to consist of bacteria, fungal hyphae, spores
of saprobes and pathogens smaller than 20 lm, and soil
chemicals.
Treatment combinations were spatially randomized
upon planting and in alternate weeks until harvest.
Because soil did not compact equally for all pots, a
depth-to-soil measurement was taken from the top of
the pot to the top of the soil to account for any size
variation due to differences in soil amount and
compaction. Pots were watered daily but not fertilized.
Seedlings dying within the first week were replaced, but
not afterwards. Mortality was recorded throughout the
experiment.
After 12 weeks of growth, shoots were cut off at the
soil, dried, and weighed. Roots were thoroughly washed
before drying. Dry roots were weighed, then rehydrated,
cleared in boiling 10% KOH, primed in 5% HCl, and
stained in 0.1% trypan blue for visualizing colonization
by AMF and other microbe morphotypes (Philips and
Hayman 1970). Stained roots were stored in deionized
water before mounting onto microscope slides using
polyvinyl alcohol lactic acid glycerol (PVLG). A subset
of at least six individuals per treatment combination was
examined with a compound microscope, and colonization was scored using the modified line-intersect method
with 100 intersections per sample (Philips and Hayman
1970, McGonigle et al. 1990). Because we were working
with dried materials, arbuscules were difficult to see;
AMF colonization was measured solely by scoring the
presence of fungal hyphae or vesicles, without distinguishing between the two. A conservative estimate of the
number of non-mycorrhizal fungal (NMF) and microbe
morphotypes per slide was made based on morphological characters, including color, presence or absence of
septae, hyphae, shape, size, or any other distinguishing
features that led us to believe that they were biological
but not AMF.
Elemental analysis of plant tissue.—A subset of five
shoot samples per treatment of D. flexuosa (5 samples 3
18 treatments ¼ 90 total) were analyzed for P and Zn
concentration; 25 mg of tissue was ground to a fine
powder and acid digested according to a method
modified from Zarcinas et al. (1987). Samples were
analyzed using inductively coupled plasma spectrometry
July 2012
CONTEXT-DEPENDENCY OF SOIL MICROBES
on a Perkin Elmer Optima 5300 DV (Waltham,
Massachusetts, USA), where three wavelengths were
measured for each element (P 178.221, P 213.617, P
214.914; Zn 202.548, Zn 206.200, Zn 213.857) and
averaged for each sample. Elemental analysis was
validated against an internationally recognized reference, National Institute of Standards and Technology
1547 peach leaves.
Statistical analysis.—JMP 8 (SAS Institute 2010) was
used for all analyses. For D. flexuosa, three-way
analyses of variance were used to examine shoot and
root biomass; shoot concentration of P and Zn;
percentage root colonization by AMF and NMF; and
number of NMF morphotypes identified per root as a
function of soil type, AMF origin, and MIC origin,
with depth to soil as a covariate. Because of high
mortality for S. nutans in the HC soil, the soil 3 AMF
3 MIC term was removed from the three-factor
ANOVA used to examine shoot and root biomass,
and depth to soil surface did not improve the model.
For S. nutans, a two-way ANOVA was used to
examine percentage root colonization by AMF and
NMF, and number of NMF morphotypes per root as a
function of AMF and MIC origin; root colonization
was measured only on S. nutans growing in LC soils
because there was negligible survival in HC soils. In
addition, survivorship analysis was conducted for S.
nutans to determine if the AMF treatment affected
plant longevity. All biomass data were multiplied by
1000, then log-transformed (log10 10003) to improve
normality and heteroscedasticity. The test slice command in JMP, a post hoc test enabled for interaction
effects that allows for multiple simultaneous contrasts,
and the Tukey HSD test were used to elucidate
differences among treatment combinations.
RESULTS
Performance of Deschampsia flexuosa.—For Deschampsia flexuosa, both shoots and roots were significantly larger in low-contaminated (LC) soil compared
to high-contaminated (HC) soil (Table 1; Fig. 1), but
there was no effect of arbuscular mycorrhizal fungi
(AMF) and microbes (MIC) or their interaction on
either biomass measure. Depth from the top of the pot
to the soil surface was nearly significant as a covariate (P
, 0.06) for shoot biomass but not for root biomass.
Survival of D. flexuosa was nearly 100%. Raw biomass
data for D. flexuosa across treatments can be found in
Appendix C.
Performance of Sorghastrum nutans.—Both soil type
and AMF origin affected survivorship of Sorghastrum
nutans. Survival was 100% in LC soil but only 20% in
HC soil, where plants inoculated with AMF of HC
origin died sooner (median 28.5 d) than plants
inoculated with AMF of LC origin (median 38.0 d;
survival analysis log-rank chi-square, 6.66; P , 0.01).
Shoot biomass and root biomass of S. nutans
responded in the same way to experimental treatments.
1553
Biomass was greater in LC soil than in HC soil overall
(shoots P , 0.001; roots P , 0.001; Fig. 2A) but did not
respond to AMF origin or MIC origin as main effects
(Table 1). The effect of AMF origin on biomass
depended on whether other soil microbes were present,
resulting in a significant AMF origin 3 MIC origin
interaction (shoots P , 0.05; roots P , 0.01; Figs. 2B
and 2C; Table 1). With AMF from LC soil, adding any
MIC increased biomass compared to no MIC added,
but with AMF from HC soil, adding any MIC decreased
biomass compared to no MIC added. Raw biomass data
for S. nutans across treatments can be found in
Appendix D.
Root colonization for Deschampsia flexuosa.—For D.
flexuosa, percentage root colonized by AMF was greater
in LC soils than in HC soils (F1, 107 ¼ 47.48, P , 0.001;
Fig. 3A; Appendix E). AMF origin was significant as a
main effect (F1, 107 ¼ 29.65, P , 0.001; Appendix E);
plants receiving no AMF inoculum had lower (but still
present) rates of AMF colonization than when any
AMF inoculum was applied, but there was no difference
in colonization between AMF inoculum from LC and
HC soils (Fig. 3B).
Percentage root colonized by non-mycorrhizal fungi
(NMF) was greater in LC soil compared to HC soil
(F1, 107 ¼ 13.77, P , 0.001; Fig. 3A) and differed among
AMF treatments (F1, 107 ¼ 5.27, P , 0.01; Fig. 3B) but
surprisingly, not among MIC treatments (Appendix E).
Colonization by NMF was greater when AMF inoculum originated from LC soil compared to AMF from
HC soil or no AMF inoculum added (Fig. 3B). Whether
MIC origin affected the number of NMF morphotypes
found in D. flexuosa roots also depended on the presence
and origin of AMF (MIC origin 3 AMF origin; F1, 107 ¼
3.56, P , 0.01; Fig. 3C). The largest number of NMF
morphotypes were found in D. flexuosa roots when
plants were amended with microbial wash from LC
origin but not inoculated with AMF, and the fewest
NMF morphotypes were found when plants were
inoculated with AMF of LC origin and no microbial
wash was added (Fig. 3C). The number of NMF
morphotypes on D. flexuosa was also greater in LC soil
(6.0 6 0.8; mean 6 SE) compared to HC soil (4.1 6 0.6;
F1, 107 ¼ 31.02, P , 0.001; Appendix E). Raw root
colonization data for D. flexuosa can be found in
Appendix F.
Root colonization for Sorghastrum nutans.—For S.
nutans, neither AMF treatment nor MIC treatment nor
their interaction significantly affected percentage root
colonized by AMF (Appendix G), which averaged 6.0 6
1.6 (mean 6 SE) across both AMF treatments.
MIC treatment but not AMF treatment affected both
percentage of root colonized by NMF and the number
of NMF morphotypes (Appendix G). Percentage of root
colonized by NMF was lower (P , 0.01) when no MIC
was added (10.4 6 3.2; mean 6 SE) compared to MIC
originating from either LC (20.2 6 3.8) or HC (22.8 6
3.7) soil, which were not different from each other.
1554
SYDNEY I. GLASSMAN AND BRENDA B. CASPER
Ecology, Vol. 93, No. 7
TABLE 1. ANOVA results for a study with experimental factors from Blue Mountain, a heavymetal-contaminated site in Carbon County, Pennsylvania, USA.
Source
df
Sum of squares
F
P
D. flexuosa
Shoot biomass
Soil type
AMF origin
MIC origin
Soil 3 AMF
Soil 3 MIC
AMF 3 MIC
Soil 3 AMF 3 MIC
Depth to soil (cm)
1
2
2
2
2
4
4
1
18.486942
0.078957
0.090387
0.676123
0.154685
0.671228
0.320076
0.428844
159.5408
0.3407
0.39
2.9174
0.6675
1.4482
0.6906
3.7009
,0.0001
0.7118
0.6777
0.057
0.5144
0.2206
0.5995
0.0562
Root biomass
Soil type
AMF origin
MIC origin
Soil 3 AMF
Soil 3 MIC
AMF 3 MIC
Soil 3 AMF 3 MIC
Depth to soil (cm)
1
2
2
2
2
4
4
1
20.156343
0.34462
0.237193
0.262389
0.362569
0.76821
0.358936
0.000368
131.8085
1.1268
0.7755
0.8579
1.1855
1.2559
0.5868
0.0024
,0.0001
0.3266
0.4622
0.426
0.3083
0.2897
0.6727
0.9609
S. nutans
Shoot biomass
Soil type
AMF origin
MIC origin
Soil 3 AMF
Soil 3 MIC
AMF 3 MIC
1
1
2
1
2
2
18.337721
0.000014
0.03856
0.015696
0.024585
0.636003
238.5496
0.0002
0.2508
0.2042
0.1599
4.1368
,0.0001
0.9892
0.779
0.653
0.8526
0.0207
Root biomass
Soil type
AMF origin
MIC origin
Soil 3 AMF
Soil 3 MIC
AMF 3 MIC
1
1
2
1
2
2
17.855694
0.059542
0.001945
0.014583
0.073793
1.437133
175.6585
0.5858
0.0096
0.1435
0.363
7.069
,0.0001
0.447
0.9905
0.7062
0.6971
0.0017
Notes: Shown are effects of soil type (low-contaminated [LC], high-contaminated [HC]), source
of arbuscular mycorrhizal fungi (AMF) (LC, HC, none), and source of non-mycorrhizal soil
microbes (MIC) (LC, HC, none) on Deschampsia flexuosa (error df ¼ 159) shoot biomass (F18, 159 ¼
9.741, P , 0.001) and root biomass (F18, 159 ¼ 8.31, P , 0.0001), and on Sorghastrum nutans (error
df ¼ 61) shoot biomass (F9,61 ¼ 39.59, P , 0.0001) and root biomass (F9,61 ¼ 27.15, P , 0.0001).
Likewise, the number of NMF morphotypes was lower
(P , 0.001) when no MIC was added (2.8 6 0.41)
compared to MIC originating from either LC (5.8 6 0.9)
or HC soil (mean: 4.3 6 0.7). Raw root colonization
data for S. nutans can be found in Appendix H.
Elemental analysis of tissue for D. flexuosa.—Overall,
levels of P and Zn in the shoots of D. flexuosa were very
high (Appendix I). P averaged 2.94 6 0.41 mg/g across
all treatments, but neither AMF nor MIC treatment
significantly affected P concentration. Plants grown in
HC soil had slightly higher levels of P in their shoots on
average (3.15 6 0.5 mg/g) than plants grown in LC soil
(2.73 6 0.32 mg/g); F1,72 ¼ 4.11; P , 0.05). Plants grown
in HC soil had much higher levels of Zn (1.80 6 0.05
mg/g) than those in LC soil (0.23 6 0.008 mg/g); F1,72 ¼
1274.8; P , 0.001). AMF, as a main effect, did not affect
Zn concentration in plant shoots, but MIC treatment
did (F2,72 ¼ 5.1; P , 0.001). Plants amended with no
microbial wash, on average, showed the highest levels of
FIG. 1. Both root and shoot biomass (mean 6 SE;
measured in kg) of Deschampsia flexuosa (DF) were significantly greater (P , 0.05; Table 1) in low-contaminated soil
(LC) than in high-contaminated soil (HC) from Blue Mountain, a heavy-metal-contaminated site in Carbon County,
Pennsylvania, USA.
July 2012
CONTEXT-DEPENDENCY OF SOIL MICROBES
1555
In HC soils, plants not amended with microbial wash
had the highest levels of Zn overall (1.96 6 0.06 mg/g);
amendment with MIC of HC origin significantly
lowered the level of Zn (1.65 6 0.08 mg/g), while plants
amended with MIC of LC origin were in between (1.79
FIG. 2. (A) Both roots and shoots of Sorghastrum nutans
(SN) produced significantly greater biomass (mean 6 SE; units
as in Fig. 1) in low-contaminated soil (LC) than in highcontaminated soil (HC); P , 0.05 (Table 1). The arbuscular
mycorrhizal fungi (AMF) and other soil microbial treatments
(AMF 3 MIC) interacted to affect (B) shoot biomass and (C)
root biomass of Sorghastrum nutans (Table 1). Uppercase
letters contrast biomass with AMF of HC origin (AMF high) in
the different MIC treatments. Lowercase letters similarly
contrast biomass with AMF of LC origin (AMF low).
Treatments with the same letter and case do not differ
significantly at P , 0.05 (test slices).
Zn (1.1 6 0.16 mg/g), and plants amended with MIC
from HC soil had the lowest levels of Zn (0.94 6 0.14
mg/g); plants amended with MIC from LC soil (1.0 6
0.16 mg/g) did not differ from either. The significant soil
3 MIC interaction (F2,72 ¼ 3.4; P , 0.04) reflects an
effect of microbial wash in HC soils but not in LC soils.
FIG. 3. (A) Root colonization for Deschampsia flexuosa by
both arbuscular mycorrhizal fungi (AMF) and non-mycorrhizal
fungi (NMF) was significantly greater in low-contaminated soil
(LC) compared to high-contaminated soil (HC); P , 0.05
(Appendix E). (B) AMF was significant as a main effect for
root colonization of D. flexuosa by both AMF and NMF.
Colonization by AMF is shown as uppercase letters, and by
NMF as lowercase letters; AMF treatments with the same letter
and case do not differ significantly at P , 0.05 (Tukey’s HSD).
(C) The AMF and other soil microbial treatments (AMF 3
MIC) interacted to affect the number of NMF morphotypes in
roots of D. flexuosa (P , 0.05). Colonization by NMF differed
between MIC low and No MIC, AMF low (Tukey’s HSD; P ,
0.05). No other pairwise comparisons differed significantly.
1556
SYDNEY I. GLASSMAN AND BRENDA B. CASPER
Ecology, Vol. 93, No. 7
PLATE 1. (Bottom) Mixed Deschampsia flexuosa and Danthonia spicata grassland that has revegetated at the low contaminated
end of the gradient on Blue Mountain in Palmerton, Pennsylvania, USA. (Top) View looking out from the field site in Palmerton
across the completely defoliated mountainside that has not been revegetated and the abandoned zinc smelter that caused the heavy
metal contamination. Photo credits: S. I. Glassman.
6 0.08 mg/g) and not significantly different from the
other two MIC treatments.
DISCUSSION
The main goal of this study was to evaluate arbuscular
mycorrhizal fungi (AMF) function within the abiotic
context of soil metal contamination level and the biotic
context of other soil microbes in the system. We
confirmed clear biological consequences of soil metal
contamination levels, and for the C4 grass Sorghastrum
nutans, the effect of AMF on plant growth depended on
the presence of other soil microbes. There are other main
findings as well. (1) We found no evidence of ecological
matching between the AMF spore community and the
soil background as would occur if AMF from highcontaminated (HC) soil allowed plants to grow larger in
HC soil, and AMF from low-contaminated (LC) soil
allowed plants to grow larger in LC soil. (2) Microbes
other than AMF lowered Zn concentration in shoot
tissues of the native C3 grass Deschampsia flexuosa. (3)
D. flexuosa, which is more tolerant of soil Zn
contamination than S. nutans, did not respond in
biomass to AMF and microbe (MIC) treatments.
The differences between S. nutans and D. flexuosa in
how they responded to AMF and MIC treatments are
similar to results of other studies showing that the
interaction between AMF and root pathogens varies
with plant species (Sikes et al. 2009), and thus is also
July 2012
CONTEXT-DEPENDENCY OF SOIL MICROBES
context dependent. For instance, plants with complex
root systems are more susceptible to infection by
common fungal pathogens such as Fusarium spp. than
plants with simpler root systems, but colonization by
AMF can make them less susceptible to such infections
(Newsham et al. 1995, Sikes et al. 2009). The ability of a
particular mycorrhizal fungus to protect the plant from
pathogens also depends on the plant host. Setaria
glauca, which has a complex root system, is better
protected from the root pathogen Fusarium oxysporum
by AMF in the family Glomeraceae, while Allium cepa,
with a simple root system, receives better pathogen
protection from members of the family Gigasporaceae
(Sikes et al. 2009). Klironomos (2003) also found that
different plant species respond differently to the same
AMF inoculum and suggested that such differential
responses may contribute to plant species coexistence
and thus biodiversity.
Our results stand in contrast to those where microbial
washes consistently depressed plant growth (Mills and
Bever 1998, Klironomos 2002). Similar to how AMF of
different origins interacted with other soil microbes to
affect biomass of S. nutans, adding a microbial wash to
mycorrhizal plants of Andropogon gerardii reduced
growth, while adding the same wash to non-inoculated
plants improved growth (Hetrick et al. 1988). Hetrick et
al. (1988) speculated that the increased growth of noninoculated plants in response to the addition of a
microbial amendment may be due to increased mineralization of nutrients by the soil microbes. Rovira and
Bowen (1966) suggested that soil microorganisms
improve plant growth by detoxifying the soil. Such a
detoxification effect may be applicable in the case of our
heavy-metal-contaminated soils, considering that the
microbial wash from HC soils decreased the level of Zn
in tissues of D. flexuosa, especially for those grown in
HC soils. The addition of a microbial wash may also
cause the entry of soilborne pathogenic or saprophytic
fungi through mycoparasitism (De Jaeger et al. 2010),
and mycorrhizal fungi can be suppressed by competition
with pathogenic or saprophytic soil microbes (Sylvia
and Schenck 1984). Here, however, the AMF treatment
affected root colonization by NMF in D. flexuosa, but
the MIC treatment did not affect AMF colonization in
either species.
Especially because our results show that soil microorganisms other than AMF can affect the levels of Zn in
plant tissues and interact with the mycorrhizal symbiosis
both positively and negatively, these diverse effects
should be investigated more fully. For natural systems,
the functional significance of either group should not be
evaluated without the other present. Microbial washes
are expected to contain bacteria, fungal hyphae, and
spores of saprobes and pathogens smaller than 20 lm, in
addition to soil chemicals, and the relative proportions
of these different groups likely vary considerably. There
might be such differences in the microbial community
composition in LC and HC soil at the Palmerton site.
1557
While plant performance consequences of AMF are
often examined in the presence of a microbial wash
(Klironomos 2003, Scheublin et al. 2007), the source of
the wash usually does not vary within the experiment as
was done here. A few studies have purposely varied
combinations of AMF and soil microbes but not the full
complement of non-AMF microbes found in the soil
(Andrade et al. 1998, Dabire et al. 2007, Siasou et al.
2009).
We had expected to find that a particular AMF
community was more effective in promoting plant
growth in the soil contamination level in which the
community was collected, as was found for serpentine
and nonserpentine AMF communities by Ji et al. (2010).
Such was not the case. In fact, the differentiation
between the AMF of LC and HC origin was more subtle
and complicated, due to AMF of LC origin prolonging
the life of S. nutans in HC soil and AMF origin
interacting with the presence of other soil microbes in
affecting biomass of S. nutans. Our estimates indicated
more abundant and diverse spores in LC soil than in HC
soil, but AMF root colonization is consistently low
across both soils and both plant species, perhaps due to
low spore viability or low AMF tolerance for even low
levels of heavy metal contamination. Interestingly, there
was no effect of AMF treatment, either presence or
location of origin, on P levels in the shoot tissues of D.
flexuosa. Although P uptake is a main function of AMF
(Smith and Read 2008), other studies, too, have shown
AMF not to affect P concentrations. Newsham et al.
(1995) similarly found that AMF did not affect P
concentrations; rather, the main effect of AMF was to
protect plants from root pathogenic fungi.
We must conclude that the native D. flexuosa is a
better species than S. nutans for restoration of the site.
The latter was chosen for restoration of Blue Mountain,
in part, because it is one of the dominant grasses in
serpentine outcrops in the region (D. Kunkle, personal
communication). C4 plants are also better able to tolerate
drought (Hartnett and Wilson 1999), which is an issue
both on shallow serpentine soils and on the highly
eroded Blue Mountain soils. Although seeds of S. nutans
have been broadcast throughout the site, it primarily
establishes where soils have been amended with compost
or lime and fertilizer. Our results corroborate that S.
nutans does not establish well in HC soils and suggest
that the native C3 grasses will probably maintain their
dominance on Blue Mountain. Our results also suggest
the need to focus on the roles of other soil microbes, in
addition to AMF, in facilitating revegetation of heavymetal-polluted sites. HC soils at Palmerton might be a
fruitful place to look for microbes with particular Zn
sequestration or remediation ability. An in-depth
characterization of the soil microbial community followed by physiology assays might elucidate the identity
of these microbes and their Zn sequestration mechanisms.
1558
SYDNEY I. GLASSMAN AND BRENDA B. CASPER
ACKNOWLEDGMENTS
We are deeply grateful to P. Petraitis and R. SalgueroGómez for help in statistical analyses, and to F. Anderson, R.
Vandegrift, S. Bentivenga, J. Doherty, and the University of
Pennsylvania Biology Department greenhouse staff for help or
advice on different aspects of the experiment. P. Brooks
provided ICP training, and D. Kunkle at the Lehigh Gap
Nature Center facilitated access to the field sites and sample
collections. Many thanks to T. Bruns, K. Crawford, K. Peay,
and two anonymous reviewers for valuable comments on the
manuscript. This work was partially supported by the Earth
and Environmental Studies Professional Program Research
Award to S. I. Glassman.
LITERATURE CITED
Abbott, L. K., A. D. Robson, and C. Gazey. 1992. Selection of
inoculant vesicular-arbuscular mycorrhizal fungi. Methods in
Microbiology 24:1–21.
Agrawal, A. A., et al. 2007. Filling key gaps in population and
community ecology. Frontiers in Ecology and the Environment 5:145–152.
Andrade, G., R. G. Linderman, and G. J. Bethlenfalvay. 1998.
Bacterial associations with the mycorrhizosphere and hyphosphere of the arbuscular mycorrhizal fungus Glomus mosseae.
Plant and Soil 202:79–87.
Andrade, S. A. L., P. L. Gratao, M. A. Schiavinato, A. P. D.
Silveira, R. A. Azevedo, and P. Mazzafera. 2009. Zn uptake,
physiological response and stress attenuation in mycorrhizal
jack bean growing in soil with increasing Zn concentrations.
Chemosphere 75:1363–1370.
Auge, R. M. 2001. Water relations, drought and vesiculararbuscular mycorrhizal symbiosis. Mycorrhiza 11:3–42.
Azcon-Aguilar, C., and J. M. Barea. 1996. Arbuscular
mycorrhizas and biological control of soil-borne plant
pathogens: an overview of the mechanisms involved.
Mycorrhiza 6:457–464.
Bradshaw, A. D. 1952. Populations of Agrostis tenuis resistant
to lead and zinc poisoning. Nature 169:1098–1098.
Buchauer, M. J. 1973. Contamination of soil and vegetation
near a zinc smelter by zinc, cadmium, copper, and lead.
Environmental Science and Technology 7:131–135.
Castelli, J. P., and B. B. Casper. 2003. Intraspecific AM fungal
variation contributes to plant–fungal feedback in a serpentine
grassland. Ecology 84:323–336.
Dabire, A. P., V. Hien, M. Kisa, A. Bilgo, K. S. Sangare, C.
Plenchette, A. Galiana, Y. Prin, and R. Duponnois. 2007.
Responses of soil microbial catabolic diversity to arbuscular
mycorrhizal inoculation and soil disinfection. Mycorrhiza
17:537–545.
De Jaeger, N., S. Declerck, and I. E. de la Providencia. 2010.
Mycoparasitism of arbuscular mycorrhizal fungi: a pathway
for the entry of saprotrophic fungi into roots. Fems
Microbiology Ecology 73:312–322.
Del Val, C., J. M. Barea, and C. Azon-Aguilar. 1999. Diversity
of arbuscular mycorrhizal fungus populations in heavymetal-contaminated soils. Applied and Environmental Microbiology 65:718–723.
Doherty, J. H. 2009. Niche partitioning among arbuscular
mycorrhizal fungi and consequences for host plant performance. University of Pennsylvania, Philadelphia, Pennsylvania, USA.
Dudka, S., and D. C. Adriano. 1997. Environmental impacts of
metal ore mining and processing: a review. Journal of
Environmental Quality 26:590–602.
Fitter, A. H., and B. Moyersoen. 1996. Evolutionary trends in
root–microbe symbioses. Philosophical Transactions of the
Royal Society B 351:1367–1375.
Frey-Klett, P., J. Garbaye, and M. Tarkka. 2007. The
mycorrhiza helper bacteria revisited. New Phytologist
176:22–36.
Ecology, Vol. 93, No. 7
Galli, U., H. Schuepp, and C. Brunold. 1994. Heavy-metal
binding by mycorrhizal fungi. Physiologia Plantarum
92:364–368.
Gamalero, E., G. Lingua, G. Berta, and B. R. Glick. 2009.
Beneficial role of plant growth promoting bacteria and
arbuscular mycorrhizal fungi on plant responses to heavy
metal stress. Canadian Journal of Microbiology 55:501–514.
Hartnett, D. C., and G. W. T. Wilson. 1999. Mycorrhizae
influence plant community structure and diversity in tallgrass
prairie. Ecology 80:1187–1195.
Hetrick, B. A. D., G. T. Wilson, D. G. Kitt, and A. P. Schwab.
1988. Effects of soil-microorganisms on mycorrhizal contribution to growth of big bluestem grass in non-sterile soil. Soil
Biology and Biochemistry 20:501–507.
Hildebrandt, U., M. Regvar, and H. Bothe. 2007. Arbuscular
mycorrhiza and heavy metal tolerance. Phytochemistry
68:139–146.
Hoeksema, J. D., et al. 2010. A meta-analysis of contextdependency in plant response to inoculation with mycorrhizal fungi. Ecology Letters 13:394–407.
Jeffries, P., S. Gianinazzi, S. Perotto, K. Turnau, and J. M.
Barea. 2003. The contribution of arbuscular mycorrhizal
fungi in sustainable maintenance of plant health and soil
fertility. Biology and Fertility of Soils 37:1–16.
Ji, B., S. P. Bentivenga, and B. B. Casper. 2010. Evidence for
ecological matching of whole AM fungal communities to the
local plant–soil environment. Ecology 91:3037–3046.
Johnson, A. H., and S. L. Richter. 2010. Organic-horizon lead,
copper, and zinc contents of Mid-Atlantic forest soils, 1978–
2004. Soil Science Society of America Journal 74:1001–1009.
Jordan, M. J. 1975. Effects of zinc smelter emissions on a
chestnut–oak woodland. Ecology 56:78–91.
Jordan, M. J., and M. P. Lechevalier. 1975. Effects of zincsmelter emissions on forest soil microflora. Canadian Journal
of Microbiology 21:1855–1865.
Kelly, J. J., M. M. Häggblom, and R. L. Tate. 2003. Effects of
heavy metal contamination and remediation on soil microbial communities in the vicinity of a zinc smelter as indicated
by analysis of microbial community phospholipid fatty acid
profiles. Biology and Fertility of Soils 38:65–71.
Klironomos, J. N. 2002. Feedback with soil biota contributes to
plant rarity and invasiveness in communities. Nature 417:67–
70.
Klironomos, J. N. 2003. Variation in plant response to native
and exotic arbuscular mycorrhizal fungi. Ecology 84:2292–
2301.
Kunkle, D. R. 2006. Refuge grass planting nearing completion.
Wildlife Activist.
Latham, R. E., D. B. Steckel, H. M. Harper, C. Steckel, and M.
Boatright. 2007. Lehigh Gap Wildlife Refuge ecological
assessment. Lehigh Gap Nature Center, Slatington, Pennsylvania, USA.
McGonigle, T. P., M. H. Miller, D. G. Evans, G. L. Fairchild,
and J. A. Swan. 1990. A new method which gives an
objective-measure of colonization of roots by vesicular
arbuscular mycorrhizal fungi. New Phytologist 115:494–501.
McKenney, M. C., and D. L. Lindsey. 1987. Improved method
for quantifying endomycorrhizal fungi spores from soil.
Mycologia 79:779–782.
Mills, K. E., and J. D. Bever. 1998. Maintenance of diversity
within plant communities: soil pathogens as agents of
negative feedback. Ecology 79:1595–1601.
Nash, T. H. 1975. Influence of effluents from a zinc factory on
lichens. Ecological Monographs 45:183–198.
Newsham, K. K., A. H. Fitter, and A. R. Watkinson. 1995.
Multi-functionality and biodiversity in arbuscular mycorrhizas. Trends in Ecology and Evolution 10:407–411.
Oyler, J. A. 1993. Use of power plant fly ash/municipal sludge
admixture to reclaim land near a smelter. Tenth International
Ash Use Symposium. The Institute, Palo Alto, California,
USA.
July 2012
CONTEXT-DEPENDENCY OF SOIL MICROBES
Pawlowska, T. E., J. Blaszkowski, and A. Ruhling. 1996. The
mycorrhizal status of plants colonizing a calamine spoil
mound in southern Poland. Mycorrhiza 6:499–505.
Philips, J. M., and D. S. Hayman. 1970. Improved procedures
for clearing roots and staining parasitic and vesicular–
arbuscular mycorrhizal fungi for rapid assessment of
infection. Transactions of the British Mycological Society 55.
Roberts, D. R., A. C. Scheinost, and D. L. Sparks. 2002. Zinc
speciation in a smelter-contaminated soil profile using bulk
and microspectroscopic techniques. Environmental Science
and Technology 36:1742–1750.
Rovira, A. D., and G. D. Bowen. 1966. Effects of microorganisms upon plant growth. 2. Detoxication of heatsterilized soils by fungi and bacteria. Plant and Soil 25:129–
142.
SAS Institute. 2010. JMP 8. Version 8.02. SAS Institute, Cary,
North Carolina, USA.
Schechter, S. P., and T. D. Bruns. 2008. Serpentine and nonserpentine ecotypes of Collinsia sparsiflora associate with
distinct arbuscular mycorrhizal fungal assemblages. Molecular Ecology 17:3198–3210.
Scheublin, T. R., R. S. P. Van Logtestijn, and M. G. A. Van der
Heijden. 2007. Presence and identity of arbuscular mycorrhizal fungi influence competitive interactions between plant
species. Journal of Ecology 95:631–638.
Siasou, E., D. Standing, K. Killham, and D. Johnson. 2009.
Mycorrhizal fungi increase biocontrol potential of Pseudomonas fluorescens. Soil Biology and Biochemistry 41:1341–
1343.
1559
Sikes, B. A., K. Cottenie, and J. N. Klironomos. 2009. Plant
and fungal identity determines pathogen protection of plant
roots by arbuscular mycorrhizas. Journal of Ecology
97:1274–1280.
Smith, S. E., and D. J. Read. 2008. Mycorrhizal symbiosis.
Academic Press, Cambridge, UK.
Stanton, M. L. 2003. Interacting guilds: moving beyond the
pairwise perspective on mutualisms. American Naturalist
162:S10–S23.
Storm, G. L., G. J. Fosmire, and E. D. Bellis. 1994. Persistence
of metals in soil and selected vertebrates in the vicinity of the
Palmerton zinc smelters. Journal of Environmental Quality
23:508–514.
Strojan, C. L. 1978. Impact of zinc smelter emissions on forest
litter arthropods. Oikos 31:41–46.
Sylvia, D. M., and N. C. Schenck. 1984. Aerated-steam
treatment to eliminate VA mycorrhizal fungi from soil. Soil
Biology and Biochemistry 16:675–676.
Wehner, J., P. M. Antunes, J. R. Powell, J. Mazukatow, and
M. C. Rillig. 2010. Plant pathogen protection by arbuscular
mycorrhizas: a role for fungal diversity? Pedobiologia
53:197–201.
Wright, S. F., and A. Upadhyaya. 1998. A survey of soils for
aggregate stability and glomalin, a glycoprotein produced by
hyphae of arbuscular mycorrhizal fungi. Plant and Soil
198:97–107.
Zarcinas, B. A., B. Cartwright, and L. R. Spouncer. 1987.
Nitric-acid digestion and multi-element analysis of plantmaterial by inductively coupled plasma spectrometry. Communications in Soil Science and Plant Analysis 18:131–146.
SUPPLEMENTAL MATERIAL
Appendix A
Soil elemental values from Doherty 2009 (Ecological Archives E093-138-A1).
Appendix B
List of AMF morphospecies found in Palmerton soils (Ecological Archives E093-138-A2).
Appendix C
Raw biomass data for Deschampsia flexuosa across treatments (Ecological Archives E093-138-A3).
Appendix D
Raw biomass data for Sorghastrum nutans across treatments (Ecological Archives E093-138-A4).
Appendix E
ANOVA results from root colonization analyses of Deschampsia flexuosa (Ecological Archives E093-138-A5).
Appendix F
Raw root colonization data for Deschampsia flexuosa across treatments (Ecological Archives E093-138-A6).
Appendix G
ANOVA results from root colonization analyses of Sorghastrum nutans (Ecological Archives E093-138-A7).
Appendix H
Raw root colonization data for Sorghastrum nutans across treatments (Ecological Archives E093-138-A8).
Appendix I
Levels of P and Zn in shoots of Deschampsia flexuosa shoots across treatments (Ecological Archives E093-138-A9).