Multimodal correlative imaging and modelling of phosphorus
uptake from soil by hyphae of mycorrhizal fungi
Sam Keyes1*, Arjen van Veelen1,2,3*, Dan McKay Fletcher1 , Callum Scotson1, Nico Koebernick1,
Chiara Petroselli1, Katherine Williams1, Siul Ruiz1, Laura Cooper1, Robbie Mayon1, Simon Duncan1,
Marc Dumont4, Iver Jakobsen5, Giles Oldroyd6, Andrzej Tkacz7, Philip Poole7, Fred Mosselmans8, Camelia Borca9,
Thomas Huthwelker9, David L. Jones10,11 and Tiina Roose1
1Bioengineering Sciences Research Group, Department of Mechanical Engineering, School of Engineering, Faculty of Engineering and Physical Sciences, University of Southampton,
Southampton, SO17 1BJ, UK; 2Material Science and Technology Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA; 3Stanford Synchrotron Radiation Lightsource,
SLAC National Accelerator Laboratory, Menlo Park, CA 94025, USA; 4School of Biological Sciences, University of Southampton, Southampton, SO17 1BJ, UK; 5Department of Plant and
Environmental Sciences, University of Copenhagen, Thorvaldsensvej 40, Frederiksberg DK-1871, Denmark; 6Crop Science Centre, University of Cambridge, 93 Lawrence Weaver Road,
Cambridge, CB3 0LE, UK; 7Department of Plant Sciences, University of Oxford, South Parks Road, Oxford, OX1 3RB, UK; 8Diamond Light Source, Diamond House, Harwell Science
& Innovation Campus, Didcot, OX11 0DE, UK; 9Swiss Light Source, PSI, Forschungsstrasse 111, Villigen 5232, Switzerland; 10School of Natural Sciences, Bangor University, Bangor,
LL57 2DG, UK; 11SoilsWest, Food Futures Institute, Murdoch University, 90 South Street, Murdoch, WA 6150, Australia
Author for correspondence:
Tiina Roose
Email: t.roose@soton.ac.uk
Received: 1 November 2021
Accepted: 3 January 2022
New Phytologist (2022)
doi: 10.1111/nph.17980
Key words: mycorrhizas, plant phosphorus
uptake, rhizosphere modelling, synchrotron,
X-ray computed tomography, X-ray
fluorescence.
Summary

 Phosphorus (P) is essential for plant growth. Arbuscular mycorrhizal fungi (AMF) aid its
uptake by acquiring P from sources distant from roots in return for carbon. Little is known
about how AMF colonise soil pore-space, and models of AMF-enhanced P-uptake are poorly
validated.

 We used synchrotron X-ray computed tomography to visualize mycorrhizas in soil and syn-
chrotron X-ray fluorescence/X-ray absorption near edge structure (XRF/XANES) elemental
mapping for P, sulphur (S) and aluminium (Al) in combination with modelling.

 We found that AMF inoculation had a suppressive effect on colonisation by other soil fungi
and identified differences in structure and growth rate between hyphae of AMF and nonmyc-
orrhizal fungi. Our results showed that AMF co-locate with areas of high P and low Al, and
preferentially associate with organic-type P species over Al-rich inorganic P.

 We discovered that AMF avoid Al-rich areas as a source of P. Sulphur-rich regions were
found to be correlated with higher hyphal density and an increased organic-associated P-
pool, whilst oxidized S-species were found close to AMF hyphae. Increased S oxidation close
to AMF suggested the observed changes were microbiome-related. Our experimentally-
validated model led to an estimate of P-uptake by AMF hyphae that is an order of magnitude
lower than rates previously estimated – a result with significant implications for the modelling
of plant–soil–AMF interactions.
Introduction
Mineral phosphorus (P) resources are sparse and unevenly dis-
tributed across the world (Gross, 2017). Arbuscular mycorrhizal
fungi (AMF) play an important role in mediating plant uptake of
P (Smith et al., 2003), which is often considered to be a growth-
limiting soil resource (Vaccari, 2009). Mycorrhizal plants supply
carbon (C) to AMF mycelia to drive hyphal growth, while in
return the hyphae provide P and other nutrients back to the
plant. Recent studies have found that in some cases a reduced
plant C allocation to AMF does not alter AMF P supply to the
plant (Charters et al., 2020), which shows promise for crop
production when P supply is limiting to growth. Much is known
about plant–AMF symbioses on the soil bulk-scale, but little is
known about the spatial distribution of AMF hyphae in soil due
to difficulties in visualizing soil pore-space in situ. However, it is
important to know how AMF hyphae interact with soil as P is
strongly bound to soil surfaces, and pore-scale processes thus gov-
ern AMF-mediated P uptake. While we can modify certain pore-
scale processes using chemical and microbial soil treatments, it is
important to establish the mechanistic basis of symbiont
behaviour within the soil matrix.
Sulphur (S), like P, is an essential macronutrient required by
plants, and it can also promote hyphal growth (Hepper, 1984). It
has the potential to be a surrogate marker for the soil microbial
community. Sulphur exists in soils in a wide variety of*These authors contributed equally to this work.

 2022 The Authors
New Phytologist 
 2022 New Phytologist Foundation
New Phytologist (2022) 1
www.newphytologist.com
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.
Research
environmentally-dependent oxidation states, ranging from 2 to
+6, with organo-S comprising > 90% of the total S pool in most
soils (Gene et al., 2002; Prietzel et al., 2007, 2011). Arbuscular
mycorrhizal fungi may play a role in plant S metabolism through
uptake and up-regulation of plant sulphate carriers, and through
their interaction with organo-S mobilizing microbes (Gahan &
Schmalenberger, 2014; Berruti et al., 2015). Arbuscular mycor-
rhizal fungi mycelia are surrounded by complex bacterial and
fungal communities that both interact with and sustain their
metabolic function. Therefore, similar to the area affected by
roots (i.e. the rhizosphere), a ‘hyphosphere’ of AMF can be a
zone of increased bacterial abundance, and a site of localised bio-
chemical activity (Rozmos et al., 2021).
Several groups have modelled the growth of hyphal networks,
as reviewed by Boswell & Davidson (2012). These models have
addressed cell physics aspects (Bartnicki-Garcia et al., 2000),
physiological population growth (Edelstein, 1982), and biochem-
ical processes controlling hyphal growth (Tlalka et al., 2003).
However, a persistent limitation in such models is a lack of vali-
dation due to experimental constraints. As soil is opaque to visi-
ble light, direct observation of hyphal morphology without
disturbing the soil matrix is challenging. Boswell & Davidson
(2012) reviewed models of mycelial development, while also
identifying the need for model validation with experimental data.
Simard et al. (2012) reviewed the ecology and modelling of myc-
orrhizal fungi and highlighted the fact that a special challenge is
the lack of techniques for observing hyphae in soil at a suitable
spatial resolution. Some of the first models to specifically include
AMF growth in soil and uptake of P were developed by our
group (Schnepf & Roose, 2006; Schnepf et al., 2008a). However,
these models were only validated against data on the bulk soil
and plant scales, and the soil P status was not monitored spa-
tiotemporally. This lack of knowledge regarding AMF mycelial
architecture on the soil pore-scale has impeded further investiga-
tion of the significance of soil–AMF–plant interactions for P
uptake. In this paper we take the first necessary step towards
building a fully validated plant–AMF P-uptake model. This is
achieved by imaging AMF structures in the soil pore-space in situ
using synchrotron X-ray computed tomography (SXRCT) in
combination with traditional hyphal length measurements as set
out by Jakobsen et al. (1992b), and correlating these data with
spatial profiles of P, S and Al on the same soil samples using X-
ray fluorescence/X-ray absorption near edge structure (XRF/
XANES) imaging. Aluminium profiles are used to distinguish
soil-mineral-associated P from the organic pool. The structural
and chemical imaging results are integrated into a plant–AMF
mathematical model of P uptake (Schnepf & Roose, 2006; Sch-
nepf et al., 2008a) to further estimate hyphal uptake rates and
predict how the plant–AMF symbiosis benefits plant P acquisi-
tion. This enables us to image and quantify AMF in soil in three
dimensions for the first time, apply finer-scale limits on P uptake
rates by AMF hyphae, and provide the first calibrated mathemati-
cal model for AMF hyphae P uptake in soil. The comparison of
hyphal length density measurements via SXRCT and a destruc-
tive approach also allows us to quantify the differences between
two mycelial detection techniques.
Materials and Methods
Plant and fungal growth assay
The growth medium was a sand-textured Eutric Cambisol soil
collected from a surface plot at Abergwyngregyn, North Wales,
UK (lat. 53°140N, long. 4°010W), for which the soil organic
matter content was 7%. This soil corresponds to ‘soil B’ in
Lucas & Jones (2006). A split-compartment system was
designed to produce soil samples of small cross-section
(∅ < 5 mm) for hyphal compartments, allowing SXRCT imag-
ing at a sufficient spatial resolution to observe hyphal structures
(see Supporting Information Methods S1: Figs S1.1, S1.2
therein). The system enabled the growth and maintenance of
mycorrhizal wheat plants under controlled conditions for up to
4 wk. A bespoke growth box was designed to maintain both
hyphal and root compartments in dark conditions under stable
and externally applied water potentials whilst allowing gas
exchange and aerial plant growth (see Methods S1 for full
design details). A mesh barrier maintained the separation
between the root and the hyphal compartments while permit-
ting the transfer of hyphae (Faber et al., 1991; Carminati et al.,
2009) and water between the two. In P+ treatments, a small
solid pellet of triple super phosphate (mean mass
0.0757 0.007 g) was added to each hyphal compartment. In
P treatments, no supplemental P was added.
Plant-inoculation treatments A wheat cultivar Triticum aes-
tivum L. cv Apache (WBCDB0003-PG-1) provided by the
Germplasm Resources Unit of the John Innes Centre (UK) was
selected due to its indicated high affinity for colonization by
AMF (Leake, 2016). Seeds were surface sterilised and germinated
in the dark for 96 h at 23°C. Seedlings were selected for uniform
radicle length and transplanted to each of 12 root compartments
(2 P treatments9 2 time points9 3 replicates).
Root compartments were inoculated with Rhizophagus irregu-
laris (BEG72; PlantWorks Ltd, Henfield, UK); we also tried
inoculation with slow growing fungus Gigaspora rosea, but as this
fungus did not grow, we did not use it for the remainder of the
study (Methods S1). The inoculum contained hyphae, spores
and colonized root fragments at 1.69 106 propagules l1 of an
inert zeolite carrier substrate. A uniform inoculating paste was
achieved by concentrating the substrate according to manufac-
turer recommendations (Methods S1: Section 1.3 therein). Dur-
ing the filling of each root compartment, 1.5 g of inoculating
paste and 0.5 g of raw inoculum (see also Methods S1: Sec-
tion 1.3 therein) was added at 25, 65 and 105 mm (d1, d2 and
d3, respectively) from the top of each root compartment (Meth-
ods S1: Fig. S1.1 therein); this was advised by the manufacturer
PlantWorks Ltd to maximise the symbiosis.
The growth boxes were maintained in a climate-controlled
growth chamber (Conviron A1000, Conviron UK, Isleham, UK)
with a 14 h : 10 h, 23°C : 18°C, light : dark photoperiod (full
light: 700 µmol m2 s1) at 75% humidity. For each treatment,
three plant replicates (R1, R2, R3) were grown to 2 and 4 wk after
transplantation (i.e. a total of six plants per P condition).
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist2
Control treatments The soil was neither autoclaved nor
gamma-irradiated, since this would dramatically alter the natural
microbiome and biogeochemical cycling processes occurring
within the soil. Hence, the soil was assumed to contain native
fungal spores. To prevent AMF propagules from developing
hyphae in the controls, these contained no plants. Therefore, the
‘plant’ compartments for the control replicates were filled as out-
lined in the previous subsection and were lightly sealed with rub-
ber bungs, with the soil water potential of the controls being kept
similar to that of the plant compartments.
Structural imaging
Following a growth period of 14 or 28 d, depending on the treat-
ment, the hyphal compartments were imaged using two separate
SXRCT beamlines: The I13 beamline at the Diamond Light
Source, UK, and the TOMCAT beamline at the Swiss Light
Source (SLS), Switzerland, using a setup described in previous
studies (Keyes et al., 2013; Koebernick et al., 2017, 2019). For
each hyphal compartment, three vertical positions were imaged:
an intermediate position (h2) was set as the approximate P pellet
position in the P+ samples, and a near-root (h1) and far (h3) posi-
tion were set 2000 µm below and above this position respectively.
The effective voxel size was c. 1.6 lm. For data collected at I13,
only absorption reconstruction was available. For data collected
at SLS, both phase/Paganin (Paganin et al., 2002) and absorption
reconstruction were available. See Methods S1 for full details.
Synchrotron X-ray computed tomography image analysis
Unless specified otherwise, image analysis was carried out using
custom scripts written in IMAGEJ/FIJI (Schindelin et al., 2012).
The soil phase was first segmented into three phases: air-filled
pore-spaces (‘pore’), primary mineral grains (‘primary’), and
mixed phase (‘mixed’) (Methods S1: Figs S1.13–S1.16 therein)
using a WEKA machine-learning approach (Daly et al., 2015;
Keyes et al., 2017; Koebernick et al., 2017). Hyphal classification
was carried out using the absorption volumes due to the greater
hypha-to-pore contrast (Methods S1: Fig. S1.12 therein), after
the soil segmentation result from the phase reconstructed volumes
had first been used to mask out mineral regions. Hyphal classifi-
cation was carried out using a custom segmentation approach
with morphology first filtered using a two-pass morphological fil-
ter to remove small noise artifacts while conserving long filamen-
tous structures of largely consistent diameter (i.e. hyphae
(Methods S1)).
The segmented and filtered hyphal structures were skele-
tonised, and the following metrics were quantified using the
BONEJ toolbox (Doube et al., 2010): total hyphal length, number
of discrete hyphal clusters, branch count per cluster, mean branch
length per cluster, tortuosity of branches, and angle of orientation
to the hyphal compartment midline (i.e. the vector normal to the
barrier mesh). These metrics were computed in MATLAB 16b
(Natick, MA, USA), along with the SD and SE in the mean
across replicates for each measure (Methods S1).
Traditional measurements
Traditional hyphal counting The fraction of hyphae captured
by the SXRCT approach was compared against that derived from
a destructive sampling method (Jakobsen et al., 1992a,b). Replicate
hyphal compartments were cut into three equal sections at the
same locations used for SXRCT imaging, and each soil sample was
divided into two subsamples for counting (Methods S1). Hyphae
were counted using digitised images collected via light microscopy
with an BX41 microscope (Olympus UK & Ireland, Southend-
on-Sea, UK) with a 920 objective magnification and bright-field
illumination. A total of 20 images were acquired from randomly
selected locations across each sample. This produced a total of 40
images for each soil sample. These images were examined to assess
the phenotypic diversity within the entire sample set (Methods S1:
Fig. S1.29 therein). Based on these classifications, a set of candi-
date AMF phenotypes was defined with reference to the literature
(Abbott & Robson, 1985; Friese & Allen, 1991; Giovannetti et al.,
1993, 2001, 2004), to which all hyphal structures were subse-
quently assigned manually during the counting stage. See Methods
S1 for full detailed methods and results.
Polymerase chain reaction Polymerase chain reaction (PCR)
analysis of soil double-stranded DNA purified from soil was car-
ried out to validate the colonisation of hyphal compartments by
AMF. Four sets of primers were used: (a) AMF primers (Kruger
et al., 2009) designed to amplify Glomeromycota fungi, (b) a
bespoke set of AMF internal transcribed spacer (ITS) primers (re-
ferred as Tkacz primers), (c) a broad range primer (Buee et al.,
2009a,b) designed to amplify Ascomycota and Basidiomycota
rather than Glomeromycota, and (d) bacterial primers (515F and
806R) targeting the prokaryotic 16S rRNA V4 region (Caporaso
et al., 2011). See Methods S1 for further details of the primers
used. We first confirmed that primers (a) and (b) target AMF,
including R. irregularis, while primers (c) and (d) target other
fungal and nonfungal species, but not AMF (Methods S1: Sec-
tion 7.3). For all primer sets, standard polymerase chain reaction
(PCR) conditions were used: 98°C for 3 min, 35 cycles of 98°C
for 15 s, 55°C for 30 s and 72°C for 100 s, followed by a final
elongation step at 72°C for 7 min using Phusion high-fidelity
polymerase (M0530L; New England Biolabs, Hitchin, UK) and
a PCR master mix with GC buffer (M0532L; New England Bio-
labs). Following PCR, electrophoresis gels were used to deter-
mine whether amplification had been successful.
Microbial diversity analysis The 18S rRNA genes were ampli-
fied from soil DNA extracts from hyphal compartments using
the PCR primers F-574 and R-962 (Hadziavdic et al., 2014), and
sequencing was performed on an Illumina MiSeq platform (Envi-
ronmental Sequencing Facility, University of Southampton).
Sequences were processed and analysed using the DADA2 pipeline
(Callahan et al., 2016). Sequences shorter than 320 bp were dis-
carded. Amplicon sequence variants (ASVs) were classified using
the Wang Bayesian classifier in DADA2 and the Silva taxonomy
(Pruesse et al., 2012). The taxonomy of selected ASVs that could
not be identified using the classifier were individually analysed by

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 3
BLAST (Altschul et al., 1990) against the nonredundant (nr)
NCBI database; the taxonomy of the closest BLAST match was
used as an indication of the identity of the ASV. The ASV counts
were subjected to a Hellinger transformation with the decostand
function in the VEGAN (Oksanen et al., 2013) software package.
Principal component analysis (PCA) was performed using the R
function prcomp. To focus on the ASVs that best explained the
difference between these samples, the 50 ASVs with the largest
absolute loadings in the first and second components were
selected. A heatmap representation of the relative abundance of
these ASVs was constructed using PHEATMAP (Kolde, 2019).
Sequence data were deposited in the NCBI Sequence Read
Archive under accession no. PRJNA498673.
Root colonization analysis To independently confirm AMF
colonisation, cleared and stained roots were imaged via microscopy
for signs of AMF colonisation based on the visible presence of
intraradical hyphae, vesicles, arbuscules and external hyphae.
Plant biomass Fresh roots and shoots from the plant compart-
ments were weighed and dried at 70°C for 48 h. Dry samples
were digested using a H2SO4/H2O2 digestion method
(Novozamsky et al., 1983) and prepared for high resolution
inductively coupled plasma mass spectrometry (HR-ICP-MS)
analysis to determine the total P content.
Chemical mapping of phosphorus and sulphur The SXRCT-
imaged hyphal compartment replicates were freeze-dried (to pre-
serve organic residues, e.g. hyphae and bacteria) and fixed in
epoxy resin (Epotek-301 (Epoxy Technology Europe Ltd, Marl-
borough, UK) diluted with ethanol; Methods S1: Fig. S1.32
therein). The samples were prepared using standard geological
thin-sectioning procedures (Camuti & McGuire, 1999;
Lanzirotti et al., 2010).
At the DLS I18 beamline hyphal compartment, longitudinal
thin-sections of hyphal compartments were mounted on a three-
axis-stage. The beamline comprises Kirkpatrick–Baez mirrors
producing a spot size of 20 lm, utilising a silicon (Si(111))
monochromator to scan the incident beam energy. Chemical
maps were acquired under a helium atmosphere at 3 and 2.7 keV
photon energies (to mitigate artifacts from chlorine present in the
epoxy resin). For each pixel, a full energy-dispersive spectrum
was recorded using a four-element Vortex silicon drift detector
(Hitachi) positioned normal to the incident beam and 45° rela-
tive to the sample. The flux was estimated to be between 1010
and 1011 photons s1. The same setup was used at SLS, except
that elemental maps were obtained under vacuum (10–6 mbar)
and X-rays were detected using a single element Ketek silicon
drift detector. All maps were fitted using the PYMCA package in
batch fitting mode (Sole et al., 2007). Concentrations (in µg g1)
of Al, Si, P and S were obtained by fitting maps obtained under
identical experimental parameters using a spessartine garnet and
Durango apatite mineral standards with known elemental con-
centrations (Methods S1: Tables S1.2, S1.3 therein).
X-ray absorption near edge structure spectra were collected in
fluorescence mode to constrain changes in P (K-edge) and S
(K-edge) speciation. The K-edge positions of P and S were cali-
brated against the first derivative of plots obtained using apatite
(Ca5(PO4)3(OH)) and sulphate (ZnSO4) standards. Spectra
from a series of P and S standards (Fig. 1) were collected to per-
form a linear combination fit.
Phosphorus concentration with distance from the root com-
partment was estimated from the XRF P intensity maps. Average
P concentrations and SD were calculated for the mixed-phase
pixels. Since both P and S are in the tender X-ray region, and the
attenuation length of P in quartz is c. 2 µm, these maps are
assumed to be 2D representations of concentration. Thus, the P,
S and Al XRF maps were aligned via visual inspection to the cor-
responding SXRCT data of the same physical samples to corre-
late the chemical and structural information. The correlated data
were used to determine if the presence of hyphae was correlated
Fig. 1 Sulphur (S) and phosphorus (P) K-edge (X-ray absorption near edge
structure) XANES spectra. The top set of spectra are from standards used for
linear combination fits; the bottom set are point measurements as a function
of distance from the segmented hyphae. The selection of collected spectra
from the samples show spectral differences. For S, the ratios between
different peaks are particularly important. The vertical dashed lines represent
the energies of L-cysteine (labelled ‘a’), sulfonate (‘b’), and sulfate (‘c’). Some
spectra show contrasts, for example IP– T4 (inoculated, no supplemental P, t
= 4 wk) at the end of the compartment closest to roots (predominantly
cysteine) and IP+ T2 (inoculated, added phosphate, t = 2 wk) at the fertilizer
pellet location (predominantly sulfate). For P, the spectral differences are
much more subtle. However, the differences are measurable. In particular,
calcium orthophosphate (CaPO4)-like species have a shifted white line (the
sharp feature in the K-edge spectrum) towards lower energy (labelled ‘d’), a
shoulder (‘e’) and oscillation (‘f’) at higher energies. Interestingly, the
shoulder for CaPO4-type species disappears with increasing distance from
the fertilizer pellet, suggesting dissolution and re-adsorption to different
mineral surfaces. Spectral changes with increasing distance from fungal
hyphae do not show significant differences. CP+, control added P; CP–,
control no added P; IP+, inoculated added P; IP–, inoculated no added P.
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist4
with local variation in elemental concentrations of P, S and Al. A
3D Euclidean distance transform of the SXRCT-derived seg-
mented hyphal geometries for each hyphal compartment was
used to determine distance from hyphae to each position on the
XRF maps. Pixels on the XRF maps were separated into two dis-
tance classes: pixels within 50 µm of the nearest hyphal surface
(‘close to hyphae’) and further than 200 µm (‘far from hyphae’).
Pixels on the XRF maps were labelled as mixed phase if they were
determined to be neither air (low total XRF signal) nor primary
mineral (high Si signal). Excluding pixels that indicated high Si
also limited the overflow of Si signal into the neighbouring P or
Al signal. Mean XRF counts of P, S and Al in the mixed phase
were measured for each distance class of pixels over all treat-
ments.
Mathematical modelling
The AMF hyphae length density data at T2 and T4 (2 and 4 wk
of plant growth) were fitted with a model developed by Schnepf
et al. (2008a) which included the simplest linear net hyphal
branching term (see Methods S2 for a full mathematical mod-
elling description). The model fitted was @tn + v@xn = bn,
@tq = n|v| dq, where n is the hyphal tip density, v is the hyphal
tip growth rate, b is the net hyphal branching rate, q is the hyphal
length density, d is the net hyphal length destruction rate, t is the
time and x is the distance from the root compartment along the
hyphal compartment midline. The model was solved with a zero
initial condition (i.e. n = 0 and q = 0 at t = 0), assuming constant
k tip production at the root surface (i.e. v = k on x = 0 for t > 0).
We estimated values using the MATLAB R2017a fmincon routine
for k, d, b and v by fitting the model to the length–density data
from both SXRCT and destructive (Jakobsen et al., 1992b) mea-
surements. This gave us the lower and upper bound for the
parameter values, since SXRCT only detects hyphae in the pore
spaces and the method described by Jakobsen et al. (1992b)
detects all hyphae.
X-ray fluorescence data for P was fitted with the model of Sch-
nepf & Roose (2006), which links the fitted hyphal length den-
sity results for q to the soil P profiles. The model consists of a soil
P movement equation that accounts for first-order binding of P
to the soil mineral surfaces and P diffusion in the soil pore-space,
that is, @tcTOT =Deff@xxcTOT 2prmkhqcTOT, where cTOT = cs +
hcl = (b + h)cl is the total amount of P in the soil, cl is the concen-
tration of P in the soil fluid/mixed phase, h is the volume fraction
of the mixed phase, b is the first-order equilibrium binding buffer
power of P in soil, kh characterizes the rate of P uptake per unit
of hyphal surface area per unit volume of total soil P, and rm is
the hyphal radius. The model was solved, setting the boundary
condition at the root/AMF hyphae compartment boundary as
Deff@xcTOT = FmaxcTOT on x = 0 and cl? c∞ as x?∞, where
c∞ is the P farfield concentration assumed to be the initial condi-
tion of P at t = 0. We used the value for Deff = (Dhf /
(h + b)) = 1.059 108 cm2 s1, as estimated by McKay Fletcher
et al. (2017) and used the fitted results of our hyphal length den-
sity measurements q from the growth model above. We esti-
mated the parameter kh against the XRF data by minimizing the
sum of squares between the data points and the model using
fmincon in MATLAB R2017a.
Statistics
Throughout this study, we used the MATLAB Statistics Toolbox to
analyse data. We used a t-test at a significance of P < 0.05, using
ttest2, for means, and the Kolmogorov–Smirnov (KS) test at
P < 0.05, using kstest2, for the distributions.
Results and Discussion
Imaging of hyphal networks
Observations from SXRCT data revealed significant differences
in hyphal morphology between treatments (Fig. 2a,b). Hyphae in
the control samples are more heterogeneous, have many more
branches and have spherical features at some termini, which are
assumed to be spores or spore-like structures. Rhizophagus irregu-
laris inoculated samples (Fig. 2b) are characterized by more lin-
ear/straight hyphae, less branching and fewer spherical/spore-like
structures.
Synchrotron X-ray computed tomography imaging is noninva-
sive and provides richer information than destructive bulk
length-density measurements, enabling estimation of the number
of branches per hyphal cluster, branch length, angle and tortuos-
ity. However, as hyphae were only detectable in the pore space
due to the contrast limitations of SXRCT, these measurements
might not be fully representative of the entire soil hyphal popula-
tion. Control P+ measurements showed significantly higher (t-
test) and differently distributed (KS-test) branch numbers per
cluster (when defining a single cluster to be all hyphae that are
continuously connected on the images) than control P and
both P+ and P inoculated treatments (Fig. 2), suggesting that
in the soil used, nonAMF hyphae are more branched in the soil
pore-space than AMF hyphae. The differences in branch number
and distribution between inoculated P+ and P treatments were
not found to be statistically significant (t-test and KS-test) poten-
tially indicating, consistent with previous work (Drew et al.,
2003), that AMF hyphal morphology is not significantly depen-
dent on P availability. Whilst we cannot fully quantify the
nonAMF contribution to the length density, it is logical to con-
clude that any correction of the length density downwards would
correct the hyphal uptake rate upwards. Further integrated stud-
ies which go hand in hand with more detailed genomic sequenc-
ing would be able to shed light on this.
Synchrotron X-ray computed tomography data enabled quan-
tification of the alignment of hyphae with respect to the major
axis of each hyphal compartment. On average, this hyphal align-
ment angle varied between 45° and 85°. The only significant dif-
ference was that the control P+ treatment had a lower mean
angle than the CP and inoculated IP+ and IP treatments
(Methods S1: Fig. S1.28 therein, t-test). Further analysis indi-
cated that at the fertilizer pellet location (h2), statistically signifi-
cant differences were recorded in the alignment of hyphae
between CP+ (49°) and both P treatments (CP– is 65° and

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 5
IP– is 67°), and when pooling the data at the same location,
hyphae in P+ treatments were found to be statistically more
aligned to the compartment axis (51°) than in P treatments
(66°). One way of explaining these differences is in the context of
the cost–benefit of the symbiosis: when the supply of P is com-
paratively more scarce, the hyphae ‘search’ for P by deviating
more from the primary growth direction, as defined by the geom-
etry of the growth assay. The SXRCT data also allowed analysis
of the (normalised) tortuosity of hyphal branches. Means of
tortuosity in the control P+ and P treatments were not signifi-
cantly different (t-test), but the distributions were significantly
different (KS-test). This could be explained by the larger
variation seen in control P+ (Fig. 2c). The opposite was found
for inoculated treatments, where P+ displayed a significantly
different/wider (normalised) hyphal tortuosity distribution (KS-
test) than P. However, when comparing the inoculated and
control P+ treatments, the inoculated P+ treatment had a signifi-
cantly different/wider tortuosity distribution than the control P+
treatment, supporting the hypothesis that AMF fungi might be
searching out P sources more aggressively than the nonmycor-
rhizal strains present in the control samples.
In the control samples with added P (Fig. 2c, CP+), the overall
mean hyphal length density was higher (t-test) and the distribu-
tion was different (KS-test) compared to inoculated samples with
plants with and without P addition (P+/) and control samples
with no P (P). Since the soil was nonsterile, this suggests that
AMF inoculation had a suppressive/allelopathic effect on the
development of other soil fungi. Adding localised P also resulted
(a)
(c) (d) (e)
(b)
Fig. 2 Synchrotron X-ray computed tomography imaging and image processing protocols enabled extraction of air-based hyphal morphology. The images
in (a) and (b) show sample images of 3D hyphal morphology for exemplary samples from control and inoculated treatments; (c) shows the mean (red line)
hyphal length density, the blue box describes the interquartile range (i.e. 25–75% of data falls within this interval), red dots indicate outliers, and black
whiskers show 2.7r, where r is the SD; (d) shows the average hyphal length density for all different treatments at t = 2wk, along with the SD; (e) shows
the average hyphal length density for all different treatments at t = 4 wk along with the SD. In (d) and (e) h1 = 13.5mm, h2 = 28.5mm, and h3 = 43.5mm
from the interface of the hyphal compartment to the adjoining root compartment along the hyphal compartment midline. CP+, control added P; CP–,
control no added P; IP+, inoculated added P; IP–, inoculated no added P.
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist6
in higher overall hyphal length in both control (CP+) and inocu-
lated (IP+) samples. Differences in the observed hyphal length
densities in time as a function of distance (h1, h2, h3) from the
mesh interface with the plant compartment can be seen in
Fig. 2(d,e). Hyphal length density in this context means the total
length of observed hyphae – of any length – per unit volume of
soil. Based on these measurements we found that the initial
growth of native soil fungi in the P+ control treatment (Fig. 2d,
CP+) was reduced such that the hyphal length density in control
(CP+) and inoculated treatments (IP) was roughly the same by
week 4. The inoculated P+ treatment (IP+) had higher hyphal
length density at all three positions at the 2-wk stage, but by week
4, the inoculated P treatment (IP) had a higher hyphal length
density than IP+ closest to the root compartment (h1) (see Meth-
ods S1: Section 5.4; Fig. S1.28 therein for supplementary analysis
of hyphal branch distance, branch alignment and tortuosity as
detected by SXRCT).
The plot shown in Fig. 2c indicates that the P+ treatments
in both control and inoculated treatments led to higher hyphal
length density and a different distribution compared to P
treatments (t-test and KS-test) when summing over all positions
in the hyphal compartments; this is consistent with the findings
of previous studies (Abbott et al., 1984; Olsson et al., 1997).
Mean hyphal length density (t-test) and distribution (KS-test)
for control P+ was significantly higher/different compared to
the inoculated P+ treatment (Fig. 2c). Similarly, the hyphal
length density closest to the root compartment (h1, Methods
S1: Fig. S1.1 therein) for control samples had a significantly
higher mean and a different distribution than inoculated sam-
ples at h2 and h3, but the differences were not found when
comparing at the closest (h1) distance from the root compart-
ment (Methods S1: Fig. S1.1 therein, t-test and KS-test). How-
ever, the inoculated samples indicated AMF colonisation,
whereas hyphae in the control samples were likely from sapro-
trophic fungal strains. Similar analysis of the results for the
Jakobsen measurements revealed an equivalent suppressive/al-
lelopathic effect of inoculum/plant presence (Methods S1:
Fig. S1.30 therein). Whilst there are many studies that describe
plant root architectural responses to P conditions/hetero-
geneities in soil, there are few studies that show such results for
AMF or other soil fungi (Olsson et al., 1997; Cavagnaro et al.,
2005).
Destructively-derived hyphal length densities were significantly
higher than those determined via SXRCT (Methods S1:
Fig. S1.30 therein). This difference can be explained by the
majority of hyphae residing in the soil mixed-phase and/or at the
soil-particle/air interfaces, where they cannot be detected by
SXRCT. We conclude that 5% of control and 2% of inoculated
sample hyphae were present in the soil pore-space and 95% of
control and 98% of inoculated hyphae were in the inter-soil
aggregate/mixed space. This is to be expected as the soil mixed-
phase has higher P and S bioavailable concentrations and greater
water availability. An alternative explanation is that the addition
of P as a fertilizer, rather than a substrate with higher naturally-
occurring P, might affect the extent to which hyphae penetrate
into the soil mixed phase.
Polymerase chain reaction, sequencing and other
traditional validation measurements
Polymerase chain reaction analysis indicated that one out of three
inoculated samples (P+/) tested at the 2-wk stage were positive
with the AMF-specific Kruger and Tkacz primers. At the 4-wk
stage, all inoculated samples were positive to Kruger primers and
2 out of 3 inoculated samples were positive to Tkacz primers.
None of the inoculated samples were positive to broad primers
(Buee et al., 2009a,b). None of the control samples at weeks 2 or
4 were positive to Kruger, Tkacz, or broad primers, indicating a
lack of AMF colonisation as expected.
Amplicon sequencing of eukaryote-specific 18S rRNA genes
verified the presence of AMF in the inoculated samples. One
sample (IP+, 4 wk) failed to sequence properly and was excluded.
Two ASVs affiliated to AMF were detected in four of the inocu-
lated samples with combined relative abundances of 0.34–1.34%,
whereas only one was detected in an uninoculated control sample
at a relative abundance of 0.02%; see Methods S1. The reverse
PCR primer for amplification had a single mismatch to the gene
in Rhizophagus species in the penultimate (30) position, which
may explain the low detection rate.
The total combined relative abundance of sequences assigned
to fungi detected in the samples (i.e. Ascomycota, Basidiomycota,
and Mucoromycota) was higher in the uninoculated (32.3%)
than inoculated treatments (24.7%). Although the sequencing
only shows relative and not absolute abundances, when inte-
grated with SXRCT and destructive imaging results, this suggests
the growth of native fungi was suppressed by AMF inoculation.
The profiles between the inoculated and uninoculated treatments
were distinct; see PCA analysis in Fig. 3. These differences were
spread across diverse phylogenetic groups of eukaryotes, includ-
ing nematodes and protozoa. The relative abundance of 18S
rRNA genes in the original soil clustered together with the pro-
files from the hyphal compartments of the inoculated micro-
cosms, implying that the eukaryotic community in the
uninoculated hyphal compartments underwent a greater degree
of change. On the basis of the genomic analyses, we conclude that
the imaged inoculated samples contained AMF hyphae and con-
trol samples did not. Moreover, root staining showed clear myc-
orrhizal colonization (Fig. 4a) in all inoculated treatments.
Elemental mapping
Synchrotron X-ray computed tomography images were correlated
with elemental maps of P, S and Al for the same samples using
XRF allowing hyphae and soil morphology to be correlated with
soil chemistry at the pore scale. Previous work has suggested
hyphal length density and soil P content should, theoretically,
show strong negative correlation (Joner et al., 1995; Schnepf
et al., 2005, 2008a,b, 2011; Schnepf & Roose, 2006; Thonar
et al., 2011). This study did not find this strong correlation.
However, previous studies carried out no chemical mapping or
hyphal P-uptake calibration; hence, the hyphal P uptake parame-
ters were highly uncertain. In this study, simultaneous measure-
ment of hyphal length density and P gradients allowed more

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 7
accurate estimation of hyphal P uptake rates. Further analysis
indicated that organic S is the main S fraction in soil, making up
95–98% of total soil S, due to it being a major component of soil
microorganisms, plants and animals (Scherer, 2009). Since the
majority of observed soil S can be assumed to be associated with
soil microorganisms due to exclusion of animals and roots from
hyphal compartments, we treated S as a ‘surrogate marker’ for
microbe-derived soil organic matter. The sequencing analysis
enables attribution of this S signal to a wide diversity of prokary-
otes (> 2000 ASVs) and microeukaryotes (> 3000 ASVs).
The effect of organic matter on hyphal densities was also
observed, and demonstrated how P and S counts were correlated
against the hyphal distance maps (Fig. 5a–c). These images show
a clear positive correlation between soil organic material and fun-
gal hyphae, in that hyphae are clearly oriented towards a structure
which is high in S and relatively low in P. Fig. 5(e) shows rela-
tively uniform S counts across the mixed phase, with the excep-
tion of an S-rich fragment of organic matter which is apparently
low in P. An SXRCT-derived map of 3D distance from hyphal
surfaces is shown in Fig. 5(f), indicating high hyphal presence in
the S-rich regions. Hyphal correlation with organic material
(Hodge, 2014) was also observed in other samples (Methods S1:
Fig. S1.26 therein). It must be noted that since SXRCT does not
exclusively detect AMF hyphae, there is a possibility that some of
the structures observed may be saprophytic in origin.
Distributions of P, S and Al XRF counts for pixels in the
mixed-phase are shown in Fig. 6(a–c), both close to hyphae
(within 50 µm of SXRCT measured hyphae) and far from
hyphae (further than 200 µm from SXRCT measured hyphea)
over all inoculated treatments. The mean P XRF concentration
close to hyphae (1.039 104 XRF-counts) was significantly higher
(P < 0.05) than far from hyphae (8.39 103 XRF-counts).
Fig. 6(a) shows the shapes of the two distributions are similar,
but close to the hyphae, the distribution is shifted to the right. A
possible explanation is preferential hyphal growth close to areas
of higher P concentration. Similar calculations for S showed no
significant difference between the means (t-test) or distributions
(KS-test) of the two S distributions (Fig. 6b). The corresponding
aluminium (Al) distributions are shown in Fig. 6(c), with the Al
mean concentration far from the hyphae being significantly
higher than that close to the hyphae (t-test).
The Al concentration close to the hyphae is generally lower
and has a narrow distribution of concentration compared to the
Al distribution far away from the hyphae (Fig. 6c). This could
suggest that AMF hyphae are less likely to grow in proximity to
Al-rich clay minerals, which is consistent with previous findings
(Vos
atka et al., 1999; G€oransson et al., 2008; Zhang et al., 2015).
This leads to the following question: do AMF hyphae seek P
from Al-rich clay minerals if they contain high P? To test this
hypothesis, XRF pixels were classified into two elemental classes:
‘high P/low Al’, and ‘high P/high Al’, and mean distance from
hyphae for these pixels was calculated. ‘High P’ pixels were
defined as having a higher P concentration than the mean con-
centration of pixels within 50 µm of hyphae (i.e. 1.039 104).
‘Low Al’ pixels were defined as having a lower Al concentration
than the mean concentration of pixels within 50 µm of hyphae
(i.e. 3.239 103). The same computation was performed for Al
(i.e. 89 103; Fig. 6c). The mean distance from hyphae for the
‘high P/low Al’ pixels was 884 µm, which was significantly lower
(t-test; P < 0.05) than the ‘high P/high Al’ mean of 1100 µm.
The distribution of distance from hyphae for the two elemental
classes is shown in Fig. 6(d); ‘High P/low Al’ has many more pix-
els that are close to hyphae (within 500 µm). Interestingly, only
0.05% of ‘high P/high Al’ pixels are within 100 µm of hyphae,
compared to 1.52% of ‘high P/low Al’ pixels. These results sug-
gest that AMF might avoid Al-rich clay minerals as a source of P.
Fig. 3 Principal component analysis (PCA) of
microbial sequencing results.
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist8
Such a result, if confirmed to be AMF-specific, could have signifi-
cant and detrimental implications for the ecological effectiveness
of the AMF–plant symbiosis in clay soils. Previous research has
also highlighted the importance of soil texture in the role of
AMF in N cycling (Zhai et al., 2021) and the establishment of
different AMF species (Mathimaran et al., 2005; Verbruggen
et al., 2012).
Results for ‘high S/high Al’ and ‘high S/low Al’ are shown in
Fig. 6(e), indicating that hyphae are closer to regions of high-S
and low-Al. This is consistent with the assumption that a high-S
signal could be associated with microbial presence in the soil and
that much of the microbial population chooses other sources to
mine, for example Fe-rich minerals (redox active) and organic
material. There were no significant differences between the mean
(a) (b)
(c)
Fig. 4 (a) Trypan-stained mycorrhizal structures (Rhizosphagus irregularis) in wheat roots from root compartments. Staining confirms that plants from all
inoculated samples were successfully colonised by mycorrhizal fungi. Red arrows point to mycorrhizas. (b) Microscope image of mycorrhizal structures
from a root compartment showing hyphae emerging from a root into the soil at the interface with a hyphal compartment. The red arrow points to
mycorrhizas. (c) Root and shoot dry weight and phosphorus (P) content measurements. The significant differences (t-test, P < 0.05) are detected for: P+
(added phosphate) vs P (no supplemental P) root and shoot weights at t = 2wk, shoot and root P content at t = 2 wk and t = 4 wk for P+ vs P, and
t = 4wk P+ vs P, and dry root weight.

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 9
concentrations of P, S or Al for pixels (t-test) close to and far
from the hyphae measured in the control samples.
When interpreting these results, it should be recalled that
SXRCT only detects hyphae in the soil pore-space, and as set out
above, the destructive Jakobsen measurement suggests the major-
ity of the hyphae are in the clay–water mixture phase (see Meth-
ods S1: Fig. S1.30 therein). Hence, mixed-phase pixels classified
as far from hyphae may in fact be near to hyphae that were unde-
tected by SXRCT. Nevertheless, it is likely that regions of mixed-
phase surrounded by hyphae in the pore-space also contain a
large number of hyphae (i.e. hyphal presence in pores is a proxy
for hyphal presence in proximate mixed phase regions and soil/
air interfaces). Pixel distance from hyphae is disregarded in
Fig. 6(f), and the S XRF concentration distribution between the
‘high P/high Al’ pixels and ‘high P/low Al’ pixels is shown,
demonstrating that the ‘high P/low Al’ pixels have a significantly
higher mean S concentration and a larger tail. Under the
assumption that S is a surrogate marker for microbial presence, a
similar conclusion as to Fig. 6(e) could be made: the microbial
community avoids Al and P rich clay minerals. Local biotoxicity
and low bioavailability are both possible explanations for this
observed behaviour.
Speciation of phosphorus and sulphur via X-ray absorption
near edge structure
The collected XANES spectra of standard samples and a selection
of spectra of the hyphal compartment samples are presented in
Fig. 1: (with various features labelled ‘a’–‘f’). For S, thiol peaks
(c. 2473 eV) and sulfonate (c. 2480.5 eV) and sulphate
(c. 2482 eV) ratios are clearly observable, which suggest differ-
ences in organic (cysteine, glutathione, sulfoxide, and sulfone)
and inorganic S (sulfate cysteic, and sulfone) pools. For P, these
spectral changes are subtle, with the largest changes observed
(a) (b) (c)
(d) (e) (f)
Fig. 5 Phosphorus (P), sulphur (S) and distance from hyphae in a 2-wk time point (T2) inoculation and added phosphate (IP+) treatment shown in their 3D
environment. The green tubes are the hyphae segmented from the synchrotron X-ray computed tomography (SXRCT) data and the brown (red-greyish
opaque) shows organic structure segmented from the SXRCT data. (a) The P heat map correlated to SXRCT. (b) Sulphur heat map correlated to SXRCT.
(c) The distance transform of the segmented hyphae correlated to the P and S XRF maps and SXRCT data. The images in (d–f) show the P, S and distance
transform from hyphae maps used in the correlation with SXRCT data. (d) The P heat map shows P concentration. (e) The S heat map shows S
concentration. (f) The distance from hyphae heat map shows the distance (µm) from the nearest segmented hyphae in 3D. The images show a clear
preferential relationship between organic material in the soil and fungal hyphae. The hyphae are pointing towards this structure, which is high in S and
relatively low in P. Concentration in (a, b, d, e) is expressed in ppm (mg kg1).
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist10
around the fertilizer pellet in the P+ samples – which showed
signs of dissolution – said changes being indicated by a shift in
the white line and disappearance of the shoulder after the white
line (features labelled ‘d’–‘f’ in Fig. 1), suggesting adsorption of P
to soil mineral surfaces. Spectral changes with increasing distance
from fungal hyphae did not show significant differences.
Linear combination fits of the P and S XANES spectra (Fig. 1)
were performed in order to constrain the compositions. The
results suggest that there is an increased organic-associated pool
of P and oxidized S species close to the interface with the root
compartment (Figs 7c, 8c,d). These results are in accordance with
those of a previous study by our group (van Veelen et al., 2020).
These changes in speciation appear to be the result of an
increased microbial community and related biochemical activity
in either the rhizosphere or hyphosphere. The increased micro-
bial biomass resulting from C input by roots (i.e. rhizodeposits)
or mycorrhizal fungi is known to play an important role in the S
cycle (Leustek & Saito, 1999). The rate at which mineralisation
of organo-sulphur in soils takes place is dependent on both the
proportions of sulphate-esters and C-bonded S, and by the type
of crop (Bertin et al., 2003). Previous studies by Vong (2003)
found that addition of glucose or other C compounds, such as
organic acids, to soil encourages rapid bacterial growth which
transforms inorganic S into organic S. Similarly, significant
organic S increase has been observed when cellulose was added as
a C source (Chowdhury et al., 2000). An increase in microbial
biomass and organic matter will also increase the concentration
of organic-associated P. There is also evidence that soil microbes
actively compete with plants for available sulphate in the rhizo-
sphere (Kertesz & Mirleau, 2004).
The presence of AMF caused an opposite trend in S speciation,
with increased oxidized sulphur closer to hyphae (Fig. 8a,b). These
changes were not observed in the controls (Methods S1: Fig. S1.37
therein), which indicates that the observed changes are microbiome
related. The S composition of the organic structure presented in
Fig. 8 was determined to comprise amino acid thiols (64.1%), sul-
fonate (12%) and sulphate (23.9%). This suggests that the hyphae,
together with bacteria, preferentially mine organic S (amino acids
and sulphate esters). In addition, this suggests that AMF hyphae
are actively involved in the conversion and transport of S to the
plant roots (Allen & Shachar-Hill, 2009). This is another indica-
tion that S could be used as a surrogate marker for hyphal abun-
dance, providing a proxy for protein, which is abundant within
hyphae (Rak et al., 2014). Finally, P speciation in both controls
(a) (b) (c)
(d) (e) (f)
Fig. 6 Histograms of X-ray fluorescence (XRF) phosphorus (P), sulphur (S) and aluminium (Al) concentrations in the clay–water mixture phase over all
inoculated samples and planting times, separated by distance from hyphae. (a–c) Phosphorus, S and Al distributions, respectively, of XRF pixels within
50 µm of a hypha (‘Close to hyphae’) and further than 200 µm from a hypha (‘Far from hyphae’). The distance class means, shown in the legend, are
significantly different for only P and Al. (d) Histogram of distances from hyphae of XRF pixels with both high P (> 9.7 g kg1)/low Al (25 g kg1),
mean = 884 µm, and high P/high Al (> 90 g kg1), mean = 1100 µm. The means are significantly different (P 0.01). (e) Histogram of distances from
hyphae of XRF pixels with both high S (> 780mg kg1)/low Al, mean = 840 µm, and high S/high Al, mean = 1055 µm. The mean values are significantly
different (P 0.01). (f) Histogram of S XRF counts with both high P/low Al, mean 980mg kg1, and high P/high Al, mean 595mg kg1. The mean values
are significantly different (P 0.01). Distances were calculated using the correlated synchrotron X-ray computed tomography (SXRCT) images. The bin
width was selected as 1000 raw XRF units for (a, b, c) and 100 µm for (d, e). The bin width for (f) was 100 raw XRF units for high P/low Al and 50 raw XRF
units for high P/high Al. The number of pixels used to create the histograms in (a, b, c) was 1207 for those within 50 µm and 603 410 for those more
distant than 200 µm. Each histogram was normalized by the total counts.

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 11
(Methods S1: Fig. S1.36 therein) and inoculated treatments
(Methods S1: Fig. S1.34 therein) did not reveal clear trends with
increasing distance from hyphae. Such a lack of clear changes in P
speciation around hyphae might have been caused by their prefer-
ential mining of P from organic material which has a low overall P
concentration. In addition, the competition and mutualism
between microbes and AMF also influence the speciation of P in
the soil microbiome. Finally, the different dynamics (i.e. competi-
tion between organisms, and mineral mining), as shown in Fig. 6,
will likely affect P speciation.
The addition of a fertilizer pellet had an overall effect on the spe-
ciation of both P and S in the surrounding soil via dissolution and
diffusion (Fig. 7). However, these effects were limited to the area
directly adjacent to the pellet due to the absence of a significant
mass flow of water and low P diffusion in soil. Consequently, it did
not influence P speciation closer to the root compartment.
Modelling results
The results of the hyphal growth model fitted to the traditional
destructive and SRCT hyphal measurements (Methods S2) are
shown in Table 1, together with a previous estimation of the
same parameters using a different dataset by Schnepf et al.
(2008a). We observed that the hyphal elongation rate
v 105 cm s1 estimated using the Jakobsen and SXRCT data
are similar, as the pattern of soil colonisation is similar, but both
are one order of magnitude faster than was estimated by Schnepf
et al. (2008a). The root hyphal infection rate estimates based on
our two methods of hyphal observation for k 104 cm2 s1 are
an order of magnitude lower than estimated by Schnepf et al.
(2008a). These differences between our hyphal parameter esti-
mates and those estimated by Schnepf et al. (2008a) might be
due to the fact that the previous estimates were based on
(a) (b)
(c) (d)
Fig. 7 Linear combination fitted X-ray absorption near edge structure (XANES) results of phosphorus (P) K-edge with respect to distance from the root
compartment, showing (a) apatite-like and soil mineral P fractions, (b) organic associated phosphate (PO4), (c) the collected elemental maps which show
how the distances are related, and (d) the postulated P interactions based on the XANES results. The blue highlighted ellipsoids in (a) and (b) and the
vertical blue bar in (c) denote the area affected by the triple super phosphate fertilizer. The data show no clear trends for apatite-like and mineral-
associated P (a). However, the organic-associated P fraction is increased closer to the roots, as depicted in (d). The units are given in mass fraction form (0–
1) with 1 meaning 100%. The residual P fractions of (a) and (b) add up to 1. Error bars represent SD. CP+, control added P; CP–, control no added P; IP+,
inoculated added P; IP–, inoculated no added P; R1, replica one; R2, replica two; R3, replica 3.
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist12
experimental measurements (Jakobsen et al., 1992a,b) gathered
using a much larger-scale assay, and whilst the fungal inoculum
was the same, the plant used in past studies (Jakobsen et al.,
1992a,b) was not a wheat plant, but clover (Trifolium subterra-
neum). Furthermore, a different soil was used between the two
experiments, and this would be expected to lead to different
AMF colonisation dynamics.
When the SXRCT-based hyphal growth model was used to
estimate the hyphal P uptake rate k, the values ranged between
10–11 and 1.869 10–5 cm s1, with a mean of 3.789 10–6
cm s1 (Fig. 9), which is in agreement with the value 3.269 10–6
cm s1 estimated by Schnepf & Roose (2006). Estimates of k
based on data from destructive hyphal analysis ranged from
4.59 10–14 to 2.199 10–7 cm s1, with a mean of 1.059 10–7
cm s1. These are much lower than both the SXRCT estimate
and that of Schnepf & Roose (2006). The destructive length den-
sity measure is probably an overestimation of the true AMF
hyphal length density, as it contains nonAMF fungi, resulting in
an underestimation of the true hyphal P uptake rate.
Conclusions
In this study we showed for the first time that AMF networks can
be visualised in 3D within soil using SXRCT, albeit as yet only in
the soil pore-spaces. We found that fungal structures and morphol-
ogy respond to local P concentrations (Cavagnaro et al., 2005) (i.e.
(a) (b)
Fig. 8 Linear combination fitted X-ray absorption near edge structure (XANES) results of sulphur (S) K-edge. (a) Reduced (i.e. amino acid thiols) and
oxidized (i.e. sulfonate/sulfate) S species distribution vs distance from segmented hyphae derived from synchrotron X-ray computed tomography (SXRCT)
data. The data show that closer to hyphal surfaces, more amino acid thiols (e.g. cysteine) and fewer oxidized S species (e.g. sulfonate and sulfate) are
detected. (b) Reduced and oxidized S species distributions with respect to distance from the interface of the hyphal compartment with the associated root
compartment. Here the data show an opposite trend: more oxidized S species closer to the roots and less amino acid thiols. Units are in are mass fraction
(0–1) with 1 meaning 100%. Error bars represent SD. IP+, inoculated added P; IP–, inoculated no added P; R1, replica one; R2, replica two.
Table 1 Estimated best-fit model parameters for hyphal length density measured using traditional Jakobsen destructive soil sample testing (first column)
and synchrotron X-ray computed tomography (SXRCT) data (middle column), in comparison to the values estimated in Schnepf et al. (2008) (last column).
Model fitting parameters for
hyphal length densities obtained
by Jakobsen destructive testing method
Model fitting parameters for
hyphal length densities obtained
by SRXCT analysis method
Values from Schnepf
et al. (2008a)
k (cm2 s1) 4.669 10–4 1.7589 10–4 1.029 10–4 1.7679 10–4 2.899 10–3
b (s1) 2.079 10–6 1.3869 10–6 5.459 10–6 1.2739 10–5 5.799 10–9
v (cm s1) 1.649 10–5 9.0479 10–6 2.069 10–5 2.5669 10–5 2.899 10–6
R2 9.899 10–1 7.6809 10–3 2.349 10–1 2.539 10–1 –
Numbers show mean SD. k, root hyphae infection rate; b, net hyphal branching rate; v, hyphae elongation rate; R2, goodness of fit measure.
Fig. 9 Model-fitted hyphal phosphorus (P) uptake rate k for the inoculated
samples with (IP+) and without (IP) fertiliser treatments using both the
synchrotron X-ray computed tomography (SRXCT) and destructively
measured hyphal parameters. Error bars represent the SD.

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 13
P+ treatments had many more hyphal structures than P treat-
ments). Whilst it has previously been shown that root architectures
respond to soil conditions in this manner (Shen et al., 2011), this
is possibly one of the first studies to visualise in situ that AMF
hyphae respond similarly to differences in soil P conditions.
The modelling results unequivocally show that hyphal P
uptake rates are an order of magnitude lower than previously
thought (Schnepf & Roose, 2006; Schnepf et al., 2008a,b,
2011) and an increase in hyphal length density does not lead to
a significant decrease in soil P content. Correlation of SXRCT
and XRF/XANES data suggested colocation of hyphal length
density and P and S species in the clay/water soil phase, and sug-
gested S is a candidate proxy marker for hyphae in this soil
phase. Our results also hint that AMF hyphae might preferen-
tially mine P from organic structures such as organic-matter
fragments, and confirm the hypothesis of microbial–fungal
mutualism. As such, this would suggest that AMF fungi (or
organisms that they support in the hyphosphere) acquire nutri-
ents from more reactive labile and/or organic sources over clay
minerals, even when soil is high in P, supporting our observa-
tion of preferential P acquisition.
Finally, whilst many challenges remain, our study provides
strong motivation for interdisciplinary groups to apply fully-
integrated experimental and modelling approaches to the charac-
terisation of plant–soil–fungi systems. Our study represents a
necessary step forward in the integration of different techniques
to unravel the processes underpinning plant–AMF interaction
across different spatiotemporal scales (Ferlian et al., 2018),
demonstrating that closer integration of theoretical and experi-
mental techniques is important in gaining deeper insights.
Acknowledgements
AvV, SK, RM, CS, DMF, SR, CP, KW and TR were funded by
ERC Consolidator Grant no. 646809-DIMR. NK, LC and TR
were funded by BBSRC SARISA BB/L02620/1, SD was funded
by BBSRC Case Studentship BB/L502625/1, DMF was funded
by an EPSRC PhD studentship, TR was also funded by BBSRC
BB/P004180/1, NERC NE/L00237/1, EPSRC EP/M020355/1,
and AT and PP were funded by BBSRC grants BB/N013387/1
and BB/R017859/1. We would also like to acknowledge the help
of Dr Keith Daily in modelling and Pengrui Cai in sequencing
analysis. Finally, we gratefully acknowledge the support of the
US Department of Energy through the LANL/LDRD Program
and the G. T. Seaborg Institute to AvV. Los Alamos National
Laboratory is operated by Triad National Security, LLC, for the
National Nuclear Security Administration of the US Department
of Energy (contract no. 89233218CNA000001).
Author contributions
TR designed the study and methodology, and led the data analy-
sis and interpretation. SK led the X-ray computed tomography
measurements and image analysis. AvV led the X-ray fluorescence
measurement and image analysis. DMF, NK, CP, CS, KW, SR
and LC worked on and assisted with XCT, XRF, data analysis
and modelling. RM, SD and DMF did model fitting to the data.
MD led the sequencing analysis. IJ and DJ assisted in data analy-
sis and interpretation. GO assisted in molecular biology and data
interpretation. AT and PP conducted the polymerase chain reac-
tion. FM, CB and TH assisted with XRF measurements. All
authors co-wrote the manuscript and Supporting Information.
SK and AvV contributed equally to this work.
ORCID
Dan McKay Fletcher https://orcid.org/0000-0001-6569-
2931
Tiina Roose https://orcid.org/0000-0001-8744-2060
Data availability
Data related to this study are available at 10.5258/SOTON/
D2088.
References
Abbott LK, Robson AD. 1985. Formation of external hyphae in soil by
four species of vesicular-arbuscular mycorrhizal fungi. New Phytologist 99:
245–255.
Abbott LK, Robson AD, De Boer G. 1984. The effect of phosphorus on the
formation of hyphae in soil by the vesicular-arbuscular mycorrhizal fungus,
Glomus fasciculatum. New Phytologist 97: 437–446.
Allen JW, Shachar-Hill Y. 2009. Sulfur transfer through an arbuscular
mycorrhiza. Plant Physiology 149: 549–560.
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic
local alignment search tool. Journal of Molecular Biology 215: 403–
410.
Bartnicki-Garcia S, Bracker CE, Gierz G, Lopez-Franco R, Lu HS. 2000.
Mapping the growth of fungal hyphae: orthogonal cell wall expansion during
tip growth and the role of turgor. Biophysical Journal 79: 2382–2390.
Berruti A, Lumini E, Balestrini R, Bianciotto V. 2015. Arbuscular mycorrhizal
fungi as natural biofertilizers: let’s benefit from past successes. Frontiers in
Microbiology 6: 1559.
Bertin C, Yang X, Weston LA. 2003. The role of root exudates and
allelochemicals in the rhizosphere. Plant and Soil 256: 67–83.
Boswell GP, Davidson FA. 2012. Modelling hyphal networks. Fungal Biology
Reviews 26: 30–38.
Buee M, De Boer W, Martin F, van Overbeek L, Jurkevitch E. 2009a. The
rhizosphere zoo: an overview of plant-associated communities of
microorganisms, including phages, bacteria, archaea, and fungi, and of some of
their structuring factors. Plant and Soil 321: 189–212.
Buee M, Reich M, Murat C, Morin E, Nilsson RH, Uroz S, Martin F. 2009b.
454 Pyrosequencing analyses of forest soils reveal an unexpectedly high fungal
diversity. New Phytologist 184: 449–456.
Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP.
2016. DADA2: high-resolution sample inference from Illumina amplicon data.
Nature Methods 13: 581–583.
Camuti KS, McGuire PT. 1999. Preparation of polished thin sections from
poorly consolidated regolith and sediment materials. Sedimentary Geology 128:
171–178.
Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA,
Turnbaugh PJ, Fierer N, Knight R. 2011. Global patterns of 16S rRNA
diversity at a depth of millions of sequences per sample. Proceedings of the
National Academy of Sciences, USA 108: 4516–4522.
Carminati A, Vetterlein D, Weller U, Vogel HJ, Oswald SE. 2009. When roots
lose contact. Vadose Zone Journal 8: 805–809.
Cavagnaro TR, Smith FA, Smith SE, Jakobsen I. 2005. Functional diversity in
arbuscular mycorrhizas: exploitation of soil patches with different phosphate
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist14
enrichment differs among fungal species. Plant, Cell & Environment 28: 642–
650.
Charters MD, Sait SM, Field KJ. 2020. Aphid herbivory drives asymmetry in
carbon for nutrient exchange between plants and an arbuscular mycorrhizal
fungus. Current Biology 30: 1801–1808.e5.
Chowdhury MAH, Kouno K, Ando T. 2000. Critical sulphur concentration and
sulphur requirement of microbial biomass in a glucose and cellulose-amended
regosol. Biology and Fertility of Soils 32: 310–317.
Daly KR, Keyes SD, Masum S, Roose T. 2015. Image based modeling of
nutrient movement in and around the rhizosphere. Journal of Experimental
Botany 67: 1059–1070.
Doube M, Kłosowski MM, Arganda-Carreras I, Cordelieres FP, Dougherty RP,
Jackson JS, Schmid B, Hutchinson JR, Shefelbine SJ. 2010. BONEJ: free and
extensible bone image analysis in IMAGEJ. Bone 47: 1076–1079.
Drew EA, Murray RS, Smith SE, Jakobsen I. 2003. Beyond the rhizosphere:
growth and function of adbuscular mycorrhizal external hyphae in sands of
varying pore sizes. Plant and Soil 251: 105–114.
Edelstein L. 1982. The propagation of fungal colonies – a model for tissue-
growth. Journal of Theoretical Biology 98: 679–701.
Faber BA, Zasoski RJ, Munns DN, Shackel K. 1991. A method for measuring
hyphal nutrient and water-uptake in mycorrhizal plants. Canadian Journal of
Botany 69: 87–94.
Ferlian O, Biere A, Bonfante P, Buscot F, Eisenhauer N, Fernandez I, Hause B,
Herrmann S, Krajinski-Barth F, Meier IC et al. 2018. Growing research
networks on mycorrhizae for mutual benefit. Trends in Plant Science 23: 975–
984.
Friese CF, Allen MF. 1991. The spread of VA mycorrhizal fungal hyphae
in the soil: inoculum types and external hyphal architecture. Mycologia 83:
409–418.
Gahan J, Schmalenberger A. 2014. The role of bacteria and mycorrhiza in plant
sulfur supply. Frontiers in Plant Science 5: 1–7.
Gene EL, Charles TD, Buso DC, Myron JM, Gary ML, Scott WB, Siccama
TG, Reiners WA, Alewell C. 2002. The biogeochemistry of sulfur at Hubbard
Brook. Biogeochemistry 60: 235–316.
Giovannetti M, Fortuna P, Citernesi AS, Morini S, Nuti MP. 2001. The
occurrence of anastomosis formation and nuclear exchange in intact arbuscular
mycorrhizal networks. New Phytologist 151: 717–724.
Giovannetti M, Sbrana C, Avio L, Citernesi AS, Logi C. 1993. Differential
hyphal morphogenesis in arbuscular mycorrhizal fungi during pre-infection
stages. New Phytologist 125: 587–593.
Giovannetti M, Sbrana C, Avio L, Strani P. 2004. Patterns of below-ground
plant interconnections established by means of arbuscular mycorrhizal
networks. New Phytologist 164: 175–181.
G€oransson P, Olsson PA, Postma J, Falkengren-Grerup U. 2008. Colonisation
by arbuscular mycorrhizal and fine endophytic fungi in four woodland grasses –
variation in relation to pH and aluminium. Soil Biology & Biochemistry 40:
2260–2265.
Gross M. 2017. Where is all the phosphorus? Current Biology 27: R1141–R1155.
Hadziavdic K, Lekang K, Lanzen A, Jonassen I, Thompson EM, Troedsson C.
2014. Characterization of the 18S rRNA gene for designing universal eukaryote
specific primers. PLoS ONE 9: e87624.
Hepper CM. 1984. Inorganic sulfur nutrition of the vesicular arbuscular
mycorrhizal fungus Glomus caledonium. Soil Biology & Biochemistry 16: 669–
671.
Hodge A. 2014. Interactions between adbuscular mycorrhizal fungi and organic
material substrates. Advances in Applied Microbiology 89: 47–99.
Jakobsen I, Abbott LK, Robson AD. 1992a. External hyphae of vasicular-
arbuscular mycorrhizal fungi associated with Trifolium subtarraneum L. 1.
Spread of hyphae and phosphorus inflow into roots. New Phytologist 120: 371–
380.
Jakobsen I, Abbott LK, Robson AD. 1992b. External hyphae of vesicular
arbuscular mycorrhizal fungi associated with Trifolium subterraneum L. 2.
Hyphal transport of P-32 over defined distances. New Phytologist 120: 509–
516.
Joner EJ, Magid J, Gahoonia TS, Jakobsen I. 1995. P depletion and activity of
phosphatases in the rhizosphere of mycorrhizal and non-mycorrhizal cucumber
(Cucumis sativus L.). Soil Biology & Biochemistry 27: 1145–1151.
Kertesz MA, Mirleau P. 2004. The role of soil microbes in plant sulphur
nutrition. Journal of Experimental Botany 55: 1939–1945.
Keyes SD, Boardman RP, Marchant A, Roose T, Sinclair I. 2013. A robust
approach for determination of the macro-porous volume fraction of soils with
X-ray computed tomography and an image processing protocol. European
Journal of Soil Science 64: 298–307.
Keyes SD, Cooper L, Duncan S, Koebernick N, Fletcher DMM, Scotson
CP, van Veelen A, Sinclair I, Roose T. 2017. Measurement of micro-
scale soil deformation around roots using four-dimensional synchrotron
tomography and image correlation. Journal of the Royal Society Interface
14: 20170560.
Koebernick N, Daly KR, Keyes SD, Bengough AG, Brown LK, Cooper LJ,
George TS, Hallett PD, Naveed M, Raffan A et al. 2019. Imaging
microstructure of the barley rhizosphere: particle packing and root hair
influences. New Phytologist 221: 1878–1889.
Koebernick N, Daly KR, Keyes SD, George TS, Brown LK, Raffan A, Cooper
LJ, Naveed M, Bengough AG, Sinclair I et al. 2017. High-resolution
synchrotron imaging shows that root hairs influence rhizosphere soil structure
formation. New Phytologist 216: 124–135.
Kolde R. 2019. pheatmap: Pretty Heatmaps. R package v.1.0.12 [WWW
document] URL https://CRAN.R-project.org/package=pheatmap.
Kruger M, Stockinger H, Kruger C, Schussler A. 2009. DNA-based species level
detection of Glomeromycota: one PCR primer set for all arbuscular
mycorrhizal fungi. New Phytologist 183: 212–223.
Lanzirotti A, Tappero R, Schulze DG. 2010. Chapter 2 – Practical application of
synchrotron-based hard X-ray microprobes in soil sciences. In: Singh B, Gr€afe
M, eds. Developments in soil science, vol. 32. Burlington, MA: Elsevier, 27–72.
Leake J. 2016.MycoRhizaSoil. [WWW document] URL https://www.soilsecurity.
org/presentations-archive-page/#MycoAC2015.
Leustek T, Saito K. 1999. Sulfate transport and assimilation in plants. Plant
Physiology 120: 637.
Lucas SD, Jones DL. 2006. Biodegradation of estrone and 17 beta-estradiol in
grassland soils amended with animal wastes. Soil Biology & Biochemistry 38:
2803–2815.
Mathimaran N, Ruh R, Vullioud P, Frossard E, Jansa J. 2005. Glomus
intraradices dominates arbuscular mycorrhizal communities in a heavy textured
agricultural soil.Mycorrhiza 16: 61–66.
McKay Fletcher DM, Keyes SD, Daly KR, Van Veelen A, Roose T. 2017. A
multi image based approach for plant fertilizer interaction. Rhizosphere 10:
100152.
Novozamsky I, Houba VJG, Vaneck R, Vanvark W. 1983. A novel digestion
technique for multi-element plant analysis. Communications in Soil Science and
Plant Analysis 14: 239–248.
Oksanen J, Blanchet FG, Kindt R, Legendre P, O’Hara RG. 2013. VEGAN:
community ecology package. R package v.2.0–10. [WWW document] URL
http://CRAN.R-project.org/package=vegan.
Olsson PA, Baath E, Jakobsen I. 1997. Phosphorus effects on the mycelium and
storage structures of an arbuscular mycorrhizal fungus as studies in the soil and
roots by analysis of fatty acid signatures. Applied and Environmental
Microbiology 63: 3531–3538.
Paganin D, Mayo SC, Gureyev TE, Miller PR, Wilkins SW. 2002.
Simultaneous phase and amplitude extraction from a single defocused image of
a homogeneous object. Journal of Microscopy 206: 33–40.
Prietzel J, Botzaki A, Tyufekchieva N, Brettholle M, Thieme J, Klysubun W.
2011. Sulfur speciation in soil by S K-edge XANES spectroscopy: comparison
of spectral deconvolution and linear combination fitting. Environmental Science
& Technology 45: 2878–2886.
Prietzel J, Thieme J, Salom
e M, Knicker H. 2007. Sulfur K-edge XANES
spectroscopy reveals differences in sulfur speciation of bulk soils, humic acid,
fulvic acid, and particle size separates. Soil Biology and Biochemistry 39: 877–890.
Pruesse E, Peplies J, Glockner FO. 2012. SINA: accurate high-throughput
multiple sequence alignment of ribosomal RNA genes. Bioinformatics 28:
1823–1829.
Rak M, Salome M, Kaminskyj SG, Gough KM. 2014. X-ray microfluorescence
(muXRF) imaging of Aspergillus nidulans cell wall mutants reveals biochemical
changes due to gene deletions. Analytical and Bioanalytical Chemistry 406:
2809–2816.

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
New Phytologist (2022)
www.newphytologist.com
New
Phytologist Research 15
Rozmos M, Bukovska P, Hrselova H, Kotianova M, Dudas M, Gancarcikova K,
Jansa J. 2021. Organic nitrogen utilisation by an arbuscular mycorrhizal fungus
is mediated by specific soil bacteria and a protist. ISME Journal. doi: 10.1038/
s41396-021-01112-8.
Scherer HW. 2009. Sulfur in soils. Journal of Plant Nutrition and Soil Science
172: 326–335.
Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T,
Preibisch S, Rueden C, Saalfeld S, Schmid B et al. 2012. FIJI: an open-source
platform for biological-image analysis. Nature Methods 9: 676–682.
Schnepf A, Himmelbauer M, Loiskandl W, Roose T. 2005. Modelling the
rhizosphere. Comparative Biochemistry and Physiology – Part A: Molecular &
Integrative Physiology 141: S222.
Schnepf A, Jones D, Roose T. 2011. Modelling nutrient uptake by individual
hyphae of arbuscular mycorrhizal fungi: temporal and spatial scales for an
experimental design. Bulletin of Mathematical Biology 73: 2175–2200.
Schnepf A, Roose T. 2006. Modelling the contribution of arbuscular mycorrhizal
fungi to plant phosphate uptake. New Phytologist 171: 669–682.
Schnepf A, Roose T, Schweiger P. 2008a. Growth model for arbuscular
mycorrhizal fungi. Journal of the Royal Society Interface 5: 773–784.
Schnepf A, Roose T, Schweiger P. 2008b. Impact of growth and uptake patterns
of arbuscular mycorrhizal fungi on plant phosphorus uptake – a modelling
study. Plant and Soil 312: 85–99.
Shen J, Yuan L, Zhang J, Li H, Bai Z, Chen X, Zhang W, Zhang F. 2011.
Phosphorus dynamics: from soil to plant. Plant Physiology 156: 997–1005.
Simard SW, Beiler KJ, BinghamMA, Diselippe JR, Philip LJ, Teste FP. 2012.
Mycorrhizal networks: mechanisms, ecology and modelling. Fungal Biology
Reviews 26: 39–60.
Smith SE, Smith FA, Jakobsen I. 2003. Mycorrhizal fungi can dominate
phosphate supply to plants irrespective of growth responses. Plant Physiology
133: 16–20.
Sole VA, Papillon E, Cotte M, Walter Ph, Susini J. 2007. A multiplatform code
for the analysis of energy-dispersive X-ray fluorescence spectra. Spectrochimica
Acta Part B 62: 63–68.
Thonar C, Schnepf A, Frossard E, Roose T, Jansa J. 2011. Traits related to
differences in function among three arbuscular mycorrhizal fungi. Plant and
Soil 339: 231–245.
Tlalka M, Hensman D, Darrah PR, Watkinson SC, Fricker MD. 2003.
Noncircadian oscillations in amino acid transport have complementary profiles
in assimilatory and foraging hyphae of Phanerochaete velutina. New Phytologist
158: 325–335.
Vaccari DA. 2009. Phosphorus: a looming crisis. Scientific American 300: 54–59.
van Veelen A, Koebernick N, Scotson CS, McKay-Fletcher D, Huthwelker T,
Borca CN, Mosselmans JFW, Roose T. 2020. Root-induced soil deformation
influences Fe, S and P: rhizosphere chemistry investigated using synchrotron
XRF and XANES. New Phytologist 225: 1476–1490.
Verbruggen E, Van Der Heijden MG, Weedon JT, Kowalchuk GA, Roling WF.
2012. Community assembly, species richness and nestedness of arbuscular
mycorrhizal fungi in agricultural soils.Molecular Ecology 21: 2341–2353.
Vong P-C, Dedourge O, Lasserre-Joulin F, Guckert A. 2003. Immobilized-S,
microbial biomass-S and soil arylsulfatase activity in the rhizosphere soil of rape
and barley as affected by labile substrate C and N additions. Soil Biology &
Biochemistry 35: 1651–1661.
Vos
atka M, Batkhuugyin E, Albrechtov
a J. 1999. Response of three arbuscular
mycorrhizal fungi to simulated acid rain and aluminium stress. Biologia
Plantarum 42: 289–296.
Zhai S, Wu Y, Xu C, Chen W, Feng J, Zheng Q, Meng Y, Yang H. 2021.
Symbiotic soil fungi suppress N2O emissions but facilitate nitrogen
remobilization to grains in sandy but not clay soils under organic amendments.
Applied Soil Ecology 167: 104012.
Zhang S, Zhou J, Wang G, Wang X, Liao H. 2015. The role of mycorrhizal
symbiosis in aluminum and phosphorus interactions in relation to aluminum
tolerance in soybean. Applied Microbiology and Biotechnology 99: 10225–10235.
Supporting Information
Additional Supporting Information may be found online in the
Supporting Information section at the end of the article.
Methods S1 Full description of experimental methods.
Methods S2 Full description of mathematical modelling.
Please note: Wiley Blackwell are not responsible for the content
or functionality of any Supporting Information supplied by the
authors. Any queries (other than missing material) should be
directed to the New Phytologist Central Office.
New Phytologist (2022)
www.newphytologist.com

 2022 The Authors
New Phytologist
 2022 New Phytologist Foundation
Research
New
Phytologist16