SÃO PAULO STATE UNIVERSITY - UNESP 
SCHOOL OF ENGINEERING 

CAMPUS OF ILHA SOLTEIRA 
 
 
 
 
 
 
 
 
 

GUSTAVO DOS SANTOS COTRIM 
 
 
 
 
 
 
 
 
 
 
 

EFFECT OF POTASSIUM AVAILABILITY ON SOYBEAN METABOLISM BY 
INTEGRATED METABOLOMICS AND IONOMICS ANALYSIS 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 
Ilha Solteira 

2022



 
 

GRADUATE PROGRAM IN AGRONOMY  

 
 
 
 
 

GUSTAVO DOS SANTOS COTRIM 
 
 
 
 
 
 
 
 
 

EFFECT OF POTASSIUM AVAILABILITY ON SOYBEAN METABOLISM BY 
INTEGRATED METABOLOMICS AND IONOMICS ANALYSIS 

 
 
 
 
 

Dissertation submitted to Engineering College of 
Ilha Solteira (FEIS - UNESP) in partial fulfillment 
of the requirements for the degree of Master of 
Science in Agronomy. Emphasis: Cropping 
Systems. 
 
Prof. Dr. Lucíola Santos Lannes 
Supervisor 
 
Prof. Dr. Clara Beatriz Hofmann Campo 
Co-supervisor  
 
 
 
 
 
 
 
 
 
 

 
 

Ilha Solteira 
2022 



Cotrim Effect of Potassium Availability on Soybean Metabolism by Integrated Metabolomics and Ionomics AnalysisIlha Solteira2022 65 Sim Dissertação (mestrado)Outros cursosSistemas de ProduçãoNão

.

.

.

FICHA CATALOGRÁFICA

Desenvolvido pelo Serviço Técnico de Biblioteca e Documentação

Cotrim, Gustavo dos Santos.
Effect of potassium availability on soybean metabolism by integrated 

metabolomics and ionomics analysis  / Gustavo dos Santos Cotrim. -- Ilha 
Solteira: [s.n.], 2022

65 f. : il.

Dissertação (mestrado) - Universidade Estadual Paulista. Faculdade de 
Engenharia de Ilha Solteira. Área de conhecimento: Sistemas de Produção, 2022

Orientadora: Lucíola Santos Lannes
Coorientadora: Clara Beatriz Hoffmann-Campo
Inclui bibliografia

1. Glycine max. 2. Fabaceae. 3. Potassium deficiency. 4. Phytoalexins. 5.
Specialised metabolism. 6. Abiotic stress.

C845e



UNIVERSIDADE ESTADUAL PAULISTA

Câmpus de Ilha Solteira

EFFECT OF POTASSIUM AVAILABILITY ON SOYBEAN METABOLISM BY

INTEGRATED METABOLOMICS AND IONOMICS ANALYSIS  
TÍTULO DA DISSERTAÇÃO:

CERTIFICADO DE APROVAÇÃO

AUTOR: GUSTAVO DOS SANTOS COTRIM

ORIENTADORA: LUCÍOLA SANTOS LANNES

COORIENTADORA: CLARA BEATRIZ HOFFMANN-CAMPO

Aprovado como parte das exigências para obtenção do Título de Mestre em Agronomia, área:

Sistemas de Produção pela Comissão Examinadora:

Profa. Dra. LUCÍOLA SANTOS LANNES (Participaçao  Virtual)
Departamento de Biologia e Zootecnia  / Faculdade de Engenharia de Ilha Solteira - UNESP

Profa. Dra. CRISTIÉLE DA SILVA RIBEIRO (Participaçao  Virtual)
Departamento de Biologia e Zootecnia / Faculdade de Engenharia de Ilha Solteira  UNESP

Prof. Dr. JHONYSON ARRUDA CARVALHO GUEDES (Participaçao  Virtual)
Química Analítica e Físico-Química / Universidade Federal do Ceará - UFC

Ilha Solteira, 02 de setembro de 2022

Faculdade de Engenharia - Câmpus de Ilha Solteira -
Av. Brasil Centro, 56, 15385000

www.feis.unesp.br/#!/ppgaCNPJ: 48.031.918/0015-20.



 
 

DEDICATION 
 

 
This work is completely dedicated to respectful memory of my grandparents  

Jacira Brizola dos Santos and Ranili Custodio Cotrim 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 
 



 
 

ACKNOWLEDGMENTS 
 

I would like to express my gratitude to: 
 
My supervisor Prof. Dr. Lucíola Santos Lannes not only enlightened me with academic 
knowledge but also gave me valuable advice. At the same time, my thanks for your productive 
discussions, encouragement, support, and opportunity conceived in the Graduate Program in 
Agronomy of São Paulo State University – UNESP.  
 
My co-supervisor Prof. Dr. Clara Beatriz Hoffmann Campo has generously provided me with 
her time, insight, continuous support, encouragement, and improving the quality of my writing. 
I express gratitude for the opportunity to research activities in Chemical Ecology Laboratory in 
Brazillian Agricultural Research Corporation – Embrapa Soybean. 
 
Dr. Adilson de Oliveira Junior, Dr. Cesar de Castro, and Dr. Guilherme Julião Zocolo for their 
technical support and encouragement.  
 
Dr. Deivid Metzker da Silva, Dr. José Perez da Graça, and Dr. Rejane Stubs Parpinelli for their 
experimental procedures support, useful discussion, suggestions, and friendship.  
 
Finally, thank you to my family and friends who have been there for me throughout graduate 
school. This especially includes my mother Sandra, my father Reginaldo, and my sister Ana 
Beatriz.  
 
This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível 
Superior - Brasil (CAPES) - Financial Code 001. 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 



 
 

RESUMO 

 
O potássio (K+) tem funções fisiológicas vitais nas plantas e sua disponibilidade pode impactar 

a tolerância de plantas a condições de estresse biótico e abiótico. Anteriormente, estudos 

limitados investigaram o efeito da adubação potássica no metabolismo da soja. Neste trabalho, 

a utilização de análises ômicas integradas, metabolômica e ionômica, permitiram investigar a 

resposta de plantas de soja (Glycine max) em campo sob quatro níveis de disponibilidade de 

K+: muito baixa (KVL), baixa (KL), média (KM) e muito alta (KVH). Folhas (V7) e tecidos da 

vagem (R5.5) coletados foram submetidos à extração e analisados por cromatografia líquida de 

ultra eficiência acoplada à espectrometria de massa de alta resolução (UPLC-HRMS) e 

espectroscopia de emissão óptica por plasma acoplado indutivamente (ICP-OES). O emprego 

de modelos multivariados (PCA-X&Y e O2PLS-DA) permitiram identificar 51 compostos 

pertencentes a 19 vias metabólicas regulados diferencialmente nos tecidos de plantas que se 

desenvolveram sob condições contrastantes de disponibilidade de K+ (KVL vs. KVH). Os níveis 

de potássio também influenciaram o rendimento de grãos com menores níveis em KVL (2211 

kg ha-1) e KL (3737 kg ha-1), diferindo dos tratamentos KM e KVH (4093 e 4096 kg ha-1, 

respectivamente). Sob baixíssima concentração de K+, os teores de Ca2+, Mg2+, Fe2+, Cu2+ e B 

aumentaram nas folhas jovens e maduras. Não somente, o conteúdo de isoflavonas, 

coumestanos, pterocarpanos e sojasaponinas mostraram-se positivamente regulados nas folhas 

severamente deficientes em K+, estando estes eventos associados a um possível estado de 

estresse oxidativo e fotodinâmico decorrentes da deficiência nutricional. A abordagem ômica 

integrada revelou que a adubação potássica é responsável por promover o aumento da 

biossíntese de carboidratos, galactolipídeos e flavonóis glicosídicos em folhas e vagens. No 

entanto, os tecidos da vagem deficientes em K+ apresentaram aumento no contéudo de 

aminoácidos, oligossacarídeos, derivados do ácido benzoico e isoflavonas. O aminoácido 

asparagina é majoritariamente acumulado nos tecidos deficientes em potássio, sinalizando 

como possível biomarcador para a deficiência deste macronutriente na soja. Além disso, os 

resultados indicam que a regulação positiva de metabólitos especializados constitutivos 

detectados por UPLC-HRMS não estão diretamente associados com o aumento nos níveis de 

K+ no solo. Em geral, a presente contribuição melhora nossa compreensão sobre como o 

metabolismo especializado da soja é dependente e/ou influenciado pela nutrição potássica.  

 

Palavras-chave: Glycine max; Fabaceae; Deficiência de Potássio; Fitoalexinas; Metabolismo 

Especializado; Estresse Abiótico; Metabolômica; Ionômica. 



 
 

ABSTRACT 

 
Potassium (K+) has vital physiological functions in plants and its availability can impact the 

tolerance of species to biotic and abiotic stress conditions. Limited studies have investigated 

the effect of K+ fertilization on soybean metabolism. Using integrated omics, ionomics and 

metabolomics, we investigated the response of field-grown soybean (Glycine max ) to four rates 

of soil K+ availability: very low (KVL), low (KL), medium (KM), and very high (KVH). 

Soybean trifoliate leaf (V7) and pod tissue (R5.5) extracts were analysed by ultra-performance 

liquid chromatography coupled to high-resolution mass spectrometry (UPLC-HRMS) and 

inductively coupled plasma optical emission spectroscopy (ICP-OES). Multivariate analyses 

showed that 51 compounds of 19 metabolic pathway maps were regulated in response to K+ 

availability. The soybean yield parameters also were influenced in plants under very low (KVL; 

2211 kg ha-1) and low (KL; 3737 kg ha-1) differing from KM and KVH (4093 and 4096 kg ha-

1, respectively) treatments. Under very low potassium availability, soybean plants promoted the 

accumulation of Ca2+, Mg2+, Fe2+, Cu2+, and B in young and old leaves. Not only, isoflavones, 

coumestans, pterocarpans, and soyasaponins also were elicited in severely K+ deficient trifoliate 

leaves, which can be associated with oxidative and photodynamic stress status. Potassium 

fertilization upregulated carbohydrate, galactolipid, and flavonol glycoside biosynthesis in 

leaves and pod valves, while K+ deficient pod tissues showed increasing contents of amino 

acids, oligosaccharides, benzoic acid derivates, and isoflavones. Additionally, results 

demonstrate that asparagine content is higher in potassium deficient tissues, which suggests 

being a biomarker of K+ deficiency in soybean plants. These results demonstrate that potassium 

soil fertilization did not linearly contribute to changes in specialised constitutive metabolites of 

soybean. Altogether, this work provides a reference for improving the understanding of soybean 

metabolism as dependent on K+ availability. 

Keywords: Glycine max; Fabaceae; Potassium deficiency; Phytoalexins; Specialised 

metabolism; Abiotic stress; Metabolomics; Ionomics. 

 

 

 
 
 
 
 

 



 
 

TABLE OF CONTENTS 

 

1 INTRODUCTION................................................................................................ 11 

   

2 MATERIAL AND METHODS............................................................................. 14 

2.1 GENERAL EXPERIMENTAL PROCEDURES..................................................... 14 

2.2 EXPERIMENTAL DESIGN AND ASSAY............................................................  14 

2.3 PLANT MATERIALS............................................................................................. 15 

2.4 SOIL CHEMICAL ANALYSIS..............................................................................  15 

2.5 PLANT EXTRACTION........................................................................................... 15 

2.5.1 Extraction Process for Plant Metabolomics.........................................................  16 

2.5.2 Extraction Process for Plant Ionomics.................................................................  16 

2.6 INSTRUMENTAL ANALYSES............................................................................. 16 

2.6.1 Chromatographic Analysis (UPLC).....................................................................  16 

2.6.2 Mass Spectrometry (ESI-QTof-MSE)...................................................................  17 

2.6.3 Inductively Coupled Plasma (ICP-OES)..............................................................  17 

2.7 YIELD PARAMETERS........................................................................................... 18 

2.8 DATA INTERPRETATION.................................................................................... 18 

2.8.1 Univariate Analysis................................................................................................  18 

2.8.2 Pre-processing and Chemometrics analysis…..................................................... 18 

2.8.3 Metabolite Annotation and Molecular Networking............................................  19 

   

3 RESULTS AND DISCUSSION............................................................................. 21 

   

3.1 SOYBEAN LEAF IONOMICS ANALYSIS........................................................... 21 

3.2 SOYBEAN POD IONOMICS ANALYSIS............................................................. 24 

3.3 YIELD PARAMETERS........................................................................................... 26 

3.4 CHEMOMETRICS ANALYSIS.............................................................................  27 

3.5 MOLECULAR NETWORKING AND METABOLIC PATHWAY......................  30 

3.6 SOYBEAN LEAVES UNDER K+ HOMEOSTASIS..............................................  33 

3.7 SOYBEAN POD TISSUES UNDER K+ HOMEOSTASIS....................................  34 

3.8 SOYBEAN LEAVES UNDER K+ DEFICIENCY..................................................  37 

3.9 SOYBEAN POD TISSUES UNDER K+ DEFICIENCY.........................................  42 



 
 

4 CONCLUDING REMARKS............................................................................... 44 

   

 REFERENCES..................................................................................................... 45 

   

 APPENDIX A – Soil Chemical Atrributes...........................................................  57 

 APPENDIX B – Leaves Tukey’s t-value...............................................................  58 

 APPENDIX C – Leaves Growth Region Tukey’s t-value...................................  59 

 APPENDIX D – Pod tissues Tukey’s t-value........................................................  60 

 APPENDIX E – Pod tissues Studen’t t-value.......................................................  61 

 APPENDIX F – Multivariate Analysis Values.....................................................  62 

 APPENDIX G – Putative Metabolite Identification............................................  63 



11 

1 INTRODUCTION 

Potassium (K+) is an important macronutrient that plays a vital role in metabolic 

processes, growth, and adaptation to environmental stresses (HASANUZZAMAN et al., 2018; 

PANDEY; MAHIWAL, 2020). Many physiological and biochemical processes are dependent 

on K+ availability, such as cell turgidity regulation, long-distance solute transport, enzymatic 

regulation (AMTMANN et al., 2008; HUBER; ARNY, 1985), photosynthesis, protein 

synthesis, and membrane transport (HAFSI et al., 2014; MARSCHNER, 2012). Balanced K+ 

nutrition can promote plant tolerance to stress, as flooding (MUGNAI et al., 2011), low 

temperatures (DEVI et al., 2012), drought (WANG et al., 2013), and salinity (ALMEIDA et 

al., 2017). In contrast, plants under K+ deficiency showed oxidative stress status (HAFSI et al., 

2014; HASANUZZAMAN et al., 2018; HERNANDEZ et al., 2012) modifying heavily 

metabolite biosynthesis (ARMENGAUD et al., 2009; CUI et al., 2019; GAALICHE et al., 

2019). 

Potassium is responsible for activation of nearly 60 enzymes taking part in carbon and 

nitrogen metabolism (AMTMANN et al., 2008). Plants under adequate K+ levels synthesise 

large biomolecules, as cellulose, starch, and proteins (HAZANUZZAMAN et al., 2018). 

However, under K+ deficiency the content of small molecules, as free sugars, amino acids, 

organic acids, and amides increases (PRASAD et al., 2010). Also, we must not disregard that 

the impact of K+ supply in plant metabolism is multifaceted, tissue and species-specific, as 

observed in the responses to stresses in leaves and roots of tomatoes (WEINERT et al., 2021) 

and oil palm sapling (CUI et al., 2019).  

The prevalent view is that a high K+ availability decreases the severity of diseases, as  

caused by Cercospora kikuchii (ITO et al., 1993; MEYER; KLEPKER, 2007) and Phakopsora 

pachyrhizi (BALARDIN et al., 2006; PINHEIRO et al., 2011) or insect incidence in plants 

(PERRENOUD, 1990; SEVERTSON et al., 2016). Therefore, farmers are generally instructed 

to increase K+ rates to improve crop health (GAO et al., 2018; POTASH AND PHOSPHATE 

INSTITUTE, 1998; SEIXAS et al., 2020). In spite of the assumption that in most related cases 

K+ fertilization decreases the incidence of the diseases, the opposite effect is also reported 

(DAVIS et al., 2018; PERRENOUD, 1990), as a 63% decrease and 28% increase in insect and 

mite incidence in plants associated to potassium nutrition (AMTMANN et al., 2008; 

PERRENOUD, 1990). 

Potassium deficiency has also a considerable effect on plant metabolism, including 

alterations in the profile, concentration, and distribution of many metabolites in different tissues 



12 

(AMTANN et al., 2008; HAZANUZZAMAN et al., 2018). Likewise, there is a consensus that 

specialised metabolism is crucial for plant susceptibility and defence against insects or 

pathogens (AMTMANN et al., 2008; DAVIS et al., 2018; GAO et al., 2018). Nevertheless, 

how plants under K+ contrasting availability can cope with biotic stresses remains unclear, and 

according to Marschner (2012) the understanding of the impacts of K+ availability on 

specialised metabolism remains incipient when compared to the K+ effect on physiological and 

primary metabolism processes.  

Increase content of reducing sugars (primary metabolism) and glucosinolates 

(specialised metabolism) are potentially relevant to insects and/or pathogens, as detected in 

Arabidopsis thaliana under K+ deficiency (ARMENGAUD et al., 2009; TROUFFLARD et al., 

2010). Similarly, plants of Hordeum vulgare under K+ deficiency increased jasmonic acid and 

other oxylipins content sensitive to the pathogen Blumeria graminis (DAVIS et al., 2018). In 

Glycine max (L.) Merr. (Fabaceae) field experiments with K+ fertilization revealed multiple 

inducing mechanisms improving plant resistance to cyst nematode Heteroda glycines via root 

exudation of phenolic acids and plant pathogen-related genes (GAO et al., 2018). Furthermore, 

the metabolomic analysis revealed differences in gentisic acid and shikimic acid content from 

exudates of soybean roots, when plants are grown under low K+ conditions (TANTRIANI et 

al., 2020).  

Previous studies of soybean metabolite content influenced by K+ fertilization used only 

methodologies that allowed the detection of few target compounds (HU et al., 2015; SEGUIN; 

ZHENG, 2006; VYN et al., 2002) or analytical methods covering primary metabolism 

(TANTRIANI et al., 2020). Certainly, these results improved our knowledge, but not enough 

to understand how soybean under K+ deficiency or adequate nutrition has constitutive 

specialised metabolites that can respond differently in plant interactions. Recently, studies 

comprising integrated omics greatly contributed to elucidating the influence of nutrients on 

plant metabolism (CUI et al., 2019; TANTRIANI et al., 2020; ZHAO et al., 2020). Ionomics 

and metabolomics analyses showed that K+ deficiency and waterlogging in sunflower are not 

simply additive stresses, but also influence respiration as well nitrogen metabolism when 

associated with biomarkers (CUI et al., 2019). However, it remains unknown whether K+ 

availability can influence the constitutive content of soybean specialised metabolites with a 

recognized role in biotic and abiotic interactions; thus, research methodologies based on 

untargeted metabolomics allow investigating of metabolome changes on a larger scale (FENG 

et al., 2020; YANG et al., 2018). 



13 

We hypothesize that K+ rates in soil fertilization can promote an increase in specialised 

metabolite profiles of soybeans. This work aimed to advance a step towards understanding the 

effect of K+ nutrition on plant interactions, and it is expected that these results can provide new 

insights into the importance of potassium in soybean metabolism. Therefore, we carried out 

untargeted metabolomics and ionomics analysis to compare the metabolic profile of trifoliate 

leaves and pod tissues (valves and immature seeds) of soybean influenced by K+ availability. 

The contributions of this work allowed us to suggest the correlation of the soybean metabolite 

phenotype with K+ status and list its importance previously reported in biotic and abiotic 

interactions. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



14 

2 MATERIAL AND METHODS 

2.1 GENERAL EXPERIMENTAL PROCEDURES 

Leucine-enkephalin (Waters Technologies, USA) and quercetin (≥ 95% HPLC, Sigma-

Aldrich, USA) were used as external control standards. Formic acid LiChropur™ (98-100% 

HPLC, Merck, Germany), acetonitrile and methanol Chromasolv™ (≥ 99.9% LC-MS, 

Honeywell, Germany) were the solvents for running the samples in chromatographic analysis. 

Reference standards were applied to compare retention time, MS1 and MS2 data including, 

daidzin, genistin, glycitin, malonyl-daidzin, malonyl-genistin, malonyl-glycitin, acetyl-daidzin, 

acetyl-genistin, acetyl-glycitin, daidzein, genistein, glycitein, and coumestrol (≥ 95% HPLC; 

Sigma-Aldrich, USA). All aqueous solutions were prepared using a Milli-Q™ system-

producing ultrapure water (18.2 MΩ cm-1; Millipore, USA). 

 

2.2 EXPERIMENTAL DESIGN AND ASSAY 

The experiments were carried out at Embrapa Soybean Experimental Station (23°18' 

37'' S, 51°17' 68'' W; 590 m asl) in Londrina, Paraná State, Brazil, set on a Rhodic Hapludox 

(Staff, 2014). This area has flat and smooth wavy topography, with 760 g kg-1 of clay in the no-

tillage system (0 to 20 cm layer) of cultivation. The region has a humid subtropical climate 

(Cfa) according to the Köppen climate classification (ALVARES et al., 2013). Potassium (K+) 

availability conditions were tested after application of 0, 40, 80, and 160 kg K2O ha-1 (KCl; 

60% K2O). Subsequently, the soils at 0 to 0.2 m depth sites were randomly collected (15 sub-

samples) and homogenized to form a composite sample at each experimental unit in October 

2019 using a soil auger (0.05 m diameter) (Appendix A). Soil samples were analysed and 

allowed us to consider four levels (treatments): very low (KVL), low (KL), medium (KM), and 

very high (KVH) K+ availability. The nutrient soil content interpretation was based on Pauletti 

and Motta (2019) for regional conditions (Paraná State, Brazil).  

 Seeds of Glycine max (L.) Merr. (Fabaceae) cv. BRS 1010 IPRO (VI maturity group) 

were sown on November 2nd, 2019, with a row spacing of 0.5 m and 15 seeds per linear metre, 

expecting a density of 300,000 plants ha-1. The experimental design was a randomized block 

design, with plots of 4 m x 8 m, composed of four lines of 6 m, and six replicates. Before 

sowing, seeds were treated with cobalt (2 g ha-1), molybdenum (20 g ha-1), and inoculated with 

2 mL kg-1 of Bradyrhizobium japonicum SEMIA 5079 (OLIVEIRA JUNIOR et al., 2020). Soil 

fertilization was carried out with 111 kg ha-1 of triple superphosphate – TSP (45% P2O5) as 

recommended by Pauletti and Motta (2019).  



15 

2.3 PLANT MATERIALS  

Potassium is a very mobile nutrient in plants (MARSCHNER, 2012); therefore, 

potential differences in leaf metabolomic profiles can vary according to tissue maturity. Thus, 

we sampled trifoliate leaves of soybean plants at the V7 stage (FEHR; CAVINESS, 1977) when 

in vegetative stage the highest K+ uptake occurs (BENDER et al., 2015), and collected old and 

young leaves. Leaf samples were classified according to the growth region, lower third (LT), 

medium third (MT), and upper third (UT). These tissues were collected 49 days after sowing 

(DAS) and immediately packed in liquid nitrogen (six replicates). The fresh mass was measured 

on an AY220 Marte™ analytical balance (Shimadzu, Brazil) and stored at -80 °C in an 

ultrafreezer (Indrel Scientific, Brazil) at the Embrapa Chemical Ecology Laboratory. 

Subsequently, leaves were lyophilized at -55 °C for 48 h in an L101 freeze-dryer (Liotop, 

Brazil). Previous studies have associated potassium-deficient soils with the incidence of 

diseases caused by the pathogen C. kikuchii in soybean pods (MEYER; KLEPKER, 2007; 

SEIXAS et al., 2020). Therefore, we also collected soybean pods (R5.5 stage) at 98 DAS, and 

immature seeds (IS) were detached from the valves (PV) and separately maintained using the 

same storage process previously described for the analysis of trifoliate leaves. 

 

2.4 SOIL CHEMICAL ANALYSIS 

The collected soil samples were oven-dried at 45 °C for 48 h and sieved through a 2 

mm. Extractable Ca2+, Mg2+, and Al3+ in 1 mol L-1 KCl (MEHLICH, 1953) and plant-available 

P and K+ were evaluated using Mehlich-1 solution (MEHLICH, 1984). Soil pH was determined 

in 0.01 mol L-1 CaCl2 and organic carbon (OC) following the procedures described by Walkley 

and Black (1934). Chemical attributes from the experimental area are available in Appendix A.  

 

2.5 PLANT EXTRACTION 

The plant material (soybean leaves, immature seeds, and pod valves) was ground in a 

mortar with liquid nitrogen for metabolomics and ionomics analyses. Ultra-performance liquid 

chromatography quadrupole-time-of-flight mass spectrometry (UPLC-QTof-MSE) was used 

for metabolomics analysis at the Embrapa Chemical Ecology Laboratory. Plant ionomics was 

carried out in inductively coupled plasma optical emission spectroscopy (ICP-OES) at the 

Embrapa Plant and Soil Analysis Laboratory. 

 



16 

2.5.1 Extraction Process for Plant Metabolomics  

Metabolomic analysis was performed using a 90 mg aliquot of plant tissues in 15 mL 

Falcon tubes. For metabolite extraction, 3 mL of 80% methanol (MeOH: H2O; 80:20; v/v) was 

added to the plant tissue. Subsequently, samples were placed in an LSUC2-500-22.5 (60 Hz 

frequency, 1300 W power; Logen Scientific, Brazil) ultrasonic bath for 20 min, vortexed for 10 

s and centrifuged at 5,500 rpm, at 4 °C for 20 min in a Sorvall™ Legend™ X1R centrifuge 

(Rotor 75003623; Thermo Fisher Scientific, USA). The supernatant (3 mL) was filtered with 

Millex-HV Durapore™ 0.45 µm Acrodisc Filter Syringes (SLHVX13NL, PVDF; Merck 

Millipore, Germany) into glass tubes for drying in N2 flux in MA 4006 dry-block (400 W 

power; Marconi Equipamentos, Brazil). Samples were placed in a freezer at -20 °C until 

performing the analyses. For chromatographic analysis, samples were solubilized in 80% 

MeOH and filtered on Millex-HV Durapore™ 0.22 µm Acrodisc Filter Syringes 

(SLHVX13TL, PVDF; Merck Millipore, Germany), aiming at a final concentration of 2.5 mg 

mL-1.  

 

2.5.2 Extraction Process for Plant Ionomics  

Plant elemental content analysis was performed using a 150 mg aliquot of plant sample 

freeze-dried and ground in a mortar. Subsequently, the samples of plant tissue were digested in 

6 mL nitric acid (HNO3: H2O; 50:50; v/v) and 2 mL hydrogen peroxide (H2O2) in a 

MARSXpress microwave oven (CEM Corporation, Canada). The conditions followed 10 min 

of the heating ramp, maintaining a constant temperature at 170 °C for 15 min under 2 MPa 

pressure and a power of 1600 W (USPEA, 1996).  

 

2.6 INSTRUMENTAL ANALYSES 

2.6.1 Chromatographic Analysis (UPLC) 

Chromatographic analyses were performed using ACQUITY UPLC™ ultra-

performance liquid chromatography (Waters Corporation, USA) equipment set in a binary 

solvent system and sample manager. Aliquots (5 µL) of plant extracts were injected through an 

autosampler at 20 °C into an ACQUITY UPLC HSS C18 SB stationary reversed-phase column 

(100 mm x 2.1 mm; 1.8 µm; Waters Corporation, USA). The column temperature was 

maintained at 40 °C, with a flow rate of 0.4 mL min-1. The mobile phase gradient was as follows: 

(A) binary elution system composed of Milli-Q™ ultrapure water and 0.1% formic acid (H2O: 



17 

CH2O2; v/v) and (B) acetonitrile with 0.1% formic acid (ACN: CH2O2; v/v). The gradient system 

was set 2-95% B (0-15 min), 100% B (15.1-17 min) and reconditioned with 2% B (17.1-19.1 

min). Samples were injected in triplicate, randomly, and included blanks.  

 

2.6.2 Mass Spectrometry (ESI-QTof-MSE) 

A Xevo™ QTof-MS mass spectrometer (Waters Corporation, USA) equipped with a 

ZSpray™ source operating in negative electrospray ionization (ESI-) was used for analyses. Data 

acquisition mass range was set at 100 to 1100 Da and scan time at 0.25 s. MS Tune was 

configured for MSE Centroid mode data acquisition, 6 V low collision energy on MS1, and MS2 

function with ramp ranging from 20 to 40 V. The following MS parameters were applied in the 

analyses: 2.6 kV capillary voltage, 35 V sampling cone voltage, 150 °C supply temperature, 

and 1.5 V extraction cone voltage.  

Nitrogen was produced by an NM30LA-MS high-purity N2 generator (Peak Scientific 

Instrument™, Scotland), which was used as the cone gas, desolvation temperature, and flow at 

350 °C and 500 L h-1, respectively. Argon ≥ 99.9% purity (White Martins, Brazil) was used as 

a collision gas at a pressure of 1.3×10−2 mbar. Mass accuracy and reproducibility were ensured 

by leucine-enkephalin infusion into the lock mass system (400 ng mL-1), [M-H]- ion m/z 

554.2615, and every 20 s on the LockSpray flow probe at 20 µL min-1. The instrument was 

calibrated using a sodium formate standard solution and MassLynx™ 4.1 software (Waters 

Corporation, USA) used for data acquisition. Quercetin-3-O-rhamnoside was used as an 

external control standard (tR = 5.64 min; m/z = [M-H-Rha]- 301.0348) analysed every 10 

injections, these data of stability were registered and reported in Appendix G. 

 

2.6.3 Inductively Coupled Plasma (ICP-OES) 

Plant tissue ionomics analyses were performed in inductively coupled plasma optical 

emission spectroscopy (ICP-OES) Optima™ 8300 (Perkin Elmer, USA). The macro and 

micronutrient contents were measured using the calibration curve processed for each batch of 

samples (Table 1 and Table 2). 

 

 

 



18 

2.7 YIELD PARAMETERS  

At maturity soybean (R8 stage, 131 DAS), three rows of soybean 6 m in length were 

mechanically harvested to estimate the 100-seed dry weight (g) and soybean grain yield (kg ha-

1). The weight and moisture of the harvested seed were recorded, and the final seed yield was 

calculated based on a standard method of seed moisture content of 130 g kg-1 (13%) water 

content (Table 3), as also described by Firmano et al. (2020). 

 

2.8 DATA INTERPRETATION  

2.8.1 Univariate Analysis  

Yield parameters and ionomics data were submitted to analysis of variance (ANOVA), 

and means were compared using Tukey´s test (p-value ≤ 0.05). Statistical comparisons between 

leaves from different growing regions (LT, MT, and UT) were performed using ANOVA + 

Tukey´s test, and pod tissues (PV and IS) by two-sample t-test. ANOVA and Student's t-test 

were analysed in Minitab™ 18.1 software (Minitab Inc., USA) 

 

2.8.2 Pre-processing and Chemometrics analysis   

Soybean leaf and pod extracts from K+ treatments (KVL, KL, KM, and KVH) were 

analysed in a UPLC-QTof-MSE system, and raw mass data were processed using MarkerLynx 

XS™ 4.1 software (Waters Corporation, USA). The processing parameters for chromatographic 

data were adjusted by retention time interval of 0 to 14 min. Mass acquisition amplitude was 

maintained from 100 to 1100 Da, mass tolerance was set at 0.04 Da, and the minimum peak 

intensity limit was considered at 50 counts. The isotopic data removal function was employed 

with smoothing (Savitzky-Golay); the noise elimination level was 9, and the window and mass 

retention times were set at 0.02 and 0.20, respectively. The preprocessing dataset was imported 

into SIMCA™ EZinfo software (Umetrics™, Sweden) for multivariate statistical analysis and 

the Pareto method was used for data scaling (SALDANHA et al., 2020).  

Metabolomics (X) and ionomics (Y) analysis data arrays were submitted to multivariate 

analysis. Firstly, the dataset was employed the unsupervised model, based on principal 

component analysis (PCA-X&Y), followed by the supervised model, two-way orthogonal 

partial least squares discriminant analysis (O2PLS-DA). These chemometric tools help to 

reduce the dimension of the dataset and enable the visualization of variability sources. These 

models can join two multivariate matrices in a single model, allowing to visualize the 



19 

interaction and variance of both datasets (X-variable: untargeted metabolomics analysis; Y-

variable: targeted ionomics analysis). Models were evaluated from the R2 values representing 

the model's explanatory capacity, while Q2 suggested its predictive capacity (Appendix F). 

Supervised multivariate models were generated for comparative treatments KVL vs. KVH, and 

S-Plot was performed for variable selection. 

Potential markers were annotated based on four conditions: (I) selection of metabolites 

observed within the quadrant p(corr) [1] ≥ 0.6 and p[1] (loadings) ≥ 0.06 on the S-Plot graph; 

(II) variable importance in projection (VIP) value ≥ 1 was used; (III) variables were selected 

when p-value ≤ 0.05 in the CV-ANOVA test; and (IV) false discovery rate (FDR), proposed by 

Benjamini and Hochberg (1995) was performed for calculating p-value transformation into q-

value (TUGIZIMANA et al., 2016; VANDERPLANCK; GLAUSER, 2018; ZHAO et al., 

2017; ZHU et al., 2018). Fold change was calculated by dividing the mean intensity values of 

metabolites in KVL vs. KVH treatments (Table 4, Table 5, Table 6, and Table 7). 

 

2.8.3 Metabolite Annotation and Molecular Networking 

Metabolomics data were imported into MS-DIAL 4.70 software (TSUGAWA et al., 

2015, 2019) for deconvolution and MSE spectra alignment; posteriorly, putative identification 

was performed using MS-FINDER 3.52 (TSUGAWA et al., 2016). Heuristic rules were 

considered (SUMNER et al., 2007), and the mass tolerance was adjusted to 0.02 Da. 

Metabolites were putatively identified according to metabolic standards initiative (MSI) of level 

2.1 (SCHRIMPE-RUTLEDGE et al., 2016; SUMNER et al., 2007; TSUGAWA et al., 2019). 

These metabolites were annotated based occur previously reported in G. max or Fabaceae plants 

using the Human Metabolome Database (HMDB; https://hmdb.ca/), PubChem Database 

(https://pubchem.ncbi.nlm.nih.gov/), LipidMaps (LMSD; https://www.lipidmaps.org/), 

MassBank Database (https://mona.fiehnlab.ucdavis.edu/), ChemSpider 

(http://www.chemspider.com/), Kyoto Encyclopedia of Genes and Genomes Database (KEGG; 

https://www.genome.jp/kegg/compound/), Chemical Entities of Biological Interest (ChEBI; 

https://www.ebi.ac.uk/chebi/), FooDB (https://foodb.ca/), KnapSAcK 

(http://www.knapsackfamily.com/), and PlantCyc database (https://plantcyc.org/). A feature-

based molecular networking (FBMN) workflow (NOTHIAS et al., 2020) on global natural 

products social molecular networking (GNPS) (WANG et al., 2016) was created, and 

Cytoscape 3.8 (SHANNON et al., 2003) using yFiled Layout Algorithms 1.1 and chemViz2 

1.1 were used for molecular network manipulation, visualization, and analysis (Figure 3). 



20 

Chemical ontology descriptions, MS2 fragments, average intensity, mass theoretical, and 

measured data are available in Appendix G.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



21 

3. RESULTS AND DISCUSSION  

3.1 SOYBEAN LEAF IONOMICS ANALYSIS 

We evaluated the profile of nutrients and metabolites in soybean tissues under four soil 

nutrient conditions, very low (KVL; 0.05 cmolc dm-3), low (KL; 0.09 cmolc dm-3), medium 

(KM; 0.13 cmolc dm-3), and very high (KVH; 0.50 cmolc dm-3) K+ availability by Mehlich-1 

extraction (Appendix A). Potassium leaf content increased in response to K+ rates applied to 

the soil, as expected (Table 1). Potassium accumulation evaluated in very high K+ soil 

availability (KVH) was 3.4-fold (16.22 g kg-1) at UT, averaging 4.3-fold (12.08 g kg-1) in MT 

and LT leaves in comparison to very low K+ availability (KVL) treatment (Table 1 and 

Appendix B). Plant K+ content was associated with soybean tissue maturity (Table 1 and 

Appendix C). We observed that the upper third (UT) leaves had the highest K+ levels compared 

to the medium (MT) and lower third (LT) leaves, respectively. Previously, the high mobility of 

potassium was reported by Marschner (2012) and Mengel et al. (2001) who also related that 

due to higher demand for K+ in metabolic activities, potassium is often transported, 

redistributed, and accumulated primarily in the young tissues.  

Leaf calcium (Ca2+) and magnesium (Mg2+) contents decreased significantly with 

increasing K+ rates; phosphorus (H2PO4
-) and sulphur (S) levels were not altered by treatments 

(Table 1). However, the mean content of S tended to be higher in MT and UT (> 2.0 g kg-1), 

compared with the LT (Table 1 and Appendix C). Other authors reported a reduction in Ca2+ 

and Mg2+ content by increases in K+ soil fertilization (DIEM; GODBOLD, 1993; FIRMANO 

et al., 2020; HAFSI et al., 2014). The high K+ level in the soil solution could competitively 

inhibit the uptake of Ca2+ and Mg2+ by the plant root absorption site (CASTRO et al., 2021; 

FIRMANO et al., 2020). In fact, the increased content of these macronutrients (Ca2+/Mg2+) was 

also detected by Hafsi et al. (2014) in leaves under K+ deficiency, which they suggested being 

associated with alleviating the effects of nutritional deficiency. Management programs to obtain 

high K+ rates or getting soils with high natural contents of K+ can lead to a lower uptake of 

Mg2+ and even Ca2+, that generally triggers yield reductions (CASTRO et al., 2021). 

The contents of micronutrients in soybean old leaves (LT), such as manganese (Mn2+), 

iron (Fe2+), copper (Cu2+), and boron (B) significantly increased by 1.04-, 1.11-, 1.23-, and 

1.88-fold under KVL compared to KVH (Table 1). Regardless of the treatment, Fe2+ and Mn2+ 

concentrations were lower in MT and UT leaves in comparison to LT (Table 1 and Appendix 

C). However, zinc (Zn2+) leaf content in the three growing regions (LT, MT, and UT) was 

reduced when soybean plants grown under very low potassium availability (KVL).  



22 

These experimental results suggest that soybean leaves can resist low potassium 

availability by accumulating Ca2+, Mg2+, Mn2+, Fe2+, Cu2+, and B. Integrated omics of wild 

soybean (Glycine soja) under low phosphate availability showed an accumulation of Ca2+, 

Mg2+, Fe3
+, NO3

- and S in young leaves, and reduced Zn2+ levels as related by Shen et al. (2021). 

These results suggest that this interaction (P and K+ deficiency) should be investigated in the 

future for a better understanding of the mechanisms of soybean tolerance to nutrient deficiency.   

Iron usually acts as a mineral regulatory element and defends plants against stresses by 

activating plant enzymatic antioxidants (TRIPATHI et al., 2018). The reduction in boron levels 

in the soybean leaves in conditions of increased K+ rates (Table 1) was probably associated with 

its low phloem mobility (MARSCHNER, 2012). Moreover, a reduction in B availability may 

also occur due to low Ca2+ uptake, revealed in plants that grown under high K+ soil content, 

which reduced the B requirement (FIRMANO, 2017; FIRMANO et al., 2020). In general, our 

results demonstrate that the content of Mn2+, Zn2+, and Cu2+ in the leaves are poorly or non-

responsive to increasing K+ levels in the soil. In fact, the concentration of these nutrients in 

soybean leaves (cv. BRS 1010 IPRO) were also not affected in the agronomic assays reported 

by Firmano (2017), being within or above the class of contents considered appropriate in 

soybean plants (HARGER, 2008). 

 

 

 

 

 

 

 

 



23 

Table 1 - Ionomic analysis of soybean trifoliate leaves (cv. BRS 1010 IPRO) in response to K+ soil fertilization 

Treatments 
 Leaf 

DW   P K+ Ca2+ Mg2+ S   Zn2+ Mn2+ Fe2+ Cu2+ B 

 (g)  (g kg-1)  (mg kg-1) 
KVL LT  0.22  2.47ªB 2.73cB 22.9ªA 11.37aA 1.67aB  43.17bA 279.72ªA 398.17aA 12.30aA 64.75ªA 
KL LT  0.24  2.44ªB 7.25bB 19.9ªA 5.78bA 1.75ªB  50.83aA 282.28ªA 389.08bA 9.95bA 38.59bB 
KM LT  0.25  2.39ªB 8.84bB 14.5bA 4.17cA 1.65ªC  51.39aA 270.50bA 353.50cA 8.95bA 37.67bB 

KVH LT  0.25  2.41ªB 12.02aB 12.0bA 3.27cA 1.62ªB  50.39aA 268.44bA 358.17cA 9.95bA 34.32bB                 
KVL MT  0.31  2.60ªA 2.87cB 20.1ªB 9.38aB 2.27ªA  35.94bB 248.22ªB 269.33ªB 10.60ªA 58.38ªA 
KL MT  0.37  2.62ªA 7.97bB 16.2bB 5.07bB 2.23ªA  44.44aB 251.61ªB 226.50bB 9.60ªA 37.33bB 
KM MT  0.35  2.59ªA 9.40bB 12.2cA 3.67cB 2.14ªB  46.50aB 246.61aB 200.67cB 8.22ªA 29.90cC 

KVH MT  0.36  2.56ªA 12.14ªB 11.5cA 3.02cAB 2.16ªA  47.06aB 254.06ªB 196.17cB 8.57aB 30.85cB                 
KVL UT  0.38  2.60ªA 4.70dA 17.2ªC 7.75aC 2.39ªA  25.06bC 137.28ªC 263.16ªB 11.07aA 61.95ªA 
KL UT  0.44  2.61ªA 9.88cA 11.0bC 3.53bC 2.37ªA  28.00aC 133.44ªC 221.83bB 9.32bA 44.48bA 
KM UT  0.44  2.57ªA 13.64bA 9.60cB 3.07bC 2.31ªA  28.56ªC 128.50ªC 182.67cB 8.18bA 41.35bA 

KVH UT   0.43   2.53aA 16.22aA 9.50cB 2.78bB 2.37aA   27.89ªC 130.06aC 183.19cB 7.48bC 39.75bA 

Note: Leaf DW – leaf dry weight; KVL – very low K+ availability; KL – low K+ availability; KM – medium K+ availability; KVH – very high K+ availability; LT – lower 
third; MT – medium third; UT– upper third; The data presented are representations of the mean (n = 6) and results of ANOVA + Tukey´s test (p ≤ 0.05); Different lower 
case letters indicate differences amongst K+ availability conditions (KVL, KL, KM, and KVH) in the same soybean tissue; Different capital letters indicate differences 
amongst soybean trifoliate leaves (LT, MT, and UT) in the same K+ treatment. 

Source: The Autor (2022).  



24 

3.2 SOYBEAN POD IONOMICS ANALYSIS 

Ionomics analysis showed that the K+ content increased 1.6-fold in immature seeds (IS), 

and 5.0-fold in pod valves (PV) when KVH treatment was compared to KVL (Table 2 and 

Appendix D). The increased content of K+ in immature seeds was similar for KVH and KM, 

but higher than that assimilated by IS of plants under KL and KVH treatment. In general, our 

results obtained for K+ accumulation in soybean seeds were similar to those reported by Parvej 

et al. (2015) and Vyn et al. (2002). The seeds, as propagation organs of the species, tend to 

receive a greater nutrient amount, even in plants with a nutritional deficiency status (Marschner, 

2012).  

The pod valves were more affected by low potassium availability (KVL and KL 

treatments); they presented mean levels of K+ accumulation > 2.50 g kg-1, compared with 

immature seeds (> 7.07 g kg-1). Other macronutrients (P, Ca2+, Mg2+, and S) were not 

significantly altered by treatments in soybean pod tissues. In general, the P content observed in 

IS (< 3.23 g kg-1) is higher than detected in PV (< 2.63 g kg-1) by the two-sample t-test, except 

in the KM-PV vs. KM-IS (Table 2 and Appendix E). However, the accumulation of Ca2+ and 

Mg2+ was greater in PV (> 7.03 and 3.73 g kg-1, respectively). 

The concentration of Zn2+ and Mn2+ were higher in the PV of soybean plants in soils 

with medium (KM) and very high K+ availability (KVH) (Table 2). The same tendency was 

observed for Mn2+ in IS. These results indicated that low K+ soil availability increased Mn2+ 

accumulation in soybean pod tissues (R5.5 stage). Copper content (Cu2+) responded only in pod 

valves when soybean plants were exposed to very low and low potassium treatments (KVL and 

KL). Immature seeds (IS) showed increased Fe2+ concentration in KVL compared to other K+ 

treatments (< 53.6 mg kg-1). Under our experimental conditions, no significant changes in the 

content of Zn2+, Cu2+, and B were observed in IS, and Fe2+ and B in PV considering the four K+ 

availability conditions (Table 2). 

The effect of potassium availability on the elemental content of vegetative and 

reproductive tissues investigated reveals that Ca2+ and Mg2+ responded to K+ rates only in 

leaves, but this interaction did not occur in soybean pod tissues (R5.5 stage). Furthermore, 

plants grown under very low and low K+ content (KVL and KL) did not accumulate Ca2+, Mg2+, 

Mn2+, and B in pod tissues, as observed in the results of trifoliate leaves (especially LT leaves). 

However, Cu2+ concentration increase occurred only in pod valves and leaves under low K+ 

availability (Table 1 and Table 2) and Fe2+ content only responded to immature seeds (IS) in 

comparison to pod valves.  



25 

Table 2 - Ionomic analysis of soybean pods (cv. BRS 1010 IPRO) sampled at R5.5 growth stage in response to K+ soil fertilization 

Treatments 
  P K+ Ca2+ Mg2+ S   Zn2+ Mn2+ Fe2+ Cu2+ B 
 (g kg-1)  (mg kg-1) 

KVL PV  2.63ªB 2.50dB 7.03aA 4.47ªA 1.06ªB  6.83cB 33.00cA 69.33ªA 10.27ªB 27.13ªA 
KL PV  2.55ªB 5.13cB 7.37aA 4.47ªA 1.13ªB  9.97bA 40.67bA 71.67ªA 10.57ªB 29.30ªA 
KM PV  2.57ªA 10.53bA 9.50ªA 3.93ªA 1.19ªB  11.20ªA 68.67ªA 68.67ªA 8.70bB 34.30ªA 

KVH PV  2.53ªB 12.57ªA 8.37ªA 3.73ªA 1.05ªB  13.60ªA 65.67aA 71.33aA 7.53bB 31.77ªA               
KVL IS  3.23ªA 7.07bA 2.37ªB 1.53ªB 1.76ªA  11.43aA 24.00bB 58.67ªA 21.73ªA 15.97ªB 
KL IS  3.14ªA 7.87bA 2.37ªB 1.43ªB 1.73ªA  10.50ªA 23.67bB 51.67bB 18.97ªA 15.73ªB 
KM IS  3.15ªA 11.47ªA 2.47ªB 1.50ªB 1.91ªA  12.83ªA 36.00aB 52.33bB 20.23ªA 17.52ªB 

KVH IS   3.12ªA 11.73ªB 2.43ªB 1.50aB 1.87aA   11.70ªA 35.00aB 53.67bA 19.50aA 17.87aB 

Note: KVL – very low K+ availability; KL – low K+ availability; KM – medium K+ availability; KVH – very high K+ availability; IS - immature seeds; PV – pod valves; 
The data presented are the mean values (n = 6); Different lower case letters indicate differences amongst K+ availability conditions (KVL, KL, KM, and KVH) in the same 
soybean tissue by Tukey´s test (p ≤ 0.05); Different capital letters indicate differences amongst soybean pod tissues (PV and IS) in the same K+ treatment by two-sample t-
test (p ≤ 0.05). 

Source: The Autor (2022). 

 

 

 

 

 



26 

 An indirect effect of the increase of K+ fertilization reported by Firmano et al. (2020)  

was greater Mn2+ accumulation in soybean seeds, with stabilization of this concentration after 

soil rate fertilization (> 120 kg K2O ha-1). The Mn2+ increase content in IS can be associated 

with its activity in energetic metabolism (RANADE-MALVI, 2011). The high K+ availability 

can also promote the major metabolic activity of plants, which leads to greater Mn2+ content in 

seeds (FIRMANO, 2017). Overall, our results are consistent in the relationship of nutritional 

balance of macro and micronutrients when soybean plants were under the contrasting conditions 

of K+ availability (MARSCHNER, 2012). 

 

3.3 YIELD PARAMETERS 

The dry weight of 100 seeds and the yield parameters increased with the K+ rates (Table 

3). The 100-seed dry weight was higher in the KVH (15.5 g) and KM (15.6 g) than in the KL 

(14.1 g) and KVL (12.4 g) treatments, as expected. The yield followed the same trend; the 

highest means were achieved by KVH (4096 kg ha-1), and KM (4093 kg ha-1) treatment plots, 

in contrast to KL (3737 kg ha-1) and KVL (2211 kg ha-1). Parvej et al. (2015) related yield 

increase of 40 and 60% when soybean grown under medium and high K+ soil fertility (81 and 

91 mg K+ kg-1, respectively), compared with the low K+ treatment (61 mg K+ kg-1). A previous 

study in the same edaphoclimatic conditions of ours conducted by Firmano et al. (2020) showed 

that soybean yield stabilised when the K+ soil content extracted by Mehlich-1 reaches 0.15 

cmolc dm-3. In our analysis, the potassium soil concentration in KM treatment was 0.13 cmolc 

dm-3 and 0.50 cmolc dm-3 in KVH (Appendix A), justifying the soybean yield similarity between 

both treatments. 

 

Table 3 - Soybean (cv. BRS 1010 IPRO) grains yield parameters as function of K+ soil 
availability 

Note: KVL – very low K+ availability; KL – low K+ availability; KM – medium K+ availability; KVH – very high 
K+ availability. The data presented are the mean values (n = 6) obtained from ANOVA. Means followed by the 
same letter in the column do not differ from each other by Tukey´s test (p ≤ 0.05).  
 

Source: The Autor (2022). 

Treatments 
 100-Seeds DW Yield 
 (g) (kg ha-1) 

KVL  12.4c 2211c 
KL  14.1b 3737b 
KM  15.6a 4093a 

KVH  15.5a 4096a 



27 

3.4 CHEMOMETRICS ANALYSIS 

The PCA-X&Y models indicate the response of the metabolite profile to K+ availability. 

In our study, we established two-blocks (X = intensities of metabolites; Y = ion concentration) 

as a bidirectional multivariate method that aims to separate the covariance between two data 

sets. Graphs of PCA-X&Y indicate the response of metabolite profile to K+ availability. PCA-

X&Y_A was created for trifoliate leaf samples and shows 404 variables (Figure 1a). The Pareto 

scale scored 21 components, explained by 83% R2X(cum) variance. The first two main 

components accounted for 45.3% of the total variance (R2X [1] = 0.317 and R2X [2] = 0.136). 

PCA-X&Y_A also showed distant points representative of the leaf growth regions (LT, MT, 

and UT) in PC1. The O2PLS-DA_B model exhibited similar clustering, with better separation 

of the treatments in leaf samples (Figure 1b). Cluster I have 9.1 to -14.5 variance in PC2, and 

in general, the O2PLS-DA_B score plot showed two main components, cluster I (KL, KM, and 

KVH) and cluster II (KVL), observable by PC2. 

 Loading plots of trifoliate leaves reveal the influence of K+ content measured by 

ionomic analysis (Figure 1c), as previously reported (2.1 section). Potassium fertilization had a 

greater contribution to cluster I (KL, KM, and KVH), predominantly on the upper third (UT) 

leaves in both treatments (Figure 1b, c); Zn2+ and Mn2+ content also contribute to PC2, but 

mostly for the lower third (LT) and medium third (MT) leaves. The concentration of Ca2+, Mg2+, 

Cu2+, Fe2+, and B were associated with the leaves of plants that grow under very low potassium 

availability (KVL).  

In the PCA-X&Y_D model, the pod tissue samples dataset used was composed of 221 

variables, with seven components explaining 89% of the variance in R2X(cum) (Figure 1d). Two 

main components explained 75.4% of the variance (R2X [1] = 0.697 and R2X [2] = 0.057), 

revealing low treatment separation characteristics for PC2. Pod valves (PV) and immature seeds 

(IS) samples were distant from the projection (PC1). The use of O2PLS-DA_E (supervised 

model) was necessary for treatment distinctions (Figure 1e) since only PCA-X&Y_D did not 

clearly explain the relationship between the K+ rates and metabolite content in pod tissues. The 

clusters in the supervised multivariate model (O2PLS-DA_E) from the pod tissues and the 

results of ANOVA (Tukey’s test) were similar in regard to K+ availability (Table 2). Four 

groups were observed in the O2PLS-DA_E score plot, cluster I (KM-IS and KVH-IS), cluster 

II (KL-IS and KVL-IS), cluster III (KM-PV and KVH-PV), and cluster IV (KL-PV and KVL-

PV), represented by Figure 1e. 



28 

Chemometric models showed that K+ rates applied in soil fertilization were represented 

by PC2, while different sampling leaf regions (LT, MT, and UT) and pod tissues (PV and IS) 

were explained by PC1 (Figure 1). The positive quadrant of PC2 in the plot of IS and PV 

metabolic profiles reveals that medium and very high potassium availability (KM and KVH) 

treatments showed a correlation of K+ content measured in ionomic analysis (Figure 1f). 

However, the variance between the K+ availability (KVL, KL, KM, and KVH) was significantly 

lower in pod tissues (3.33% by PC2; Figure 1e) compared to trifoliate leaves (11.5% by PC2; 

Figure 1b). These results suggest that potassium rates promote more variance in leaf metabolite 

content than in pod valves and immature seeds, respectively. 

The O2PLS-DA models (Appendix F), established from the metabolomics/ionomics 

datasets for soybean leaves and pods, explained 99% R2Y(cum)
 and predicted 98% Q2Y(cum) of 

the total variance, indicating the adequacy of the model, and allowing the establishment of S-

Plot graphs, using only KVL vs. KVH treatments. We observed that soybean under very high 

potassium availability (KVH) accumulated differentially 17 and 7 compounds in leaves and 

pods, respectively, compared with the lowest K+ availability (Table 4 and 5). Under very low 

potassium availability, O2PLS-DA models (KVL vs. KVH) revealed that 21 and 18 metabolites 

were upregulated in trifoliate leaves and pod tissues, respectively (Table 6 and 7). These 

experimental results indicated that K+ rates did not contribute to linearly increasing constitutive 

specialised metabolites in soybean development stages sampled, rejecting our initial 

hypothesis. However, our data showed that under very low potassium availability soybean 

metabolism had the highest metabolic shift, as also observed previously by Shen et al. (2021) 

with G. soja deficient in phosphorus. 

 

Figure 1 - Unsupervised and supervised chemometric models of soybean plants under four soil 

K+ availability. These models allowed us to correlate the metabolomics data (404 and 221 

molecular features to soybean leaves and pod tissues, respectively as X input) with ionomics 

(10 factors as Y input) coherently. (a) PCA-X&Y_A scatter plot of trifoliate leaves. (b) O2PLS-

DA_B score plot highlighting the identified two clusters (C-I and C-II) of trifoliate leaves. (c) 

O2PLS-DA_B loading plot of trifoliate leaves, the loadings (factor) in the graph represent 

macro and micronutrients quantified by ICP-OES that contribute to the O2PLS-DA model. (d) 

PCA-X&Y_D scatter plot of soybean pod tissues. (e) O2PLS-DA_E score plot highlighting the 

identified four clusters (C-I, C-II, C-III, and C-IV) of soybean pod tissues. (f) O2PLS-DA_E 

loading plot of soybean pod tissues, the loadings (factor) in the graph represent nutrients 

quantified by ICP-OES that contribute to the O2PLS-DA model. List of abbreviations: LT – 



29 

lower third leaves; MT – medium third leaves; UT– upper third leaves; IS - immature seeds; 

PV – pod valves; KVL – very low K+ availability; KL – low K+ availability; KM – medium K+ 

availability; KVH – very high K+ availability. 

 

 
Source:  The Autor (2022). 

 

 



30 

The multivariate analysis grouped the leaf samples according to the growth region (LT, 

MT, and UT). The progressive metabolite accumulation may be a response to observed 

differences in tissue maturity, also reported by Song et al. (2014) and Yuk et al. (2011). These 

authors associated the expressive accumulation of coumestrol and glyceollin VII in soybean 

leaves occur at R7/R8 stage. Likewise, wild soybean under low phosphate (Pi) conditions 

promoted the transportation and reuse of sugars and amino acid metabolites and mobilize Pi 

from hexose-phosphate from old to young leaves (SHEN et al., 2021). Unlike K+, phosphorus 

is a structural component of many molecules in plants, which justifies major metabolome 

reprogramming. However, potassium in plants is more associated with catalysis or activation 

of enzymes (CUI; TCHERKEZ, 2021; MARSCHNER, 2012). Soybean under potassium 

homeostasis should have a sufficient level of this nutrient for enzymatic activity, not resulting 

in an increase in the contents of constitutive specialised metabolite, as shown by our 

experiment. Otherwise, K+ deficiency leads to relevant metabolic changes (LU et al., 2019; 

TROUFFLARD et al., 2010), as also observed to low nitrogen (LIU et al., 2019) or phosphorus 

(SHEN et al., 2021) conditions.  

 

3.5 MOLECULAR NETWORKING AND METABOLIC PATHWAY  

Feature-based molecular networking analysis was carried out to gain a comprehensive 

overview of the differentially regulated metabolites in response to K+ availability, 

complementing the data indicated by the base peak intensity (BPI) in chromatograms and the 

differential chemical profile of soybean tissues (Figure 2). Molecular networking showed the 

connection of  902 nodes, 1430 edges, and 311 components that resulted in 89 spectral 

molecular families putative annotated through an automated library spectral matching (Figure 

3). The features highlighted by the chemometric analysis were identified and color-coded in the 

molecular network. This strategy allowed us to highlight the compounds impacted by K+ 

availability and relate them to chemical structures. Spectra not clustered into molecular families 

were represented as self-loop nodes at the bottom of the network; colored nodes represent 

ontology groups of soybean metabolites regulated by K+ fertilization. The lack of 

comprehensive spectral libraries associated with characteristics of our MS1 and MS2 data 

(MSE/DIA – data-independent acquisition) can explain the considerable number of spectra not 

clustered. 

Chemometrics analysis showed 51 metabolites differently regulated in response to K+ 

availability (Figure 2). The chemical ontology analysis of metabolites influenced by K+ 



31 

fertilization showed regulation of sixteen phenolic compounds, members of the groups of 

flavonols, isoflavones, coumestans, and pterocarpans; eight lipids classified as jasmonic acid 

conjugates, fatty acids, and membrane lipids; Additionally, nine terpenoids were putatively 

identified as sesquiterpenoid and triterpenoid saponins, and seven were carbohydrates; besides 

of eleven compounds classified as organic acids, amino acids, and nucleic acids (Appendix G).  

 

Figure 2 - Soybean tissues representation and base peak intensity (BPI) mass chromatograms 

(UPLC-QTof-MSE) in negative ion mode (ESI-) displaying comparative metabolomic profile 

differences. The corresponding list of metabolites annotated in the chromatograms is available 

in Appendix G 

 

Fifty-one annotated metabolites of soybean plants influenced by contrasting soil 

potassium availability were correlated using the KEGG database. These compounds were 

linked to nineteen metabolic pathways, such as alanine, aspartate, glutamate, histidine, β-

alanine, citrate cycle, glyoxylate and glucuronate interconversion, nitrogen, galactolipid, 

glycerophospholipid, α-linolenic acid, purine and pyrimidine metabolism, arginine, aminoacyl-

tRNA, galactosylcyclitol, phenylpropanoid, flavonol, isoflavonoid, sesquiterpenoid, and 

triterpenoid biosynthesis. Overall, these KEGG pathways belong to the pathway maps that 

include carbohydrates, amino acids, nucleotides, lipids, specialised metabolites, and energetic 

metabolism (Appendix G).  



32 

Figure 3 - Complete feature-based molecular network (FBMN) in global natural product social 

molecular networking (GNPS) of soybean trifoliate leaves and pod tissues influenced by K+ 

availability. Nodes represent MS2 spectra and are connected based on spectral similarity defined 

(cosine score ≥ 0.8), matched fragment ion (5), and network TopK (10), encompassing 902 

nodes and 1430 edges organized in 89 spectral molecular families. Large coloured nodes 

represent metabolites influenced by K+ nutrition identified by chemometrics models and 

representative chemical ontology 

 
Source:  The Autor (2022). 



33 

3.6 SOYBEAN LEAVES UNDER K+ HOMEOSTASIS  

Flavonol glycosides, as kaempferol 3-glucosyl-(1→2)-gentiobioside, kaempferol-3-O-

α-L-rhamnopyranosyl-(1→6)-β-D-galactopyranoside, and kaempferol-3-O-rutinoside were 

upregulated in soybean leaves of fertilized plants (KVH) when compared to very low potassium 

availability (KVL) treatment (Table 4). Flavanols are usually associated with pathogen, insect 

resistance (DILLON et al., 2017), and plant protection from UV-B effects (HARBORNE; 

WILLIAMS, 2000). Previous studies also associated that K+ nutrition provoked an increase in 

phenolic compounds, such as ferulic, cinnamic, salicylic, and rosmarinic acids (GAO et al., 

2018; NGUYEN et al., 2010; PRASAD et al., 2010). Indeed, K+ favours metabolic reactions 

and the turgidity of cells, maintaining the leaf orientation for major light interception 

(MALAVOLTA, 1980); these events suggest that differential regulation of phenolic 

compounds can be associated with UV-B. 

Monogalactosyldiacylglycerol (18:3) and lyso-monogalactosyldiacylglycerol 

(16:0/0:0) content also were upregulated (6.1- and 3.3-fold, respectively) in leaves of upper 

third, when soybean plants grown under KVH compared to KVL treatment (Table 4). In rice, 

galactolipid biosynthesis was downregulated in seedlings deprived of K+ nutrition (SHANKAR 

et al., 2013). Monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG) 

are major galactolipids constituents of chloroplast photosynthetic membranes (MURAKAWA 

et al., 2014), deficiency of these metabolites is associated with plant light sensitivity (JARVIS 

et al., 2000). Nitrogen deficiency in A. thaliana led to a decrease in MGDG and a concomitant 

increase in DGDG, resulting in a coordinated breakdown of galactolipids and chlorophyll 

(GAUDE et al., 2007).  

Soyasaponins can occur constitutively in any soybean tissue (TSUNO et al., 2018; 

YATES et al., 2021). In our analyses, soybean plants under potassium homeostasis (KVH) 

showed differential regulation (upregulation) of three triterpenoid saponins, as soyasaponin Ba 

(3.4-fold), soyasaponin γg (3.9-fold), and soyasaponin Bc´ (3.4-fold) in the UT leaves (Table 

4). Nevertheless, soybean leaves under very low potassium availability (KVL) also revealed 

upregulation of four additional soyasaponins. Soyasaponin Bc´ and soyasaponin γg have the 

same structure, except for the conjugate 2,3-dihydro-2,5-dihydroxy-6-methyl-4H-pyran-4-one 

(DDMP) moiety. Upregulation of soyasaponin γg and soyasaponin Bc´ biosynthesis (> 3.3-

fold) were also detected after abiotic stress, as occurred in common bean seeds of plants grown 

under severe drought conditions (HERRERA et al., 2019). Shankar et al. (2013) revealed that 

when K+ privation is imposed on rice seedlings, gene expression of terpenoid biosynthesis is 



34 

reduced. Moreover, other macronutrients, such as nitrogen and phosphate availability can alter 

saponin content (SZAKIEL et al., 2011). 

 

3.7 SOYBEAN POD TISSUES UNDER K+ HOMEOSTASIS 

Metabolomics analysis of pod tissues showed increased concentration of kaempferol-3-

O-digalactopyranoside (1.5-fold), malonyl-genistin (1.5-fold), β-D-glucopyranosyl-11-

hydroxyjasmonic acid (1.7-fold), trihydroxy-octadecadienoic acid (2.7-fold), and soyasaponin 

Bb´ (1.8-fold) in plants grown under KVH in contrast with KVL treatment (Table 5). Isoflavone 

content increases in soybean seeds when plants are grown under K+ rates were reported by 

Bellaloui et al. (2013) and Vyn et al. (2002). Especially, Seguin and Zheng (2006) described 

the difference in isoflavonoid content was not significant when K+ was applied in medium-to 

high-fertility soils. In fact, we expected that the upregulation of flavonoids and isoflavones in 

soybean plants caused by K+ fertilization would be higher in immature seed extracts. 

Nevertheless, it is necessary to consider that this regulation is also dependent on yearly 

environmental factors (drought and temperature), edaphic conditions, and potassium soil 

content. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



35 

Table 4 - Putative metabolite identification from soybean leaves upregulated under very high potassium (KVH) availability by UPLC-QTof-MSE 

Note: Metabolites were annotated by level 2.1 of the metabolomics standards initiative; tR – retention time; KVL – very low K+ availability; KVH – very high K+ availability; LT – lower third; MT – 
medium third; UT– upper third; “ * ” - identified using authentic standards; “ – ” - not significant; O2PLS-DA and S-Plot models were performed using p-value ≤ 0.05 and false discovery rate (FDR) 
with q-value ≤ 0.05 (Appendix G); Importance variable in score projection (VIP ≥ 1.0) were considered; Fold change (FC) refers to metabolite relative intensity values between KVH vs. KVL 
treatments (metabolite content is higher in KVH than KVL treatment); MS1 and MS2 data, chemical ontology and KEGG Pathway are available in Appendix G. 

Source:  The Autor (2022). 

tR 
(min) Formula Identification Chemical Ontology 

 Lower Third (LT)  Medium Third (MT)  Upper Third (UT) 

KVL-LT vs. KVH-LT KVL-MT vs. KVH-MT KVL-UT vs. KVH-UT 
 FC FDR VIP  FC FDR VIP  FC FDR VIP 

0.66 C12H22O11 Sucrose Carbohydrates −− −− −− 2.1 4.17E-02 2.2 3.3 4.00E-02 1.8 

0.73 C5H10O5 D-xylose Carbohydrates  16.1 3.57E-03 1.6  15.1 4.17E-03 1.7  11.7 3.33E-03 1.2 

0.73 C4H6O5 Malic acid Beta hydroxy acids  7.8 7.14E-03 1.2  9.9 1.25E-02 1.3  2.1 3.33E-02 0.7 

3.40 C33H40O21 Kaempferol-3-glucosyl-(1→2)-gentiobioside Flavonols  1.4 1.79E-02 2.2  1.4 3.33E-02 2.2  1.4 5.00E-02 1.1 

4.35 C27H30O15 Kaempferol-3-O-α-L rhamnopyranosyl-(1→6)-β-D-
galactopyranoside Flavonols  4.9 1.07E-02 1.7  4.8 8.33E-03 2.1  5.0 4.67E-02 0.6 

4.51 C27H30O15 Kaempferol-3-O-rutinoside Flavonols  4.6 1.43E-02 1.8  3.5 2.50E-02 2.1  2.9 1.67E-02 1.8 

5.03 C24H22O13 Malonyl-genistin* Isoflavones  1.4 3.93E-02 2.4  −− −− −−  −− −− −− 

6.30 C18H32O5 Trihydroxy-octadecadienoic acid Fatty acids  1.3 4.64E-02 1.3  1.3 4.58E-02 1.9  1.3 4.33E-02 1.4 

6.49 C15H10O5 Genistein* Isoflavones  1.6 2.86E-02 3.0  −− −− −−  −− −− −− 

6.61 C48H78O19 Soyasaponin Ba (V) Triterpenoid saponins  1.4 2.14E-02 1.2  −− −− −−  3.4 3.67E-02 1.1 

8.70 C48H86O18P2 Phosphoinositide phosphate Glycerophosphoinositol 
phosphates 

 −− −− −−  2.4 2.92E-02 1.4  2.5 2.33E-02 1.3 

9.57 C48H74O17 Soyasaponin γg Triterpenoid saponins  −− −− −−  −− −− −−  3.9 3.00E-02 1.5 

9.90 C45H74O11 Monogalactosyldiacylglycerol (18:3) Monogalactosyldiacylglycerols  2.8 2.50E-02 2.1  8.7 1.67E-02 4.1  6.1 6.67E-03 3.7 

10.15 C31H58O14 Lyso-monogalactosyldiacylglycerols (16:0/0:0) Lyso-
monogalactosyldiacylglycerols 

 −− −− −−  2.2 2.08E-02 1.9  3.3 1.00E-02 2.7 

11.74 C22H43O9P Phosphatidylglycerol (16:1(9Z)/0:0) Phosphatidylglycerols  1.9 3.21E-02 1.8  −− −− −−  2.2 2.00E-02 2.4 

11.90 C41H66O13 Soyasaponin Bc' (IV) Triterpenoid saponins  2.1 3.57E-02 1.2  2.0 3.75E-02 1.5  3.4 1.33E-02 1.8 

13.64 C46H75O10P Phosphatidylglycerol (18:2/22:6) Phosphatidylglycerols  −− −− −−  −− −− −−  2.9 2.67E-02 2.0 



36 

Table 5. Putative metabolite identification from soybean pod tissues upregulated under very high potassium (KVH) availability by UPLC-QTof-MSE 

tR 
(min) Formula Identification Chemical Ontology FC FDR VIP 

  Immature Seeds  
6.33 C18H32O5 Trihydroxy-octadecadienoic acid Fatty acids 2.7 4.58E-02 1.7 
8.61 C42H68O14 Soyasaponin Bb' (III) Triterpenoid saponins 1.8 8.33E-03 1.9 

  Pod Valves  
0.70 C6H8O7 Citric acid Tricarboxylic acids 1.6 2.00E-02 1.1 
0.70 C4H4N2O2 Uracil Pyrimidones 1.5 4.00E-02 2.2 
2.79 C18H28O9 β-D-glucopyranosyl-11-hydroxyjasmonic acid Fatty acyl glycosides 1.7 3.33E-03 1.3 
3.86 C27H30O16 Kampferol-3-O-digalactopyranoside Flavonols 1.5 4.33E-02 1.8 
5.00 C24H22O13 Malonyl-genistin* Isoflavone 1.5 4.67E-02 2.3 

Note: Metabolites were annotated by level 2.1 of the metabolomics standards initiative; tR – retention time; KVL – very low K+ availability; KVH – very high K+ availability; “ * ” – 
identified using authentic standards; O2PLS-DA and S-Plot models were performed using p-value ≤ 0.05 and false discovery rate (FDR) with q-value ≤ 0.05 (Appendix G); Importance 
variable in score projection (VIP ≥ 1.0) were considered; Fold change (FC) refers to metabolite relative intensity values between KVH vs. KVL treatments (metabolite content is higher 
in KVH than KVL treatment); MS1 and MS2 data, chemical ontology and KEGG Pathway are available in Appendix G. 

Source:  The Autor (2022). 

 

 



37 

3.8 SOYBEAN LEAVES UNDER K+ DEFICIENCY 

The metabolomic profile of leaf samples from plants grown in KVL showed differential 

regulation at 21 metabolites, compared to KVH (Table 6). Under K+ deficiency, soybean leaves 

revealed upregulation of amino acids, benzoic acid derivates, pyrimidines, tricarboxylic acids, 

isoflavones, coumestans, pterocarpans, phosphatidylglycerols, and triterpenoid saponins 

content. The phytoalexins (isoflavonoids and soyasaponins) were the major class of specialised 

metabolites elicited in K+ deficient soybean leaves (Table 6 and Figure 4). 

The amino acids mostly accumulated in UT trifoliate leaves of plants in KVL were 

allantoic acid (3.3-fold) and L-asparagine (16.6-fold) compared to KVH treatment (Table 6). 

L-asparagine was also a differentially regulated metabolite in PV (2.4-fold) and IS (1.6-fold) in 

KVL treatment (Table 6 and 7). Both metabolites are participants in soybean nitrogen 

metabolism (alanine, aspartate, glutamate, and purine metabolism) by KEGG Metabolic 

Pathway (Appendix G). Previously, L-asparagine was differentially regulated in Fabaceae plants 

under potassium deficiency (STEWART; LARHER, 1980). L-asparaginases (EC 3.5.1.1) have 

been associated with K+ dependence for catalytic activity, promoting L-asparagine hydrolysis 

in L-aspartate and ammonium (BRUNEAU et al., 2006; LEA et al., 2007). However, changes 

in the content of L-asparagine can favour pest infestation, as Aphis glycines in soybeans 

(SEVERTSON et al., 2016; WALTER; DIFONZO, 2007) and Myzus persicae in Brassica 

napus (SEVERTSON et al., 2016). Therefore, based on previous studies (AMTMANN et al., 

2008; WALTER; DIFONZO, 2007) and our results, we consider that L-asparagine and other 

amino acids changes can affect soybean pest infestation and such susceptibility are important 

issues for future research. 

Four soyasaponins were upregulated in leaves from plants under very low potassium 

availability (KVL), as malonyl-dHex-Hex-HexA-soyasapogenol B (≤ 2.6-fold), soyasaponin 

Be (≤ 2.6-fold), soyasaponin βa (≤ 4.9-fold), and soyasaponin βg (≤ 4.7-fold). However, the 

regulation of soyasaponin Be and malonyl-dHex-Hex-HexA-soyasapogenol was restricted to 

MT and UT leaves, being undetectable in the LT (Table 6). Interestingly, the soyasaponins 

elicited in plants under deficiency were heterosides containing soyasapogenol B (aglycone) plus 

three moiety units conjugated to position C-3. Soyasaponin βa and soyasaponin βg have a 

DDMP moiety and pisumsaponin I, a malonyl group, both in the C-22 hydroxyl position. 

DDMP-conjugated saponins showed pro-oxidative and superoxide radical (O2
-) scavenging 

activity, and this property was attributed to the DDMP moiety (OKUBO; YOSHIKI, 2000; 

YOSHIKI; OKUBO, 1995). DDMP saponins accumulation in K+ deficient plants can be 



38 

associated with oxygen radical controllers in soybean leaves under oxidative stress (YOSHIKI 

et al., 1998). 

Isoflavonoids (isoflavones, coumestans, and pterocarpans) were predominantly 

upregulated compounds in soybean plants under KVL treatment (Table 6). The isoflavonoids 

have been frequently detected in soybean leaves (HOFFMANN-CAMPO et al., 2001; 

PIUBELLI et al., 2005) and pods (DA GRAÇA et al., 2016; PIUBELLI et al., 2005) under 

biotic interactions. This class of flavonoid can be observed, as constitutive (phytoanticipins) or 

induced (phytoalexins) metabolites (GRAHAM et al., 1990; SIMONS et al., 2011; TIKU, 

2020). Our results indicated upregulation of isoflavones genistein (1.3-fold), malonyl-genistin 

(≤ 2.9-fold), acetyl-genistein (≤ 2.5-fold), and daidzein (≤ 16.1) in leaves of plants under K+ 

soil deficiency (Table 6). 

 Chemometrics models showed that glycosylated isoflavones were predominantly 

accumulated in leaf extracts from medium (MT) and upper third (UT) of soybean plants. The 

biosynthesis of aglycone daidzein in LT leaves was 2.7-fold upregulated in KVL compared 

with KVH. Interestingly, the differential regulation of this compound was 16.1-fold greater in 

UT, and no changes were detected in MT leaves. These experimental results suggest that the 

LT leaves primarily had to cope with potassium deficiency at V7 stage, when soybean samples 

were collected. Potassium is a highly mobile nutrient in the plant, its redistribution was 

therefore expected and revealed by the ionomic analysis. In addition, although different tissues 

had different metabolites depending on the tissue maturity, they are part of the same organism 

that has complex and induced responses. The high daidzein upregulation in upper third (UT) 

leaves can be associated with induced mechanisms. This condition suggests that UT leaves were 

preparing to deal with the stress caused by K+ lacking, a situation that already occurred in the 

older leaves. Our results demonstrate that K+ deficiency accumulated not only the aglycones 

but also glycosidic isoflavones, corroborating previous investigations after stress by herbicide 

acifluorfen (COSIO et al., 1985), ethylene application (BAN et al., 2020), Nezara viridula 

(PIUBELLI et al., 2003), and Spodoptera litura herbivory (MURAKAMI et al., 2014).  

 

 

 

 

 



39 

Table 6 - Putative metabolite identification from soybean leaves upregulated under very low potassium (KVL) availability by UPLC-QTof-MSE 

tR  
(min) Formula Identification Chemical Ontology 

 Lower Third (LT)  Medium Third (MT)  Upper Third (UT) 
 KVH-LT vs. KVL-LT KVH-MT vs. KVL-MT KVH-UT vs. KVL-UT 
 FC FDR VIP  FC  FDR VIP  FC  FDR  VIP  

0.64 C4H8N2O3 L-asparagine Amino acids −− −− −− 8.0 2.31E-02 1.0 16.6 3.57E-03 1.8 

0.68 C4H8N4O4 Allantoic acid Amino acids  −− −− −−  −− −− −−  3.3 2.86E-02 1.3 

0.71 C4H4N2O2 Uracil Pyrimidones  3.1 3.33E-03 1.7  2.1 3.85E-03 1.7  2.4 1.07E-02 1.6 

0.72 C6H8O7 Citric acid Tricarboxylic acids  2.4 6.67E-03 2.6  1.7 7.69E-03 2.6  2.0 7.14E-03 2.7 

1.60 C13H16O9 Gentisic acid 2-β-D-glucoside Benzoic acids  1.3 4.67E-02 1.0  2.6 1.15E-02 2.6  3.0 1.79E-02 2.3 

4.60 C21H20O10 Genistin* Isoflavones  −− −− −−  1.2 5.00E-02 1.9  1.5 1.43E-02 3.7 

5.03 C24H22O13 Malonyl-genistin* Isoflavones  −− −− −−  1.6 3.85E-02 2.6  2.9 3.21E-02 3.1 

5.14 C23H22O11 Acetyl-genistin* Isoflavones  −− −− −−  1.6 3.08E-02 1.3  2.5 3.57E-02 1.2 

5.47 C15H10O4 Daidzein* Isoflavones  2.7 5.00E-02 1.0  −− −− −−  16.1 4.64E-02 1.1 

6.49 C15H10O5 Genistein* Isoflavones  −− −− −−  −− −− −−  1.3 5.00E-02 2.4 

6.51 C16H10O6 Isotrifoliol Coumestans  2.1 4.00E-02 1.4  −− −− −−  ND ND ND 

6.61 C15H8O5 Coumestrol Coumestans  2.4 3.67E-02 4.5  −− −− −−  −− −− −− 

7.04 C16H10O6 Trifoliol Coumestans  3.5 3.33E-02 2.3  −− −− −−  ND −− ND 

8.15 C20H16O6 Unknown Unknown  9.2 2.00E-02 1.5  −− −− −−  ND −− ND 

8.40 C51H80O21 Malonyl-dHex-Hex-HexA-soyasapogenol B (Pisumsaponin I) Triterpenoid saponins  −− −− −−  1.8 2.69E-02 1.1  2.6 2.50E-02 1.5 

8.50 C48H76O18 Soyasaponin Be (Dehydrosoyasaponin I) Triterpenoid saponins  −− −− −−  2.1 1.54E-02 1.0  2.6 2.14E-02 1.3 

8.52 C21H22O5 Glyceollin IV Pterocarpans  7.0 2.67E-02 1.3  −− −− −−  −− −− −− 

8.95 C20H16O5 Phaseol (4-prenyl-coumestrol) Coumestans  3.4 2.33E-02 5.2  −− −− −−  −− −− −− 

8.98 C53H82O20 Soyasaponin βa Triterpenoid saponins  4.9 1.33E-02 2.2  2.9 3.46E-02 2.4  2.3 3.93E-02 2.1 

9.11 C54H84O21 Soyasaponin βg Triterpenoid saponins  4.7 1.00E-02 2.1  3.0 1.92E-02 2.3  2.0 4.29E-02 1.8 

12.47 C22H45O9P Phosphatidylglycerol (16:0/0:0) Phosphatidylglycerols  4.8 1.67E-02 1.2  −− −− −−  −− −− −− 

Note: Metabolites were annotated by level 2.1 of the metabolomics standards initiative; tR – retention time; KVL – very low K+ availability; KVH – very high K+ availability; LT – lower 
third; MT – medium third; UT– upper third; “ * ” – identified using authentic standards; “ – ” - not significant; “ND” – not detected; O2PLS-DA and S-Plot models were performed using 
p-value ≤ 0.05 and false discovery rate (FDR) with q-value ≤ 0.05 (Appendix G); Importance variable in score projection (VIP ≥ 1.0) were considered; Fold change (FC) refers to metabolite 
relative intensity values between KVL vs. KVH treatments (metabolite content is higher in KVL than KVH treatment); MS1 and MS2 data, chemical ontology and KEGG Pathway are 
available in Appendix G. Source: The Autor (2022). 



40 

Figure 4 - Isoflavonoids (isoflavones, coumestans, and pterocarpans) and triterpenoid saponins (soyasaponins) as phytoalexins upregulated in 

soybean leaves under very low potassium availability 

 

Source: The Autor (2022).



41 

Leaves under very low K+ availability (KVL) showed a higher daidzein content that was 

exclusively differently regulated under stress (KVL vs. KVH). This aglycone is a direct 

precursor of soybean phytoalexin biosynthesis, as coumestans and pterocarpans (GRAHAM et 

al., 1990; KUC, 1995). Chemometrics models revealed that leaves under nutritional stress 

upregulated biosynthesis of coumestans, as coumestrol (2.4-fold), isotrifoliol (2.1-fold), 

trifoliol (3.5-fold), phaseol (3.4-fold), as well as, gliceollin IV (7.0-fold) a pterocarpan (Table 

6). The upregulation of these phytoalexins is likely associated with metabolism disorders and 

plant oxidative stress status (AHMAD et al., 2014; HERNANDEZ et al., 2012). Our results 

suggest that this also occurs in soybean leaves under severe K+ deficiency. Phytoalexins can be 

autotoxic to cells, therefore upregulation of biosynthesis is generally associated with defence 

mechanisms. Previously, coumestans elicitation occur when Phaseolus vulgaris leaves were 

injured by sulphur dioxide, ozone, and herbicides (RUBIN et al., 1983). Ozone fumigation, a 

highly reactive and cytotoxic agent, increased daidzein and coumestan contents in soybean 

(KEEN; TAYLOR, 1975). Phaseol and other prenylated isoflavones were also detected in 

soybean cotyledons after lactofen (diphenyl ether) treatment (CHENG et al., 2011). 

Previously published investigations were consistent in associating that potassium 

deficiency leads to status for the uncontrolled generation of reactive oxygen species (ROS) in 

plants (AHMAD et al., 2014; CAKMAK, 2005; HERNANDEZ et al., 2012; MIAO et al., 

2010). ROS are mainly produced in chloroplasts and peroxisomes (ASADA, 2000; POSPÍŠIL, 

2016) when plants are exposed to nutrient deficiency, salinity, and water deficit (CAKMAK, 

1994; FOYER et al., 1994). Interestingly, ROS generation is enhanced when plants are under 

abiotic stress and also exposed to light, causing photo-oxidative damage to chloroplasts 

(CAKMAK; MARSCHNER, 1992; ISSAWI et al., 2018; MARSCHNER; CAKMAK, 1989), 

and photodynamic stress status (ISSAWI et al., 2018). Under low K+ availability and sufficient 

contents of other nutrients, plants usually showed chlorotic and necrotic tissues when exposed 

to light (CAKMAK, 2005; MARSCHNER et al., 1996), however, these symptoms were absent 

when plants were maintained under shading (MARSCHNER; CAKMAK, 1989). Phosphorus 

and iron deficiency typically cause chlorotic symptoms (MARSCHNER, 2012). Nevertheless, 

in our experiment, independently of K+ availability, phosphorus content remained unaltered, 

and iron content increased in leaves severely K+ deficient, supporting our findings that leaf 

symptoms were only associated with K+ deficiency. 

The metabolomic profile of soybean plants under potassium deficiency showed 

differential regulation of DDMP saponins that are oxygen radical controllers in G. max 

(YOSHIKI et al., 1998). Isoflavones and pterocarpans (e.g., daidzein, isotrifoliol, and phaseol) 



42 

are excellent free radical inhibitors and lipid oxidation reductors (BOUÉ et al., 2008; LEE et 

al., 2005; LI et al., 2017; TODA; SHIRATAKI, 1999). Based on previous works and our 

results, we suggest that soybean metabolism was reprogrammed for the biosynthesis of non-

constitutive specialised metabolites (phytoalexins) for handling oxidative damage caused by 

potassium starvation. 

 

3.9 SOYBEAN POD TISSUES UNDER K+ DEFICIENCY  

 Eighteen metabolites were differentially regulated in K+ deficient soybean pod tissues 

(Table 7). Unlike in leaves, amino acids were the main class of metabolites differentially 

regulated in the pod tissues. L-glutamine and L-asparagine were upregulated (≤ 2.4-fold) in 

immature seeds (IS) and pod valves (PV), and the amino acids belonging to the nitrogen 

metabolism pathway, as ureidoglycine (1.9-fold in PV) and allantoic acid (1.6-fold in PV) were 

not differentially regulated in IS. 

Metabolomic profile analyses indicated positive regulation of the isoflavones daidzein 

(1.5-fold) and genistein (17.8-fold) in IS, and daidzin (2.3-fold) in PV tissues when plants grew 

under K+ deficiency (Table 7). In our analysis, genistein is highly regulated in IS; the conjugated 

metabolite genistein (genistin) was also induced by the stink bug, Nezara viridula, whose 

damage (seed punctures) are directly imposed in seeds (PIUBELLI et al., 2003), but the 

concentration increases 3.1-fold. Additionally, higher regulation of genistein was detected 

when soybean plants were under abiotic stress, such as salt stress, flooding, and shading (LIU 

et al., 2017; VANTOAI et al., 2012; WU et al., 2008). 

Low K+ availability also showed upregulation of protocatechuic acid-4-O-β-glucoside 

(1.4-fold) and gentisic acid-5-O-β-glucoside (2.1-fold) in PV, and L-histidine (1.8-fold) only in 

IS (Table 7). Protocatechuic and gentisic acids (benzoic acid derivates) are effective 

antioxidants and inhibitors of lipid peroxidation (MORAN et al., 1997; RICE-EVANS et al., 

1997), whereas L-histidine is associated with energetic storage compounds, control of hydroxyl 

radicals (ZS-NAGY; FLOYD, 1984) and singlet oxygen (1O2) (WADE; TUCKER, 1998). Our 

results demonstrate that K+ status promoted upregulation of many compounds in soybean pods 

and leaves (reproductive and vegetative structures), many of them are considered antioxidants; 

therefore, the biosynthesis of these compounds might have similar purposes, as a mechanism 

to control oxidative stress status in soybean plants under very low K+ availability.



43 

Table 7. Putative metabolite identification from soybean pod tissues upregulated under very low potassium (KVH) availability by UPLC-QTof-MSE 

 
tR 

(min) Formula Identification Chemical Ontology FC FDR VIP 

                                          Immature Seeds  
0.62 C6H9N3O2 L-histidine Amino acids 1.8 2.50E-02 2.4 
0.64 C5H10N2O3 L-glutamine Amino acids 2.4 3.33E-02 1.4 
0.65 C4H8N2O3 L-asparagine Amino acids 1.6 4.17E-02 1.8 
0.66 C18H32O15 Oligosaccharide unknown Oligosaccharides 2.8 2.92E-02 2.3 
0.66 C17H30O15 Oligosaccharide unknown Oligosaccharides 2.1 3.75E-02 1.8 
0.67 C6H12O7 Galactonic acid Hydroxy acids 3.9 1.25E-02 1.8 
0.67 C19H34O16 Ciceritol Oligosaccharides 1.7 5.00E-02 1.3 
0.69 C19H20O12 3,5-dihydroxyphenyl 1-O-(6-O-galloyl-β-D-glucopyranoside) Oligosaccharides 1.6 1.67E-02 4.0 
5.41 C15H10O4 Daidzein* Isoflavones 1.5 2.08E-02 2.4 
6.50 C15H10O5 Genistein* Isoflavones 17.8 4.17E-03 2.8 

                                          Pod Valves  
0.64 C5H10N2O3 L-glutamine Amino acids 2.1 2.67E-02 1.1 
0.65 C4H8N2O3 L-asparagine Amino acids 2.4 1.33E-02 2.1 
0.66 C11H20O10 3-O-β-D-galactopyranosyl-L-arabinose Monosaccharides 2.9 1.67E-02 1.3 
0.68 C3H7N3O3 Ureidoglycine Amino acids 1.9 3.67E-02 2.0 
0.69 C4H8N4O4 Allantoic acid Amino acids 1.6 2.33E-02 2.0 
0.73 C4H6O5 Malic acid Beta hydroxy acids 2.2 5.00E-02 1.2 
0.74 C13H16O9 Gentisic acid 2-β-D-glucoside Benzoic acids 2.1 6.67E-03 1.4 
1.62 C13H16O9 Protocatechuic acid 4-O-β-glucoside Benzoic acids 1.4 3.00E-02 1.5 
2.34 C21H32O10 Dihydrophaseic acid 4-O-β-D-glucoside Sesquiterpenoids 1.6 1.00E-02 2.3 
3.64 C21H20O9 Daidzin* Isoflavones 2.3 3.33E-02 1.3 

Note: Metabolites were annotated by level 2.1 of the metabolomics standards initiative; tR – retention time; KVL – very low K+ availability; KVH – very high K+ availability; “ * ” – 
identified using authentic standards; O2PLS-DA and S-Plot models were performed using p-value ≤ 0.05 and false discovery rate (FDR) with q-value ≤ 0.05 (Appendix G); Importance 
variable in score projection (VIP ≥ 1.0) were considered; Fold change (FC) refers to metabolite relative intensity values between KVL vs. KVH treatments (metabolite content is 
higher in KVL than KVH treatment); MS1 and MS2 data, chemical ontology and KEGG Pathway are available in Appendix G. Source: The Autor (2022). 



44 

4 CONCLUDING REMARKS  

Untargeted metabolomics reveals 51 primary and specialised metabolites that act in 19 

biochemical pathways that were influenced by K+ availability. Potassium is not a structural 

component of plant metabolites in comparison with phosphorus and nitrogen, but under very 

high K+ availability, the upregulation of flavonoids, galactolipids, and carbohydrates may be 

resulting from enzymatic activity increasing. In contrast, the biochemical mechanisms of 

soybeans in response to very low K+ availability are associated to: (I) accumulation of Ca2+, 

Mg2+, Fe2+, Cu2+, and B in young and old leaves; (II) increase of Fe2+ and Cu2+ content in 

immature seeds and pod valves, respectively; (III) leaf metabolism reprogramming for the 

biosynthesis of phytoalexins, such as isoflavones, coumestans, pterocarpans, and soyasaponins; 

(IV) enhancement of the amino acids, mono and oligosaccharides, benzoic acid derivates, and 

isoflavones content in soybean pods (IS and PV); (V) L-asparagine accumulation in vegetative 

(leaves) and reproductive (pod tissues) structures. 

Overall, the results suggest that soybean metabolism reprogramming is associated with 

oxidative stress caused by low K+ availability, and shows an increased content of phytoalexins 

(antioxidant agents), as isoflavonoids and triterpenoid saponins. Additionally, results suggest 

that L-asparagine is a promising soybean biomarker under K+ deficiency.  

Although the literature shows that plants fertilized with K+ are more tolerant/resistant to 

biotic and abiotic stresses, our data reveals that increasing fertilization rates did not promote a 

substantial increase in the content of constitutive specialised metabolites in soybean, refuting 

our initial hypothesis. We understood that this interaction is more complex and thus further 

work using integrated omics is certainly required, to understand whether the profile of 

metabolites induced by biotic elicitors (e.g., insects or pathogens) may be associated to K+ 

content in plants. 

 

 

 

 

 

 

 

 



45 

REFERENCES 

AHMAD, P.; ASHRAF, M.; HAKEEM, K. R.; AZOOZ, M.; RASOOL, S.; CHANDNA, R.; 
AKRAM, N.A. Potassium starvation-induced oxidative stress and antioxidant defense 
responses in Brassica juncea. Journal of Plant Interact, Abingdon, v. 9, p. 1–9, 2014. DOI: 
https://doi.org/10.1080/17429145.2012.747629 

ALMEIDA, D. M.; OLIVEIRA, M. M.; SAIBO, N. J. M. Regulation of Na+ and K+ 
homeostasis in plants: towards improved salt stress tolerance in crop plants. Genetics and 
Molecular Biology, Ribeirão Preto, v. 40, p. 326–345, 2017. DOI: 
https://doi.org/10.1590/1678-4685-gmb-2016-0106 

ALVARES, C. A.; STAPE, J. L.; SENTELHAS, P. C.; GONÇALVES, J. L. M.; SPAROVEK, 
G. Köppen’s climate classification map for Brazil. Meteorologische Zeitschrift, Stuttgart, v. 
22, p.711–728, 2013. DOI: https://doi.org/10.1127/0941-2948/2013/0507 

AMTMANN, A.; TROUFFLARD, S.; ARMENGAUD, P. The effect of potassium nutrition 
on pest and disease resistance in plants. Physiologia Plantarum, Hoboken, v. 133, p. 682–
691, 2008. DOI: https://doi.org/10.1111/j.1399-3054.2008.01075.x 

ARMENGAUD, P.; SULPICE, R.; MILLER, A. J.; STITT, M.; AMTMANN, A.; GIBON, Y. 
Multilevel analysis of primary metabolism provides new insights into the role of potassium 
nutrition for glycolysis and nitrogen assimilation in Arabidopsis roots. Plant Physiology, 
Rockville, v. 150, p. 772–785, 2009. DOI: https://doi.org/10.1104/pp.108.133629 

ASADA, K. The water–water cycle as alternative photon and electron sinks. Philosophical 
Transactions of the Royal Society of London, Series B: Biological Sciences, London, v. 
355, p. 1419–1431, 2000. DOI: https://doi.org/10.1098/rstb.2000.0703 

BALARDIN, R. S.; DALLAGNOL, L. J.; DIDONÉ, H. T.; NAVARINI, L.. Influência do 
fósforo e do potássio na severidade da ferrugem da soja Phakopsora pachyrhizi. 
Fitopatologia Brasileira, Brasília, v. 31, p. 462–467, 2006. DOI: 
https://doi.org/10.1590/S0100-41582006000500005 

BAN, Y. J.; SONG, Y. H.; KIM, J. Y.; BAISEITOVA, A.; LEE, K. W.; KIM, K. D.; PARK, 
K. H. Comparative investigation on metabolites changes in soybean leaves by ethylene and 
activation of collagen synthesis. Industrial Crops and Products, Amsterdam, v. 154, p. 
112743, 2020. DOI: https://doi.org/10.1016/j.indcrop.2020.112743 

BELLALOUI, N.; YIN, X.; MENGISTU, A.; MCCLURE, A. M.; TYLER, D. D.; REDDY, 
K. N. Soybean seed protein, oil, fatty acids, and isoflavones altered by potassium fertilizer 
rates in the Midsouth. American Journal of Plant Sciences, Irvine, v. 4, p. 976–988, 2013. 
DOI: DOI: https://doi.org/10.4236/ajps.2013.45121 

BENDER, R. R.; HAEGELE, J. W.; BELOW, F. E. Nutrient uptake, partitioning, and 
remobilization in modern soybean varieties. Agronomy Journal, Madison, v. 107, p. 563–
573, 2015. DOI: https://doi.org/10.2134/agronj14.0435 

BENJAMINI, Y.; HOCHBERG, Y. Controlling the false discovery rate: a practical and 
powerful approach to multiple testing. Journal of the Royal Statistical Society, Series B 
(Methodology), London, v. 57, p. 289–300, 1995. DOI: https://doi.org/10.1111/j.2517-
6161.1995.tb02031.x 



46 

BOUÉ, S. M.; SHIH, F. F.; SHIH, B. Y.; DAIGLE, K. W.; CARTER-WIENTJES, C. H.; 
CLEVELAND, T. E. Effect of biotic elicitors on enrichment of antioxidant properties and 
induced isoflavones in soybean. Journal of Food Science, Chicago, v. 73, p. H43–H49, 
2008. DOI: DOI: https://doi.org/10.1111/j.1750-3841.2008.00707.x 

BRUNEAU, L.; CHAPMAN, R.; MARSOLAIS, F. Co-occurrence of both L-asparaginase 
subtypes in Arabidopsis: At3g16150 encodes a K+-dependent L-asparaginase. Planta, 
Heidelberg, v. 224, p. 668–679, 2006. DOI: https://doi.org/10.1007/s00425-006-0245-9 

CAKMAK, I.; MARSCHNER, H. magnesium deficiency and high light intensity enhance 
activities of superoxide dismutase, ascorbate peroxidase, and glutathione reductase in bean 
leaves. Plant Physiology, Rockville, v. 98, p. 1222–1227, 1992. DOI: 
https://doi.org/10.1104/pp.98.4.1222 

CAKMAK, I. Activity of ascorbate-dependent H2O2-scavenging enzymes and leaf chlorosis 
are enhanced in magnesium- and potassium-deficient leaves, but not in phosphorus-deficient 
leaves. Journal of Experimental Botany, London, v. 45, p. 1259–1266, 1994. DOI: 
https://doi.org/10.1093/jxb/45.9.1259 

CAKMAK, I. The role of potassium in alleviating detrimental effects of abiotic stresses in 
plants. Journal of Plant Nutrition and Soil Science, Weinheim, v. 168, p. 521–530, 2005. 
DOI: https://doi.org/10.1002/jpln.200420485 

CHENG, J.; YUAN, C.; GRAHAM, T. L. Potential defense-related prenylated isoflavones in 
lactofen-induced soybean. Phytochemistry, Oxford, v. 72, p. 875–881, 2011. DOI: 
https://doi.org/10.1016/j.phytochem.2011.03.010 

COSIO, E. G.; WEISSENBÖCK, G.; MCCLURE, J. W. Acifluorfen-induced isoflavonoids 
and enzymes of their biosynthesis in mature soybean leaves. Plant Physiology, Rockville, v. 
78, p. 14–19, 1985. DOI: https://doi.org/10.1104/pp.78.1.14 

CUI, J.; ABADIE, C.; CARROLL, A.; LAMADE, E.; TCHERKEZ, G. Responses to K 
deficiency and waterlogging interact via respiratory and nitrogen metabolism. Plant, Cell & 
Environment, Chichester, v. 42, p. 647–658, 2019. DOI: https://doi.org/10.1111/pce.13450 

CUI, J.; TCHERKEZ, G. Potassium dependency of enzymes in plant primary metabolism. 
Plant Physiology and Biochemistry, Issy-les-Moulineaux, v. 166, p. 522–530, 2021. DOI: 
https://doi.org/10.1016/j.plaphy.2021.06.017 

DA GRAÇA, J. P.; UEDA, T. E.; JANEGITZ, T.; VIEIRA, S. S.; SALVADOR, M. C.; DE 
OLIVEIRA, M. C. N.; ZINGARETTI, S. M.; POWERS, S. J.; PICKETT, J. A.; BIRKETT, 
M. A.; HOFFMANN-CAMPO, C. B. The natural plant stress elicitor cis-jasmone causes 
cultivar-dependent reduction in growth of the stink bug, Euschistus heros and associated 
changes in flavonoid concentrations in soybean, Glycine max. Phytochemistry, Oxford, v. 
131, p. 84–91, 2016. DOI: https://doi.org/10.1016/j.phytochem.2016.08.013 

DAVIS, J. L.; ARMENGAUD, P.; LARSON, T. R.; GRAHAM, I. A.; WHITE, P. J.; 
NEWTON, A. C.; AMTMANN, A. Contrasting nutrient-disease relationships: potassium 
gradients in barley leaves have opposite effects on two fungal pathogens with different 
sensitivities to jasmonic acid. Plant, Cell & Environment, Chichester, v.  41, p. 2357–2372, 
2018. DOI: https://doi.org/10.1111/pce.13350 



47 

DEVI, B. S. R.; KIM, Y. J.; SELVI, S. K.; GAYATHRI, S.; ALTANZUL, K.; PARVIN, S.; 
YANG, D. U.; LEE, O. R.; LEE, S.; YANG, D. C. Influence of potassium nitrate on 
antioxidant level and secondary metabolite genes under cold stress in Panax ginseng. Russian 
Journal of Plant Physiology, Moscow, v. 59, p. 318–325, 2012. DOI: 
https://doi.org/10.1134/S1021443712030041 

DIEM, B.; GODBOLD, D. L. Potassium, calcium and magnesium antagonism in clones of 
Populus trichocarpa. Plant and Soil, Dordrecht, v. 155, p. 411–414, 1993. DOI: 
https://doi.org/10.1007/BF00025070 

DILLON, F. M.; CHLUDIL, H. D.; ZAVALA, J. A. Solar UV-B radiation modulates 
chemical defenses against Anticarsia gemmatalis larvae in leaves of field-grown soybean. 
Phytochemistry, Oxford, v. 141, p. 27–36, 2017. DOI: 
https://doi.org/10.1016/j.phytochem.2017.05.006 

FEHR, W. R.; CAVINESS, C. E. Stages of soybean development. Special Report, 80, p. 11, 
1977. 

FENG, Z.; DING, C.; LI, W.; WANG, D.; CUI, D. Applications of metabolomics in the 
research of soybean plant under abiotic stress. Food Chemistry, Amsterdam, v. 310, p. 
125914, 2020. DOI: https://doi.org/10.1016/j.foodchem.2019.125914 

FIRMANO, R. F. Efeito residual da adubação potássica nas formas do nutriente no solo. 
2017. 120 p. Dissertação (Mestrado em Solos e Nutrição de Plantas) - Escola Superior de 
Agricultura Luiz de Queiroz, Universidade de São Paulo, Piracicaba, 2017. DOI: 
https://doi.org/10.11606/D.11.2017.tde-07062017-090108 

FIRMANO, R. F.; OLIVEIRA, A.;  CASTRO, C.; ALLEONI, L. R. F. Potassium rates on the 
cationic balance of an Oxisol and soybean nutritional status after 8 years of K deprivation. 
Experimental Agriculture, Cambridge, v. 56, p. 293–311, 2020. DOI: 
https://doi.org/10.1017/S0014479719000371 

FOYER, C. H.; LELANDAIS, M.; KUNERT, K. J. Photooxidative stress in plants. 
Physiologia Plantarum, Hoboken, v. 92, p. 696–717, 1994. DOI: 
https://doi.org/10.1111/j.1399-3054.1994.tb03042.x 

GAALICHE, B.; LADHARI, A.; ZARRELLI, A.; BEN MIMOUN, M. Impact of foliar 
potassium fertilization on biochemical composition and antioxidant activity of fig (Ficus 
carica L.). Scientia Horticulturae, Amsterdam, v. 253, p. 111–119, 2019. DOI: 
https://doi.org/10.1016/j.scienta.2019.04.024 

GAO, X.; ZHANG, S.; ZHAO, X.; WU, Q. Potassium-induced plant resistance against 
soybean cyst nematode via root exudation of phenolic acids and plant pathogen-related genes. 
PLoS One, San Francisco, v. 13, p. e0200903, 2018. DOI: 
https://doi.org/10.1371/journal.pone.0200903 

GAUDE, N.; BRÉHÉLIN, C.; TISCHENDORF, G.; KESSLER, F.; DÖRMANN, P. Nitrogen 
deficiency in Arabidopsis affects galactolipid composition and gene expression and results in 
accumulation of fatty acid phytyl esters. Plant Journal, Chichester, v. 49, p. 729–739, 2007. 
DOI: https://doi.org/10.1111/j.1365-313X.2006.02992.x 

GRAHAM, T. L.; KIM, J. E.; GRAHAM, M. Y. Role of constitutive isoflavone conjugates in 
the accumulation of glyceollin in soybean infected with Phytophthora megasperma. 



48 

Molecular Plant-Microbe Interactions, Saint Paul, v. 3, p. 157, 1990. DOI: 
https://doi.org/10.1094/MPMI-3-157 

HAFSI, C.; DEBEZ, A.; ABDELLY, C. Potassium deficiency in plants: effects and signaling 
cascades. Acta Physiologiae Plantarum, Heidelberg, v. 36, p. 1055–1070, 2014. DOI: 
https://doi.org/10.1007/s11738-014-1491-2 

HARBORNE, J. B.; WILLIAMS, C. A. Advances in flavonoid research since 1992. 
Phytochemistry, Oxford, v. 55, p. 481–504, 2000. DOI: https://doi.org/10.1016/S0031-
9422(00)00235-1 

HARGE, N. Faixas de suficiência para teores foliares de nutrientes em soja, definidas 
pelo uso do método DRIS, para solos de origem basáltica. 2008. 88 p. Tese (Doutorado em 
Agronomia) – Universidade Estadual de Londrina, Londrina, 2008 

HASANUZZAMAN, M.; BHUYAN, M.; NAHAR, K.; HOSSAIN, MD.; MAHMUD, J.; 
HOSSEN, MD.; MASUD, A.; MOUMITA; FUJITA, M. Potassium: a vital regulator of plant 
responses and tolerance to abiotic stresses. Agronomy, Basel, v. 8, p. 31, 2018. DOI: 
https://doi.org/10.3390/agronomy8030031 

HERNANDEZ, M.; FERNANDEZ-GARCIA, N.; GARCIA-GARMA, J.; RUBIO-
ASENSIO, J. S.; RUBIO, F.; OLMOS, E. Potassium starvation induces oxidative stress in 
Solanum lycopersicum L. roots. Journal of Plant Physiology, Muenchen, v. 169, p. 1366–
1374, 2012. DOI: https://doi.org/10.1016/j.jplph.2012.05.015 

HERRERA, M. D.; ACOSTA-GALLEGOS, J. A.; REYNOSO-CAMACHO, R.; PÉREZ-
RAMÍREZ, I. F. Common bean seeds from plants subjected to severe drought, restricted- and 
full-irrigation regimes show differential phytochemical fingerprint. Food Chemistry, 
Amsterdam, v. 294, p. 368–377, 2019. DOI: https://doi.org/10.1016/j.foodchem.2019.05.076 

HOFFMANN-CAMPO, C. B.; HARBORNE, J. B.; MCCAFFERY, A. R. Pre-ingestive and 
post-ingestive effects of soya bean extracts and rutin on Trichoplusia ni growth. Entomologia 
Experimentalis et Applicata, Chichester, v. 98, p. 181–194, 2001. DOI: 
https://doi.org/10.1046/j.1570-7458.2001.00773.x 

HU, W.,; YANG, J.; MENG, Y.; WANG, Y.; CHEN, B.; ZHAO, W.; OOSTERHUIS, D. M.; 
ZHOU, Z. Potassium application affects carbohydrate metabolism in the leaf subtending the 
cotton (Gossypium hirsutum L.) boll and its relationship with boll biomass. Field Crops 
Research, Amsterdam, v. 179, p. 120–131, 2015. DOI: 
https://doi.org/10.1016/j.fcr.2015.04.017 

HUBER, D. M.; ARNY, D. C. Interactions of potassium with plant disease. In: MUNSON, R. 
D. (Org.), Potassium in Agriculture. American Society of Agronomy, 1985. p. 467–488. 
DOI: https://doi.org/10.2134/1985.potassium.c20 

ISSAWI, M.; SOL, V.; RIOU, C. Plant photodynamic stress: what’s new?. Frontiers in Plant 
Science, Lausanne, v. 9, p. 1–9, 2018. DOI: https://doi.org/10.3389/fpls.2018.00681 

ITO, M. F.; TANAKA, M. A. S.; MASCARENHAS, H. A. A.; TANAKA, R. T.; 
DUDIENAS, C.; GALLO, P. B. Efeito residual da calagem e da adubação potássica sobre a 
queima foliar (Cercospora kikuchii) da soja. Summa Phytopathologica, Botucatu, v. 19, p. 
21–23, 1993. 



49 

JARVIS, P.; DORMANN, P.; PETO, C. A.; LUTES, J.; BENNING, C.; CHORY, J. 
Galactolipid deficiency and abnormal chloroplast development in the Arabidopsis MGD 
synthase 1 mutant. Proceedings of the National Academy of Sciences of the United States 
of America, Washington, v. 97, p. 8175–8179, 2000. DOI: 
https://doi.org/10.1073/pnas.100132197 

KEEN, N. T.; TAYLOR, O. C. Ozone injury in soybeans. Plant Physiology, Rockville, v. 55, 
p. 731–733, 1975. DOI: https://doi.org/10.1104/pp.55.4.731 

KUC, J. Phytoalexins, stress metabolism, and disease resistance in plants. Annual Review of 
Phytopathology, Palo Alto, v. 33, p. 275–297, 1995. DOI: 
https://doi.org/10.1146/annurev.py.33.090195.001423 

LEA, P. J.; SODEK, L.; PARRY, M. A. J.; SHEWRY, P. R.; HALFORD, N. G. Asparagine 
in plants. Annals of Applied Biology, Chichester, v. 150, p. 1–26, 2007. DOI: 
https://doi.org/10.1111/j.1744-7348.2006.00104.x 

LEE, C. H.; YANG, L.; XU, J. Z.; YEUNG, S. Y. V. Y.; HUANG, Y.; CHEN, Z.-Y. Relative 
antioxidant activity of soybean isoflavones and their glycosides. Food Chemistry, 
Amsterdam, v. 90, p. 735–741, 2005. DOI: https://doi.org/10.1016/j.foodchem.2004.04.034 

LI, H.; YOON, J. -H.; WON, H. -J.; JI, H. -S.; YUK, H. J.; PARK, K. H.; PARK, H. -Y.; 
JEONG, T. -S. Isotrifoliol inhibits pro-inflammatory mediators by suppression of TLR/NF-κB 
and TLR/MAPK signaling in LPS-induced RAW264.7 cells. International 
Immunopharmacology, London,