Presence of the microbiome consistently decreased growth rates and
modified phenotypes of the host plants
The main aim of our experiment was to assess the fitness and phenotypic
consequences of the microbiome for L. minor . Although
considerable recent work has investigated the importance of certain
microbes (mostly bacteria) for L. minor growth (Ishizawa et al.
2017a, 2017b, 2019, Gilbert et al. 2018, Chen et al. 2019, Acosta et al.
2020, Iwashita et al. 2020, O’Brien et al. 2020a, 2020b, Tan et al.
2021), the aim has largely been to isolate certain PGPB that increase
plant fitness, with few studies characterising the impact of the entire
intact natural microbiome on plant performance. Here we isolate plants
from eight different genotypes with their full natural microbiomes and
assess the impact of the microbiome on host fitness and phenotype.
The effect of the microbiome on plant growth rates was strong and
consistent. Contrary to our expectation, the presence of the microbiome
decreased plant fitness, on average by 12%. This was the case across
all environmental conditions (Fig. 1), and for seven of the eight
genotypes (Fig. 2a). Although several important plant-bacteria and
plant-fungi mutualisms have been identified for L. minor that
increase plant fitness (Acosta et al. 2020, O’Brien et al. 2020a, 2020b,
Tan et al. 2021), our results suggest that the importance of pathogens,
parasites, and competitors in the microbial assemblage far surpass that
of any mutualistic microbes. This is not necessarily surprising given
the rich literature documenting the importance of fungal and bacterial
pathogens (Rejmankova et al. 1986, Underwood and Baker 1991, Zhang et
al. 2010, Ishizawa et al. 2017a, 2017b) and algal competition (van
Moorsel 2022), on L. minor growth. In land plants, assemblages of
PGPB are often unstable in the field (Parnell et al. 2016), and inL. minor , the effects of fitness-enhancing strains can be lost
with the inclusion of additional strains due to non-additive effects
(Ishizawa et al. 2017b).
The effects of the microbiome on plant phenotype were equally clear. The
presence of the microbiome resulted in plants with shorter roots and
smaller fronds across all genotypes (Fig. 3). One explanation for
smaller fronds would be the presence of many microbes, including
photosynthetic algae, that decrease nutrient availability through direct
competition with L. minor . However, if this were the main
mechanism through which the microbiome modified L. minorphenotype, then it would result in increased root length, the ubiquitous
plastic response to decreased nutrient availability. However, we found
the opposite, i.e. shorter roots, perhaps a plant response to limit the
available surface area for microbes to colonize. In addition, although
decreased nutrient availability results in an increase in colony size,
we find that the presence of the microbiome increases frond abscission
resulting in smaller colonies (Fig. 4). This is consistent with other
work that has found microbially-mediated shifts in average colony size
in L. minor (O’Brien et al. 2020a). We therefore conclude that
the mechanism by which the microbiome supressed plant fitness in our
experiment goes beyond changes in resource levels. Frond abscission in
response to heavy metals has been extensively studied in L. minorin the ecotoxicology literature and it is well known that toxic stress
generally decreases colony size (Severi 2001, Li and Xiong 2004a, 2004b,
Henke et al. 2011, Topp et al. 2011, O’Brien et al. 2020a). The decrease
in colony size we observe when the microbiome was present could be due
to a similar phenomenon, resulting from toxic microbial secondary
metabolites.
One reason for the apparent inconsistency of our results with studies
that report a fitness enhancing effect of many microbes (eg. O’Brien et
al. 2020a, 2020b, Tan et al. 2021), is that the majority of these
studies are explicitly looking to identify only these mutualistic
associations. Due to its extremely rapid growth rate, among the fastest
of all plants, L. minor research is often in the context of its
many industrial applications which include waste water remediation
(Landesman et al. 2011, Iqbal and Baig 2016), biomass production as
biofuel (Verma and Suthar 2015), animal feed (Islam et al. 2004, Cheng
and Stomp 2009), and human consumption (Sree et al. 2016, Appenroth et
al. 2017). For all these applications, there is a keen incentive to
further enhance growth rate. Much research has focussed on identifying
and selecting the most productive genetic strains of L. minor(Bergmann et al. 2000), and much of the work on the microbiome has been
done in the same vein, aiming to identify and isolate specific strains
of PGPB (Yamaga et al. 2010, Tang et al. 2015, Appenroth et al. 2016).
This bias in the literature could lead to a general impression that the
microbiome is dominated by mutualistic fitness-enhancing associations,
despite a general lack of evidence. Most studies intentionally isolate
strains of bacteria that are good candidates to promote plant growth,
which are then artificially inoculated to the axenic plants. Here we
take the opposite approach, to estimate the overall effect on their host
of the large and diverse assemblages of microbes that make up the
microbiome. There are few studies that have tested the effect of entireL. minor microbiome on plant fitness instead of just a small
subset of carefully chosen bacteria, and those that did found
conflicting results. The study that most resembles ours in design,
reinoculated the full microbial community to axenic L. minor , and
concluded that the microbiome increased frond senescence (Underwood and
Baker 1991).
A limitation of this study is the fact that we did not characterize the
microbial community and thus, we can only speculate on the mechanisms
responsible for our results. Variation in the phenotypic and fitness
consequences of the microbiome was surprisingly consistent across all
genotypes. This is notable since our genotype treatment included not
just different plant clones, but also independent microbiomes from each
genotype. Despite the possibility of strong differences in microbial
community composition among genotypes, their overall effect on each
plant genotype was overwhelmingly uniform. This is consistent with work
that has shown the absence of plant-microbe specialisation among
genotypes in L. minor by manipulating plant genotype and
microbial community source independently (O’Brien et al. 2020a). It
appears that in our experiment, the eight independent microbial
communities were of similar composition, at least in terms of board
functional groups and their interactions with the plant host.