A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Me allic mic onu ien s a e associa ed wi h
he s uc u e and unc ion o he soil
mic obiome
Zhongmin Dai
1,2,3
,XuGuo
1,2
, Jiahui Lin
1,2
,XiuWang
1,2
,DanHe
1,2
, Rujiong Zeng
1,2
,
Jun Meng
4
,JipengLuo
5
, Manuel Delgado-Baque izo
6
,
Edua do Mo eno-Jiménez
7,8
, Philip C. B ookes
1,2
& Jianming Xu
1,2,3
The ela ionship be ween me allic mic onu ien s and soil mic oo ganisms,
and he eby soil unc ioning, has been li le explo ed. He e, we in es iga e he
ela ionship be ween me allic mic onu ien s (Fe, Mn, Cu, Zn, Mo and Ni) and
he abundance, di e si y and unc ion o soil mic obiomes. In a su ey ac oss
180 si es in China, co e ing a wide ange o soil condi ions he s uc u e and
unc ion o he soil mic obiome a e highly co ela ed wi h me allic mic o-
nu ien s, especially Fe, ollowed by Mn, Cu and Zn. These esul s a e obus o
con olling o soil pH, which is o en epo ed as he mos impo an p e-
dic o o he soil mic obiome. An incuba ion expe imen wi h Fe and Zn
addi ions o fi e di e en soil ypes also shows ha inc eased mic onu ien
concen a ion a ec s mic obial communi y composi ion and unc ional
genes. In addi ion, s uc u al equa ion models indica e ha mic onu ien s
posi i ely con ibu e o he ecosys em p oduc i i y, bo h di ec ly (mic o-
nu ien a ailabili y o plan s) and, o a lesse ex en , indi ec ly ( ia a ec ing
he mic obiome). Ou findings highligh he impo ance o mic onu ien s in
explaining soil mic obiome s uc u e and ecosys em unc ioning.
Mic onu ien s such as i on (Fe), manganese (Mn), coppe (Cu), zinc
(Zn), molybdenum (Mo), and nickel (Ni) a e c i ical egula o s o
mic obial-d i en p ocesses such as pho osyn hesis, espi a ion, bio-
molecule syn hesis, edox homeos asis, and cell g ow h and immune
sys em unc ioning1–3. Al hough hey a e equi ed by o ganisms in a
small amoun , he deficiency o mic onu ien s significan ly limi s
o ganism g ow h and biological p ocesses.Soil pH and mac onu ien s
a e ega ded as he majo p edic o s o he s uc u e and unc ion o
soil mic obiome4,5. S ikingly, e y li le is known on how me allic
mic onu ien s co ela e wi h he soil mic obiome ac oss en i on-
men al g adien s.
Mic onu ien s may help explain soil mic obiomes o h ee main
easons. Fi s , mos cells and enzymes associa ed wi h mic obial
ep oduc ion and egula ed biological p ocesses equi e soil mic o-
nu ien s, e.g. Fe and Mn o mic obial espi a ion, Cu and Zn o
immunocompe ence, and Fe, Mo and Ni o N fixa ion, e c. Second,
Recei ed: 6 Ap il 2023
Accep ed: 4 Decembe 2023
Check o upda es
1
Ins i u e o Soil and Wa e Resou ces and En i onmen al Science, College o En i onmen al and Resou ce Sciences, Zhejiang Uni e si y, 866 Yuhang ang
Road, Hangzhou 310058, China.
2
Zhejiang P o incial Key Labo a o y o Ag icul u al Resou ces and En i onmen , Zhejiang Uni e si y, 866 Yuhang ang Road,
Hangzhou 310058, China.
3
The Ru al De elopmen Academy a Zhejiang Uni e si y, Zhejiang Uni e si y, Hangzhou 310058, China.
4
Zhejiang P o ince Key
Labo a o y o Recycling and Ecological T ea men o Was e Biomass, School o En i onmen al and Na u al Resou ces, Zhejiang Uni e si y o Science and
Technology, Hangzhou 310023, China.
5
Minis y o Educa ion Key Labo a o y o En i onmen al Remedia ion and Ecological Heal h, College o En i onmen al
and Resou ce Sciences, Zhejiang Uni e si y, Hangzhou 310058, China.
6
Labo a o io de Biodi e sidad y Funcionamien o Ecosis émico. Ins i u o de Recu sos
Na u ales y Ag obiología de Se illa (IRNAS), CSIC, A . Reina Me cedes 10, E-41012 Se illa, Spain.
7
Depa men o Ag icul u al and Food Chemis y, Facul y o
Sciences, Uni e sidad Au ónoma de Mad id, 28049 Mad id, Spain.
8
Ins i u e o Ad anced Resea ch in Chemical Sciences, Facul y o Sciences, Uni e sidad
Au ónoma de Mad id, 28049 Mad id, Spain. e-mail: [email protected]
Na u e Communica ions | (2023) 14:8456 1
1234567890():,;
1234567890():,;
many soil C, N, and S edox eac ions a e coupled wi h mic onu ien
elemen s ha a e able o p o ide and accep elec ons (e.g. Fe and
Mn)6–8. Thi d, soils e ol ed om pa en ma e ials o seconda y
mine als ha con ains di e en mic onu ien s p oduce con as ing
habi a s, which would a o some mic oo ganisms bu no o he s9.
The e o e, he mic onu ien s p obably influence all aspec s o
mic obiome such as mic obial abundance, di e si y, ne wo k con-
nec i i y (i.e. po en ial in e ac ion be ween mic oo ganisms) and
associa ed unc ions in ol ed in nu ien cycling. Finally, soil pH is
ega ded as he maind i e o soil mic obial communi ies, bu i is also
ac i icalinfluence o mic onu ien a ailabili y. In his espec ,
mic onu ien s may help o explain he s ill poo ly unde s ood
mechanisms behind he well-desc ibed ela ionship be ween pH and
he s uc u e and unc ion o mic obiomes.
Mic oo ganisms also o m s able mu ualis ic ela ionships wi h
plan oo s, con ibu ing o plan immuni y as well as nu ien up ake10.
Fo example, mic obial abundance, di e si y, and ne wo k complexi y
ha e been epo ed o be posi i ely co ela ed wi h ecosys em
p oduc i i y11–13. Mo eo e , mic oo ganisms pa icipa e in elemen
cycling such as o ganic C decomposi ion, N fixa ion, ni ifica ion and
deni ifica ion, P solubiliza ion, and S sulphidisa ion14–16. These unc-
ions influence he ca bon seques a ion and nu ien a ailabili y o
plan s, also benefi ing ecosys em p oduc ion. Recen s udies epo
ha soil mic onu ien a ailabili y oge he wi h soil o he p ope ies
(e.g. pH, mac onu ien , e c) a ec ecosys em p oduc i i y17,while
whe he mic onu ien s con ibu e o ecosys em p oduc ion also by
a ec ing he soil mic obiome emains o be es ed. Because o his,
iden i ying he ole o mic onu ien s in explaining soil mic obiomes is
c i ical o inc ease ou capaci y o p edic ecosys em esponses o
en i onmen al changes.
To be e unde s and he ela ionship be ween mic onu ien s
and he s uc u e and unc ion o soil mic obiome, and how hese
changes a ec he ecosys em p oduc ion, we conduc ed a la ge
spa ial collec ion o soil samples om 180 si es in China, wi h di e se
ecosys ems including con as ing soil p ope ies, clima ic condi ions
and ege a ion ypes om opical o cold egions (Fig. S1–3). The
mean annual p ecipi a ion and empe a u e ange om 395 o
2486 mm and −2.7 o 27.9 °C, and hese si es a e ep esen a i e o
o he pa s o he globe. The a ailabili y and o al amoun o Fe, Mn,
Cu, Zn, Mo, and Ni, bac e ial and ungal abundance, di e si y, ne -
wo k connec i i y and associa ed C, N, P and S gene abundances we e
e alua ed. We aimed o examine he ela ionship be ween mic o-
nu ien s and mic obial communi y composi ion and unc ioning,
u he he ecosys em p oduc ion, and compa e he e ec s wi h he
well-known p edic o s o soil pH and mac onu ien s. We hypo he-
sized ha mic onu ien s can explain a unique po ion o he a ia-
ion in soil mic obiome s uc u e and unc ion in addi ion o pH and
mac onu ien s. Ou findings highligh ha mic onu ien s a e
associa ed wi h he s uc u e and unc ion o soil mic obiomes, and
also wi h he ecosys em p oduc ion.
Resul s
Mic obial o al abundance
Mic onu ien con en was significan ly co ela ed wi h bac e ial
abundance, bu was no associa ed wi h ungal abundance. The a ail-
able PC1 (wi h co ela ion coe ficien o 0.27) and o al PC2 (0.23) had
posi i e ela ionships wi h bac e ial o al abundance(Fig. S4). Bac e ial
o al abundance was also posi i ely co ela ed by he a ailable Fe
(0.27), Zn (0.28), Cu (0.15) and Ni (0.15), and o al Mn (0.16) (Fig. S4).
Howe e , he ungal o al abundance was no influenced by ei he
a ailable o o al p incipal componen s (Fig. S4) and was only posi-
i ely co ela ed wi h a ailable Zn (0.18) and Mo (0.17), and o al Zn
(0.17) (Fig. S4). Soil mac onu ien s and pH we e posi i ely co ela ed
wi h bac e ial o al abundance (excep soil C:N), while ungal o al
abundance was only influenced by soil pH and o al P (Fig. S5).
Abundance o specific gene a
Mic onu ien s had bo h nega i e and posi i e co ela ed axa a he
genus le el (Figs. S6 and S7). A ailable Fe had he la ges numbe o
co ela ed bac e ial (i.e. 129) and ungal (i.e. 43) gene a compa ed o
he o he mic onu ien s (Fig. S6). Fo example, he lowe abundances
o axa such as Pi 4_lineage and Gibbe ella and he highe abundances
o axa such as SHA-26 and Scolecobasidium we e associa ed wi h
highe a ailable Fe (Fig. 1and Table S1). To al Mn had he la ges
numbe o he sum o posi i e and nega i e co ela ed bac e ial (i.e.
114) and ungal (i.e. 55) gene a compa ed o he o he mic onu ien s
(Fig. S7). Fo example, he lowe abundances o axa such as FCPS473
and Sai ozyma and he highe abundances o axa such as P omic o-
monospo a and Gibbe ella we e associa ed wi h highe o al Mn (Fig. 1
and Table S2). Fu he , he highe abundance o Th ichode ma and
Rub obac e we e associa ed wi h lowe o al Cu and a ailable Zn
(Fig. 1, Tables S1 and S2). The highe abundance o Fusa ium was
associa ed wi h lowe o al Mo and highe a ailable Ni (Fig. 1, Tables S1
and S2).
Mic obial di e si y and co-occu ence ne wo ks
The land use and i s in e ac ion wi h clima e a ec ed he mic obial
communi y (Fig. S8). Excep o hese ac o s, he mic onu ien con-
en was significan ly co ela ed wi h bac e ial and ungal commu-
ni ies, and he co ela ions wi h bac e ia we e s onge han hose wi h
ungi (Fig. 2).The a ailable PC1 had nega i e co ela ions wi h bac e ial
alpha di e si y, be a di e si y, and he abundance o dominan phyla
such as Acidobac e ia,Ac inobac e ia and posi i e co ela ions wi h
he abundance o P o eobac e ia (Fig. 2a). Especially, he a ailable Fe
co ela ed mos s ongly wi h mos o he abo e pa ame e s and
had a significan ela ionship wi h bac e ial ne wo k connec i i y
(bac e ia ha e mo e ela ionships among hemsel es when connec-
ions inc ease) shown by he o al deg ee, a e age deg ee and a e age
pa h leng h. The a ailable Cu, Zn, and Mo nega i ely co ela ed
wi h bac e ial alpha di e si y, be a di e si y, he abundance o
Acidobac e ia and posi i ely wi h he abundance o P o eobac e ia.The
numbe o co ela ed pa ame e s wi h o al mic onu ien con en s
was ewe han wi h hei a ailable o ms. Soil pH, C:P and N:P a ios
significan ly a ec ed he bac e ial di e si y, dominan phyla abun-
dance, and bac e ial ne wo k connec i i y (Fig. S9). In con as , ungal
pa ame e s we e ba ely co ela ed wi h soil mic onu ien s, ega dless
o he soil o al o a ailable pools, excep o ungal ne wo k con-
nec i i y and be a di e si y (Fig. 2b). Soil pH, C:P and N:P a ios
a ec ed ungal be a di e si y and ne wo k connec i i y (Fig. S9).
Mic obial unc ional genes in ol ed in C, N, P, and S cycling
Mic onu ien s, oge he wi h soil pH and mac onu ien s, g ea ly
a ec ed he mic obial gene abundance in C, N, P, and S cycling (Fig. 3
and Table S3). The a ailable PC1 (wi h he numbe o 48 co ela ed
genes), a ailable Fe (42) and a ailable Zn (60) had he la ges numbe
o co ela ed genes, ollowed by o al Mn (24), a ailable Cu (17), o al
PC2 (11), a ailable Mo (9), o al Zn (5) and a ailable Ni (4) (Fig. 3a). The
numbe o genes which we e co ela ed wi h soil mac onu ien s and
pHwassligh lyla ge hanwi hmic onu ien s(Fig.3b).
Specific mic onu ien s also posi i ely a ec ed he mic obial-
egula ed nu ien pa hways. A ailable Fe was associa ed wi h bo h
s a ch and cellulose deg ada ion by posi i ely co ela ing wi h he
amyX gene and cdh gene. To al Mn was significan ly posi i ely co e-
la ed wi h he manB,mpn and pox genes ha we e esponsible o
he deg ada ion o hemicellulose and lignin, espec i ely (Fig. 3c).
A ailable Fe, Cu, and Mo a ec ed he mic obial deni ifica ion p ocess
by co ela ing wi h he na G and ni S/K genes. The Zn significan ly
co ela ed wi h he napA and ni K genes esponsible o ni ogen
educ ion, amoA gene o ni ifica ion and he bpp gene esponsible
o o ganic P mine aliza ion (Fig. 3d, e). The N
2
fixa ion was influenced
by he a ailable Fe (Fig. 3d), while a ailable Fe had no co ela ions on
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 2
P ans o ma ion p ocesses (Fig. 3e). A ailable Fe and Cu also a ec ed
he mic obial S educ ion p ocess by co ela ed wi h he ds A and ds B
genes (Fig. 3 ). No ably, he Rals onia (wi h he a e age ela i e
abundance o 0.4% in a whole communi y) ha co ela ed wi h Fe also
con ained he Fe- esponsi e genes such as amyX and na G (Table S4).
The s ep omyces ha co ela ed wi h Zn (2.3%) con ained he Zn-
esponsi e genes such as napA,ni K and bbp (Table S4).
E ec s o Fe and Zn addi ion on soil mic obiome
The Fe and Zn addi ions g ea ly a ec ed he s uc u e and unc ion o
he soil mic obiome in he incuba ion expe imen s. The Fe addi ion
significan ly dec eased he Shannon index and changed bac e ial
communi y composi ion in mos soils, excep o he soil om GD si e
(Fig. 4a), which was consis en wi h nega i e co ela ion be ween
a ailable Fe and Shannon index/PCoA1 om he obse a ional s udy
(Fig. 2a). Simila finding o alpha di e si y and communi y composi ion
was obse ed in Zn addi ion ea men s (Fig. 4b), in ag eemen wi h
he nega i e co ela ion be ween Shannon index/PCoA1/PCoA2
(Fig. 2a). Simila o he posi i e co ela ion wi h he ela i e abundance
o P o eobac e ia (Fig. 2a), he Fe and Zn addi ions inc eased he
ela i e abundance o P o eobac e ia in mos soils (Fig. 4). Wi h
mic obial unc ion, he Fe addi ion inc eased he ela i e abundances
o cdh,ni H, ds B and ni S genes in he soils om NM and YN, while had
no inc easing e ec s on he gene abundances in he soils om GD and
JS, bo h soils ha ing highe ini ial Fe concen a ion han NM and YN
(Fig. 4a). The Zn addi ion inc eased he ela i e abundances o bbp
gene in he soils om GD, SD and YN si es, and also inc eased he gene
abundance o amoA gene in he soils om GD, NM and YN si es
(Fig. 4b). The inc eases in hese gene abundances we e consis en wi h
hei posi i e co ela ions wi h a ailable Fe/Zn concen a ions om
he obse a ional s udy (Fig. 3).
Con ibu ion o mic onu ien s o ecosys em p oduc ion
Mic onu ien s con ibu ed o ecosys em p oduc ion by bo h di ec
and indi ec e ec s (Fig. 5). Wi h he bac e ial model, ecosys em p o-
duc ion was co-co ela ed by soil pH, bac e ial abundance, and
mic onu ien s. Soil mic onu ien s a ec ed he ecosys em p oduc-
ion mainly by di ec e ec s and also by inc easing bac e ial abun-
dance (Fig. 5a). The mic onu ien -a ec ed bac e ial di e si y and
ne wo k connec i i y we e no co ela ed wi h he ecosys em p o-
duc ion (Fig. 5a). Wi h he ungal model, ecosys em p oduc ion wasco-
co ela ed by soil pH, o al C, ungal abundance and mic onu ien s.
Soil mic onu ien s di ec ly and posi i ely co ela ed wi h he eco-
sys em p oduc ion, bu no by a ec ing ungal abundance, di e si y,
and ne wo k connec i i y (Fig. 5b). Wi h he unc ion model, he eco-
sys em p oduc ion was co-co ela ed by soil pH, mic onu ien s and
mic obial unc ions in ol ed in C and N cycling. Howe e , mic o-
nu ien s did no a ec ecosys em p oduc ion by al e ing mic obial
unc ion. Ins ead, soil pH and o al C ha posi i ely co ela ed wi h he
mic obial unc ions con ibu ed o he ecosys em p oduc ion (Fig. 5c).
Fig. 1 | E ec s o o al and a ailable mic onu ien s (Fe, Mn, Cu, Zn, Mo, and Ni)
on he abundance o soil mic obial axa. Mic oo ganisms a he genus le el we e
iden ified by Deseq2, ha ing ela i e abundances ha we e significan ly highe o
lowe in ela i ely low mic onu ien s as compa ed wi h he high mic onu ien
g oup (Figs. S6 and S7). Two nega i ely and posi i ely iden ified gene a wi h he
mos significance (i.e. lowes adjus ed p- alue) a e p esen ed in he Figu e. The
solid and dashed lines ma ked wi h “B”and “F”p esen he bac e ial and ungal
gene a espec i ely. The co ela ion coe ficien s wi h p- alue adjus men a e p e-
sen ed in Tables S1 and S2.
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 3
Discussion
The influence o mic onu ien s as main egula o s o mic obial com-
muni ies is being inc easingly acknowledged in ma ine ecosys ems18
and in he human gu mic obiome19. Ye , he con ibu ion o mic o-
nu ien s o explain he s uc u e and unc ion o soil mic obiomes in
e es ial ecosys ems emains la gely unde s udied.
Ou findings, based on an obse a ional s udy ac oss China and a
labo a o y alida ion, highligh he impo ance o soil mic onu ien s
in explaining he di e si y and communi y composi ion o soil
mic obes. Theo e ically, mic onu ien s pa icipa e in he auxilia y
me abolism and soil edox eac ion, and he mine als con aining
mic onu ien s c ea e di e en habi a s. These mechanisms would
c ea e a mo e specialized mic obial communi y, causing o e all
di e si y and communi y composi ion changes. P e ious s udies ha e
epo ed ha mic onu ien s explained a la ge a ia ion in he s uc-
u e o mic obial communi ies in ag icul u al soils ecei ing specific
mic onu ien s20, while we u he confi m he ela ionship be ween
mic onu ien s and mic obial s uc u e in a wide ange o soils wi h
mo e con as ing p ope ies and unde di e en land uses. Ou finding
is also consis en wi h some expe imen s showing ha i on mine als
shaped soil mic obial communi y21 and Zn e iliza ion dec eased soil
mic obial di e si y and al e ed communi y composi ion in paddy
soil22. In gene al, soil pH is p e iously epo ed o be posi i ely linked
o soil bac e ial di e si y4,23 and also nega i ely linked o mic onu ien
a ailabili y24. The ela ionship be ween mic onu ien s and mic obial
di e si y can be masked by he pH e ec s. Thus, ou analyses con-
olled he pH e ec s on he soil mic obiome s a is ically and p o ided
solid e idence ha he mic onu ien s, especially hei a ailabili y, can
help explain mic obial di e si y (Fig. 2). The di ec s ong ela ionship
be ween mic onu ien s and he s uc u e o he soil mic obiome can
pa ly explain p e ious e idence o en ichmen o specific mic o-
o ganisms in Fe-Mn nodules25 and he high Fe associa ed-mic obial
occu ence26. I may also assis in helping o explain he unknown
d i en ac o s in bac e ial communi ies whe e he soil pH and
Fig. 2 | E ec s o o al and a ailable mic onu ien s (Fe, Mn, Cu, Zn, Mo, Ni) on
soil mic obial communi ies. a Hea map shows he ela ionship be ween alpha,
be a di e si y, phylum abundance, and ne wo k connec i i y o bac e ia and
mic onu ien s, conduc ed by pa ial co ela ion. bHea map shows he ela ion-
ship be ween alpha, be a di e si y, phylum abundance, and ne wo k connec i i y
o ungi and mic onu ien s, conduc ed by pa ial co ela ion.
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 4
mac onu ien s a e conside ed27. Ou s andingly, he ungal commu-
ni y (e.g. abundance and di e si y) was less sensi i e o soil mic o-
nu ien s compa ed o bac e ia. Being bo h he e o ophic and
euka yo ic o ganisms, hey a e p obably mo e dependen on ene gy
esou ces (C and N) a he han mic onu ien s and mo e esis an
o en i onmen al changes caused by mic onu ien addi ion han
bac e ia5.
Ou s udy also ex ended he field s udy o gene abundance and
expec ed soil unc ions, on he basis o mic obiome s uc u e20.We
e ealed ha mic onu ien s we e co ela ed wi h he abundance o
mic obial unc ional genes associa ed wi h he C, N, P, and S cycle.
Coupled eac ions o mac o and mic onu ien s a e expec ed o be
essen ial o suppo soil mac onu ien u no e , flux and pool sizes,
and nu ien a ailabili y o plan s (e.g. labile C, NO
3
-,e c.).Specifically,
one majo finding shows mic onu ien s, especially Fe and Mn, pa i-
cipa ed in mic obial- egula ed soil nu ien cycling, such as o ganic C
decomposi ion, C fixa ion, deni ifica ion, S educ ion, and po en ially
influencing ecosys em unc ioning (Fig. 3). Mic onu ien s such as Fe
and Mn usually pa icipa e in mic obially- egula ed edox eac ions o
C, N and S in soils8, because mic onu ien s a e in ol ed in elec on
a ailabili y and ans e in biochemical eac ions6,28. Fo ins ance, soil
Fe and Mn specia ion change om educed o ms unde anae obic
condi ions o oxidized o ms unde ae obic condi ions, and his cou-
ples wi h C and N ans o ma ions29, while sul a e- educing bac e ia
a e in ol ed in Fe and Mn oxida ion30. In addi ion, mic onu ien s such
as Cu and Zn posi i ely explained some mic obial pa hways o soil
o ganic C decomposi ion, ni ifica ion, and deni ifica ion ha a e
ela ed o soil espi a ion and g eenhouse gas emissions (e.g. CO
2
and
N
2
O). This indica es he po en ial influence o hese mic onu ien s on
soil-clima e change eedback ha should be u he e alua ed2,23,31.
Ano he impo an finding is ha he mic onu ien s always posi i ely
co ela ed wi h gene abundances (Fig. 3). O e he mode a e ange o
Fig. 3 | E ec s o o al and a ailable mic onu ien s (Fe, Mn, Cu, Zn, Mo, Ni) on
soil mic obial egula ed unc ions in ol ed in C, N, P, and S cycles. a Numbe o
genes in ol ed in mic obial C, N, P, and S cycles whose absolu e abundances sig-
nifican ly co ela ed wi h soil mic onu ien s by pa ial co ela ion. bNumbe o
genes whose absolu e abundances significan ly co ela ed wi h soil chemical
p ope ies by pa ial co ela ion. The “mean” ep esen ed he a e age numbe o
genes co ela ed wi h all mic onu ien s o soil mac onu ien +pH.
cMic onu ien -pa icipa ed mic obial pa hways o he C cycling. dMic onu ien -
pa icipa ed mic obial pa hways o he N cycling. eMic onu ien -pa icipa ed
mic obial pa hways o he P cycling. Mic onu ien -pa icipa ed mic obial pa h-
ways o he S cycling. The labels o mic onu ien s p esen ed in he pa hways we e
s ic ly fil e ed i hei co ela ions wi h he absolu e abundance and ela i e
abundance o co esponding genes we e bo h significan (p< 0.05). Labels wi h ed
and blue colo s p esen ed ha nega i e and posi i e co ela ions be ween mic o-
nu ien s and gene abundance. The mic oo ganisms con aining he genes in he
pa hways a he phylum and genus le el a e p esen ed in Table S4.
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 5

me als in ou soils (i.e. non-con amina ed soils), he supply o mic o-
nu ien s may no cause oxici y o inhibi ion o mic obial g ow h in
any case. Ins ead, hey will enhance mic obial g ow h and enzyma ic
ac i i ies, indica ing he mode a e inc ease o mic onu ien s can
accele a e he mac onu ien cycling in e es ial ecosys ems. Ou
incuba ion expe imen s also e ealed he beneficial impac s o
mic onu ien s Fe and Zn on mic obial-specific unc ions. The added
Fe and Zn demons a ed he po en ial o se e as coenzymes o induce
al e a ions in soil condi ions6,32. Howe e , he Fe and Zn e ec s
depended on soil ypes. Fo example, he Fe addi ion in he soil om
GD did no change mic obial unc ions (Fig. 4), p obably a ibu ed o
he high ini ial Fe con en in ed soil. On he o he hand, ce ain
mic obial unc ions migh exclusi ely exhibi changes wi hin specific
incuba ion condi ions, such as unde highly anae obic condi ions o in
he p esence o o he coexis ing elemen s.
Fu he analyzes iden ified bo h bac e ial and ungal gene a ha
we e highly co ela ed o di e en mic onu ien s, e.g. 0319-7L14 o Fe,
Aspe gillus o Mn, and Fusa ium o Mo and Ni ha ha e much lowe
concen a ions in soils (Fig. 1). As he ela ionship o mos iden ified
gene a ha e no been p e iously epo ed, hese gene a can be used as
bioma ke s p edic ing he specific habi a s ha ing one ex emely high
o low mic onu ien . In addi ion, some Fe posi i ely-co ela ed gene a
would be po en ial magne o ac ic mic oo ganisms ha p edic he high
con en o magne i e, goe hi e, o e ihyd a e in soil33. Some gene a
pa icipa ed in impo an soil p ocesses such as Fe-co ela ed s a ch
decomposi ion ( egula ed by Mic oluna us), Zn-co ela ed ni ifica ion
(Candida us Ni osocosmicus), and deni ifica ion (Soli ub obac e ),
Zn-co ela ed o ganic P mine aliza ion (Soli ub obac e ), and Cu-
co ela ed S educ ion (Piscinibac e )(TableS4).Theiden ified mic o-
o ganisms will be impo an mic obial esou ces o egula ing soil
nu ien cycling and main aining ecological unc ions, especially in
mic onu ien -abundan habi a s.
We u he demons a ed ha he con ibu ion o mic o-
nu ien s o mic obial communi ies depends on he a ailabili y o
mic onu ien s. Rega ding mic onu ien pools o biological sig-
nificance, o al s ocks a e mo e s able o e ime, while he a ailable
pools a e mo e mobile and soluble in soils and expec ed o be mo e
accessible o mic oo ganisms9. Thus, in ou s udy, mic obial di e -
si y and he abundance o specific phyla we e mo e co ela ed wi h
mic onu ien a ailabili y han i s o al s ocks (excep o Mn and
Mo). Also, some mic onu ien s a e p esen in soils in biologically
ine solid phases. Fo ins ance, soil o al Fe accoun s o mo e han 3
% o soil mass bu mos o i is fixed as i on oxides o o he mine als34,
ha a e included in he o al ac ion. Only a small pa o i is eadily
bio-a ailable. In con as , o al Mn and Mo had ela ionships wi h
mo e mic obial pa ame e s ela i e o hei a ailable o m and we
a ibu ed i o he mo e compa able o de o magni udes in o al
concen a ion and he espec i e a ailable concen a ion o bo h
elemen s. Mo eo e , we ound ha Fe had he la ges numbe o
co ela ions wi h mic obial pa ame e s, ollowed by Mn, Zn, and Cu,
and Mo and Ni had he lowes numbe . This was pa ly expec ed
because p e ious wo k sugges s he ole o Fe, Mn, Cu, and Zn in
many impo an enzyma ic p ocesses in soils and ac as eac an s o
C, N, and S coupled eac ions, while Mo and Ni a e only specialized o
less p ocesses6,29,31,35. Al hough each single mic onu ien plays spe-
cific biological oles, he in eg a ed e ec s o mic onu ien s (i.e.
e ealed by PCA) explained mo e a iances o bac e ial abundance,
di e si y and unc ional genes compa ed o he specific e ec o each
mic onu ien (Fig. 2a and Fig. 3a).
Recen s udies showed ha plan p oduc i i y is no only influ-
enced byclima e and edaphic ac o s, bu alsois highly associa ed wi h
belowg ound mic obial communi ies ( o example, globally, 64% o
plan biomass is p omo ed by soil mic obiome es o a ion)13.Gi en
ha mic onu ien s a e co ela ed wi h mic obial communi y and
unc ion, we specula e ha mic onu ien -d i en mic obiomes con-
ibu e o ecosys em p oduc ion. Ou SEM model e ealed ha soil
mic onu ien s explain ecosys em p oduc ion mainly by a di ec
e ec , p obably supplemen ing mic onu ien s o plan g ow h17,36.
Howe e , he SEM model also sugges s ha highe mic onu ien
concen a ions benefi plan p oduc i i y by inc easing soil bac e ial
biomass, shown by he high posi i e ela ionship be ween bac e ial
biomass and plan p oduc i i y (Fig. 5). This is p obably associa ed
wi h he e idence ha mic oo ganisms can inc ease plan nu ien
acquisi ion and esis ance o s esses10. Howe e , he mic onu ien -
d i en mic obial di e si y and ne wo k connec i i y (i.e. indi ec
e ec ) did no con ibu e o ecosys em p oduc ion, sugges ing ha
he e ec s o mic onu ien -d i en mic obiome s uc u e on ecosys-
em p oduc ion should no be o e s a ed. Mic onu ien s-d i e unc-
ion also did no con ibu e o he ecosys em. This is p obably
a ibu ed o he much lowe concen a ions o mic onu ien s com-
pa ed o mac onu ien s in soil. Fo example, he con ibu ion o
mic obiome unc ion (e.g. C and N cycles) d i en by o al C o eco-
sys em p oduc ion was mo e dominan (Fig. 5). The inc eased C
cycling may inc ease he soil C fluxes and o ganic C loss, he e o e
nega i ely a ec ing ecosys em p oduc ion. While he inc eased N
cycle may lead o mo e ino ganic N eleased o he en i onmen ,
he e o e inc easing plan g ow h. O e all, i is impo an o conside
all hese ac o s oge he o explain ecosys em p oduc ion including
mac onu ien s, soil pH, and spa ial and clima ic ac o s17,37.
In conclusion, mic onu ien s a e associa ed wi h he s uc u e
and unc ion o soil mic obiomes, highligh ing he impo ance o
mic onu ien s on ecosys em unc ioning (Fig. 6). The e ec s o
Fig. 4 | E ec s o Fe and Zn addi ions on soil mic onu ien a ailabili y and
mic obial pa ame e s in soil incuba ion expe imen s. a Changes in soil a ailable
Fe concen a ion, mic obial di e si y, P o eobac e ia, and ela i e abundances o
ni H,ni S,ds B,cdh, and amyX a e Fe addi ion. bChanges in he soil a ailable Zn
concen a ion, mic obial di e si y, P o eobac e ia and ela i e abundances o
amoA,bbp and ni K a e Zn addi ion. The di e ences in mic obial pa ame e s
no malized by Z-sco e be ween con ol and he Fe/Zn ea men s we e conduc ed
by independen - es . The “*” ep esen s he di e ences we e significan
a p < 0.05.
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 6
mic onu ien s on mic obialcommuni ies we e la ge o bac e ia han
o ungi, wi h he soil-a ailable pool o mic onu ien s mo e s ongly
associa ed han he o al pool. No ably, he Fe co ela ed mos wi h he
mic obial communi y compa ed o Mn, Cu, Zn, Mo, and Ni. While he
ole o mic onu ien s should no be o e es ima ed, ou findings
sugges ha i should be aken in o accoun in u u e esea ch and
global sus ainable de elopmen goals.
Me hods
Si e desc ip ion and sample collec ion
Su ace soils we e collec ed om 180 si es ac oss China, wi h a span o
32.5olongi ude and 33.4ola i ude. These loca ions a y wi h di e en
clima ic condi ions ( om cold o opical zones), soil condi ions, and
ege a ion co e (c ops, g ass, and o es s). The mean annual p e-
cipi a ion and empe a u e anges om 395 o 2486 mm and −2.7 o
Fig. 5 | Con ibu ions o en i onmen al a iables o ecosys em p oduc i i y.
aE ec s o en i onmen al a iables (i.e. clima e, geo-spa ial a iance, soil basic
p ope ies, soil mac onu ien s, and soil mic onu ien s) on ecosys em p oduc ion
ia al e ing bac e ial communi y. bE ec s o en i onmen al a iables on ecosys-
em p oduc ion ia al e ing ungal communi y. cE ec s o en i onmen al a iables
on ecosys em p oduc ion ia mic obial unc ions. Bac e ial communi y and ungal
communi y include he abundance, di e si y, and ne wo k connec i i y
pa ame e s. Mic obial unc ion includes he abundance o mic obial genes
in ol ed in C, N, P, and S cycling. The ed and blue a ows ep esen ed he sig-
nifican (p< 0.05) nega i e and posi i e ela ionships be ween a iables espec-
i ely, while he a ows wi h non-significan ela ionshipswe e no shown. Adjacen
alues nea he a ows indica e he pa h coe ficien s. 2 alue indica es he p o-
po ion o ecosys em p oduc ion explained by each a iable.
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 7
27.9 °C, espec i ely. The NDVI alue anged 120 om 0.22 o 0.90
ac oss si es. The sample dis ibu ion and associa ed in o ma ion such
as land use a e p esen ed in Fig. S1.
Su ace soils (0–20 cm), i.e. bulk soils, om each si e we e sam-
pled. Sample si es we e chosen in a cen al a ea wi h he same land use
o a oid he dis u bance o o he land uses and ce ain ea u es (e.g.
oads, buildings). In each si e, fi e soil co es we e collec ed and hen
pooled o o m a composi e sample wi hin a size o 15 m x 15 m. The
soil samples we e immedia ely anspo ed o he labo a o y on ice.
The soils we e hen sie ed h ough 2 mm a e he emo al o plan
esidues and s ones. Each soil sample was di ided in o wo pa s; one
wasai -d ied, ball-milled, sie ed o <0.15 mm, and s o ed a 4 °C be o e
chemical analyzes, and he o he was s o ed a −80 °C p io o DNA
ex ac ion. All soil samples we e synch onously analyzed o chemical
p ope ies and DNA ex ac ion.
Soil chemical p ope y cha ac e iza ion
Soil pH was measu ed wi h a pH me e in a soil- o-wa e suspension o
1:2.5. Soil o al ca bon (C) and ni ogen (N) we e de e mined by an
elemen al analyze (Va io EL II, Ge many). The molybda e blue col-
o ime ic me hod was used o measu e soil o al P a e he diges ion
o he soil samples wi h H
2
SO
4
–HClO
4
. Soil pH and o al C con en
ange om 4.4 o 9.9 and 0.3% o 7.4%, espec i ely. The de ails o he
ange o soil pH and mac onu ien con en s ac oss he 180 soil sam-
ples a e p esen ed in Fig. S2.
Soil o al concen a ion o me allic mic onu ien s including
Fe, Mn, Cu, Zn, Mo, and Ni we e de e mined a e diges ion wi h
HF-HNO
3
-H
2
O
2
(1:2:1, / / ) a 210 °C o 60 min using a mic owa e38.
All he diges ion solu ions we e fil e ed h ough fil e pape s and hen
adjus ed o a olume o 50 mL wi h dis illed wa e . Soil a ailable
mic onu ien s we e ex ac ed by an ex ac an con aining 0.005 mol/
L DTPA, 0.01 mol/L CaCl
2
, and 0.1 mol/L ie hanolamine a a pH o
7.30 and he ex ac ion solu ions we e fil e ed h ough a g ade 42
Wha man fil e pape 39. This ex ac ion has been commonly used o
di e en soils du ing he las decades. The concen a ions o o al and
a ailable mic onu ien s we e measu ed by induc i ely coupled
plasma op ical emission spec ome y (ICP-OES) (iCCP 6300, The mo
Scien ific, USA). The ange o o al and a ailable mic onu ien con-
cen a ions ac oss 180 soil samples a e p esen ed in Fig. S3.
Mic obial DNA ex ac ion and sequencing
Soil mic obial DNA was ex ac ed by he Fas DNA SPIN ki (MP Bio-
medicals, Solon, OH, USA) ollowing he manu ac u e ’sp o ocol.Soil
DNA samples we e hen sen o o lib a y p epa a ion and amplicon
sequencing a e quali y checked by aga ose gel elec opho esis. The
p ime s o 515F-907R (5’-GTGCCAGCMGCCGCGGTAA-3’,5’-CCGTC
AATTCCTTTGAGTTT-3’) we e a ge ed o he 16 S V4 e sion o soil
bac e ia40, and he p ime s o ITS3F-ITS4R (5’-GCATCGATGAAGAA
CGCAGC-3’,5’-TCCTCCGCTTATTGATATGC-3’) we e a ge ed o he
ITS2 e sion o soil ungi41.PCR eac ionswe ep epa edas ollows:a
o al olume o 50 μl eac ion mix u e con aining 25 μlo 2xP emix
Taq (ob ained om Taka a Bio echnology, Dalian Co. L d., China), 1 ul
o each p ime (10 μmol/L), and 3 μlo DNA.The he malcycling
p og am consis ed o : 1) an ini ial dena u a ion a 94 °C o 5 min, 2) 30
Fig. 6 | Concep ual diag am illus a ing he impac s o soil mic onu ien s and
soil mac onu ien on he s uc u e and unc ion o soil mic obiome. Mic o-
nu ien s a e highly co ela ed o he s uc u e and unc ion o soil mic obiomes,
and a e compa able o he e ec s o soil mac onu ien s and pH. Mic onu ien s
posi i ely con ibu ed o he ecosys em p oduc i i y by di ec e ec (nu ien
supply o plan s) a he han indi ec e ec (associa ed mic obiome). The icons
e e ing o ege a ion used in he figu e we e applied wi hou endo semen om
he websi e o h ps://www.icon on .cn/.
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 8
cycles o dena u a ion a 94 °C o 30s, annealing a 52 °C o 30 s, and
ex ension a 72 °C o 30 s and 3) final elonga ion a 72 °C o 10 min.
The PCR amplifica ion was ca ied ou using he BioRad S1000 PCR
ins umen (Bio-Rad Labo a o y, CA, USA). Amplified PCR p oduc s
we e sequenced on he Illumina pai ed-end pla o m wi h he ins u-
men o No aSeq 6000 (Illumina, San Diego, CA, USA). The quali y o
aw da a was con olled by he Fas p (h ps://gi hub.com/OpenGene/
as p, e sion 0.20.0) wi h sliding window (-W 4, -M 20) and he p i-
me s we e emo ed by cu adap (h ps://gi hub.com/ma celm/
cu adap /). In a e age, 65866 and 73408 good-quali y eads we e
e ained o bac e ia and ungi espec i ely. Using he QIIME 2 pipeline
(2022.11), he pai ed-end eads we e me ged and he ea u e able was
gene a ed using he dada2 denoise-pai ed plugin wi h chime ic eads
immed42. The q2- ea u e classifie plugin was used o assign ax-
onomy o sequences using he Sil a 138 ( he confidence h eshold was
a de aul alue o 0.7). In o al, 98805 and 20903 amplicon sequence
a ian s we e gene a ed o bac e ia and ungi espec i ely.
High- h oughpu qPCR based chip
The absolu e abundance o genes in ol ed in C, N, P, and S cycling
we e measu ed using a high- h oughpu qPCR-based chip (QMEC) on
Sma Chip Real-Time PCR Sys em (Wa e Gen Biosys ems, F emon ,
USA) ollowing manual ins uc ions43. The gene names and unc ion
anno a ions a e p esen ed in Table S5. The 16 S RNA gene was used
as he e e ence gene. In b ie , he qPCR amplifica ion p ocedu e was
an ini ial dena u a ion a 95 °C o 10 min, ollowed by 40 cycles o
30 s dena u a ion a 95 °C, 30 s annealing a 58 °C and 30 s ex ension
a 72 °C. Each DNA sample was amplified in iplica e. I he esul s
amplifica ion e ficiencies o <1.8 and > 2.2, nega i e con ol amplified
and a h eshold cycle (CT) > 31, hey we e excluded om u he
analyzes. In addi ion, bac e ial and ungal absolu e copy numbe s
we e measu ed by qPCR using he same p ime s o 16 S and ITS
sequencing, ep esen ing soil o al bac e ial and ungal abundance,
espec i ely.
Me agenomics
To examine he mic obial axa con aining he specific genes in ol ed
in C, N, P, and S cycling, we sen he DNA samples o me agenomics
sequencing. Pai ed-end sequencing was pe o med on he ins umen
No aSeq 6000. The quali y o aw eads wi h a sequencing dep h o
20 G was con olled by using he Fas p ( e sion 0.20.0)44.Thefil e ed
eads we e assembled o con igs using he Megahi ( e sion 1.1.2) ia
de B uijn g aph and wi h he minimum and maximum k-me sizes o 21
and 121 espec i ely and s ep size o 1045. Con igs wi h he leng h >
500 bp we e e ained. P odigal( e sion 2.6.3) was hen used o iden i y
he open eading ames in con igs and p edic he p o ein-coding
genes46. The KEGG unc ional anno a ions we e conduc ed using Dia-
mond ( e sion 0.8.35) agains he Kyo o Encyclopedia o Genes and
Genomes da abase ( e sion 94.2) wi h an e- alue cu o o (<1 × 10−5)47.
To ob ain he axonomic assignmen s co esponding o specific
unc ional genes, he sequences o p edic ed genes om he KEGG
da abase we e anno a ed based on he NCBI NR da abase using blas p
in Diamond wi h e- alue cu o o <1 × 10−5.
Soil incuba ion expe imen s
To e i y he ela ionship be ween soil mic onu ien s and soil
mic obiome, we conduc ed a soil incuba ion expe imen by adding Fe
and Zn in o fi e soils wi h con as ing chemical p ope ies. The Fe and
Zn we e chosen o alida ion expe imen s because hei concen a-
ions co ela ed wi h mos mic obial pa ame e s. Soils we e collec ed
om geog aphically dis an si es ac oss China, specifically Guangdong
(GD, 110°6'6“E, 21°8'2“N), Jiangsu (JS, 118°38'4“E, 32°28'52“N), Shan-
dong (SD, 116°19'50“E, 35°59'44“N), Yunnan (YN, 100°20'6“E,
25°34'47“N) and Neimenggu (NM, 117°44'18“E, 44°12'58“N) p o inces.
The o al Fe concen a ions in he soils collec ed om GD, JS, SD, YN,
and NM si es we e 83231, 34381, 26920, 28500 and 10193 mg kg−1
espec i ely. The o al Zn concen a ions we e 73, 55, 82, 42, and
24 mg kg−1 espec i ely. The concen a ions o e ic (FeCl
3
)and e -
ous chlo ides (FeCl
2
) (1:1) and zinc chlo ide (ZnCl
2
) dissol ed in wa e
we e 1000 mg/kg and 100 mg/kg espec i ely, and he soils wi hou
mic onu ien addi ion we e he con ols. The soil samples we e
incuba ed in plas ic bags unde da kness a 25 °C, while main aining
he mois u e a 60% o he wa e -holding capaci y. All he ea men s
had h ee eplica es. A e a wo-week incuba ion, samples we e col-
lec ed o DNA ex ac ion, 16 S sequencing, and gene abundance
measu emen s. The me hods employed o DNA ex ac ion, sequen-
cing, and qPCR we e consis en wi h he p ocedu es desc ibed abo e.
The a ailable Fe and Zn concen a ions we e measu ed acco ding o
he me hod desc ibed abo e39. While unspiked soils had Fe and Zn
a e age concen a ions o 8.7 and 1.2 mg/kg espec i ely a he end o
he incuba ion, he Fe spiked soils had he Fe a e age concen a ion o
22.6 mg/kg and he Zn spiked soils had a concen a ion o 4.9 mg/kg (3
and 4 imes g ea e a ailabili y in Fe o Zn spiked soils).
Da a analysis
P incipal componen analysis (PCA) in SPSS s a is ics so wa e ( e -
sion 24.0)48 was used o g oup he o al and a ailable concen a ions
o Fe, Mn, Cu, Zn, Mo, and Ni. Co ela ions be ween hese mic o-
nu ien s a e p esen ed in Table S6. The p incipal componen 1
(TPC1) and p incipal componen 2 (TPC2) explained 51.4% and 19.5%
o o al mic onu ien s espec i ely (Fig. S10). The TPC1 mainly
explained he a iables o o al Fe, Cu, Zn, and Ni, and he TPC2
explained he o al Mn and Mo. Simila ly, he p incipal componen 1
(APC1) and p incipal componen 2 (APC2) explained 43.1% and 25.8%
o a ailable mic onu ien s espec i ely (Fig. S10). The APC1 mainly
explained he a iables o a ailable Fe, Cu, and Zn, and he APC2
explained he a ailable Mn, Mo, and Ni.
Bac e ial and ungal alpha di e si y, i.e. Shannon and Chao 1
index, we e ob ained be o e a e ying all sequences a a minimum
numbe o sequences pe sample, 33018 and 35115, espec i ely40.
Mic obial be a di e si y, i.e. PCoA1 and PCoA2, based on he weigh ed
uni ac dis ance ma ix was pe o med using he R package
“phyloseq”49. The di e ences in he o e all mic obial communi y
be ween di e en clima e zones and land uses we e es ed by wo-way
ANOVA (Fig. S8). The di e ences in he o e all mic obial communi y
be ween he soils wi h low, medium, and high concen a ions o o al
and a ailable mic onu ien s we e in es iga ed using PERMANOVA50.
Gi en ha he o e all mic obial communi y be ween soils wi h lowand
high mic onu ien concen a ions in mos cases we e significan
(Tables S7 and S8), he “DESeq2”( e sion 1.40.2) was pe o med o
iden i y he highly co ela ed bac e ial and ungal axa a he genus
le el (Figs. S6 and S7). Gene a wi h a log2- old change in ela i e
abundance >1 and an adjus ed p< 0.05 we e selec ed as he gene a ha
we e highly co ela ed o mic onu ien s51.Then, he ela ionship
be ween co ela ed axa (iden ified by Deseq2 wi h he lowes adjus ed
p- alue) and mic onu ien concen a ions we e conduc ed by Spea -
man co ela ions. Due o he po en ial he e ogenei y o he samples
collec edac ossChina,weusedda ano maliza iono log10 ans o -
ma ion o mic onu ien s o mi iga e he impac o sample he e o-
genei y on he ela i e abundances o co ela ed gene a. Resul s
showed simila co ela ions as compa ed o no da a no maliza ion
(Tables S1 and S2).
Bac e ial and ungal co-occu ence ne wo ks we e s uc u ed
sepa a ely based on he Spea man co ela ion ma ix wi h a cu -o
co ela ion coe ficien de e mined by RMT heo y in an au oma ic
ashion52 and he p- alue was co ec ed by he FDR me hod53.The
opological pa ame e s o o al deg ee, a e age deg ee and a e age
pa h leng h ha ep esen ed mic obial ne wo k connec i i y we e
A icle h ps://doi.o g/10.1038/s41467-023-44182-2
Na u e Communica ions | (2023) 14:8456 9