UNDERSTANDING THE MICROBIOME OF COMMON BEAN PLANTS (PHASEOLUS VULGARIS L.): 
DROUGHT STRESS, MICROBIOME TRANSMISSION AND ASSEMBLY, AND OUTLOOKS FOR 
AGRICULTURAL APPLICATIONS 

By 

Abby Grieb 

A DISSERTATION 

Submitted to 
Michigan State University 
in partial fulfillment of the requirements   
for the degree of 

Microbiology and Molecular Genetics – Doctor of Philosophy 
Ecology, Evolutionary Biology, and Behavior – Dual Major 

2024 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

The plant microbiome is composed of diverse bacterial, archaeal, and fungal species 

that support the health of the plant in various plant compartments, including the root and 

surrounding soil, leaf surfaces, and internal tissues. These microbial species play an important 

role in plant health such as protecting the plant from pathogenic species and assisting with 

water and nutrient assimilation. With the ongoing climate crisis, it is important to understand 

how repeated seasons of stress, such as drought, are impacting the plant microbiome of 

agricultural crops, particularly staple food crops like common bean (dry beans). Work is being 

done to improve the resilience of common bean to environmental stress, including utilizing 

beneficial microbiome members to support the plant. However, the effect of drought on the 

common bean plant microbiome is not well understood, and it is important to select microbial 

inoculants that are also resistant to abiotic stress in the environment. 

Chapter 2 describes a multigenerational experiment conducted to study the impact of 

repeated drought exposure to common bean over two plant generations and in two common 

bean genotypes, Red Hawk and Flavert. We identified more significant effects of the drought 

treatments and legacy effects in the microbiome of the Flavert plants, while the microbiome of 

Red Hawk was more stable. Additionally, we identified bacterial orders that are consistently 

associated with the drought treatment across generations and genotypes, particularly 

Xanthomonadales and Rhizobiales, which may contain target bacterial inocula for microbiome 

modification under drought stress. In Chapter 3 the seeds of the Red Hawk plants in the 

multigenerational experiment were investigated, with an additional treatment condition of 

increased fertilizer concentration. The aim of this study was to identify vertically transmitted 

 
 
 
taxa in common bean seeds under abiotic treatment conditions. The stress treatments had a 

negligible effect on the resulting seed microbiomes, but we identified a significant impact of 

parental plant line and a signature of stable transmission of 22 prevalent seed microbiome 

members. These prevalent taxa included previously identified core taxa for common bean, and 

could be a valuable point of interest for the development of beneficial bacteria applications in 

agriculture. 

Chapter 4 investigates the development of the rhizoplane and rhizosphere microbiome 

in common bean roots across plant growth stages. An innovative experiment utilizing a growth 

delay in common bean plants allowed the influence of plant growth stage and time across the 

common bean lifecycle to be investigated separately. We identified closely aligned bacterial 

communities in the rhizoplane of common bean based on growth stage in plants, despite 

differing growth rates. Indicator taxa associated with plant growth stages were identified and 

found to be under a selective pressure by the plant, and included known beneficial plant 

microbiome taxa. 

This work provides important knowledge in understanding the impacts of repeated 

seasons of drought on plant microbiomes, exposes the importance of parental line in seed 

microbiome studies and the maintenance of the plant microbiome across generations, and 

provides insight into beneficial bacteria and the assembly processes of the plant microbiome. 

The need for sustainable solutions to support agricultural crops continues to rise, and this 

thesis contributes to our understanding of the processes shaping the beneficial plant 

microbiome that could be utilized in agricultural applications. 

 
This thesis is dedicated to my grandmother, Mary Thomas. 
My first teacher. 

iv 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ACKNOWLEDGEMENTS 

I am deeply grateful for the support and encouragement I received throughout my 

graduate school career that made this work possible. First, I would like to thank my PhD 

advisor, Dr. Ashley Shade. Ashley accepted me directly into her lab at the start of my PhD, and 

trusted me with a large and complex research project in my very first year as a PhD student. 

She has always been so supportive and helpful, even when conducting research was difficult 

during the COVID-19 pandemic. She has always supported all the extracurricular and career-

development activities I have pursued and taught me how to be a great researcher and 

scientist. Ashley has been an incredible mentor over the past five years, and I am so grateful to 

have been able to work with her.  

Additionally, I would like to thank my guidance committee, Dr. Matthew Schrenk, Dr. 

Greg Bonito, and Dr. Marjorie Weber, for their support, insight, and feedback as I completed 

my dissertation research. Marjorie also mentored me as an undergraduate researcher at 

Michigan State, and my experience doing research in her lab inspired me to pursue a PhD after 

completing my bachelor’s degree.  

I would also like to thank all the current and former members of the Shade lab who 

encouraged me and helped me complete my research projects. In particular, I would like to 

thank Dr. Fina Bintarti for teaching me how to grow beans in the growth chamber facilities and 

assisting with designing our multigenerational experiments, Keara Grady for her amazing help 

in showing me the ropes in the Shade lab and making sure everyone’s science ran smoothly, 

and my undergraduate assistants, Bryce Davis and Joanna Colovas, who helped me keep my 

plants alive and complete the large growth chamber experiments. I could not have harvested all 

v 

 
 
 
those plants and done all the DNA extractions without their help. Additionally, I would like to 

thank Dr. Samuel Barnett for his help in completing my data analysis.  

I was also fortunate in my graduate research to work with many amazing collaborators 

who assisted with experimental design and guidance on protocols for my work. I would like to 

thank Dr. Matthieu Barret, Dr. Marie Simonin, and Brice Marolleau from the INRAE in 

Beaucouzé, France. I had the wonderful opportunity to visit them at the INRAE in 2023 and 

learned so much from everyone there, especially Marie, who helped me with my data analysis 

skills. I have greatly enjoyed working with them on our collaborative project. Additionally, I 

would like to thank Dr. Chad Niederhuth and Hannah Cushman in the MSU Plant Biology 

department, who assisted with advice and materials for the zebularine experiment. 

Outside of my research activities, I also had the opportunity to participate in many 

different organizations and career development activities. I would like to thank the MSU Plant 

Resilience Institute, which has been a wonderful group to be a part of and I have learned so 

much from the people I’ve been able to interact with through the institute. Also, I’d like to 

thank Julie Rojewski and the MSU BEST program for their helpful programming on non-

academic career paths. I would also like to thank Dr. Robert Last and the Plant Biotechnology 

for Health and Sustainability Program for allowing me to join the program later in my PhD and 

giving me the opportunity to explore plant biotechnology and develop skills and interests for 

my future career. Additionally, I’d like to thank the Graduate Women in Science organization 

for all of the wonderful opportunities, and the Mid-Michigan chapter for their friendship and 

support throughout my time in graduate school.   

vi 

 
 
 
 
Lastly, I would not have been able to complete my PhD without the love and support of 

my family and friends. Thank you to my wonderful parents, Bill and Patty Sulesky, for always 

supporting my dreams and encouraging me to be the best in all that I do. I will never forget my 

dad bringing me to visit Michigan State for the first time in 2013. I fell in love with MSU and that 

visit changed my life forever. Also, thank you to my in-laws, Chuck and Lynn Grieb, who have 

always been there for me and given me a home away from home for the past 10 years. Lastly, 

thank you to my husband, Thomas, for his endless support and encouragement. I could not 

have completed graduate school without him.   

vii 

 
 
 
 
TABLE OF CONTENTS 

CHAPTER 1: Literature Review – Plant Microbiomes and their Role in                              
Sustainable Agriculture ........................................................................................................... 1 
EXPLORING THE PLANT MICROBIOME ........................................................................................ 2 
PLANT MICROBIOME ASSEMBLY AND INTERACTIONS ................................................................ 6 
APPLIED RESEARCH IN PLANT MICROBIOMES FOR AGRICULTURAL APPLICATIONS ................. 10 
STUDY SYSTEM - COMMON BEAN ............................................................................................. 12 
SUMMARY AND RESEARCH AIMS ............................................................................................. 13 
REFERENCES ............................................................................................................................. 15 

CHAPTER 2: Belowground plant microbiome resistance and response after two generations of 
drought exposure in common bean ...................................................................................... 28 
ABSTRACT ................................................................................................................................. 29 
INTRODUCTION ........................................................................................................................ 30 
METHODS ................................................................................................................................. 33 
RESULTS .................................................................................................................................... 43 
DISCUSSION .............................................................................................................................. 60 
REFERENCES ............................................................................................................................. 65 
APPENDIX A: CHAPTER 2 SUPPLEMENTARY MATERIALS ........................................................... 72 

CHAPTER 3: Stable transmission of common bean seed microbiome members over three plant 
generations despite exposure to abiotic stress ...................................................................... 76 
ABSTRACT ................................................................................................................................. 77 
INTRODUCTION ........................................................................................................................ 78 
METHODS ................................................................................................................................. 81 
RESULTS .................................................................................................................................... 89 
DISCUSSION ............................................................................................................................ 102 
REFERENCES ........................................................................................................................... 108 
APPENDIX B: CHAPTER 3 SUPPLEMENTARY MATERIALS ......................................................... 117 

CHAPTER 4: Elucidating the recruitment timing and development of the root microbiome over 
the common bean lifecycle ................................................................................................. 121 
ABSTRACT ............................................................................................................................... 122 
INTRODUCTION ...................................................................................................................... 123 
METHODS ............................................................................................................................... 125 
RESULTS .................................................................................................................................. 133 
DISCUSSION ............................................................................................................................ 148 
REFERENCES ........................................................................................................................... 152 
APPENDIX C: CHAPTER 4 SUPPLEMENTARY MATERIALS ......................................................... 158 

CHAPTER 5: Conclusions and future directions .................................................................... 167 
REFERENCES ........................................................................................................................... 173 

viii 

 
 
Literature Review – Plant Microbiomes and their Role in Sustainable Agriculture 

CHAPTER 1: 

1 

 
 
 
 
EXPLORING THE PLANT MICROBIOME 

Microorganisms exist in every environment on earth, from the guts of animals and 

insects, to freshwater environments, to soils and subsurface sediments1–5. Microorganisms are 

ubiquitous in the environment and span a vast range of biodiversity6,7. It is likely that the known 

species of microorganisms currently documented account for only a small portion of the true 

biodiversity of bacteria, archaea, and fungi8. One area of particular interest in microbial ecology 

research is the plant microbiome. Plant microbiomes play an important role in plant health and 

provide many ecosystem services to their plant hosts and the environment at large9. Many 

bacterial, archaeal, and fungal species can inhabit the plant in various plant compartments and 

there is considerable work being done to investigate the membership of plant microbiomes in a 

wide range of plant species and environments10. Understanding key members of the 

microbiome is not only useful ecologically, but also allows researchers to harness the beneficial 

properties of these microbes for use in agricultural and environmental applications.  

Plant microbiome diversity and characteristics 

Microorganisms are found in all compartments of the plant (the various tissues and 

structures that make up the plant) and require unique traits and life strategies to survive in 

these environments. The most diverse community associated with the plant is the rhizosphere. 

The rhizosphere is defined as the soil directly surrounding the roots that is impacted by the 

plant, as compared to bulk soil communities that are not directly affected by the plant11. The 

rhizosphere environment can contain a wide range of microorganisms, including bacteria, 

archaea, and fungi, and larger organisms like nematodes and small insects12,13.  The rhizosphere 

can contain tens of thousands of microbial species, with some estimates stating that there are 

2 

 
 
 
over 50,000 species of microbes in one gram of soil14. There are often tradeoffs between 

microorganisms in different niches in these communities, where one species may produce a 

byproduct that is consumed by another, and species utilize a wide range of energy sources15–17. 

The rhizosphere is also susceptible to dynamic changes in resources and water availability, and 

microbial species must be able to adjust to their ever-changing environment18. Many factors 

can influence the microbial community in the rhizosphere, including soil moisture and abiotic 

properties, the presence of other organisms like protists, and the progression of the plant over 

time19–23. Rhizosphere communities include many taxa that are categorized as beneficial to the 

plant, including plant growth promoting bacteria, nitrogen fixing rhizobia, mycorrhizal fungi, 

and bacteria that provide biocontrol of pests24–27. 

Another compartment of the plant microbiome associated with roots is the rhizoplane. 

The rhizoplane community is defined as the microbial species that are closely associated with 

the surface of plant roots, and can be more or less diverse than the rhizosphere depending on 

the plant and environment28–31. These microbial species are under greater influence from the 

plant roots at small scales, where species are recruited to the rhizoplane via plant-microbe 

feedbacks and further symbioses can develop32,33. The composition of the rhizoplane is similar 

to the rhizosphere, but often contains more beneficial microbes over commensal organisms 

and increased functional diversity30. Microbial species in the rhizoplane often include bacteria 

from the phyla Firmicutes, Proteobacteria, Actinobacteria, and Bacillota, among others, as well 

as fungal taxa34–38.  

The plant microbiome also contains many endophytic taxa, which are microbial cells 

living inside plant tissues both intercellularly and intracellularly39,40. Endophytic microbes can be 

3 

 
 
found throughout the plant in all plant structures, such as stems, leaves, vascular tissues, and 

seeds, and are typically less diverse than those living in the soil39–41. Plant tissues and 

intercellular spaces can have highly variable nutrient availability and present difficult living 

conditions for bacteria42. Many endophytic species typically have unique traits that allow them 

to colonize the plant more easily, such as lipopolysaccharides and motility which assist with 

active colonization, and plant cell-wall degrading enzymes which can enable passive 

colonization of plant tissues39.  

Many rhizosphere and rhizoplane taxa are commonly found inside the plant root as 

endophytes, and are typically more specialized and provide benefits to the plant28,42–44. Root 

endophytes can colonize the root tissues through the root tips, hairs or cracks at secondary 

root emergence sites39. Stem and vascular tissue endophytes are often closely linked to root 

communities, and these compartments can serve as a colonization route for bacteria as water 

moves throughout the plant39. Endophytes are also found within the leaf and seed tissues of 

the plant. Seed endophytes can be linked to the microbial community in the vascular system 

and can also be deposited by pollinators through floral compartments45,46. Seeds are a 

particularly challenging environment for microbial species to inhabit because they become 

dormant at maturity47–49. Seed endophytes need strategies to survive in dormant seeds, such as 

sporulation, desiccation stress tolerance and dormancy of their metabolic activities while 

nutrients are sparse50. Seeds typically harbor the lowest microbial diversity within the plant 

microbiome, containing tens to hundreds of taxa51.  

Finally, the phyllosphere microbiome is comprised of microbial species that colonize the 

surfaces of the plant that are impacted by environmental factors10. Bacteria living in the 

4 

 
 
phyllosphere need additional strategies that allow them to survive there, including desiccation, 

heat and UV stress tolerance, and specialized metabolisms for low-nutrient environments52. 

The phyllosphere is the harshest of the plant microbiome environments, but contains a wide 

diversity of species that may be selected for by the plant or deposited by stochastic processes 

such as irrigation and wind53–56. Plants may also contain specialized structures on their surfaces 

that harbor unique microbial communities, such as aerial roots or epicuticular wax57. Overall, 

the plant microbiome is a highly diverse environment with many specialized species that have 

adapted to thrive both on and inside of plant compartments.  

Beneficial properties of the plant microbiome 

The bacterial and fungal species living in and on plants can include commensal 

organisms that utilize resources provided by the plant without affecting the plant in any way, 

pathogenic species that can cause harm to the plant, or beneficial species that provide a wide 

range of valuable properties that support plant growth58,59. There are many ways that beneficial 

species can support the plant, including providing protection from pathogenic species, 

providing valuable nutrients to the plant, releasing plant growth-promoting factors, or assisting 

with water uptake and drought resistance. For example, Pseudomonas strains have been shown 

to be effective in controlling blight in potato, and the co-inoculation of common bean with 

mycorrhiza and Rhizobium symbionts increased plant tolerance to Fusarium root rot60,61. Many 

studies have also found that the inoculation of beneficial bacteria in root systems increases 

nutrient uptake from the soil in plants such as tomato, soybean and common bean62–64. The 

hyphal structures of mycorrhizal fungi are also known to play a role in water uptake by roots, 

and inoculants of beneficial bacteria and fungi can improve plant growth under drought 

5 

 
 
    
stress65,66. As discussed, these beneficial microbes contribute to the health of plants in a wide 

range of environments. This area of research is particularly useful in agricultural applications, 

and there is a growing literature investigating the use of beneficial microbes for inoculations in 

the field to replace exogenous irrigation, fertilizer, and pesticide use. 

PLANT MICROBIOME ASSEMBLY AND INTERACTIONS 

Horizontal assembly and plant-soil feedbacks  

Plant microbiome members can be assembled in the microbiome through a variety of 

processes. Horizontal assembly involves microorganisms colonizing the plant from external 

sources such as soil, water, and air55. Soil microbes can disperse to above and below-ground 

plant compartments via stochastic processes and are also enriched and recruited by the plant in 

these environments37,67–69. As plants are growing, they release chemicals such as 

phytohormones and produce immune responses that can enrich or modulate species of 

bacteria and fungi70–72. These plant-microbe feedbacks play an important role in shaping the 

microbiome, particularly in root communities, and communities can shift dramatically over 

time73–75. It has been demonstrated that root exudates such as malic acid recruit beneficial 

rhizobacterium species in Arabidopsis, and other organic acid exudates are linked to bacterial 

responses over the wheat lifecycle76,77. Changes in root exudation have also been investigated 

for various plant species under altered environmental conditions, such as in maize and locust 

trees20,78.  In addition to horizontal accumulation of taxa on plant surfaces and in roots, 

microbiome members can be deposited into floral compartments which can impact the 

establishing seed microbiome46,79.  Microbiome members can also be stochastically deposited 

6 

 
 
 
on the surface of seeds as seeds are dispersing from the parent plant, which can become 

available in the next plant generation as seeds germinate80. 

Vertical transmission 

In contrast to recruitment from the environment, there is also evidence of vertical 

transmission of microbiome members from parent plants to offspring through seeds and clonal 

plants80–85. As seeds are developing in the parental plant, endophytic microbiome members can 

assemble within the seed coat in the internal tissues of the seed embryo and endosperm80. This 

has been documented in many plant species and can happen through multiple pathways, such 

as through the vascular tissue of the parent plant in common bean and rice, or through floral 

compartments as has been demonstrated in other cultivated and wild species45,46,86–88. The 

microbiome members that are packaged within seed compartments become the starting point 

for the microbiome of germinating seedlings, and priority effects may allow them to establish in 

the plant microbiome before horizontally acquired species80,89–92. Microorganisms that are 

vertically transmitted through seeds have been shown to possess plant-beneficial properties 

and may be key taxa for the health of the next plant generation93–97. Additionally, plants that 

reproduce through clonal mechanisms have been documented as passing microbiome 

members from parent plant to offspring through vertical transmission85,98. Both the horizontal 

transmission and assembly from the environment through plant-soil feedbacks, and vertical 

transmission from parent plant to offspring, can play a major role in shaping the plant 

microbiome. 

7 

 
 
 
 
 
Symbiotic relationships within plant compartments 

While assembly processes of the plant microbiome can accumulate various commensal 

or harmful microbial taxa to plant compartments, plants often recruit specific microbiome 

members that provide important beneficial functions in the microbiome ecosystem. These 

plant-microbe interactions are known to enhance the growth of the plant and perform 

important ecosystem functions. 

One example of an important plant-microbe symbiosis is the formation of root nodules 

in leguminous plants. Specialized signals between the plant roots and diazotrophic bacteria in 

the family Rhizobiaceae regulate the process of root nodule formation, allowing bacteria to 

invade the root via root hairs99,100. Once inside the root, nodule structures form around the 

bacteria and the rhizobia species perform biological nitrogen fixation, which converts 

dinitrogen gas (N2) from the air into ammonia (NH3) which can be utilized by the plant101. Plant 

species often recruit specific rhizobia symbionts, but nodules can also contain diverse non-

rhizobial endophytes102–105. The fixation of nitrogen by rhizobia in legume root nodules also 

provides biologically available nitrogen for other plant species, and rotational cropping with 

legumes is an important aspect of modern agriculture to maintain nitrogen fertility in 

agricultural soils106. Free living bacteria are also capable of performing biological nitrogen 

fixation, and diazotrophic rhizobia species are being investigated for use as inoculants to 

improve nitrogen availability not only in legumes, but other crops as well107–111. 

Further, symbioses with mycorrhizal fungi also play an important role in the plant 

microbiome25,112.  The most prevalent mycorrhizal fungi in agricultural systems are arbuscular 

mycorrhizal fungi (AMF) and ectomycorrhizal fungi (EMF), and these fungal symbionts are 

8 

 
 
 
 
present in almost all major groups of plants25. Associations of mycorrhizal fungi with plants has 

evolved across multiple lineages of fungi, and fungal symbionts play an important role in the 

health and stress tolerance of plants112–115. Naturally occurring symbioses are highly important 

to plant growth and survival under stress conditions, and these relationships are being 

investigated for exploitation in agricultural applications. 

Recruitment of core taxa 

Many microbial species have been documented as “core” taxa for various plant species. 

Core taxa have had many different definitions in the literature, such as taxa that are highly 

abundant, present in all samples, or enriched in a condition of interest116. These definitions 

have been used to investigate core taxa in a wide range of plant microbiomes, including 

Arabidopsis, common bean, poplar, maize and sugarcane36,117–120.  Core taxa in these studies 

are typically consistently identified in the plant microbiome based on abundance, occupancy, or 

a combination of methods, and in some studies are found associated with the same plant 

species despite being grown in different locations117,121. Since a hallmark of core identification is 

consistent detection in the plant microbiome, it is likely that these taxa provide a benefit to the 

plant. However, the mechanism of core taxa association is not fully understood, and horizontal 

or vertical assembly processes are both possible as discussed previously. Additionally, seed 

microbiomes have been identified as potential reservoirs of core taxa in a variety of plant 

species122–125. Regardless of recruitment method, it is likely that these microbial species play an 

important role in plant health, and may be under a selective pressure from the plant as root-

associated or endophytic taxa. The core plant microbiome is of particular interest as a target for 

beneficial crop treatments with microbiome modification.  

9 

 
 
 
APPLIED RESEARCH IN PLANT MICROBIOMES FOR AGRICULTURAL APPLICATIONS 

Current threats to global agriculture 

The exponential growth of the human population combined with anthropogenic climate 

change has created a crisis for current agricultural systems. While the global food demand 

continues to grow, arable land area is decreasing, and climate change is significantly impacting 

agricultural production. Deforestation to convert landscapes to farmland is not a viable option, 

as it contributes to the climate crisis with the loss of forests as carbon sinks and increased land 

temperature126. Extreme weather events are also on the rise, particularly hydrological events 

such as flooding and drought127. It is estimated that dry growing seasons reduce the yield of 

major crops in Australia by 25-45%, and a meta-analysis of 30 plant species found significant 

negative impacts of drought and high temperatures on crop yield and seed quality128,129. 

Irrigation during drought and applications of artificial fertilizers also contribute to climate 

concerns, particularly with greenhouse gas emissions from fertilizer production130. Developing 

countries are at exceptionally high risk of negative climate change impacts131,132.  Many people 

in rural areas are dependent on subsistence farming, which is being threatened by changing 

temperatures, extreme weather events, droughts, soil degradation and pests and disease133–135. 

As these challenges to agriculture persist, it is important to find sustainable solutions that 

increase crop yield without contributing further to the climate crisis. 

Effects of abiotic stress on the microbiome 

Climate change not only has a profound effect on agricultural systems, it also 

significantly impacts the microbial life on our planet. Soil microbiomes contribute to 

biogeochemical cycling, and changing environmental conditions will result in significant impacts 

10 

 
 
 
 
on these processes, particularly in ecosystems that are more sensitive to climate change4. It has 

been estimated that current climate projections could result in significant declines of beneficial 

plant microbiome members, particularly bacteria that assist with biocontrol of pests and stress 

resistance136. The plant microbiome is important for plant resilience against changing climate 

conditions, and it will be critical to understand how these microbial communities are impacted 

by environmental conditions in the coming decades137. While there are many knowledge gaps 

in this area, some work has been done to investigate the impacts of abiotic stress on plant 

microbiomes. Abiotic stress conditions such as drought can affect root exudation in plants, 

which in turn has been demonstrated to affect the development of root microbiomes in maize, 

sorghum and rice138–141. The phyllosphere and endosphere can also be altered by abiotic stress 

such as drought affecting the seed microbiome of common bean, or extreme heat altering floral 

communities in wildflowers79,142,143. This area of research continues to grow, and understanding 

the impacts of changing climate conditions on plant-microbe interactions will provide 

important knowledge to the field to improve plant resilience under stress144. 

Potential solutions harnessing the plant microbiome 

While changing climate conditions is an area of significant concern for agriculture, there 

is also noteworthy research being done to find sustainable solutions that harness beneficial 

microbes to support agricultural crops145,146. Utilizing microbial inoculants to improve plant 

health and resistance to stress and disease has been investigated in a wide range of plant 

systems, including barley, soybean, grapevine and many others 62,63,66,147–149. It has been 

demonstrated that applications of bacteria or introductions of native microbial consortia can 

assist plants in responding to altered environmental conditions150,151. These processes are being 

11 

 
 
 
widely investigated to exploit for the support of agricultural crops in commercial 

applications152–155. Further, as evidence for plants enriching beneficial microbiomes increases, 

researchers are now considering the plant microbiome when breeding new varieties of 

crops156,157. Application methods of beneficial microbes, such as seed treatments and foliar 

sprays, are also a burgeoning area of research, as timing and persistence of inoculants in the 

plant microbiome are also important considerations when determining their effectiveness158–

160. Finding solutions to the challenges being faced by agricultural systems is of utmost 

importance for the future of food security. Although much work is being done in this area, 

there are also significant knowledge gaps around which microbial species are the best to use for 

inoculations, and how these species should be applied in the field161. 

STUDY SYSTEM - COMMON BEAN 

This work focuses on the microbiome of the plant species Phaseolus vulgaris L., also 

known as common bean or dry beans in agricultural contexts162. Beans were domesticated 

around 8,000 years ago in two different regions in central and South America, and have been an 

important food crop for centuries163,164. In 2020, there were 34.8 million hectares of beans 

harvested around the world, and production continues to increase annually165. Developing 

regions rely on dry beans as a major food crop, as they are nutrient dense and dormant beans 

can be stored for long periods of time47,166.  

Climate impacts on dry bean production may be severe in coming years, and research is 

being done in plant breeding, genetics, and microbiome amendment to improve the growth of 

common bean under stress conditions126,167,168. Investigating the specific rhizobia species that 

nodulate dry bean has been a growing area of research, with studies looking to utilize rhizobia 

12 

 
 
 
from primitive common bean landraces as a source for potential inoculants of agricultural 

varieties169,170. Other studies have tested various inoculants to improve the growth of beans 

under abiotic stress 110,171–173. The work presented here investigates the impact of 

environmental stress on the root and seed microbiome, and the root microbiome development 

of the kidney bean variety “Red Hawk”, in order to contribute to the growing literature of 

potential uses for the plant microbiome to support important agricultural systems174.  

SUMMARY AND RESEARCH AIMS 

Research on the beneficial properties of plant microbiomes is a growing field, and many 

studies have investigated the identity and potential functions of beneficial bacteria and fungi in 

plant microbiome compartments. Microorganisms play an important role in plant growth 

promotion, protecting plants against pathogens, and improving water and nutrient assimilation 

in plant roots. The microbiome can be assembled both horizontally from the environment and 

through vertical transmission from parent plants, and plants are able to recruit specific 

microbial taxa to their microbiome that likely benefit the health of the plant. While this has 

been an exciting area of research over the past few decades, many questions remain about the 

impacts of changing environmental conditions on the plant microbiome and how the 

microbiome can be utilized in sustainable agriculture systems.  

This thesis will investigate the microbiome of common bean plants, Phaseolus vulgaris 

L., in the context of drought stress, and the development of the root microbiome over the 

lifecycle of the plant. Although both bacterial and fungal species are important in plant 

microbiomes, this work focuses on the bacterial portion of these communities. In chapter 2, a 

multigenerational experiment will be described where two common bean varieties were 

13 

 
 
exposed to drought stress over two plant generations, and the root and rhizosphere 

microbiomes were assessed to address the impacts of repeated stress exposure on the 

microbiome. Chapter 3 investigates the seeds of the plants in the multigenerational 

experiment, with an additional treatment condition of increased fertilizer concentration. The 

aim of this study was to identify vertically transmitted taxa in common bean seeds under 

abiotic treatment conditions. Lastly, chapter 4 investigates the development of the rhizoplane 

and rhizosphere microbiome in common bean roots across plant growth stages. The 

overarching goal of this work is to identify bacterial taxa that are responsive to plant stress or 

are consistently found in common bean, and propose some target species for use in agricultural 

applications. This thesis will contribute to the growing area of research in beneficial plant 

microbiomes and will provide important knowledge for the development and application of 

beneficial microbial treatments in the field.  

14 

 
 
 
REFERENCES 

1. 

2. 

3. 

4. 

5. 

6. 

7. 

8. 

9. 

10. 

11. 

12. 

13. 

Rinninella, E. et al. The role of diet in shaping human gut microbiota. Best Pract Res Clin 
Gastroenterol 62–63, (2023). 

Gupta, A. & Nair, S. Dynamics of Insect–Microbiome Interaction Influence Host and 
Microbial Symbiont. Front Microbiol 11, (2020). 

Nnadozie, C. F. & Odume, O. N. Freshwater environments as reservoirs of antibiotic 
resistant bacteria and their role in the dissemination of antibiotic resistance genes. 
Environmental Pollution 254, (2019). 

Jansson, J. K. & Hofmockel, K. S. Soil microbiomes and climate change. Nat Rev Microbiol 
18, 35–46 (2020). 

Brazelton, W. J., Nelson, B. & Schrenk, M. O. Metagenomic evidence for H2 oxidation 
and H2 production by serpentinite-hosted subsurface microbial communities. Front 
Microbiol 2, (2012). 

Thompson, L. R. et al. A communal catalogue reveals Earth’s multiscale microbial 
diversity. Nature 551, 457–463 (2017). 

Flemming, H. C. & Wuertz, S. Bacteria and archaea on Earth and their abundance in 
biofilms. Nat Rev Microbiol 17, 247–260 (2019). 

Lloyd, K. G., Steen, A. D., Ladau, J., Yin, J. & Crosby, L. Phylogenetically Novel Uncultured 
Microbial Cells Dominate Earth Microbiomes. mSystems 3, (2018). 

Cordovez, V., Dini-Andreote, F., Carrión, V. J. & Raaijmakers, J. M. Ecology and Evolution 
of Plant Microbiomes. Annual Reviews Microbiology 73, 69–88 (2019). 

Dastogeer, K. M. G., Tumpa, F. H., Sultana, A., Akter, M. A. & Chakraborty, A. Plant 
microbiome–an account of the factors that shape community composition and diversity. 
Curr Plant Biol 23, 100161 (2020). 

Uroz, S., Buée, M., Murat, C., Frey-Klett, P. & Martin, F. Pyrosequencing reveals a 
contrasted bacterial diversity between oak rhizosphere and surrounding soil. Environ 
Microbiol Rep 2, 281–288 (2010). 

Bonkowski, M., Villenave, C. & Griffiths, B. Rhizosphere fauna: The functional and 
structural diversity of intimate interactions of soil fauna with plant roots. Plant Soil 321, 
213–233 (2009). 

Bintarti, A. F., Wilson, J. K., Quintanilla-Tornel, M. A. & Shade, A. Biogeography and 
Diversity of Multi-Trophic Root Zone Microbiomes in Michigan Apple Orchards: Analysis 
of Rootstock, Scion, and Local Growing Region. Phytobiomes J 4, 122–132 (2020). 

15 

 
 
14. 

15. 

16. 

17. 

18. 

Banerjee, S. & van der Heijden, M. G. A. Soil microbiomes and one health. Nat Rev 
Microbiol 21, 6–20 (2023). 

Zhou, J. et al. The microplastisphere: Biodegradable microplastics addition alters soil 
microbial community structure and function. Soil Biol Biochem 156, (2021). 

Nuccio, E. E. et al. Niche differentiation is spatially and temporally regulated in the 
rhizosphere. ISME Journal 14, 999–1014 (2020). 

Pétriacq, P. et al. Metabolite profiling of non-sterile rhizosphere soil. Plant Journal 92, 
147–162 (2017). 

Fierer, N. Embracing the unknown: Disentangling the complexities of the soil 
microbiome. Nat Rev Microbiol 15, 579–590 (2017). 

19.  Mavrodi, D. V. et al. Long-term irrigation affects the dynamics and activity of the wheat 

rhizosphere microbiome. Front Plant Sci 9, (2018). 

20. 

21. 

Tiziani, R. et al. Drought, heat, and their combination impact the root exudation 
patterns and rhizosphere microbiome in maize roots. Environ Exp Bot 203, (2022). 

Ceja-Navarro, J. A. et al. Protist diversity and community complexity in the rhizosphere 
of switchgrass are dynamic as plants develop. Microbiome 9, (2021). 

22.  Wei, G. et al. Temporal Dynamics of Rhizosphere Communities Across the Life Cycle of 

Panax notoginseng. Front Microbiol 13, (2022). 

23. 

24. 

25. 

26. 

27. 

Dove, N. C. et al. Assembly of the Populus Microbiome Is Temporally Dynamic and 
Determined by Selective and Stochastic Factors. mSphere 6, e01316-20 (2021). 

Bulgarelli, D. et al. Revealing structure and assembly cues for Arabidopsis root-
inhabiting bacterial microbiota. Nature 488, 91–95 (2012). 

van der Heijden, M. G. A., Martin, F. M., Selosse, M. A. & Sanders, I. R. Mycorrhizal 
ecology and evolution: The past, the present, and the future. New Phytologist 205, 
1406–1423 (2015). 

Liu, X. et al. Limited Impact of Soil Microorganisms on the Endophytic Bacteria of Tartary 
Buckwheat (Fagopyrum tataricum). Microorganisms 11, (2023). 

Lu, J. et al. Co-existence of rhizobia and diverse non-rhizobial bacteria in the rhizosphere 
and nodules of Dalbergia odorifera seedlings inoculated with Bradyrhizobium elkanii, 
Rhizobium multihospitium-Like and Burkholderia pyrrocinia-like strains. Front Microbiol 
8, (2017). 

16 

 
 
28. 

29. 

30. 

31. 

32. 

33. 

34. 

35. 

36. 

37. 

38. 

39. 

40. 

41. 

Chi, F. et al. Ascending migration of endophytic rhizobia, from roots to leaves, inside rice 
plants and assessment of benefits to rice growth physiology. Appl Environ Microbiol 71, 
7271–7278 (2005). 

Addison, S. L. et al. Getting to the Root of Tree Soil Microbiome Sampling. Phytobiomes J 
(2023) doi:10.1094/pbiomes-09-22-0060-r. 

Tang, L. et al. Microbial community assembly and functional profiles along the soil-root 
continuum of salt-tolerant Suaeda glauca and Suaeda salsa. Front Plant Sci 14, (2023). 

Edwards, J. et al. Structure, variation, and assembly of the root-associated microbiomes 
of rice. Proc Natl Acad Sci U S A 112, E911–E920 (2015). 

Dennis, P. G. et al. Linking rhizoplane pH and bacterial density at the microhabitat scale. 
J Microbiol Methods 76, 101–104 (2009). 

Feng, H. et al. Listening to plant’s Esperanto via root exudates: reprogramming the 
functional expression of plant growth-promoting rhizobacteria. New Phytologist 239, 
2307–2319 (2023). 

Singh, T. & Singh, D. K. Assessing the Bacterial Community Structure in the Rhizoplane of 
Wetland Plants. Bull Environ Contam Toxicol 101, 521–526 (2018). 

Akosah, Y. et al. The potato rhizoplane actively recruits Fusarium taxa during flowering. 
Rhizosphere 20, (2021). 

Lundberg, D. S. et al. Defining the core Arabidopsis thaliana root microbiome. Nature 
488, 86–90 (2012). 

Xiong, C. et al. Host selection shapes crop microbiome assembly and network 
complexity. New Phytologist 229, 1091–1104 (2021). 

Diaz-Garza, A. M. et al. Temporal Dynamics of Rhizobacteria Found in Pequin Pepper, 
Soybean, and Orange Trees Growing in a Semi-arid Ecosystem. Front Sustain Food Syst 4, 
(2020). 

Compant, S. et al. The plant endosphere world – bacterial life within plants. Environ 
Microbiol (2020) doi:10.1111/1462-2920.15240. 

Santoyo, G., Moreno-Hagelsieb, G., del Carmen Orozco-Mosqueda, M. & Glick, B. R. 
Plant growth-promoting bacterial endophytes. Microbiol Res 183, (2016). 

Fu, S. et al. Dynamic variation of Paris polyphylla root-associated microbiome assembly 
with planting years. Planta 257, (2023). 

17 

 
 
42. 

43. 

44. 

45. 

Kandel, S., Joubert, P. & Doty, S. Bacterial Endophyte Colonization and Distribution 
within Plants. Microorganisms 5, 77 (2017). 

Schlaeppi, K., Dombrowski, N., Oter, R. G., Ver Loren Van Themaat, E. & Schulze-Lefert, 
P. Quantitative divergence of the bacterial root microbiota in Arabidopsis thaliana 
relatives. Proc Natl Acad Sci U S A 111, 585–592 (2014). 

Liu, H. et al. Inner plant values: Diversity, colonization and benefits from endophytic 
bacteria. Front Microbiol 8, (2017). 

Chesneau, G. et al. Temporal dynamics of bacterial communities during seed 
development and maturation. FEMS Microbiol Ecol 96, (2020). 

46.  Mitter, B. et al. A new approach to modify plant microbiomes and traits by introducing 
beneficial bacteria at flowering into progeny seeds. Front Microbiol 8, (2017). 

47. 

48. 

49. 

50. 

51. 

52. 

Soltani, A. et al. The genetics and physiology of seed dormancy, a crucial trait in 
common bean domestication. BMC Plant Biol 21, (2021). 

Goggin, D. E., Emery, R. J. N., Kurepin, L. V. & Powles, S. B. A potential role for 
endogenous microflora in dormancy release, cytokinin metabolism and the response to 
fluridone in Lolium rigidum seeds. Ann Bot 115, 293–301 (2015). 

Nelson, E. B. The seed microbiome: Origins, interactions, and impacts. Plant Soil 422, 7–
34 (2018). 

Abdelfattah, A., Tack, A. J. M., Lobato, C., Wassermann, B. & Berg, G. From seed to seed: 
the role of microbial inheritance in the assembly of the plant microbiome. Trends 
Microbiol 31, 346–355 (2023). 

Bintarti, A. F., Sulesky-Grieb, A., Stopnisek, N. & Shade, A. Endophytic microbiome 
variation among single plant seeds. Phytobiomes J 1–57 (2021). 

Lindow, S. E. Role of Immigration and Other Processes in Determining Epiphytic Bacterial 
Populations. in Aerial Plant Surface Microbiology (eds. Morris, C. E., Nicot, P. C. & 
Nguyen-The, C.) 155–168 (Spring, Boston, MA, 1996). 

53. 

Vorholt, J. A. Microbial life in the phyllosphere. Nat Rev Microbiol 10, 828–840 (2012). 

54. 

55. 

Finkel, O. M., Burch, A. Y., Lindow, S. E., Post, A. F. & Belkin, S. Geographical location 
determines the population structure in phyllosphere microbial communities of a salt-
excreting desert tree. Appl Environ Microbiol 77, 7647–7655 (2011). 

Grady, K. L., Sorensen, J. W., Stopnisek, N., Guittar, J. & Shade, A. Assembly and 
seasonality of core phyllosphere microbiota on perennial biofuel crops. Nat Commun 10, 
(2019). 

18 

 
 
56. 

Santoyo, G. How plants recruit their microbiome? New insights into beneficial 
interactions. J Adv Res 40, 45–58 (2022). 

57.  Mechan-Llontop, M. E., Mullet, J. & Shade, A. Phyllosphere Exudates Select for Distinct 
Microbiome Members in Sorghum Epicuticular Wax and Aerial Root Mucilage. 
Phytobiomes J (2023) doi:10.1094/pbiomes-08-22-0046-fi. 

58. 

Rascovan, N. et al. Integrated analysis of root microbiomes of soybean and wheat from 
agricultural fields. Sci Rep 6, (2016). 

59.  Mansfield, J. et al. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol 

Plant Pathol 13, 614–629 (2012). 

60. 

61. 

62. 

63. 

64. 

65. 

66. 

67. 

68. 

De Vrieze, M., Germanier, F., Vuille, N. & Weisskopf, L. Combining Different Potato-
Associated Pseudomonas Strains for Improved Biocontrol of Phytophthora infestans. 
Front Microbiol 9, (2018). 

Dar, G. H., Zargar, M. Y. & Beigh, G. M. Biocontrol of Fusarium Root Rot in the Common 
Bean (Phaseolus vulgaris L.) by using Symbiotic Glomus mosseae and Rhizobium 
leguminosarum. Microb Ecol 34, 74–80 (1997). 

Eltlbany, N. et al. Enhanced tomato plant growth in soil under reduced P supply through 
microbial inoculants and microbiome shifts. FEMS Microbiol Ecol 95, (2019). 

Liu, J. et al. Effect of Bacillus paralicheniformis on soybean (Glycine max) roots 
colonization, nutrient uptake and water use efficiency under drought stress. J Agron 
Crop Sci 209, 547–565 (2023). 

Elkoca, E., Turan, M. & Donmez, M. F. Effects of single, dual and triple inoculations with 
bacillus subtilis, bacillus megaterium and rhizobium leguminosarum bv. phaseoli on 
nodulation, nutrient uptake, yield and yield parameters of common bean (phaseolus 
vulgaris l. cv. ’Elkoca-05’). J Plant Nutr 33, 2104–2119 (2010). 

Ruiz-Lozano, J. & Azcon, R. Hyphal contribution to water uptake in mycorrhizal plants as 
affected by the fungal species and water status. Physiol Plant 95, 472–478 (1995). 

Igiehon, O. N. & Babalola, O. O. Rhizobium and Mycorrhizal Fungal Species Improved 
Soybean Yield Under Drought Stress Conditions. Curr Microbiol 78, 1615–1627 (2021). 

Howe, A. et al. Seasonal activities of the phyllosphere microbiome of perennial crops. 
Nat Commun 14, (2023). 

Liu, F. et al. Soil indigenous microbiome and plant genotypes cooperatively modify 
soybean rhizosphere microbiome assembly. BMC Microbiol 19, (2019). 

19 

 
 
69. 

70. 

71. 

72. 

73. 

74. 

75. 

76. 

77. 

78. 

79. 

80. 

81. 

82. 

Noman, M. et al. Plant–microbiome crosstalk: Dawning from composition and assembly 
of microbial community to improvement of disease resilience in plants. Int J Mol Sci 22, 
(2021). 

Bai, B. et al. The root microbiome: Community assembly and its contributions to plant 
fitness. J Integr Plant Biol 64, 230–243 (2022). 

Haichar, F. el Z., Santaella, C., Heulin, T. & Achouak, W. Root exudates mediated 
interactions belowground. Soil Biol Biochem 77, 69–80 (2014). 

Rodriguez, P. A. et al. Systems Biology of Plant-Microbiome Interactions. Mol Plant 12, 
(2019). 

Steinauer, K. et al. Root exudates and rhizosphere microbiomes jointly determine 
temporal shifts in plant-soil feedbacks. Plant Cell Environ 46, 1885–1899 (2023). 

Hu, L. et al. Root exudate metabolites drive plant-soil feedbacks on growth and defense 
by shaping the rhizosphere microbiota. Nat Commun 9, (2018). 

Lareen, A., Burton, F. & Schäfer, P. Plant root-microbe communication in shaping root 
microbiomes. Plant Mol Biol 90, 575–587 (2016). 

Rudrappa, T., Czymmek, K. J., Paré, P. W. & Bais, H. P. Root-secreted malic acid recruits 
beneficial soil bacteria. Plant Physiol 148, 1547–1556 (2008). 

Chen, S. et al. Root-associated microbiomes of wheat under the combined effect of 
plant development and nitrogen fertilization. Microbiome 7, (2019). 

Li, J., Fan, M., Yang, L., Yang, Z. & Shangguan, Z. Temporal shifts in root exudates driven 
by vegetation restoration alter rhizosphere microbiota in Robinia pseudoacacia 
plantations. Tree Physiol 43, 1081–1091 (2023). 

Russell, K. A. & McFrederick, Q. S. Floral nectar microbial communities exhibit seasonal 
shifts associated with extreme heat: Potential implications for climate change and plant-
pollinator interactions. Front Microbiol 13, (2022). 

Shade, A., Jacques, M. A. & Barret, M. Ecological patterns of seed microbiome diversity, 
transmission, and assembly. Curr Opin Microbiol 37, 15–22 (2017). 

Rezki, S. et al. Assembly of seed-associated microbial communities within and across 
successive plant generations. Plant Soil 422, 67–79 (2018). 

Abdelfattah, A., Wisniewski, M., Schena, L. & Tack, A. J. M. Experimental evidence of 
microbial inheritance in plants and transmission routes from seed to phyllosphere and 
root. Environ Microbiol (2021) doi:10.1111/1462-2920.15392. 

20 

 
 
83. 

84. 

85. 

86. 

87. 

88. 

89. 

90. 

91. 

92. 

93. 

94. 

95. 

Shahzad, R., Khan, A. L., Bilal, S., Asaf, S. & Lee, I. J. What is there in seeds? Vertically 
transmitted endophytic resources for sustainable improvement in plant growth. Front 
Plant Sci 9, (2018). 

Truyens, S., Weyens, N., Cuypers, A. & Vangronsveld, J. Bacterial seed endophytes: 
Genera, vertical transmission and interaction with plants. Environ Microbiol Rep 7, 40–
50 (2015). 

Vannier, N. et al. A microorganisms’ journey between plant generations. Microbiome 6, 
79 (2018). 

Tarquinio, F., Attlan, O., Vanderklift, M. A., Berry, O. & Bissett, A. Distinct Endophytic 
Bacterial Communities Inhabiting Seagrass Seeds. Front Microbiol 12, (2021). 

Shade, A., McManus, P. S. & Handelsman, J. Unexpected diversity during community 
succession in the apple flower microbiome. mBio 4, (2013). 

Kim, H., Jeon, J., Lee, K. K. & Lee, Y. H. Longitudinal transmission of bacterial and fungal 
communities from seed to seed in rice. Commun Biol 5, 772 (2022). 

Barret, M. et al. Emergence shapes the structure of the seed microbiota. Appl Environ 
Microbiol 81, 1257–1266 (2015). 

Fukami, T. Historical Contingency in Community Assembly: Integrating Niches, Species 
Pools, and Priority Effects. Annu Rev Ecol Evol Syst 46, 1–23 (2015). 

Johnston-Monje, D., Gutiérrez, J. P. & Lopez-Lavalle, L. A. B. Seed-Transmitted Bacteria 
and Fungi Dominate Juvenile Plant Microbiomes. Front Microbiol 12, (2021). 

Vujanovic, V., Islam, M. N. & Daida, P. Transgenerational role of seed mycobiome – an 
endosymbiotic fungal composition as a prerequisite to stress resilience and adaptive 
phenotypes in Triticum. Sci Rep 9, 18483 (2019). 

Bergna, A. et al. Tomato seeds preferably transmit plant beneficial endophytes. 
Phytobiomes J 2, 183–193 (2018). 

Rybakova, D. et al. The structure of the Brassica napus seed microbiome is cultivar-
dependent and affects the interactions of symbionts and pathogens. Microbiome 5, 104 
(2017). 

Adam, E., Bernhart, M., Müller, H., Winkler, J. & Berg, G. The Cucurbita pepo seed 
microbiome: genotype-specific composition and implications for breeding. Plant Soil 
422, 35–49 (2018). 

96. 

Bziuk, N. et al. The treasure inside barley seeds: microbial diversity and plant beneficial 
bacteria. Environ Microbiome 16, 1–21 (2021). 

21 

 
 
97.  Malinich, E. A. & Bauer, C. E. The plant growth promoting bacterium Azospirillum 

brasilense is vertically transmitted in Phaseolus vulgaris (common bean). Symbiosis 76, 
97–108 (2018). 

98. 

Swift, J. F., Migicovsky, Z., Trello, G. E. & Miller, A. J. Grapevine bacterial communities 
display compartment-specific dynamics over space and time within the Central Valley of 
California. Environ Microbiome 18, (2023). 

99. 

Stacey, G. Biology of the Nitrogen Cycle - Chapter 10: The Rhizobium-Legume Nitrogen-
Fixing Symbiosis. in (Elsevier B. V., 2007). 

100.  Crespi, M. & Gálvez, S. Molecular mechanisms in root nodule development. J Plant 

Growth Regul 19, 155–166 (2000). 

101.  Coskun, D., Britto, D. T., Shi, W. & Kronzucker, H. J. How Plant Root Exudates Shape the 

Nitrogen Cycle. Trends in Plant Science vol. 22 661–673 Preprint at 
https://doi.org/10.1016/j.tplants.2017.05.004 (2017). 

102.  Andrews, M. & Andrews, M. E. Specificity in legume-rhizobia symbioses. International 

Journal of Molecular Sciences vol. 18 Preprint at https://doi.org/10.3390/ijms18040705 
(2017). 

103.  Muresu, R. et al. Coexistence of predominantly nonculturable rhizobia with diverse, 

endophytic bacterial taxa within nodules of wild legumes. FEMS Microbiol Ecol 63, 383–
400 (2008). 

104. 

Favero, V. O. et al. Bradyrhizobium as the Only Rhizobial Inhabitant of Mung Bean (Vigna 
radiata) Nodules in Tropical Soils: A Strategy Based on Microbiome for Improving 
Biological Nitrogen Fixation Using Bio-Products. Front Plant Sci 11, (2021). 

105.  Cordero, I. et al. Rhizobial diversity, symbiotic effectiveness and structure of nodules of 

Vachellia macracantha. Soil Biol Biochem 96, 39–54 (2016). 

106. 

Tiemann, L. K., Grandy, A. S., Atkinson, E. E., Marin-Spiotta, E. & Mcdaniel, M. D. Crop 
rotational diversity enhances belowground communities and functions in an 
agroecosystem. Ecol Lett 18, 761–771 (2015). 

107.  Gen-Jiménez, A. et al. Enhance of tomato production and induction of changes on the 

organic profile mediated by Rhizobium biofortification. Front Microbiol 14, (2023). 

108.  Hungria, M., Campo, R. J. & Mendes, I. C. Benefits of inoculation of the common bean 
(Phaseolus vulgaris) crop with efficient and competitive Rhizobium tropici strains. Biol 
Fertil Soils 39, 88–93 (2003). 

109.  Pastor-Bueis, R., Jiménez-Gómez, A., Barquero, M., Mateos, P. F. & González-Andrés, F. 

Yield response of common bean to co-inoculation with Rhizobium and Pseudomonas 

22 

 
 
endophytes and microscopic evidence of different colonised spaces inside the nodule. 
European Journal of Agronomy 122, (2021). 

Yadegari, M., Asadi Rahmani, H., Noormohammadi, G. & Ayneband, A. Plant growth 
promoting rhizobacteria increase growth, yield and nitrogen fixation in Phaseolus 
vulgaris. J Plant Nutr 33, 1733–1743 (2010). 

Yurgel, S. N., Ajeethan, N. & Smertenko, A. Response of Plant-Associated Microbiome to 
Plant Root Colonization by Exogenous Bacterial Endophyte in Perennial Crops. Front 
Microbiol 13, (2022). 

110. 

111. 

112.  Chen, M., Arato, M., Borghi, L., Nouri, E. & Reinhardt, D. Beneficial services of arbuscular 

mycorrhizal fungi – from ecology to application. Front Plant Sci 9, (2018). 

113.  Miyauchi, S. et al. Large-scale genome sequencing of mycorrhizal fungi provides insights 

into the early evolution of symbiotic traits. Nat Commun 11, (2020). 

114.  Abdalla, M., Bitterlich, M., Jansa, J., Püschel, D. & Ahmed, M. A. The role of arbuscular 
mycorrhizal symbiosis in improving plant water status under drought. J Exp Bot (2023) 
doi:10.1093/jxb/erad249. 

115.  Powell, J. R. & Rillig, M. C. Biodiversity of arbuscular mycorrhizal fungi and ecosystem 

function. New Phytologist 220, 1059–1075 (2018). 

116. 

Shade, A. & Stopnisek, N. Abundance-occupancy distributions to prioritize plant core 
microbiome membership. Curr Opin Microbiol 49, 50–58 (2019). 

117. 

Stopnisek, N. & Shade, A. Persistent microbiome members in the common bean 
rhizosphere: an integrated analysis of space, time, and plant genotype. ISME Journal 15, 
2708–2722 (2021). 

118. 

Timm, C. M. et al. Abiotic Stresses Shift Belowground Populus -Associated Bacteria 
Toward a Core Stress Microbiome . mSystems 3, (2018). 

119.  Walters, W. A. et al. Large-scale replicated field study of maize rhizosphere identifies 

heritable microbes. Proc Natl Acad Sci U S A 115, 7368–7373 (2018). 

120.  Hamonts, K. et al. Field study reveals core plant microbiota and relative importance of 

their drivers. Environ Microbiol 20, 124–140 (2018). 

121.  Pérez-Jaramillo, J. E. et al. Deciphering rhizosphere microbiome assembly of wild and 
modern common bean (Phaseolus vulgaris) in native and agricultural soils from 
Colombia. Microbiome 7, (2019). 

122. 

Simonin, M. et al. Seed microbiota revealed by a large-scale meta-analysis including 50 
plant species. New Phytologist 234, 1448–1463 (2022). 

23 

 
 
123.  Ancousture, J., Durand, A., Blaudez, D. & Benizri, E. A reduced but stable core 

microbiome found in seeds of hyperaccumulators. Science of the Total Environment 887, 
(2023). 

124.  Chen, H. et al. Core microbiome of medicinal plant Salvia miltiorrhiza seed: A rich 
reservoir of beneficial microbes for secondary metabolism? Int J Mol Sci 19, 1–15 
(2018). 

125. 

Eyre, A. W., Wang, M., Oh, Y. & Dean, R. A. Identification and characterization of the 
core rice seed microbiome. Phytobiomes J 3, 148–157 (2019). 

126.  Gourdji, S., Läderach, P., Valle, A. M., Martinez, C. Z. & Lobell, D. B. Historical climate 

trends, deforestation, and maize and bean yields in Nicaragua. Agric For Meteorol 200, 
270–281 (2015). 

127. 

Furtak, K. & Wolińska, A. The impact of extreme weather events as a consequence of 
climate change on the soil moisture and on the quality of the soil environment and 
agriculture – A review. Catena (Amst) 231, (2023). 

128.  Madadgar, S., AghaKouchak, A., Farahmand, A. & Davis, S. J. Probabilistic estimates of 
drought impacts on agricultural production. Geophys Res Lett 44, 7799–7807 (2017). 

129.  Maity, A. et al. Climate Change Impacts on Seed Production and Quality: Current 

Knowledge, Implications, and Mitigation Strategies. Seed Science and Technology 51, 
65–96 (2023). 

130.  Gao, Y. & Cabrera Serrenho, A. Greenhouse gas emissions from nitrogen fertilizers could 

be reduced by up to one-fifth of current levels by 2050 with combined interventions. 
Nat Food 4, 170–178 (2023). 

131.  Ouko, K. O. & Odiwuor, M. O. Contributing factors to the looming food crisis in sub-

Saharan Africa: Opportunities for policy insight. Cogent Soc Sci 9, (2023). 

132.  Ackerl, T., Weldemariam, L. F., Nyasimi, M. & Ayanlade, A. Climate change risk, 

resilience, and adaptation among rural farmers in East Africa: A literature review. 
Regional Sustainability vol. 4 185–193 Preprint at 
https://doi.org/10.1016/j.regsus.2023.05.004 (2023). 

133. 

134. 

Tesfamichael, S. G., Shiferaw, Y. A. & Woldai, T. Forecasting monthly soil moisture at 
broad spatial scales in sub-Saharan Africa using three time-series models: evidence from 
four decades of remotely sensed data. Eur J Remote Sens 56, (2023). 

Shrestha, R. P. & Nepal, N. An assessment by subsistence farmers of the risks to food 
security attributable to climate change in Makwanpur, Nepal. Food Secur 8, 415–425 
(2016). 

24 

 
 
135.  Kundu, A., Mukhopadhyay, S. & Basu, P. Climate Change Impacts on Subsistence Paddy 
Farming on the Indian East Coast: A Geospatial and Community Perception Assessment. 
Weather, Climate, and Society 15, 571–586 (2023). 

136. 

Li, P. et al. Fossil-fuel-dependent scenarios could lead to a significant decline of global 
plant-beneficial bacteria abundance in soils by 2100. Nat Food 4, 996–1006 (2023). 

137. 

Trivedi, P., Batista, B. D., Bazany, K. E. & Singh, B. K. Plant–microbiome interactions 
under a changing world: responses, consequences and perspectives. New Phytologist 
234, 1951–1959 (2022). 

138.  Williams, A. & de Vries, F. T. Plant root exudation under drought: implications for 

ecosystem functioning. New Phytologist 225, 1899–1905 (2020). 

139.  Xu, L. et al. Drought delays development of the sorghum root microbiome and enriches 

for monoderm bacteria. Proc Natl Acad Sci U S A 115, E4284–E4293 (2018). 

140.  Wang, P., Marsh, E. L., Kruger, G., Lorenz, A. & Schachtman, D. P. Belowground microbial 
communities respond to water deficit and are shaped by decades of maize hybrid 
breeding. Environ Microbiol 22, 889–904 (2020). 

141. 

Santos-Medellín, C., Edwards, J., Liechty, Z., Nguyen, B. & Sundaresan, V. Drought stress 
results in a compartment-specific restructuring of the rice root-associated microbiomes. 
mBio 8, (2017). 

142. 

Zhu, Y. G. et al. Impacts of global change on the phyllosphere microbiome. New 
Phytologist 234, 1977–1986 (2022). 

143.  Bintarti, A. F., Kearns, P. J., Sulesky-Grieb, A. & Shade, A. Abiotic Treatment to Common 

Bean Plants Results in an Altered Endophytic Seed Microbiome. Microbiol Spectr 10, 
(2022). 

144.  Cheng, Y. T., Zhang, L. & He, S. Y. Plant-Microbe Interactions Facing Environmental 

Challenge. Cell Host Microbe 26, 183–192 (2019). 

145. 

Tosi, M., Mitter, E. K., Gaiero, J. & Dunfield, K. It takes three to tango: The importance of 
microbes, host plant, and soil management to elucidate manipulation strategies for the 
plant microbiome. Can J Microbiol 66, 413–433 (2020). 

146.  Berg, G., Kusstatscher, P., Abdelfattah, A., Cernava, T. & Smalla, K. Microbiome 
Modulation—Toward a Better Understanding of Plant Microbiome Response to 
Microbial Inoculants. Front Microbiol 12, (2021). 

147.  Bziuk, N., Maccario, L., Sørensen, S. J., Schikora, A. & Smalla, K. Barley Rhizosphere 

Microbiome Transplantation – A Strategy to Decrease Susceptibility of Barley Grown in 
Soils With Low Microbial Diversity to Powdery Mildew. Front Microbiol 13, (2022). 

25 

 
 
148.  Rolli, E. et al. Improved plant resistance to drought is promoted by the root-associated 

microbiome as a water stress-dependent trait. Environ Microbiol 17, 316–331 (2015). 

149. 

Sangiorgio, D., Cellini, A., Spinelli, F. & Donati, I. Promoting Strawberry (Fragaria × 
ananassa) Stress Resistance, Growth, and Yield Using Native Bacterial Biostimulants. 
Agronomy 13, (2023). 

150. 

Lau, J. A. & Lennon, J. T. Rapid responses of soil microorganisms improve plant fitness in 
novel environments. 109, (2012). 

151.  Vurukonda, S. S. K. P., Vardharajula, S., Shrivastava, M. & SkZ, A. Enhancement of 

drought stress tolerance in crops by plant growth promoting rhizobacteria. Microbiol 
Res 184, 13–24 (2016). 

152.  Arif, I., Batool, M. & Schenk, P. M. Plant Microbiome Engineering: Expected Benefits for 
Improved Crop Growth and Resilience. Trends Biotechnol 38, 1385–1396 (2020). 

153.  Chaudhary, P. et al. Application of Synthetic Consortia for Improvement of Soil Fertility, 
Pollution Remediation, and Agricultural Productivity: A Review. Agronomy vol. 13 
Preprint at https://doi.org/10.3390/agronomy13030643 (2023). 

154.  Goyal, R. K. & Habtewold, J. Z. Evaluation of Legume–Rhizobial Symbiotic Interactions 

Beyond Nitrogen Fixation That Help the Host Survival and Diversification in Hostile 
Environments. Microorganisms 11, 1454 (2023). 

155.  Mitter, E. K., Tosi, M., Obregón, D., Dunfield, K. E. & Germida, J. J. Rethinking Crop 

Nutrition in Times of Modern Microbiology: Innovative Biofertilizer Technologies. Front 
Sustain Food Syst 5, (2021). 

156.  Gopal, M. & Gupta, A. Microbiome selection could spur next-generation plant breeding 

strategies. Front Microbiol 7, (2016). 

157.  Kusstatscher, P. et al. Microbiome-Assisted Breeding to Understand Cultivar-Dependent 

Assembly in Cucurbita pepo. Front Plant Sci 12, (2021). 

158. 

Timmusk, S., Behers, L., Muthoni, J., Muraya, A. & Aronsson, A. C. Perspectives and 
challenges of microbial application for crop improvement. Front Plant Sci 8, (2017). 

159.  Preininger, C., Sauer, U., Bejarano, A. & Berninger, T. Concepts and applications of foliar 
spray for microbial inoculants. Appl Microbiol Biotechnol 102, 7265–7282 (2018). 

160.  Bennett, A. J. & Whipps, J. M. Dual application of beneficial microorganisms to seed 

during drum priming. Applied Soil Ecology 38, 83–89 (2008). 

161. 

Santos, L. F. & Olivares, F. L. Plant microbiome structure and benefits for sustainable 
agriculture. Curr Plant Biol 26, 100198 (2021). 

26 

 
 
162.  Graham, P. & Ranalli, P. Field Crops Research Common bean (Phaseolus vulgaris L.). Field 

Crops Res 53, 131–146 (1997). 

163.  Castro-Guerrero, N., Isidra-Arellano, M., Mendoza-Cozatl, D. & Valdés-López, O. 

Common Bean: A Legume Model onthe Rise for Unraveling Responsesand Adaptations 
to Iron, Zinc, andPhosphate Deficiencies. Front Plant Sci 7, 391–417 (2016). 

164. 

Schmutz, J. et al. A reference genome for common bean and genome-wide analysis of 
dual domestications. Nat Genet 46, 707–713 (2014). 

165.  Uebersax, M. A. et al. Dry beans (Phaseolus vulgaris L.) as a vital component of 
sustainable agriculture and food security—A review. Legume Science 5, (2023). 

166. 

Siddiq, M., Uebersax, M. A. & Siddiq, F. Global Production, Trade, Processing and 
Nutritional Profile of Dry Beans and Other Pulses. in Dry Beans and Pulses 1–28 (Wiley, 
2022). doi:10.1002/9781119776802.ch1. 

167.  Bayati, K., Hosseini, N., Moghadam, H. & Basiri, R. Effects of drought stress and nitrogen 
on grain yield and some agronomic traits of red kidney bean cultivars. Iranian Journal of 
Field Crop Science 48, Pe1069–Pe1081 (2018). 

168.  Dramadri, I. O., Nkalubo, S. T., Kramer, D. M. & Kelly, J. D. Genome-wide association 
analysis of drought adaptive traits in common bean. Crop Sci 61, 3232–3253 (2021). 

169.  Grange, L. & Hungria, M. Genetic diversity of indigenous common bean (Phaseolus 

vulgaris) rhizobia in two Brazilian ecosystems. Soil Biol Biochem 36, 1389–1398 (2004). 

170.  De Ron, A. M. et al. Diversity and conservation of wild and primitive common bean 
germplasm and their associated rizobia in the andean region. in Legumes for Global 
Food Security (2017). 

171.  Da Silva, P. M., Tsai, S. M. & Bonetti, R. Response to inoculation and N fertilization for 

increased yield and biological nitrogen fixation of common bean (Phaseolus vulgaris L.). 
Plant Soil 152, 123–130 (1993). 

172.  Gabre, V. V. et al. Multiple Effect of Different Plant Growth Promoting Microorganisms 
on Beans (Phaseolus vulgaris L.) Crop. Brazilian Archives of Biology and Technology 63, 
1–11 (2020). 

173. 

Talaat, N. B. Effective microorganisms: An innovative tool for inducing common bean 
(Phaseolus vulgaris L.) salt-tolerance by regulating photosynthetic rate and endogenous 
phytohormones production. Sci Hortic 250, 254–265 (2019). 

174.  Kelly, J. D. et al. Registration of ‘Red Hawk’ Dark Red Kidney Bean. Crop Sci 38, 280–281 

(1998). 

27 

 
 
Belowground plant microbiome resistance and response after two generations of drought 
exposure in common bean 

CHAPTER 2: 

28 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

Climate change is a topic of significant concern for the global food supply, and 

agricultural crops are experiencing unprecedented environmental conditions such as flooding, 

heat stress and drought. Drought stress can have severe impacts on crops, particularly in 

developing countries where people rely on staple food crops like common bean (Phaseolus 

vulgaris L.). Work is being done to improve the resilience of common bean to environmental 

stress, including utilizing beneficial microbiome members to support the plant. However, the 

effect of drought on the common bean plant microbiome is not well understood, and it is 

important to select microbial inoculants that are also resistant to abiotic stress in the 

environment. In this study, we conducted a multigenerational experiment to study the impact 

of repeated drought exposure to common bean over two plant generations and in two common 

bean genotypes, Red Hawk and Flavert. We identified more significant effects of the drought 

treatments and legacy effects in the microbiome of the Flavert plants, while the microbiome of 

Red Hawk was more stable. Additionally, we identified bacterial orders that are consistently 

associated with the drought treatment across generations and genotypes, particularly 

Xanthomonadales and Rhizobiales, which may contain target bacterial inocula for microbiome 

modification under drought stress. This work provides important knowledge in understanding 

the impacts of repeated seasons of drought on plant microbiomes and potential taxa to 

investigate for plant microbiome engineering.  

29 

 
 
 
 
 
INTRODUCTION 

Environmental changes are becoming an increasingly important factor in today’s 

agricultural systems1–3. Changing weather patterns have resulted in severe drought in many 

regions of the world, which are particularly devastating in developing countries facing severe 

food insecurity4–8.  An important and growing area of research is focused on the impact of these 

environmental changes on agricultural crops, with particular interest in the plant microbiome9–

12. The plant microbiome comprises millions of bacterial, archaeal, and fungal cells that inhabit 

the plant’s surfaces, tissues and root structures13,14. The beneficial microbiome plays an 

important role in plant health and provides many important services in agricultural systems, 

including water and nutrient assimilation to the plant, nutrient cycling in the soil and root 

system, and pathogen resistance13–16.  

Many studies have been conducted to improve our knowledge of the plant microbiome 

and its role in agriculture for various crops17–19. While some studies in recent years have 

addressed the impacts of environmental stress on the microbiome of plants, few have studied 

this over multiple generations of repeated stress exposure20–26. With the continuing trends of 

anthropomorphic climate change, repeated exposure to drought conditions each growing 

season will become common for many agricultural areas. These repeated stress conditions over 

multiple growing seasons may create unique challenges for crop plants, such as negative 

impacts on seed quality and long-term alterations of soil conditions1,2. Our study aimed to 

understand the impacts of repeated drought exposure on the microbiome of common bean 

(Phaseolus vulgaris L.). 

30 

 
 
 
 
Common beans are a vital food source around the world, with 27.5 million metric tons 

of food produced annually and cultivated across 34.8 million hectares of land as of 202027,28. 

The common bean species, Phaseolus vulgaris L., includes a wide range of bean varieties that 

are considered “dry edible bean” and fresh “garden bean” crops, including kidney beans, black 

beans, navy beans, green beans, and others29. These varieties have been bred over 8,000 years 

of domestication from two ancestral lines originating in Mesoamerica and the Andes 

Mountains27,30. Beans not only provide valuable nutrition, but also perform important symbiotic 

nitrogen fixation that provides usable nitrogen to other crops31. These services are particularly 

valuable in developing countries that are most greatly impacted by environmental stress, due 

to a lack of infrastructure to manage poor growing conditions and reliance on subsistence 

farming, and the ability to continue to produce dry beans in these areas is of vital importance32.  

Work has been done to improve the drought tolerance of common bean through 

breeding and plant genetics, but utilizing the plant microbiome to confer stress tolerance is a 

growing area of research33,34. Additionally, various studies have been conducted to understand 

the bean microbiome, bean plants’ association with nitrogen-fixing rhizobia, and the impact of 

abiotic stress, such as drought, on the bean microbiome35–42. However, few studies have 

investigated the impact of repeated stress exposure on the microbiome of common bean or 

other crops43. The study presented here will contribute to our understanding of how plant 

microbiomes can be affected by stress over generations, and may provide insights into 

potential management strategies that harness the benefits of the plant microbiome. 

We conducted a multigenerational experiment exposing common bean plants to 

drought stress in agricultural soil under controlled growth chamber conditions. We 

31 

 
 
 
hypothesized that (1) the bean plant health would be impacted by the drought treatment, (2) 

the microbiome of the plant roots and rhizosphere would be altered by the drought treatment 

in one or both generations in either alpha or beta diversity measures, and (3) that repeated 

exposure to drought over two generations would have consequences, either positive or 

negative, for plant health in the second generation. Specifically, we predicted that exposure to 

drought in the first generation may enrich beneficial microbiome members that could improve 

plant outcomes under drought when passed to the second generation. Alternatively, plant 

outcomes could be negatively impacted by repeated drought exposure. We tested our 

hypotheses in two different genotypes of Phaseolus vulgaris L., the Red Hawk variety, a 

representative kidney bean cultivar bred in North America, and Flavert, a representative 

European bean cultivar, in order to assess the impact of drought on the species broadly44–46. 

We also conducted the experiment in two geographic locations, Pays de la Loire, France, and 

Michigan, USA, with agricultural soil from these regions, in order to understand if the region 

and soil play a role in how the microbiome responds to stress. Common beans are regularly 

produced in both of these regions. Since the experiments and DNA sequencing were performed 

separately for the two locations with some differing parameters (see methods), the datasets 

were analyzed independently and potential differences between the datasets were expected. 

The multigenerational experiment consisted of seeds provided from a “Generation 0 

(zero)” seed pool planted in a controlled growth chamber and exposed to either a control 

condition with ample water and nutrients, or a drought condition with 60% decreased water 

but equivalent nutrients provided (Figure S1.1). Seeds from these “Generation 1” (G1) plants 

were harvested and used for growth in “Generation 2” (G2), where plants were exposed to 

32 

 
 
 
stress in a factorial design, either receiving the same or opposite treatment in generation 2. The 

plant growth, yield, and 16S V4 rRNA gene amplicon sequencing of the root and rhizosphere 

microbiome compartments were assessed in order to understand the impact of the first- and 

second-generation drought treatments on the plants and their root microbiome community. 

METHODS 

Bean cultivars  

Two cultivars of common bean were used to complete this experiment. Phaseolus 

vulgaris L. var. Red Hawk, developed by the Michigan State University Bean Breeding program, 

was selected as a representative dry bean crop46. Phaseolus vulgaris L. var. Flavert was selected 

as a representative flageolet bean variety common in French agricultural production45. Red 

Hawk seeds in the Michigan location were obtained from the Michigan State University Bean 

Breeding Program from their 2019 harvest and stored at 4°C until ready for use in experiments. 

Flavert and Red Hawk seeds in the France location were purchased from Vilmorin-Mikado 

(Limagrain group, France). These seeds were obtained as “Generation 0 (zero)” and used to 

plant the first generation of the experiment. 

Soil preparation 

For growth experiments performed in Michigan, agricultural field soil was collected from 

the Michigan State University Agronomy Farm from a field that had previously grown common 

bean in 2019 (42°42'57.4"N, 84°27'58.9"W, East Lansing, MI, USA). The soil was a sandy loam 

with an average pH of 7.2 and organic matter content of 1.9%. Soil was collected before each 

planting group from the same field location. Five-gallon buckets were used to collect enough 

soil needed for the planting experiment, avoiding the dry top layer of soil and plant debris. 

33 

 
 
Buckets of soil were covered with air-tight lids and stored at 4˚C until ready for use in the 

growth experiment. Immediately before use in planting, soil was passed through a 4mm sieve 

to remove rocks and plant debris, and the soil was mixed with autoclaved coarse vermiculite at 

a 50% v/v ratio for use in growth chamber pots. 

For Generation 1 experiments performed in Pays de la Loire, soil was collected from the 

experimental station of the National Federation of Seed Multipliers (FNAMS, 47°28’012.42”N – 

0°23’44.30”W, Brain-sur-l’Authion, France) where common bean had been cultivated in 2016. 

This soil was a clayey sand limestone with pH 7.1 and 1.9% organic matter content. Soil was 

sieved and mixed with coarse vermiculite in the same method as above. Soil materials for 

Generation 2 were collected from the experimental fields belonging to the Institut National de 

Recherche pour l’Agriculture, l’Alimentation et l’Environnement (INRAE) in Angers, France      

(47°28’50.7”N, 0°36’31.4”W). A different soil was used in G2 to simulate a realistic scenario 

during which seed multiplication occurs at a different site (seed producer) than the bean 

production (farm). This soil had a sandy-loam texture with a pH of 6.5 and 1.9% organic matter 

content. Soil was sieved and mixed with vermiculite in the same fashion as G1. 

Surface sterilization and seed germination 

For each planting group in Michigan, Red Hawk seeds were surface sterilized prior to 

planting with a solution of 10% bleach and 0.1% Tween20. Seeds were randomly selected from 

the bulk Gen0 seed supply, or the harvested Gen1 seed supply, avoiding seeds that were visibly 

cracked or moldy, and 20 seeds were placed in a sterile 50 mL conical tube. 20-30 mL of bleach 

solution was added to the tube and the seeds were left to soak for 10 minutes, swirling halfway 

through. After soaking, the seeds were rinsed 5 times with sterile DI H2O. On the final rinse, 

34 

 
 
approximately 100 µL of rinse water was spread onto TSA (Tryptic Soy Agar) and PDA (Potato 

Dextrose Agar) plates for sterility assessment. TSA plates were incubated overnight at 28°C, and 

PDA plates were incubated at room temperature for 48 hours. Seeds corresponding to plates 

with microbial growth were discarded. After the final rinse water was plated, seeds were placed 

in a Petri dish lined with sterile filter paper. 1-2 mL sterile DI H2O was applied to the filter paper 

and Petri dishes were stored in the dark at room temperature for 3-4 days for seeds to 

germinate, with an additional 2 mL sterile DI H2O added halfway through the germination 

period. Once seeds had sprouted radicle roots they were transferred to soil for growth.  

Seeds planted for growth in Pays de la Loire were not surface sterilized prior to planting, 

and were germinated directly in the prepared pots of soil. The experiment in Pays de la Loire 

included both Red Hawk and Flavert cultivars under the same soil and treatment conditions for 

each generation. 

Growth conditions 

For Generation 1 in Michigan, 1-gallon pots were filled with the soil-vermiculite mixture 

and three germinated seeds were planted per pot in a high-light BioChambers FLEX™ LED 

growth chamber with a 16-hour day/8-hour night cycle at 26°C and 22°C, respectively, and 50% 

relative humidity. Once seedlings emerged and reached the VC growth stage (two cotyledons 

and primary leaves expanded), they were culled to one seedling per pot. Plants were watered 

every other day with 300 mL 0.05% 15-10-30 water-soluble fertilizer solution (control 

condition) (Masterblend International, Morris, IL, USA). At the V3 stage (third trifoliate leaves 

expanded), stress treatments began for the drought treated plants. Drought plants received 

100 mL of 0.15% 15-10-30 fertilizer solution every other day (66% less water than control with 

35 

 
 
the same concentration of nutrients). After approximately 14 days of treatment period, when 

plants reached the R1 stage (first open flowers), they were returned to the control watering 

every other day until senescence. As plants began to dry, they were watered less frequently as 

needed. There were 20 plants grown per treatment in G1. Five plants per treatment were 

destructively harvested for plant phenotypic measurements at the R5 stage (most pods at the 

seed filling stage), 5 plants per treatment were used for root and rhizosphere microbiome 

analysis at senescence, and mature seeds were collected from 12 plants per treatment for 

planting in G2. Seeds from the additional plants were harvested and stored as reserve. 

For Generation 2, seeds from the 24 G1 parental lines that received either control or 

drought conditions were planted in a full factorial design and grown under both treatment 

conditions in G2. There were four treatment combinations total, Control_Control, 

Control_Drought, Drought_Control, and Drought_Drought (G1_G2). Eight seeds from each 

parental line were surface sterilized and germinated as above, then planted in the field soil-

vermiculite mixture in seedling trays in the growth chamber under the conditions stated above. 

Once plants reached the VC stage, four viable seedlings per parent line were transferred to 1-

gallon pots for the remainder of the experiment. Each parental line had two offspring treated 

with each of the treatments. One of those two offspring was used for plant phenotypic 

measurements while the other was harvested for microbiome analysis. Plants were watered 

according to the conditions and treatment timeline in G1, above. 

Plants grown in Pays de la Loire were grown under the same growth chamber, control 

and drought conditions above. The experiment in Pays de la Loire was replicated with both Red 

Hawk and Flavert seeds. After three weeks of growth, day 18 after sowing (V3 stage), replicate 

36 

 
 
plants (n= 5) were exposed to control treatment (300mL of 0.05% nutritive solution) or drought 

stress (66% water-withholding, 100mL of 0.15% nutritive solution) for a period of four weeks 

until Day 56, the R5 growth stage (pods at 8-10 cm long with discernible seeds), a slightly longer 

stress period than applied to the plants in Michigan. Five plant parental lines from each G1 

condition (10 lines total) from both genotypes were planted in Generation 2 in Pays de la Loire, 

and were also planted in a full factorial design with four treatment conditions in G2. See figure 

S1.1 for experimental design schematic. 

Plant phenotypic data collection 

Plants dedicated for phenotypic trait measurements at MSU were analyzed with a LI-

COR LI-6800 instrument to measure photosynthetic rate and stomatal conductance the day 

before the stress treatment period ended to assess the impact of the treatment condition (LI-

COR Biosciences, Lincoln, NE, U.S.A). Plants were then grown until they reached approximately 

the R6 growth stage where pods were developed with discernable seeds, but pods were green 

and were not yet drying. Pods were removed and placed in a paper bag, plants were cut at the 

base of the stem and shoots were placed in a large envelope. Roots were removed from the 

pot, shaken to remove excess soil, collected, rinsed, and then placed in a separate paper bag. 

Pods and seeds per plant were counted, and then all three compartments were placed in a 50 

°C drying oven for two days. After drying, the shoot, root and pod dry biomass was measured. 

In Pays de la Loire, plants were harvested for trait measurements at approximately the 

R5 stage. The entire root system was gently separated from the soil in the pot and placed in a 

plastic bag. The total fresh weight of each plant (both above- and below-ground tissues) was 

measured and the number of pods and seeds per plant were counted.  

37 

 
 
Microbiome compartment harvest 

Once the Michigan plants had senesced and pods were dried, plants were harvested for 

microbiome analysis. Seed pods were removed and stored in a sterile Whirl-pakâ bag. Root 

systems were removed from the pot and shaken to remove loose soil. Roots were collected in a 

Whirl-pakâ bag and associated rhizosphere soil was collected in a sterile 50 mL conical tube. 

Prior to analysis, pods and seeds per plant were counted, and then seeds were removed from 

the pods and pooled by plant in 50 mL conical tubes and were stored at 4°C for planting. Roots 

and rhizosphere soil were stored at -80°C until further analysis. 

The microbiome analysis in Pays de la Loire was performed on fresh plants harvested at 

approximately the R5 stage. The entire root system was gently separated from the soil in the 

pot and placed in a plastic bag. The rhizosphere soil was collected by shaking the root system in 

a plastic bag and stored at -80°C. The root system was then washed in sterile distilled water and 

transferred to 50 mL tubes and stored at -80°C.   

DNA extractions 

The Michigan root samples were thawed at room temperature and a 1-2 inch section of 

the main root system was cut and used for root DNA extraction which combined the rhizoplane 

and endophytic bacteria inside the root tissues. The selected sections were briefly rinsed in 

sterile DI water then placed in a clean mortar. Liquid nitrogen was added to the mortar and the 

roots were finely ground with a pestle. The ground root material was transferred to a 1.5 mL 

microcentrifuge tube and used for DNA extraction with the DNeasy PowerSoil Pro DNA Kit 

(Qiagen, Germantown, MD, USA) following manufacturer’s instruction with the following 

modifications. In step one, 750 μL solution CD1 was used with 50 μL ATL buffer (Qiagen, 

38 

 
 
Germantown, MD, USA). The bead beating step was performed for 15 minutes on a vortex 

genie 24-tube adapter at maximum speed. Lastly, 60 μL of the final elution buffer C6 was used, 

and tubes were incubated for 10 minutes before centrifugation. 

Rhizosphere soil was thawed at room temperature and DNA extracted with the DNeasy 

PowerSoil Pro DNA Kit with the same ATL buffer variation as the roots, above. Negative controls 

were included with each batch of DNA extractions from the start of the extraction procedure, 

and one positive mock community control was included with each compartment sample set 

(roots or rhizosphere)47. Control extraction samples were sequenced with the experimental 

samples for use in data processing. 

French samples were processed in the same fashion and extracted using the DNeasy 

PowerSoil Kit (Qiagen, Germantown, MD, USA, discontinued) following manufacturer’s 

instructions.  A blank extraction kit control, a PCR-negative control and PCR-positive control 

(Lactococcus piscium, a fish pathogen that is not plant-associated) were included in each PCR 

plate. 

Sequencing 

Sequencing of the 16S V4 rRNA gene for the Michigan root and rhizosphere samples was 

performed at the Argonne National Laboratory Environmental Sample Preparation and 

Sequencing Facility (Lemont, IL, USA). The DNA was PCR amplified with region-specific primers 

that include sequencer adapter sequences used in the Illumina MiSeq ; 

FWD:GTGYCAGCMGCCGCGGTAA; REV:GGACTACNVGGGTWTCTAAT 48–52. Each 25 µL PCR 

reaction contained 9.5 µL of MO BIO PCR Water (Certified DNA-Free), 12.5 µL of QuantaBio’s 

AccuStart II PCR ToughMix (2x concentration, 1x final), 1 µL Golay barcode tagged Forward 

39 

 
 
Primer (5 µM concentration, 200 pM final), 1 µL Reverse Primer (5 µM concentration, 200 pM 

final), and 1 µL of template DNA. The conditions for PCR were as follows: 94 °C for 3 minutes to 

denature the DNA, with 35 cycles at 94 °C for 45 s, 50 °C for 60 s, and 72 °C for 90 s; with a final 

extension of 10 min at 72 °C to ensure complete amplification. Amplicons were then quantified 

using PicoGreen (Invitrogen) and a plate reader (InfiniteÒ 200 PRO, Tecan). Once quantified, 

volumes of each of the products were pooled into a single tube, in equimolar amounts, this 

pool was then cleaned up using AMPure XP Beads (Beckman Coulter), and then quantified using 

a fluorometer (Qubit, Invitrogen). After quantification, the pool was diluted to 2 nM, 

denatured, and then diluted to a final concentration of 6.75 pM with a 10% PhiX spike for 

sequencing. Amplicons were sequenced on a 251bp x 12bp x 251bp MiSeq run using 

customized sequencing primers and procedures 50. 

For the 16S V4 rRNA gene sequencing in Pays de la Loire, PCR reactions were performed 

with a high-fidelity Taq DNA polymerase (AccuPrimeTM Taq DNA Polymerase System, 

Invitrogen) using 5μL of 10X Buffer, 1μL of forward and reverse primers (10μM), 0,2μL of Taq 

and 10μL of DNA. A first PCR amplification was performed with the primer sets V4 515f/806r 

(5‘-GTGCCAGCMGCCGCGGTAA-3’and 5’-GGACTACHVGGGTWTCTAAT-3’49. Cycling conditions 

were composed with an initial denaturation at 94°C during 3 minutes followed by 35 cycles of 

denaturation at 94°C (30 seconds), primer annealing at 55°C (45 seconds) and extension at 68°C 

(90 seconds), with a final elongation at 68°C for 10 minutes. Amplicon purification was 

performed with a ratio at 0.8 of magnetic beads (Sera-MagTM, Merck). A second PCR 

amplification was performed to incorporate Illumina adapters and barcodes: a first 

denaturation a at 94°C (1 minute), followed by 12 cycles of denaturation at 94°C (60 seconds), 

40 

 
 
primer annealing at 55°C ( 60 secondes) and extension at 68°C (60 seconds) with a final 

elongation at 68°C for 10 minutes. Amplicons were purified with a ratio at 0.7 of magnetic 

beads and quantified with the Quant-iTTM PicoGreen® dsDNA Assay Kit (Invitrogen). All the 

amplicons were pooled in equimolar concentrations and the concentration of the equimolar 

pool was monitored with quantitative PCR (KAPA SYBR® FAST, Merck). Amplicons libraries were 

mixed with 10% PhiX and sequences with a MiSeq reagent kit v3 600 cycles (Illumina).  

Sequence data processing 

Sequencing fastq files were processed in Qiime2 after primer removal (Qiime2 version: 

2022.8.0)53 . Sample fastqs were imported to Qiime2 format and demultiplexed, if necessary, 

with the qiime demux emp-paired protocol. Samples were denoised, truncated and merged at 

100% sequence identity using Dada2 in Qiime2, with the truncation lengths found in 

supplemental Table S1.154. 16S rRNA gene taxonomy was assigned with the Silva database 

release 132 for French datasets, and release 138 for Michigan datasets, and taxonomy and ASV 

tables were exported for further analysis in R55.  

Analysis 

Data analyses were performed in R version 4.3.1 and R Studio version 2023.06.1+52456. 

ASV, taxonomy and metadata tables were imported into the phyloseq package 1.44.057. 

Sequences derived from 16S rRNA genes that were unclassified at the phylum-level, affiliated to 

Chloroplasts and Mitochondria were removed. The identification of sequence contaminants 

was assessed with decontam v 1.20.058 using the prevalence of ASVs in samples and negative 

controls (threshold employed listed in Table S1.1).  

41 

 
 
 
Rarefaction curves for each dataset were made using the rarecurve() command in the 

vegan package version 2.6-459 and datasets were rarefied using the phyloseq 

rarefy_even_depth() command (rarefaction levels listed in Table S1.1). Datasets were then 

subset for further analysis using the ps_filter() command in the microViz package version 

0.10.1060. 

Alpha diversity was assessed using estimate_richness() in phyloseq and a T-test or 

ANOVA with Tukey Honest Significant Difference post-hoc test where applicable. Alpha diversity 

figures were created using the plot_richness() command from phyloseq with the ggplot2 

package version 3.4.261. Bray-Curtis distances were calculated with vegdist(), and PERMANOVA 

statistical tests were performed with adonis2(), both from the vegan package. Post-hoc analysis 

on the PERMANOVA results was performed with pairwise.adonis2() from pairwiseAdonis 

version 0.4.162. Beta diversity ordinations were created with ordinate() from phyloseq with 

ggplot2. Beta dispersion was assessed with betadisper() and permutest() from the vegan 

package with the spatial median method63. Indicator species analysis was performed with 

indicspecies package version 1.7.1464,65 and Venn Diagrams were produced with the 

VennDiagram package version 1.7.366. Plant health data was analyzed with T-Tests or ANOVA 

with TukeyHSD post-hoc tests where applicable, and figures made with ggplot2. Additional data 

wrangling and statistics were performed in the tidyverse package version 2.0.067 and R stats 

package, and figures panels were assembled with patchwork version 1.1.368.  

Data Availability 

Data analysis code can be found at 

(https://github.com/ShadeLab/Drought_multigeneration_study_common_bean). Raw 

42 

 
 
 
 
 
sequence data for the French samples can be found on the European Nucleotide Archive under 

accession number PRJEB65346. Raw sequences from the Michigan samples can be found on the 

NCBI Sequence Read Archive under BioProject accession number PRJNA1058980. 

RESULTS 

Plant phenotypic traits respond differently to drought stress in different locations 

 The plant biomass, yield and photosynthetic traits after drought treatment were 

analyzed in order to assess the impact of the stress treatment. For plants grown in Pays de la 

Loire, France, with a slightly longer stress period, both bean genotypes (Red Hawk and Flavert) 

had a significant decrease in the number of pods on the drought-treated plants (Figure 1.1, 

A.1). However, the decrease in pod number did not result in a significant decrease in seed yield 

(Figure 1.1, A.2). For Red Hawk beans grown in Michigan, the number of pods was not impacted 

by drought treatment. However, photosynthetic rate, stomatal conductance and above ground 

mass were significantly decreased following application of this abiotic stress. These data and 

trends indicate that the plants were negatively affected by the stress treatments applied in G1.  

The plant phenotypic traits after two generations of treatment were analyzed to assess 

the impact of single-generation or consecutive stress treatments. There was a general impact of 

the G2 treatment on the plants grown in Pays de la Loire, independent of G1 stress. The 

number of pods, seeds and above ground biomass were significantly decreased for both 

genotypes grown in Pays de la Loire (Figure 1.2, A). In addition, the root mass of Red Hawk was 

significantly increased by drought stress.  (Figure 1.2, A.4).  For Red Hawk plants grown in 

Michigan, only the photosynthetic rate and stomatal conductance were significantly altered 

following drought stress, similar to what was observed during the first stress treatment (Figure 

43 

 
 
1.2, B.2). These data indicate that the Flavert and Red Hawk plants in Pays de la Loire were 

greatly impacted by the drought treatment in G2. The Red Hawk plants in Michigan were 

affected by the drought immediately following treatment, but did not have any significant 

decrease in biomass as a result of the drought compared to control. The legacy of G1 treatment 

did not have a significant impact on most plant outcomes in G2, except for Flavert biomass in 

Pays de la Loire. Flavert plants that received drought treatment in both generations had 

significantly lower above-ground biomass than plants that were not droughted in G1 (Figure 

1.2, A.3). Overall, one generation of stress had a similar impact on the plants as two 

generations of stress in Red Hawk, but prolonged drought exposure over multiple generations 

was more detrimental to Flavert plants. 

44 

 
 
Figure 1.1.  Plant phenotypic measurements in Generation 1. A. Phenotypic measurements 
taken for common bean plants of both Flavert and Red Hawk genotypes grown in Pays de la 
Loire, France. Above-ground and root biomass measurements were taken on fresh plant tissue.  

45 

 
 
 
Figure 1.1 (cont’d)  
B. Phenotypic measurements taken for common bean Red Hawk plants grown in Michigan, 
USA. Above-ground and root biomass measurements were taken on dry plant tissue. All above-
ground biomass measurements include the total mass of stems, leaves and pods. Photosynth. 
Rate = Photosynthetic Rate, Stomatal Conduct. = Stomatal Conductance. Welch Two Sample t-
test, * = p-value < 0.05, ** = p-value < 0.01, *** = p-value < 0.001, n=5 per treatment. 

46 

 
 
Figure 1.2. Plant phenotypic measurements in Generation 2. A. Phenotypic measurements 
taken for common bean plants of both Flavert and Red Hawk genotypes grown in Pays de la 
Loire, France. B. Phenotypic measurements taken for common bean Red Hawk plants grown in 
Michigan, USA. Above-ground and root biomass measurements were taken on dry plant tissue. 
Above-ground biomass measurements include the total mass of stems, leaves and pods.  

47 

 
 
 
Figure 1.2 (cont’d)  
Photosynth. Rate = Photosynthetic Rate, Stomatal Conduct. = Stomatal Conductance. ANOVA 
with post-hoc Tukey HSD test, * = p-value < 0.05, ** = p-value < 0.01, *** = p-value < 0.001, 
**** = p-value < 1e-4. Non-annotated significance lines have the same p-value as lines above, 
n=5 per treatment. 

Bacterial richness is affected by drought treatment in Flavert plants  

Overall, the bacterial richness observed in samples from Pays de la Loire was much 

lower than alpha diversity observed in Michigan samples. When sequencing coverage was 

assessed for the two datasets, both datasets were proficiently sampled and rarefaction curves 

reach asymptote. When analyzing the datasets separately, significant differences in alpha 

diversity can be observed between the two genotypes grown in Pays de la Loire in both root 

and rhizosphere samples in generation 1 (Flavert vs. Red Hawk root samples p < 0.001, Flavert 

vs. Red Hawk rhizosphere samples p < 0.05, Figure 1.3A, 1.3C). When assessing the influence of 

drought treatment on the alpha diversity in G1, most samples were not affected by the 

treatment, other than a statistically significant increase in alpha diversity in Flavert drought 

roots compared to the control treatment (Figure 1.3A). There was also an overall trend of 

increased variability in G1 alpha diversity in the drought treatment across all rhizosphere 

samples.  

In generation 2 alpha diversity, root samples overall were not significantly impacted by 

the drought treatment in G2 or the legacy of G1 drought (Figure 1.4A, B). There was a trend in 

increased root alpha diversity in French plants that received drought in G1 and control in G2, 

but this was not statistically supported (Figure 1.4A). The two genotypes grown in Pays de la 

Loire were statistically different from each other in the rhizosphere samples (Flavert vs. Red 

Hawk rhizosphere samples p < 0.001, Figure 1.4C). Additionally, Flavert G2 rhizosphere samples 

48 

 
 
had significantly decreased alpha diversity in plants that received drought in G1 compared to 

G1 control (Flavert G1 Control vs. Flavert G1 Drought p < 0.001, Figure 1.4C). Flavert plants that 

received drought for the first time in G2 also had significantly decreased alpha diversity 

compared to Control_Control plants, and higher variability (Figure 1.4C). However, the legacy of 

G1 drought on the G2 outcome was not reflected in either Red Hawk rhizosphere dataset, and 

the Red Hawk rhizosphere alpha diversity was not significantly affected by the drought in 

generation 2 (Figure 1.4C, D). Overall, genotype was found to play a significant role in 

determining root and rhizosphere alpha diversity, and Flavert plants were the only plants with 

significantly altered alpha diversity under drought conditions. Red Hawk plants had relatively 

consistent alpha diversity in both compartments, and despite Red Hawk plant traits being 

impacted by drought in Pays de la Loire, the alpha diversity of these plants was not affected.  

49 

 
 
Figure 1.3. Alpha diversity observed for root and rhizosphere samples in Generation 1. A. Root 
samples of plants grown in Pays de la Loire, France. B. Root samples of plants grown in 
Michigan, USA. C. Rhizosphere samples of plants grown in France. D. Rhizosphere samples of 
plants grown in USA. Plant genotype is indicated by bars above the panels. Tukey HSD test, * = 
p-value < 0.05, n=5 per treatment. 

50 

 
 
 
 
Figure 1.4. Alpha diversity observed for root and rhizosphere samples in Generation 2. A. Root 
samples of plants grown in Pays de la Loire, France. B. Root samples of plants grown in 
Michigan, USA. C. Rhizosphere samples of plants grown in France. D. Rhizosphere samples of 
plants grown in USA. Plant genotype is indicated by bars above the panels. ANOVA with Tukey's 
Honest Significant Difference test, * = p-value < 0.05, **** = p-value < 1e-4. France n=5 per 
treatment, USA n=12 per treatment. 

Drought legacy effects are observed in bacterial community structure   

Bray-Curtis dissimilarity was used to assess beta diversity across the root and 

rhizosphere microbial communities, and was analyzed separately for plants grown in Pays de la 

Loire and Michigan. In G1, there were significant differences in the bacterial community 

structure of the two genotypes grown in Pays de la Loire in both root (p < 0.001, Figure 1.5A, B) 

and rhizosphere (p < 0.05, Figure 1.5D, E) samples (PERMANOVA). When the genotypes were 

analyzed separately, there was no influence of the drought treatment in G1 on the root sample 

51 

 
 
 
beta diversity in Pays de la Loire, but there were significant differences between drought and 

control in the rhizosphere communities for both Flavert (p < 0.01, Figure 1.5D) and Red Hawk (p 

< 0.05, Figure 1.5E) genotypes (PERMANOVA). The drought treatment did not have a significant 

effect on the root or rhizosphere communities in Michigan plants in G1 (Figure 1.5C, F).  

In G2 there were also significant differences between the two plant genotypes in Pays 

de la Loire for both compartments (root p-value < 0.001 Figure 1.6A, B, rhizosphere p-value < 

0.01 Figure 1.6D, E, PERMANOVA). Flavert roots were significantly affected by the interaction of 

the G1 – G2 treatments (PERMANOVA, p < 0.01, Figure 1.6A), with significantly different 

communities identified between the Control_Control and Drought_Control plants, which is 

consistent with alpha diversity results (Pairwise multilevel comparison with pairwiseAdonis in R, 

p < 0.01) and between the Drought_Control and Drought_Drought plants (p < 0.05), with 

variability observed between the Control_Drought and Drought_Drought plants (p=0.06). Red 

Hawk roots in G2 were unaffected by either generation of treatment in Pays de la Loire (Figure 

1.6B), but the Michigan Red Hawk plants had significantly different root communities based on 

the G2 treatment (PERMANOVA, p < 0.01, Figure 1.6C). In the G2 rhizosphere samples, Red 

Hawk plants in Michigan were not affected by either generation of treatment (Figure 1.6F), but 

in Pays de la Loire, Flavert plants had significantly different communities based on the legacy 

treatment from G1 (p < 0.001, Figure 1.6D, PERMANOVA), and Red Hawk plants were slightly 

influenced by the treatment in G2 (p = 0.053).  

These results indicate that the beta diversity of the rhizosphere samples in Pays de la 

Loire was significantly impacted by the drought treatment in G1 in both genotypes, while 

Michigan samples were not affected by the drought treatment until the second generation. 

52 

 
 
Flavert plants were affected by the stress treatment in the second generation in both roots and 

rhizosphere and presented legacy effects from the G1 treatment on the bacterial communities 

in G2, while Red Hawk plant communities were only affected by the G2 drought treatment.   

Figure 1.5. PCoA ordinations of beta diversity Bray-Curtis distances in Generation 1. A, B. Root 
samples of Flavert and Red Hawk plants grown in Pays de la Loire, France. C. Root samples of 
plants grown in Michigan, USA. D, E. Rhizosphere samples of Flavert and Red Hawk plants 
grown in Pays de la Loire, Frace. F. Rhizosphere samples of plants grown in Michigan, USA. 
PERMANOVA analysis found significant differences between Genotypes in both root (p-value < 
0.001) and rhizosphere (p-value < 0.05) samples in France. Plant genotype is indicated by bars 
above the panels. * = p-value < 0.05, ** = p-value < 0.01, n=5 per treatment. 

53 

 
 
 
 
 
Figure 1.6. PCoA ordinations of beta diversity Bray-Curtis distances in Generation 2. A, B. Root 
samples of Flavert and Red Hawk plants grown in Pays de la Loire, France. C. Root samples of 
plants grown in Michigan, USA. D, E. Rhizosphere samples of Flavert and Red Hawk plants 
grown in France. F. Rhizosphere samples of plants grown in the USA. PERMANOVA analysis 
found significant differences between Genotypes in both root (p < 0.001) and rhizosphere (p < 
0.01) samples in France. Plant genotype is indicated by bars above the panels. * = p-value < 
0.05, ** = p-value < 0.01, *** = p-value < 0.001. France n=5 per treatment, USA n=12 per 
treatment. 

Bacterial community variability is altered by treatments in Generation 2 

The dispersion of the Bray-Curtis dissimilarity was assessed using the spatial median 

between ordination points in order to assess the similarity between samples within treatments. 

Drought treatment in generation 1 did not have a significant impact on the dispersion of the 

root and rhizosphere communities (data not shown). In generation 2, Flavert plants in Pays de 

la Loire had significant changes in community dispersion, with Control_Drought plants having 

greater dispersion than Control_Control in the root samples (Permutational pairwise analysis, p 

54 

 
 
 
< 0.05, Figure 1.7A), and greater dispersion than all other treatment groups in the rhizosphere 

samples (Tukey HSD, C_D vs. C_C p < 0.001, C_D vs D_C p < 0.05, C_D vs. D_D p < 0.05, Figure 

1.7B). Dispersion of communities in Red Hawk G2 plants was unaffected by drought in Pays de 

la Loire, but Red Hawk plants in Michigan showed a statistically significant decrease in 

dispersion in the rhizosphere of plants treated with drought stress in G2, regardless of G1 

treatment (Permutational pairwise analysis, p < 0.05). Consistent with the effects observed in 

alpha and beta diversity, the dispersion of the bacterial community in the Control_Drought 

Flavert plants were significantly altered in G2 indicating legacy effects for these plants, and the 

Red Hawk rhizosphere community dispersion was affected by the G2 drought in Michigan. 

Figure 1.7. Beta dispersion calculated on the spatial median between samples. Only figures 
showing groups with significance are shown, all other ordinations had no significant differences 
in dispersion between groups. A. Distance to median in Generation 2 Flavert roots in Pays de la 
Loire, France. B. Distance to median in G2 Flavert rhizosphere in France. C. Distance to median 
in G2 Red Hawk rhizosphere in Michigan, USA. Permutational pairwise analysis with Tukey HSD 
posthoc test, * = p-value < 0.05, *** = p-value < 0.001. France n=5 per treatment, USA n=12 per 
treatment. 

55 

 
 
 
 
Indicator taxa are associated with drought treatment 

Indicator taxa were analyzed to identify ASVs that were significantly associated with the 

treatment groups and compartments within each generation. A greater number of indicator 

taxa were found in the Red Hawk plants grown in Michigan than the plants grown in Pays de la 

Loire, which aligns with the higher overall species richness in the Michigan plants (Table 1.1). In 

generation 1, there were indicator taxa identified between the control and drought treatment 

for all plant genotypes and compartments, while generation 2 plants in Pays de la Loire had 

fewer distinct groups of indicator taxa (Table 1.1). For plants grown in Michigan, indicator taxa 

could be associated with all G1_G2 treatment groups, as well as the control or drought 

treatment applied in G1 and G2 (Table 1.1, Figure S1.2, S1.3). Plants grown in Pays de la Loire 

only had indicator taxa associated with some of the treatment groups in either genotype or 

compartment (Table 1.1).  

56 

 
 
 
 
 
Table 1.1 
Indicator taxa associated with each treatment group. Numbers represent the number of 
bacterial ASVs identified that are significantly associated with each treatment group or 
treatment combination for each location, genotype, and compartment. 

Location: 
Genotype: 
Compartment: 
Control 
Drought 

Gen 1 

Gen 2  G2 Control 
G2 Drought 
Control_Control 
Control_Drought 
Drought_Control 
Drought_Drought 
Control in G1 
Drought in G1 

France 
Flavert 
Root 
0 
28 
1 
2 
0 
0 
1 
1 
0 
0 

Rhizo 
28 
5 
17 
0 
63 
3 
0 
0 
120 
0 

Red Hawk 
Root 
3 
3 
0 
0 
0 
0 
0 
0 
0 
0 

Rhizo 
1 
50 
3 
2 
1 
2 
0 
1 
5 
0 

USA 
Red Hawk 
Root 
8 
21 
112 
43 
42 
2 
62 
18 
12 
48 

Rhizo 
11 
10 
53 
67 
32 
17 
23 
43 
83 
49 

There was no overlap of specific indicator taxa ASVs between the two geographic 

locations, so ASVs were identified to the taxonomic order level and the overlap of orders that 

were associated with the drought treatment across generations was assessed. For Red Hawk 

plants grown in Michigan, there were 3 orders identified that were associated with the root 

samples in G1 drought, G2 overall drought, and also G2 Drought_Drought plants, specifically. 

These orders were Xanthomonadales, Pseudonocardiales, and Sphingomonadales. Additionally, 

there was one order found as indicator taxa in all three drought groups in the rhizosphere, 

Rhizobiales (Figure 1.8). In Flavert plants grown in Pays de la Loire, there was one order 

identified as indicator taxa in the root samples in all three groups, Pseudomonadales, and 

between the G1 drought and overall G2 drought roots, Streptomycetales. Flavert rhizosphere 

samples did not have any indicator taxa in the G2 drought treatment, but did have 4 orders 

identified in G1 drought: Xanthomonadales, Cytophagales, Micrococcales, and Salinisphaerales 

57 

 
 
  
 
 
 
(Figure 1.8). Red Hawk plants grown in Pays de la Loire had 2 orders identified in the 

rhizosphere that overlapped between G1 drought and G2 overall drought, Xanthomonadales 

and Rhizobiales, and one order overlapping between G1 and the G2 Drought_Drought plants 

identified in the phylum Acidobacteria. French Red Hawk plants did not have any indicator taxa 

in the generation 2 root samples, but 2 orders, Bacillales and Xanthomonadales, were 

associated with the G1 drought plants (Figure 1.8). Overlap of taxonomic orders significantly 

associated with the drought treatment can be identified in each plant genotype and 

compartment, as well as across the two genotypes and experimental locations. This provides 

evidence that these bacterial orders may play an important role in the plant microbiome under 

drought. 

Specifically at the ASV level, there were a small number of ASVs that were significantly 

associated with the drought treatment that overlapped between both generations in the 

Michigan root samples and between compartments in the Michigan rhizosphere samples (Table 

1.2). There were two ASVs identified in the Michigan root samples that were associated with 

both the G1 and G2 drought treatments, both from the genus Lysobacter. Additionally, two 

ASVs were identified that were associated with both the root and rhizosphere drought 

treatments in the Michigan G2 plants, from the genera Promicromonospora and Fluviicola. The 

overlap of specific ASVs across generations and compartments may suggest that these specific 

taxa are important for the Red Hawk plants under drought stress in Michigan.  

58 

 
 
 
 
Figure 1.8. Number of indicator taxa at the taxonomic order level that are associated with G1 
Drought plants, all G2 Drought plants, and Drought_Drought plants in G2, in each location, 
genotype and compartment. Flavert plants in France did not have any significant indicator taxa 
in the G2 Drought treatments in the rhizosphere. There were 4 taxonomic orders identified in 
the Flavert G1 Drought Rhizosphere. Red Hawk plants in France also did not have any significant 
indicator taxa in the G2 Drought treatments in the root samples. There were 2 taxonomic 
orders identified in the Red Hawk G1 Drought Roots. 

59 

 
 
 
 
 
 
 
 
Table 1.2  
Indicator taxa ASVs that were found overlapping between generations or compartments in 
Michigan drought treatments. 

Phylum 
Proteo-
bacteria 

Proteo-
bacteria 

Actino-
bacteriota 

Class 
Gamma-
proteobacteria 

Order 
Xantho-
monadales 

Gamma-
proteobacteria 

Xantho-
monadales 

Family 
Xanthomonadaceae 

Genus 
Lysobacter 

Xanthomonadaceae 

Lysobacter 

Actinobacteria 

Micrococcales 

Promicro-
monosporaceae 

Promicro-
monospora 

Bacteroidota 

Bacteroidia 

Flavobacteriales 

Crocinitomicaceae 

Fluviicola 

ASV ID (Silva)  Association 
G1 and G2 
3a66a38841b
drought, 
a6877568198
roots 
b9dac3a1dc 
G1 and G2 
f5c494fd1da6
drought, 
eac3404c69c
roots 
863c35c07 
G2 drought, 
47ee012e6bf
roots and 
01ba25c1e9d
rhizo 
73e2241846 
G2 drought, 
7e2b7a47670
roots and 
a5cb9656744
rhizo 
205ad8b31f 

DISCUSSION 

Our study contributes to a significant knowledge gap in investigating the effects of 

drought on the plant microbiome over consecutive plant generations. Overall, the common 

bean plants in our experiment were negatively impacted by the drought stress in both 

generations. The stress did not have as great of an impact on the plants grown in Michigan, 

which was likely due to the stress period being about 1 week shorter in the experimental design 

than it was in Pays de la Loire. The Flavert and Red Hawk plants in Pays de la Loire experienced 

a longer drought period, which had a large effect on plant traits in the second generation. The 

Red Hawk plants that received drought in both generations had lower above-ground biomass 

than any other treatment combination. Paired with an increase in root biomass in the 

generation 2 plants under drought stress, this aligns with other studies that have observed 

increased root length, depth, and biomass in plants under drought stress34. These results 

support our hypothesis that the plants would be significantly impacted by the drought 

treatment, and partially support the hypothesis that repeated drought exposure has 

60 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
consequences for the plant health in the second generation, as only the Red Hawk plants in 

Pays de la Loire showed evidence of a legacy of G1 drought affecting G2 plant outcomes. 

In addition to the impacts of the drought stress on the plant biomass and yield, the 

microbiome communities of the plant roots and rhizosphere were also significantly altered by 

the drought treatments in one or more compartments in both generations, which supports our 

second hypothesis. An intriguing element to our experiment was completing the experiment 

with two genotypes of common bean to understand the effects of the drought in two cultivars 

that have different domestication histories45,69. The two genotypes grown in Pays de la Loire 

had significantly different community diversity in the roots and rhizosphere communities, even 

when grown in the same soil and growth conditions. This is aligned with other evidence in the 

literature that different plant species harbor distinct core microbiota, and may suggest that 

microbial treatments in agriculture need to be specifically aligned to the plant genetic 

background to which they are being applied42,70,71. Over the two generations, the bacterial 

communities of the Red Hawk plants appeared to be more resistant to drought stress 

compared to the Flavert plants, as the Flavert root and rhizosphere microbiome was 

significantly altered in both generations and plant compartments. This may be a result of kidney 

beans being bred for more drought resistance traits than flageolet beans throughout their 

domestication history, leading to reduced stress signaling from the plants under drought which 

may recruit a more resistant microbiome as well72–74.  

While we can identify significant effects of the drought treatment on the microbiomes 

in both generations, the root and rhizosphere microbiomes are also highly variable between 

plants, and greater replication in our study may have allowed for more significant trends in the 

61 

 
 
 
data. Despite this, there is a clear effect of the legacy of G1 drought on the microbiome of the 

Flavert plants in generation 2, and this is also reflected in the plant trait data. When designing 

our experiment, we predicted that there could be beneficial plant microbiome members 

enriched in the first generation of drought, that could be passed on to the following generation 

via vertical transmission, thus resulting in better health outcomes in repeated stress when 

primed with a helpful microbiome from parental plants35,43,75–79. This is not evident in our 

experiment, as the G2 Flavert plants that received drought in both generations fared the worst 

in plant health outcomes, and plants that received control treatment followed by a drought had 

significant shifts in their rhizosphere community structure and dispersion. However, these 

results still support the hypothesis that repeated drought exposure would have consequences 

for plant health in the second generation.  

While the consequences to plant health under repeated drought exposure were 

negative, it is intriguing that when indicator taxa are analyzed for significant associations with 

the drought treatment, we can identify taxonomic orders that are consistently enriched in the 

drought treatment across generations, and even in plants grown in different soils and between 

the two different genotypes. The taxonomic orders Xanthomonadales and Rhizobiales were 

strongly associated with the drought-treated plants in the Red Hawk roots and rhizosphere, and 

Flavert rhizosphere. Both orders contain many taxa that have been identified as having 

beneficial impacts on plant health, such as Xanthomonas sacchari, Rhizobium calliandrae, 

Rhizobium phaseoli, prevalent in common bean root nodules, and Stenotrophomonas 

rhizophila, which has been identified as a core seed microbiome member in common bean38,79–

82. These orders are also known to contain some plant pathogen species such as Xanthomonas 

62 

 
 
citri83. Additionally, in the Michigan Red Hawk plants there were specific ASVs that were 

consistently associated with the drought treatment from the genera Lysobacter, 

Promicromonospora and Fluviicola, which could be investigated further to understand their role 

in the plant microbiome under drought stress. The taxa identified in this study may provide 

important insight into how the root-associated microbiome is supporting the plant during 

repeated drought exposure, and further investigation is necessary to understand the specific 

bacterial strains associated with the drought stress and their role in plant health. 

In summary, our study sheds light on how repeated exposure to abiotic stress over 

multiple generations can impact the plant microbiome, and the importance of understanding 

the impacts of climate change on microbial communities. Alpha and beta diversity of the root 

and rhizosphere communities were significantly altered by drought stress in both common 

bean genotypes, with greater effects in the Flavert genotype. It was apparent that the plant 

microbiomes in the two locations also had a significantly different alpha diversity in both the 

root and rhizosphere communities, which may have influenced how the microbial communities 

responded to the drought treatment. Many variables could have been attributed to these 

differences, including the soil source region and type, the slight differences in stress treatment 

period, and the different DNA extraction kits used in the two locations. However, trends in the 

bacterial communities’ responses to the drought treatment were still evident despite these 

differences. Soil type may have played a role in how the Red Hawk plants responded to the 

drought treatment in the two locations, and future studies could incorporate soils from 

different regions when conducting plant microbiome growth experiments to address these 

potential differences.  

63 

 
 
Drought in agricultural areas is becoming more and more prevalent around the world, 

and is particularly devastating to developing countries that rely on staple food crops like 

common bean. It is important that we find sustainable solutions to support agriculture under 

changing climate conditions, and applications of beneficial microbes may hold great promise for 

improving our agricultural systems84–86. This study provides evidence that repeated seasons of 

drought stress will likely be detrimental to agricultural crops and their microbiomes over time 

as climate change continues to affect agricultural regions. The microbial taxa identified here 

present a potential area of investigation into bacterial taxa that could be harnessed as 

biological products to support beans, or agricultural crops more broadly, and future research in 

this area may allow for improved crop production under drought stress in the future. 

64 

 
 
 
REFERENCES 

1. 

2. 

3. 

4. 

5. 

6. 

7. 

8. 

9. 

Furtak, K. & Wolińska, A. The impact of extreme weather events as a consequence of 
climate change on the soil moisture and on the quality of the soil environment and 
agriculture – A review. Catena (Amst) 231, (2023). 

Maity, A. et al. Climate Change Impacts on Seed Production and Quality: Current 
Knowledge, Implications, and Mitigation Strategies. Seed Science and Technology 51, 
65–96 (2023). 

Semeraro, T., Scarano, A., Leggieri, A., Calisi, A. & De Caroli, M. Impact of Climate 
Change on Agroecosystems and Potential Adaptation Strategies. Land (Basel) 12, 1117 
(2023). 

Ackerl, T., Weldemariam, L. F., Nyasimi, M. & Ayanlade, A. Climate change risk, 
resilience, and adaptation among rural farmers in East Africa: A literature review. 
Regional Sustainability vol. 4 185–193 Preprint at 
https://doi.org/10.1016/j.regsus.2023.05.004 (2023). 

Chivangulula, F. M., Amraoui, M. & Pereira, M. G. The Drought Regime in Southern 
Africa: A Systematic Review. Climate 11, (2023). 

Gourdji, S., Läderach, P., Valle, A. M., Martinez, C. Z. & Lobell, D. B. Historical climate 
trends, deforestation, and maize and bean yields in Nicaragua. Agric For Meteorol 200, 
270–281 (2015). 

Madadgar, S., AghaKouchak, A., Farahmand, A. & Davis, S. J. Probabilistic estimates of 
drought impacts on agricultural production. Geophys Res Lett 44, 7799–7807 (2017). 

Ouko, K. O. & Odiwuor, M. O. Contributing factors to the looming food crisis in sub-
Saharan Africa: Opportunities for policy insight. Cogent Soc Sci 9, (2023). 

Cheng, Y. T., Zhang, L. & He, S. Y. Plant-Microbe Interactions Facing Environmental 
Challenge. Cell Host Microbe 26, 183–192 (2019). 

10. 

Hartman, K. & Tringe, S. G. Interactions between plants and soil shaping the root 
microbiome under abiotic stress. Biochemical Journal 476, 2705–2724 (2019). 

11. 

Shade, A. Understanding Microbiome Stability in a Changing World. (2018). 

12. 

Trivedi, P., Batista, B. D., Bazany, K. E. & Singh, B. K. Plant–microbiome interactions 
under a changing world: responses, consequences and perspectives. New Phytologist 
234, 1951–1959 (2022). 

13. 

Compant, S., Samad, A., Faist, H. & Sessitsch, A. A review on the plant microbiome: 
Ecology, functions, and emerging trends in microbial application. J Adv Res 19, (2019). 

65 

 
 
14. 

15. 

16. 

17. 

18. 

Trivedi, P., Leach, J. E., Tringe, S. G., Sa, T. & Singh, B. K. Plant–microbiome interactions: 
from community assembly to plant health. Nat Rev Microbiol 18, 607–621 (2020). 

Noman, M. et al. Plant–microbiome crosstalk: Dawning from composition and assembly 
of microbial community to improvement of disease resilience in plants. Int J Mol Sci 22, 
(2021). 

Santos, L. F. & Olivares, F. L. Plant microbiome structure and benefits for sustainable 
agriculture. Curr Plant Biol 26, 100198 (2021). 

Chen, S. et al. Root-associated microbiomes of wheat under the combined effect of 
plant development and nitrogen fertilization. Microbiome 7, (2019). 

Liu, F. et al. Soil indigenous microbiome and plant genotypes cooperatively modify 
soybean rhizosphere microbiome assembly. BMC Microbiol 19, (2019). 

19.  Walters, W. A. et al. Large-scale replicated field study of maize rhizosphere identifies 

heritable microbes. Proc Natl Acad Sci U S A 115, 7368–7373 (2018). 

20.  Mavrodi, D. V. et al. Long-term irrigation affects the dynamics and activity of the wheat 

rhizosphere microbiome. Front Plant Sci 9, (2018). 

21. 

22. 

23. 

24. 

Russell, K. A. & McFrederick, Q. S. Floral nectar microbial communities exhibit seasonal 
shifts associated with extreme heat: Potential implications for climate change and plant-
pollinator interactions. Front Microbiol 13, (2022). 

Santos-Medellín, C., Edwards, J., Liechty, Z., Nguyen, B. & Sundaresan, V. Drought stress 
results in a compartment-specific restructuring of the rice root-associated microbiomes. 
mBio 8, (2017). 

Timm, C. M. et al. Abiotic Stresses Shift Belowground Populus -Associated Bacteria 
Toward a Core Stress Microbiome . mSystems 3, (2018). 

Veach, A. M. et al. Plant Hosts Modify Belowground Microbial Community Response to 
Extreme Drought. mSystems 5, (2020). 

25.  Wang, P., Marsh, E. L., Kruger, G., Lorenz, A. & Schachtman, D. P. Belowground microbial 
communities respond to water deficit and are shaped by decades of maize hybrid 
breeding. Environ Microbiol 22, 889–904 (2020). 

26. 

27. 

Tiziani, R. et al. Drought, heat, and their combination impact the root exudation 
patterns and rhizosphere microbiome in maize roots. Environ Exp Bot 203, (2022). 

Castro-Guerrero, N., Isidra-Arellano, M., Mendoza-Cozatl, D. & Valdés-López, O. 
Common Bean: A Legume Model onthe Rise for Unraveling Responsesand Adaptations 
to Iron, Zinc, andPhosphate Deficiencies. Front Plant Sci 7, 391–417 (2016). 

66 

 
 
28. 

29. 

30. 

31. 

32. 

33. 

34. 

35. 

36. 

37. 

38. 

39. 

40. 

Uebersax, M. A. et al. Dry beans (Phaseolus vulgaris L.) as a vital component of 
sustainable agriculture and food security—A review. Legume Science 5, (2023). 

Graham, P. & Ranalli, P. Field Crops Research Common bean (Phaseolus vulgaris L.). Field 
Crops Res 53, 131–146 (1997). 

Schmutz, J. et al. A reference genome for common bean and genome-wide analysis of 
dual domestications. Nat Genet 46, 707–713 (2014). 

Stacey, G. Biology of the Nitrogen Cycle - Chapter 10: The Rhizobium-Legume Nitrogen-
Fixing Symbiosis. in (Elsevier B. V., 2007). 

Shrestha, R. P. & Nepal, N. An assessment by subsistence farmers of the risks to food 
security attributable to climate change in Makwanpur, Nepal. Food Secur 8, 415–425 
(2016). 

Petrushin, I. S., Vasilev, I. A. & Markova, Y. A. Drought Tolerance of Legumes: Physiology 
and the Role of the Microbiome. Curr Issues Mol Biol 45, 6311–6324 (2023). 

Sofi, P. A. et al. Integrating root architecture and physiological approaches for improving 
drought tolerance in common bean (Phaseolus vulgaris L.). Plant Physiology Reports 26, 
4–22 (2021). 

Bintarti, A. F., Kearns, P. J., Sulesky-Grieb, A. & Shade, A. Abiotic Treatment to Common 
Bean Plants Results in an Altered Endophytic Seed Microbiome. Microbiol Spectr 10, 
(2022). 

Chesneau, G. et al. Temporal dynamics of bacterial communities during seed 
development and maturation. FEMS Microbiol Ecol 96, (2020). 

Da Silva, P. M., Tsai, S. M. & Bonetti, R. Response to inoculation and N fertilization for 
increased yield and biological nitrogen fixation of common bean (Phaseolus vulgaris L.). 
Plant Soil 152, 123–130 (1993). 

De Ron, A. M. et al. Diversity and conservation of wild and primitive common bean 
germplasm and their associated rizobia in the andean region. in Legumes for Global 
Food Security (2017). 

Grange, L., Hungria, M., Graham, P. H. & Martínez-Romero, E. New insights into the 
origins and evolution of rhizobia that nodulate common bean (Phaseolus vulgaris) in 
Brazil. Soil Biol Biochem 39, 867–876 (2007). 

Nishu, S. Das, No, J. H. & Lee, T. K. Transcriptional Response and Plant Growth 
Promoting Activity of Pseudomonas fluorescens DR397 under Drought Stress 
Conditions. Microbiol Spectr 10, (2022). 

67 

 
 
41. 

42. 

43. 

44. 

45. 

46. 

47. 

48. 

49. 

50. 

51. 

Pérez-Jaramillo, J. E. et al. Linking rhizosphere microbiome composition of wild and 
domesticated Phaseolus vulgaris to genotypic and root phenotypic traits. ISME Journal 
11, 2244–2257 (2017). 

Stopnisek, N. & Shade, A. Persistent microbiome members in the common bean 
rhizosphere: an integrated analysis of space, time, and plant genotype. ISME Journal 15, 
2708–2722 (2021). 

Lau, J. A. & Lennon, J. T. Rapid responses of soil microorganisms improve plant fitness in 
novel environments. 109, (2012). 

Bitocchi, E. et al. Beans (Phaseolus ssp.) as a model for understanding crop evolution. 
Front Plant Sci 8, (2017). 

Carrère, S. et al. First whole genome assembly and annotation of a European common 
bean cultivar using PacBio HiFi and Iso-Seq data. Data Brief 48, (2023). 

Kelly, J. D. et al. Registration of ‘Red Hawk’ Dark Red Kidney Bean. Crop Sci 38, 280–281 
(1998). 

Colovas, J., Bintarti, A. F., Mechan Llontop, M. E., Grady, K. L. & Shade, A. Do-it-Yourself 
Mock Community Standard for Multi-Step Assessment of Microbiome Protocols. Curr 
Protoc 2, (2022). 

Apprill, A., Mcnally, S., Parsons, R. & Weber, L. Minor revision to V4 region SSU rRNA 
806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquatic 
Microbial Ecology 75, 129–137 (2015). 

Caporaso, J. G. et al. Global patterns of 16S rRNA diversity at a depth of millions of 
sequences per sample. Proc Natl Acad Sci U S A 108, 4516–4522 (2011). 

Caporaso, J. G. et al. Ultra-high-throughput microbial community analysis on the 
Illumina HiSeq and MiSeq platforms. ISME Journal 6, 1621–1624 (2012). 

Parada, A. E., Needham, D. M. & Fuhrman, J. A. Every base matters: Assessing small 
subunit rRNA primers for marine microbiomes with mock communities, time series and 
global field samples. Environ Microbiol 18, 1403–1414 (2016). 

52.  Walters, W. et al. Improved Bacterial 16S rRNA Gene (V4 and V4-5) and Fungal Internal 
Transcribed Spacer Marker Gene Primers for Microbial Community Surveys. mSystems 
1, (2016). 

53. 

Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data 
science using QIIME 2. Nature Biotechnology vol. 37 852–857 Preprint at 
https://doi.org/10.1038/s41587-019-0209-9 (2019). 

68 

 
 
54. 

55. 

56. 

Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon 
data. Nat Methods 13, 581–583 (2016). 

Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data 
processing and web-based tools. Nucleic Acids Res 41, (2013). 

R Core Team. R: A Language and Environment for Statistical Computing. Preprint at 
(2023). 

57.  McMurdie, P. J. & Holmes, S. Phyloseq: An R Package for Reproducible Interactive 

Analysis and Graphics of Microbiome Census Data. PLoS One 8, (2013). 

58. 

59. 

60. 

Davis, N. M., Proctor, Di. M., Holmes, S. P., Relman, D. A. & Callahan, B. J. Simple 
statistical identification and removal of contaminant sequences in marker-gene and 
metagenomics data. Microbiome 6, (2018). 

Oksanen, J. et al. vegan: Community Ecology Package. R package version 2.5-7. 
https://CRAN.R-project.org/package=vegan (2019). 

Barnett, D., Arts, I. & Penders, J. microViz: an R package for microbiome data 
visualization and statistics. J Open Source Softw 6, 3201 (2021). 

61.  Wickham, H. Ggplot2: Elegant Graphics for Data Analysis. (Springer-Verlag, New York, 

2016). 

62.  Martinez Arbizu, P. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis. 

(2017). 

63. 

64. 

65. 

66. 

Anderson, M. J. Distance-based tests for homogeneity of multivariate dispersions. 
Biometrics 62, 245–253 (2006). 

De Cáceres, M., Jansen, F. & Dell, N. indicspecies: Relationship Between Species and 
Groups of Sites. https://cran.r-project.org/web/packages/indicspecies/index.html 
(2023). 

De Cáceres, M. & Legendre, P. Associations between species and groups of sites: Indices 
and statistical inference. Ecology 90, 3566–3574 (2009). 

Chen, H. VennDiagram: Generate High-Resolution Venn and Euler Plots. Preprint at 
https://CRAN.R-project.org/package=VennDiagram (2022). 

67.  Wickham, H. et al. Welcome to the Tidyverse. J Open Source Softw 4, 1686 (2019). 

68. 

Pedersen, T. L. patchwork: The Composer of Plots. https://CRAN.R-
project.org/package=patchwork (2023). 

69 

 
 
69. 

70. 

71. 

72. 

73. 

74. 

75. 

76. 

77. 

78. 

79. 

80. 

Beebe, S., Rengifo, J., Gaitan, E., Duque, M. C. & Tohme, J. Plant genetic resources: 
Diversity and origin of Andean landraces of common bean. Crop Sci 41, 854–862 (2001). 

Abdelfattah, A. et al. Evidence for host–microbiome co-evolution in apple. New 
Phytologist 234, 2088–2100 (2022). 

Cordovez, V. et al. Successive plant growth amplifies genotype-specific assembly of the 
tomato rhizosphere microbiome. Science of the Total Environment 772, (2021). 

Bayati, K., Hosseini, N., Moghadam, H. & Basiri, R. Effects of drought stress and nitrogen 
on grain yield and some agronomic traits of red kidney bean cultivars. Iranian Journal of 
Field Crop Science 48, Pe1069–Pe1081 (2018). 

Beaver, J. S. & Osorno, J. M. Achievements and limitations of contemporary common 
bean breeding using conventional and molecular approaches. Euphytica 168, 145–175 
(2009). 

Dramadri, I. O., Nkalubo, S. T., Kramer, D. M. & Kelly, J. D. Genome-wide association 
analysis of drought adaptive traits in common bean. Crop Sci 61, 3232–3253 (2021). 

Abdelfattah, A., Wisniewski, M., Schena, L. & Tack, A. J. M. Experimental evidence of 
microbial inheritance in plants and transmission routes from seed to phyllosphere and 
root. Environ Microbiol (2021) doi:10.1111/1462-2920.15392. 

Barret, M. et al. Emergence shapes the structure of the seed microbiota. Appl Environ 
Microbiol 81, 1257–1266 (2015). 

Shade, A., Jacques, M. A. & Barret, M. Ecological patterns of seed microbiome diversity, 
transmission, and assembly. Curr Opin Microbiol 37, 15–22 (2017). 

Rezki, S. et al. Assembly of seed-associated microbial communities within and across 
successive plant generations. Plant Soil 422, 67–79 (2018). 

Simonin, M. et al. Seed microbiota revealed by a large-scale meta-analysis including 50 
plant species. New Phytologist 234, 1448–1463 (2022). 

Acuña, J. J. et al. Diversity, Community Structure, and Potential Functions of Root-
Associated Bacterial Communities of Different Wheat (Triticum aestivum) Cultivars 
under Field Conditions. Agronomy 13, (2023). 

81. 

Gen-Jiménez, A. et al. Enhance of tomato production and induction of changes on the 
organic profile mediated by Rhizobium biofortification. Front Microbiol 14, (2023). 

82.  Wang, X. et al. Early inoculation of an endophyte alters the assembly of bacterial 

communities across rice plant growth stages. Microbiol Spectr (2023) 
doi:10.1128/spectrum.04978-22. 

70 

 
 
83. 

84. 

85. 

86. 

Huang, F. et al. Canker Disease Intensifies Cross-Kingdom Microbial Interactions in the 
Endophytic Microbiota of Citrus Phyllosphere. Phytobiomes J (2023) 
doi:10.1094/pbiomes-11-22-0091-r. 

Berg, G., Kusstatscher, P., Abdelfattah, A., Cernava, T. & Smalla, K. Microbiome 
Modulation—Toward a Better Understanding of Plant Microbiome Response to 
Microbial Inoculants. Front Microbiol 12, (2021). 

Finkel, O. M., Castrillo, G., Herrera Paredes, S., Salas González, I. & Dangl, J. L. 
Understanding and exploiting plant beneficial microbes. Curr Opin Plant Biol 38, 155–
163 (2017). 

Vurukonda, S. S. K. P., Vardharajula, S., Shrivastava, M. & SkZ, A. Enhancement of 
drought stress tolerance in crops by plant growth promoting rhizobacteria. Microbiol 
Res 184, 13–24 (2016). 

71 

 
 
 
 
APPENDIX A: 
CHAPTER 2 SUPPLEMENTARY MATERIALS 

Table S1.1 
Parameters used in 16S sequence processing. 

Dada2 trim 
parameters 

Taxonomy 
assignment 

Decontam 
parameters 

Rarefaction 

silva-138-99-515-806-
nb-classifier.qza 

Threshold 0.1 

20,976 

silva-138-99-515-806-
nb-classifier.qza 

Threshold 0.1 

20,976 

Silva 132 99% 

Threshold 0.2 

Silva 132 99% 

Threshold 0.2 

--p-trunc-len-f 
114 \ 
--p-trunc-len-r 
161 \ 

--p-trunc-len-f 
103 \ 
--p-trunc-len-r 
172 \ 

--p-trunc-len-f 
200 \ 
--p-trunc-len-r 
180- \ 

--p-trunc-len-f 
200 \ 
--p-trunc-len-r 
130- \ 

Dataset 

MSU Roots 16S 

MSU Rhizosphere 16S 

INRAE Root and 
Rhizosphere 16S - 
Generation 1 

INRAE Root and 
Rhizosphere 16S - 
Generation 2 

72 

 
 
 
 
 
 
 
 
 
Figure S1.1. Experimental design. Phaseolus vulgaris L. (common bean) plants were grown over 
two generations in a factorial design under either control treatment with ample water and 
nutrients, or drought treatment with 66% less water and equal nutrient concentration to 
control plants. The common bean variety “Red Hawk” was grown in Michigan, and both “Red 
Hawk” and “Flavert” were both grown in Pays de la Loire. Drought treatments were started at 
the V3 growth stage when plants had three trifoliate leaves expanded, and were concluded at 
the R1 stage (first open flowers) in Michigan, and the R5 stage (half of the pods filled with 
discernable seeds) in Pays de la Loire. 

73 

 
 
 
Figure S1.2. Indicator taxa identified in roots and rhizosphere of Generation 1 plants grown in 
Michigan, USA, grouped by taxonomic order. 

74 

 
 
 
 
Figure S1.3. Indicator taxa identified in roots and rhizosphere of Generation 2 plants grown in 
Michigan, USA, grouped by taxonomic class. 

75 

 
 
 
 
 
 
 
Stable transmission of common bean seed microbiome members over three plant 
generations despite exposure to abiotic stress 

CHAPTER 3:  

76 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

The seed microbiome plays an important role in plant microbiome assembly, serving as 

the starting point for microbiome members to establish in the germinating seedling. Although 

there is a broad area of work being done on the plant microbiome and its role in supporting 

plants under biotic and abiotic stress conditions, few studies have addressed the impact of 

plant stress on the resulting seed microbiome that is passed to the next generation. We 

conducted a multigenerational study to investigate the impact of drought stress or excess 

nutrients on the endophytic bacterial community in common bean seeds (Phaseolus vulgaris L. 

var. Red Hawk) while tracking specific parental lines through two generations. The stress 

treatments had a negligible effect on the resulting seed microbiomes. However, we identified a 

significant impact of parental line in the offspring seeds and discovered a signature of stable 

transmission of 22 prevalent seed microbiome members, regardless of plant treatment. These 

prevalent taxa included previously identified core taxa for common bean, and could be a 

valuable point of interest for the development of beneficial bacterial applications in agriculture. 

This study provides insights into the importance of parental line in seed microbiome studies 

and the maintenance of the plant microbiome across generations, even under challenging 

environmental conditions. 

77 

 
 
 
 
 
 
INTRODUCTION 

The plant microbiome is comprised of bacterial, archaeal, and fungal cells that inhabit all 

plant compartments, including the plant’s surfaces, internal tissues, and root structures1,2. 

These microbial species play an important role in plant health through water and nutrient 

assimilation, conferral of stress tolerance, pathogen defense, and other services2–4. Many 

studies have investigated the unique microbiomes of various plant species, either wild or 

domesticated. There is an extensive collection of literature on the beneficial rhizosphere, 

phyllosphere and endosphere microbiome in various plant species, which continue to be 

growing areas of research as the need for sustainable solutions to improve crop health 

increases5–12. One area of particular interest is the seed microbiome and its role in both 

pathogen transmission and shaping the beneficial microbiome of the plant13–15. While plant 

microbiomes have been well-documented, few studies have investigated the inheritance of 

microbiome members over plant generations, especially when mediated through seed 

transmission16–22.  

Plant microbiomes can be assembled via both horizontal transmission from the 

environment, and vertical transmission from the parent plant through the seed microbiome23–

28. Microbial cells can move through the parent plant via vascular tissue or floral compartments 

and become packaged within the internal seed tissues, and can also colonize seed surfaces28–32. 

When the seed is germinating, priority effects may allow the seed microbiome members to 

establish first in the microbiome of the new plant, potentially affecting the trajectory of the 

plant’s microbial community33–35. Additionally, seed-derived bacteria and fungi have been 

identified in seedling roots, shoots and rhizosphere compartments in a variety of plant 

78 

 
 
 
 
species34. These “inherited” microbiome members enriched by the parent plant may provide 

important benefits for the offspring plants in the next generation16,17,36–39. 

While the recruitment mechanisms of the seed microbiome have been well 

documented, only a handful of studies have investigated if the environmental conditions of the 

parent play a role in determining the microbiome of the resulting seeds14,40–42. Parental plant 

line effects on the seed microbiome have been investigated in a limited number of studies19,43. 

Bintarti et al. found that in common bean plants exposed to drought or provided extra nutrients 

in environmental growth chambers, differences in the bacterial community of the seeds could 

be identified based on the treatment that the parent plant received44. However, this study had 

low replication and the soil used was not representative of a true field system. Additionally, 

fluctuations in wheat seed fungal endophytes have been reported under drought stress 

conditions, as well as significant shifts in seed bacterial endophytes in rice under salt-stress and 

Brassica in reduced nitrogen environments45–47. Consequently, if the seed microbiome can be 

altered by environmental conditions, then the altered microbiome members that are 

establishing in the next plant generation may have consequences, either positive or negative, 

for the next plant’s ability to survive in similar conditions48. However, because the seed 

microbiome has been reported to have low diversity and small community size (tens to 

hundreds of taxa) combined with high compositional variability for many plant species, a stable 

signature of vertically inherited taxa may not be expected18,49,50. Instead, the seed microbiome 

may represent an accumulation of stochastically dispersed cells and taxa that are not 

necessarily plant-supportive, plus some inconsistently detected plant-associated (beneficial or 

79 

 
 
 
pathogenic) taxa. Thus, our study aimed to understand how parental plant exposure to abiotic 

treatments impacted any potential vertical transmission of seed microbiome members.  

We conducted a multigenerational experiment in which we exposed common bean 

plants (Phaseolus vulgaris L. var. Red Hawk) to either drought or high nutrients during their 

early vegetative growth in agricultural soil under controlled growth chamber conditions. We 

were motivated to study common bean because it is a critical legume for global food security, 

supporting the health and livelihood of millions of people worldwide51–53. The two abiotic 

treatments were chosen because of their relevance to bean production given the climate crisis. 

Beans are threatened by drought associated with warming and changes in precipitation 

patterns54–56.  Furthermore, beans in the United States and in other areas of high bean 

production are often managed with excess fertilizer, which can impact the selection and 

stability of the plant microbiome, such as limiting the diversity of nitrogen-fixing rhizobium 

found in root nodules57. Exogenous nitrogen fertilizers contribute significantly to global 

greenhouse gas emissions, and it is important to limit any excess use of these products58. 

We hypothesized that (1) Abiotic treatment to the parent plant would characteristically 

alter the seed endophytic microbiome in alpha and/or beta diversity measures, and (2) 

Characteristic taxa would be vertically transmitted to the next plant generation that is exposed 

to the same treatment. Seeds provided from a “Generation 0 (zero)” (G0) seed pool were 

planted in a growth chamber and exposed to either a control condition with ample water and 

nutrients, a drought condition with 66% decreased water but equivalent nutrients, or a nutrient 

condition with 3X concentrated nutrient solution and equivalent water to control. Seeds from 

these “Generation 1” (G1) plants were harvested and used for growth in “Generation 2” (G2), 

80 

 
 
 
where plants were exposed to the treatments in a factorial design, with offspring receiving one 

of each of the three treatments in G2. 16S V4 rRNA gene amplicon sequencing of the seed 

bacterial endophyte communities was performed for each of the three generations to assess 

the impact of the treatments on the seed microbiome and the vertical transmission of the 

bacterial taxa. This work contributes to a significant knowledge gap in understanding how 

shifting environmental conditions affect the vertically transmitted seed microbiota over 

multiple plant generations. 

METHODS 

Bean cultivar 

Phaseolus vulgaris L. var. Red Hawk, developed by the Michigan State University Bean 

Breeding program was selected as a representative dry bean crop59 Red Hawk seeds were 

obtained from the Michigan State University Bean Breeding Program from their 2019 harvest 

and stored at 4°C until ready for use in experiments. These seeds were obtained as “Generation 

0” and used to plant the first generation of the experiment. 

Soil preparation 

Agricultural field soil was collected for each planting group from the same field location 

in September 2019, December 2020, May 2021 and September 2021, from a Michigan State 

University Agronomy Farm field that was growing common bean in 2019 (42°42'57.4"N, 

84°27'58.9"W, East Lansing, MI, USA). The soil was a sandy loam with an average pH of 7.2. 

Five-gallon buckets were used to collect enough soil needed for the planting experiment, 

avoiding the dry top layer of soil and plant debris. Buckets were covered with air-tight lids and 

stored at 4°C until the experiment began. Before use in planting, soil was passed through a 

81 

 
 
4mm sieve to remove rocks and plant debris, and the soil was mixed with autoclaved coarse 

vermiculite at a 50% v/v ratio for use in growth chamber pots. 

Surface sterilization and seed germination 

For each planting group, Red Hawk seeds were surface sterilized prior to planting with a 

solution of 10% bleach and 0.1% Tween20. Seeds were randomly selected from the bulk G0 

seed supply, or the harvested G1 seed supply, avoiding seeds that were visibly cracked or 

moldy. Approximately 20 seeds were placed in sterile 50 mL conical tubes, 20-30 mL of bleach 

solution was added to the tubes and the seeds were soaked for 10 minutes, with agitation at 5 

minutes. After soaking, the seeds were rinsed 5 times with sterile DI H2O. On the final rinse, 

approximately 100 µL of rinse water was spread onto Tryptic Soy Agar (TSA) and Potato 

Dextrose Agar (PDA) plates to assess the efficacy of the seed surface sterilization. TSA plates 

were incubated overnight at 28°C, and PDA plates were incubated at room temperature for 48 

hours. Seeds corresponding to plates with microbial growth were discarded from the 

experiment and replaced with seeds that were surface-sterile. For germination, seeds were 

placed in a Petri dish lined with sterile filter paper wetted with 1-2 mL sterile DI H2O. Petri 

dishes were stored in the dark at room temperature for 3-4 days, with an additional 2 mL sterile 

DI H2O added after 2 days. Once seeds had sprouted radicle roots, they were transferred to 

soil.  

Growth conditions 

See Figure S2.1 for experimental design schematic. For G1, three germinated seeds were 

planted into 3.78 L pots were filled with the soil-vermiculite mixture and placed in a high-light 

BioChambers FLEX™ LED growth chamber with a 16-hour day/8-hour night cycle at 26°C and 

82 

 
 
22°C, respectively, and 50% relative humidity. Once seedlings emerged and reached the VC 

growth stage (vegetative growth with two cotyledons and primary leaves expanded), they were 

thinned to one seedling per pot. Plants were watered every other day with 300 mL 0.05% 15-

10-30 water-soluble fertilizer solution (control condition) (Masterblend International, Morris, IL, 

USA). At the V3 stage (vegetative growth with third trifoliate leaves expanded), treatments 

began for the drought- and nutrient-treated plants. Drought plants received 100 mL of 0.15% 

15-10-30 fertilizer solution every other day (66% less water than control with the same 

concentration of nutrients), and nutrient plants received 300 mL of 0.15% 15-10-30 fertilizer 

solution every other day (3X concentrated nutrients with the same volume of water as control). 

After approximately 14 days of treatment period, when plants reached the R1 stage 

(reproductive stage, first open flowers), they were returned to the control-treatment watering 

regime until senescence. As plants began to dry, they were watered less frequently as needed. 

Mature seeds were collected from 12 plants per treatment for seed microbiome assessment 

and for growth for the next generation. 

For Generation 2 (G2), seeds from the 36 G1 parental lines that received either control, 

drought or nutrient conditions were planted in a full factorial design and grown under each of 

three treatment conditions in G2. There were nine cross-generational treatment combinations 

total (G1_G2, n=12 plants per treatment combination): Control_Control, Control_Drought, and 

Control_Nutrient; Drought_Control, Drought_Drought, and Drought_Nutrient; 

Nutrient_Control, Nutrient_Drought, and Nutrient_Nutrient. Six seeds from each parental line 

were surface sterilized and germinated as described above, then planted in the field soil-

vermiculite mixture in seedling trays in the growth chamber under the conditions stated above. 

83 

 
 
G2 plants were grown in three randomized planting groups, with each planting group 

containing parental lines from all three treatments. Once plants reached the VC stage, three 

healthy seedlings per parent line were transferred to 3.78 L pots. Each G1 parental line 

provided one offspring per G2 treatment, for a total of 108 plants in G2. Plants were watered 

according to the conditions and treatment timeline in G1. 

Seed harvest 

Once the plants had senesced and pods were dried, seeds were harvested for planting 

or microbiome analysis. Seed pods were removed from each plant and stored in sterile Whirl-

pakâ bags for transport to the lab. Pods and seeds per plant were counted, and then seeds 

were removed from the pods and pooled by plant in 50 mL conical tubes and stored at 4°C for 

use in planting. 5 seeds per plant were selected for microbiome analysis and stored in a 

separate 15 mL conical tube at -80°C until DNA extraction was performed. 

DNA extractions 

Seed microbiome DNA extractions were performed on sets of five seeds, which is our 

unit of microbiome sampling. For the G0 bulk seed, twenty sets of five randomly selected seeds 

were analyzed from the G0 seed pool. In G1 and G2, seeds were pooled by plant and five seeds 

per plant were analyzed from the 36 G1 parent plants and 108 G2 offspring plants. Seeds were 

thawed and surface sterilized according to the method above, and then microbial DNA was 

extracted from the endophytic compartment using a protocol adapted from Barret et al. 2015 

and Bintarti et al. 202124,49. Following surface sterilization, the seeds were sliced in half 

lengthwise along the natural division of the cotyledons with a sterile razor blade. Sliced seeds 

were then placed in a 50 mL conical tube and 20-30 mL of sterile Phosphate-buffered Saline 

84 

 
 
(PBS) with 0.05% Tween 20 was added. Seeds were soaked overnight at 4°C with constant 

agitation on an orbital shaker at 160 rpm to allow microbial material to be released from the 

seed tissues. After soaking, tubes were centrifuged at 4500´g, 4°C, for one hour. Seed tissue 

and supernatant were removed, and the remaining pellet was transferred to a 1.5 mL 

microcentrifuge tube. Pellets were stored at -80°C until extraction with the E.Z.N.A. Bacterial 

DNA kit (Omega Bio-tek, Inc., Norcross, GA, USA) following manufacturer’s protocol with the 

following modifications. To begin the protocol, the seed material pellet was resuspended in 100 

µL TE Buffer, 10 µL kit-provided Lysozyme was added, the samples were vortexed thoroughly, 

and incubated at 37°C for 1 hour. The glass bead step from the E.Z.N.A. kit was utilized with 25-

30 mg glass beads, provided, and samples were vortexed at maximum speed for 10 minutes in a 

24-tube vortex adapter. After adding the Proteinase K, the samples were incubated in a shaking 

heat block at 55°C for 2 hours. In the final step, DNA was eluted in 60 µL Elution Buffer and 

incubated at 65°C for 10 minutes before centrifuging into the final tube. 

DNA extractions were performed in randomized batches within each generation (Table 

S2.1). For each batch, a negative and positive control were included. The negative control was 3 

mL sterile PBS+Tween buffer, and the positive control was an aliquot of a mixture of cells from 

a custom-made mock bacterial community suspended in 3 mL PBS+Tween buffer 60. These 

controls were shaken overnight alongside the seed samples and then processed and sequenced 

as described for the seeds, and then ultimately used to perform batch-informed bioinformatic 

sequence decontamination 61. 

Sequencing 

Sequencing of the V4 region of the 16S rRNA gene (515F-806R)62,63 was performed at 

85 

 
 
the Environmental Sample Preparation and Sequencing Facility (ESPSF) at Argonne National 

Laboratory (Lemont, IL, USA). The DNA was PCR amplified with region-specific primers that 

include sequencer adapter sequences used in the Illumina Nextseq2K flowcell; 

FWD:GTGYCAGCMGCCGCGGTAA; REV:GGACTACNVGGGTWTCTAAT62–66. Each 25 µL PCR 

reaction contained 9.5 µL of MO BIO PCR Water (Certified DNA-Free), 12.5 µL of QuantaBio’s 

AccuStart II PCR ToughMix (2x concentration, 1x final), 1 µL Golay barcode tagged Forward 

Primer (5 µM concentration, 200 pM final), 1 µL Reverse Primer (5 µM concentration, 200 pM 

final), and 1 µL of template DNA. The conditions for PCR were as follows: 94 °C for 3 minutes to 

denature the DNA, with 35 cycles at 94 °C for 45 s, 50 °C for 60 s, and 72 °C for 90 s; with a final 

extension of 10 min at 72 °C to ensure complete amplification. Amplicons were then quantified 

using PicoGreen (Invitrogen) and a plate reader (InfiniteÒ 200 PRO, Tecan). Once quantified, 

volumes of each of the products were pooled into a single tube, in equimolar amounts, this 

pool was then cleaned up using AMPure XP Beads (Beckman Coulter), and then quantified using 

a fluorometer (Qubit, Invitrogen). After quantification, the pool was diluted to 2 nM, 

denatured, and then diluted to a final concentration of 6.75 pM with a 10% PhiX spike for 

sequencing. Amplicons were sequenced on a 251bp x 12bp x 251bp Nextseq2000 run65. 

Sequence data processing 

Sequencing fastq files were processed in QIIME2 after primer removal by the 

sequencing center (QIIME2 version: 2022.8.0)67. Sample fastqs were imported to QIIME2 

format, and samples were denoised, truncated and merged using Dada2, with a forward 

truncation length of 191, and reverse truncation length of 8468. Amplicon sequence variants 

86 

 
 
(ASVs) were defined at 100% sequence identity and 16S taxonomy was assigned with the Silva 

database release 138, and taxonomy and ASV tables were exported for further analysis in R69.  

Sequence Decontamination  

Data analyses were performed in R version 4.3.1 and R Studio version 2023.06.1+52470. 

There were 126.8 million merged DNA reads in the dataset prior to host removal and 

decontamination. ASV, taxonomy, and metadata tables, and phylogenetic tree files, were 

imported into the phyloseq package and host reads classified as chloroplast and mitochondria 

were removed using the subset_taxa() command in the phyloseq package version 1.44.071. 90% 

of the total DNA reads and 13% of the ASVs were removed as host reads, leaving 12.3 million 

total bacterial DNA reads. Datasets were decontaminated with the decontam package version 

1.20.0 at the 0.1 threshold utilizing the specific negative and positive controls from each 

extraction group72. After decontamination, there were 422,719 total DNA reads with a range of 

456-5788 reads per sample in the full dataset. Rarefaction curves were created using the 

rarecurve() command in the vegan package version 2.6-473. Since seed microbiomes typically 

have low bacterial diversity containing tens to hundreds of taxa, and vertical transmission of 

specific ASVs was a primary area of investigation in this study, the full dataset was preserved to 

ensure full observation ASVs49. 

Analysis 

Alpha diversity species richness was assessed using estimate_richness() in phyloseq, and 

figures were created using the plot_richness() command from phyloseq with the ggplot2 

package version 3.4.274. Faith’s Phylogenetic Diversity was calculated with calculatePD() from 

the biomeUtils package version 0.02275. ANOVAs were performed with the base R stats 

87 

 
 
 
 
command aov(). Weighted UniFrac distances were calculated with distance() in phyloseq and 

used for all analyses of beta diversity,  and PERMANOVA statistical tests were performed with 

adonis2() from the vegan package. Post-hoc analysis on the PERMANOVA results was 

performed with pairwise.adonis2() from pairwiseAdonis version 0.4.176. Beta diversity 

ordinations were created with ordinate() from phyloseq with ggplot2. Beta dispersion was 

assessed with the betadisper() and permutest() commands from the vegan package. Datasets 

were divided into generation or treatment groups where applicable using the ps_filter() 

command in the microViz package version 0.10.1077. Figure 4A was created with the UpSetR 

package version 1.4.0, and statistical analyses were performed with leveneTest() from the car 

package and kruskal_test() and dunn_test() from the rstatix package78–80. Additional data 

analysis was performed in the tidyverse package version 2.0.0 and dplyr package version 1.1.2 

(Wickham et al. 2019, Wickham et al. 2023). ASV transmission rates were analyzed as count 

data and Pearson's Chi-squared Tests were performed with chisq.test() from the base R stats 

package. Seed microbiota were compared to core ASVs identified in Simonin et al. 202250 and 

Venn Diagrams were produced with the VennDiagram package version 1.7.383. To compare the 

taxa in this study to the 48 core bean rhizosphere microbiome taxa identified by Stopnisek and 

Shade 202184, the fasta sequences for each core OTU were used as a query set in a two 

sequence nucleotide BLAST on the National Center for Biotechnology Information (NCBI) 

database website, and the fasta sequences from the seed ASVs were compared to the 48 core 

taxa at >96% identity85.  

88 

 
 
 
 
Data Availability 

Data analysis code can be found at 

(https://github.com/ShadeLab/Seed_transmission_Common_Bean). Raw sequences can be 

found on the NCBI Sequence Read Archive under BioProject number PRJNA1058980. 

RESULTS 

Bean seed microbiomes are stable under environmental treatment in agricultural soil. 

The 16S V4 amplicon sequencing of the seed bacterial endophytes was assessed for 

each of the three generations based on the Weighted Unifrac distance matrix. Since seed 

microbiomes typically have low bacterial diversity containing tens to hundreds of taxa, and 

vertical transmission of specific ASVs was a primary area of investigation in this study, the full 

dataset was preserved49. After host removal and decontamination relative to matched DNA 

extraction controls, there was a range of 456 to 5,788 bacterial DNA reads in the seed samples, 

which was evidenced to be sufficient to exhaustively observe the low-richness seed 

microbiomes (Figure S2.2) given the method and its known biases (e.g. DNA extraction kit and 

sample contamination)49,61. Observed ASV richness ranged from approximately 20 to 80 ASVs 

across all three generations, which is in line with the expected alpha diversity in the samples, 

and there is no significant influence of treatment on the species richness in either G1 or G2 

(ANOVA, G1_treatment: F= 0.150, p-value=0.861. G1_G2: F=0.393, p-value=0.923, Figure 2.1A) 

(ANVOA, G2 Richness: F=1.122, p-value=0.334, Figure 2.1B), or Faith’s Phylogenetic Diversity in 

G2 (ANOVA, PD: F=1.111, p-value=0.346, Figure 2.1C).  

89 

 
 
 
 
Figure 2.1. Alpha Diversity. A. Number of ASVs observed in each of the seed microbiome 
samples across all generations. Five seeds were used for each sample. G0 n=20 samples, G1 and 
G2 n=12 samples per treatment group. There is no influence of treatment groups on the species 
richness observed in either G1 or G2 (ANOVA, G1_treatment: F= 0.150, p-value=0.861. G1_G2: 
F=0.393, p-value=0.923). B. Number of ASVs observed, and C. Faith’s Phylogenetic Diversity of 
G2 seed samples by parental line. Gray bars indicate treatment applied to the parent plant in 
G1. There are no significant differences between parent lines in either Richness or PD measure 
(ANOVA, Richness: F=1.122, p-value=0.334; PD: F=1.111, p-value=0.346). 

When beta diversity of the Weighted Unifrac distance is examined in the full dataset, 

there is no significant influence of the plant generation on the bacterial communities in the 

seed samples (PERMANOVA, DF=2, R2=0.0160, F-value=1.3173, p-value=0.1967, data not 

shown). Additionally, when the G1 seed microbiomes are analyzed, there is no significant 

90 

 
 
 
influence of the plant treatment on the bacterial communities (PERMANOVA, DF=2, 

R2=0.06375, F-value=1.1236, p-value=0.2948, data not shown). In generation 2, there was no 

influence of the legacy of the G1 environmental treatment, or the treatment applied in G2, on 

the resulting seed microbiome of G2 (PERMANOVA, G1 treatment: DF=2, R2=0.0194, F-

value=1.1149, p-value=0.3038; G2 treatment: DF=2, R2=0.01504, F-value=0.8646, p-

value=0.6496) (Figure 2.2A). These results suggest that with high replication within treatments 

to account for variation between plants, and growing plants in a representative agricultural soil, 

the seed microbiome communities are not directly affected by the abiotic treatments applied in 

this study. 

91 

 
 
Figure 2.2. Beta diversity in Generation 2 seed samples. A. Weighted Unifrac distance of G2 
seed samples. Points represent three offspring from each parent line, each of which received a 
different treatment in G2. Parent plant line is the only significant explanatory variable in the G2 
samples (PERMANOVA, r2=0.356, F= 1.2421, p-value=0.0065**). B. Beta dispersion around the 
spatial median of Weighted Unifrac distances in G2 seed samples. Lines are grouped by G1 
parent treatment, represented by black boxplots. G1 treatment and parent line are not 
significant. (ANOVA, G1 Treatment: DF=2, F-value= 1.2835, p-value= 0.2835. Line: DF: 35, F-
value=1.0825, p-value=0.3714). Sample G2_9, the line C13 Nutrient offspring, was removed 
from the figures as an outlier. However, statistics were performed with this sample included.  

92 

 
 
 
 
Parental plant line drives variation in the Generation 2 seed communities. 

While there was no influence of the abiotic treatments, the parental plant line that 

beget the G2 plants was highly significant in explaining the variation in the G2 beta diversity 

(PERMANOVA: DF= 33, R2= 0.3568, F-value= 1.12421, p-value=0.0065**) (Figure 2.2A). 

Additionally, beta dispersion of the Weighted Unifrac distances in the G2 seed communities 

was consistent between abiotic treatments and parental lines (ANOVA, G1 Treatment: DF=2, F-

value= 1.2835, p-value= 0.2835. Line: DF: 35, F-value:1.0825, p-value=0.3714) (Figure 2.2B). 

When the G2 samples are grouped by the treatment the G1 parent plants received, the 

parental lines are not significantly different in control and nutrient lines, while parent line is 

significant in explaining differences in the drought lines (PERMANOVA, Control: DF=11, 

R2=0.3532, F-value=1.2189, p-value=0.0526; Drought: DF=11, R2= 0.4032, F-value= 1.2353, p-

value=0.0021**; Nutrient: DF=11, R2= 0.3474, F-value= 1.1142, p-value=0.2687) (Figure 2.3A, B, 

C). These results suggest that the parental line is the most important factor determining the 

seed microbiome membership in the G2 offspring, and that the differences in community 

composition between parent lines are increased when the parent plants are exposed to abiotic 

stress conditions. 

93 

 
 
 
Figure 2.3. PCoA plots of Weighted Unifrac beta diversity in Generation 2, subset by parent 
plant treatment. When plant lines are grouped by G1 parent treatment, parent plant line is not 
significant in the control and nutrient lines, but is significant in the drought lines (PERMANOVA, 
Control: r2= 0.35319, F=1.2189, p=0. 0526, Nutrient: r2= 0.34740, F=1.1142, p= 0.2687, 
Drought: r2= 0.40311 F=1.4637 p=0.0021**). Bars above the ordinations indicated the 
treatment applied to the parent plants in G1: A. Control, B. Drought, C. Nutrient. The three 
offspring from each parent are connected by triangles to aid in visualization. Sample G2_9, the 
line C13 Nutrient offspring, was removed from figure A as an outlier. However, statistics were 
performed with this sample included. 

Overlapping seed microbiome taxa are present in each parent line and all three generations.  

Out of 658 total ASVs detected, there were 128 ASVs shared across all three generations 

(Figure 2.4A). These ASVs that were detected in all G0, G1 and G2 seeds were in higher relative 

abundance than the ASVs that were only found in one or two generations (Figure 2.4B) 

94 

 
 
 
(Kruskal-Wallis test: test-statistic=363.59, df=2, p-value<0.0001; Post-hoc Dunn’s test with 

Benjamini-Hochberg correction: 1 vs 3 generations: test-statistic=17.78, adjusted p-

value<0.0001, 2 vs 3 generations: test-statistic=5.879, adjusted p-value<0.0001). Furthermore, 

the Genus-level taxonomic profiles of G0, G1, and G2 microbiomes were highly comparable 

(Figure S2.3), suggesting a consistent taxonomic signature of the ASVs detected across seed 

generations. 

Since vertical transmission is a primary focus of this study, ASVs overlapping between a 

G1 parent plant and at least one of its offspring in G2, referred to as “overlapping taxa” within 

lines, were investigated. Overlapping taxa between G1 and G2 can be identified in all 36 

parental lines, along with taxa that are unique to one generation (Figure 2.4C). There were 99 

taxa identified as overlapping within specific plant lines, 43 of which were common to all G1 

treatments (Figure S2.4). The proportion of the ASVs in each line that are overlapping between 

G1 and G2 range from 17% to 45%, with an average of 29% of ASVs overlapping between 

parent and offspring in each line (Figure 2.4D). 

95 

 
 
 
A

B

Figure 2.4. Unique and overlapping taxa between generations. A. Number of ASVs and relative 
abundance of ASVs per generation intersect across all samples. G0 n=20 samples (five seeds per 
sample), G1 n =36, G2 n=108. B. Log relative abundance of ASVs based on how many 
generations in which they are found. Out of 658 total ASVs, taxa found in all three generations 
are significantly more abundant in the dataset than ASVs found in only one or two generations. 
Asterisks indicate p-value<0.0001, black squares indicate the mean value. (Kruskal-Wallis test: 
test-statistic=363.59, df=2, p-value<0.0001; Post-hoc Dunn’s test with Benjamini-Hochberg 
correction, 1 vs 3 generations: test-statistic=17.78, adjusted p-value<0.0001; 2 vs 3 generations: 
test-statistic=5.879, adjusted p-value<0.0001). C. Total number of ASVs per parent line and 
number of taxa found in G1, G2, or overlapping. “G1_G2 Overlap” is defined as ASVs present in 
both the G1 sample and at least one G2 offspring within a parent line. 99 ASVs were identified 
as overlapping within parent lines, and there were overlapping ASVs identified in all 36 lines. D. 
proportion of the total ASVs per line found in G1, G2, or overlapping. Boxplots represent the 
median values and first and third quartiles, and whiskers represent the 95% confidence interval. 

96 

 
 
 
 
Of the 99 overlapping taxa identified within parental lines, 29 ASVs were found in only 

one parental line. The 70 ASVs that were found in at least two lines were investigated to 

understand their prevalence (occupancy in the lines) in the dataset and their rate of 

transmission in the three offspring in G2 (Figure 2.5, Table S2.2). There were nine ASVs present 

in all 36 parental lines, and an additional 13 ASVs were found in at least half of parental lines. 

These 22 prevalent taxa generally had 100% transmission to all three G2 offspring within lines, 

while less prevalent taxa were not transmitted as consistently in the G2 offspring (Figure 2.5). 

The transmission of the ASVs to G2 offspring is not significantly impacted by the treatment the 

G1 parent received, and there is no significant difference in transmission between lines 

(Pearson's Chi-squared Test. G1 Treatment: X2 = 0.67413, df = 4, p-value = 0.9545. Line: X2 = 

63.48, df = 70, p-value = 0.6958) (Figure 2.5). The 70 overlapping taxa found in at least two 

parental lines are also highly detected in the G0 dataset, with 61 of these ASVs found in the G0 

samples, and they are very taxonomically diverse, identifying with 32 different bacterial Genera 

(16 ASVs were unresolved at the Genus level or labelled “Uncultured”), 39 bacterial Families, 

and 26 Orders (Figure 2.5, Table S2.2). The most prevalent overlapping ASVs are also highly 

abundant in the combined G1-G2 dataset (Figure 2.5) and the community profiles of these 70 

ASVs at the Genus level are very similar across parent lines in G2 (Figure S2.5).      

97 

 
 
G1 Control Lines

G1 Drought Lines

G1 Nutrient Lines

Figure 2.5. Prevalence in lines, transmission, core taxa identity and relative abundance of 
overlapping ASVs in G2 offspring. Of the 99 taxa that were found overlapping between G1 and 
G2 within parent lines, 29 ASVs that were only in one parent line were removed, and the 
remaining 70 ASVs are listed above, ordered from presence in the highest number of lines to 
lowest number of lines. Blue dots represent the number of G2 offspring containing the ASV in 
each line. 61 of these taxa are also found in the G0 dataset indicated by black dots. The 
taxonomy of each ASV identified at the Order level is indicated by colored dots in the Order 
column (26 Orders). Blue and brown squares in the Core_Taxa column indicate identity with 
seed or rhizosphere core taxa, respectively. Red boxes in the right-most column represent the 
log relative abundance of the ASV overall across G1 and G2. There is no significant difference in  

98 

 
 
 
Figure 2.5 (cont’d)  
ASV transmission in G2 offspring between G1 treatments or parental lines (Pearson's Chi-
squared Test. G1 Treatment: X2 = 0.67413, df = 4, p-value = 0.9545. Line: X2 = 63.48, df = 70, p-
value = 0.6958). 

While parental plant line was a strong driver of differences in the beta diversity of the 

entire G2 dataset, it is more pronounced in the drought-treated parental lines than the control 

and nutrient lines. To investigate a potential mechanism of this variation, the ASVs were 

queried for taxa that are only found in one parental line from each treatment. Of these taxa, 

there were 22 taxa unique to only one G1-G2 overlap in control lines, 26 unique taxa in nutrient 

lines, and 38 unique taxa in drought lines, suggesting that there are more unique taxa and more 

distinct communities transmitted from parent plants that experienced the drought stress 

treatment (Figure S2.6). Of these unique taxa within lines, there were multiple ASVs that were 

only found in lines from one G1 parent treatment (Figure S2.6). These unique taxa in each 

treatment condition were diverse at the Genus and Family level, but the taxa found only in the 

control lines were mostly from the phylum Firmicutes, while the nutrient and drought 

treatments had more unique Actinobacteriota and Proteobacteria (See Table S2.3 for a list of 

the taxonomy for all unique taxa within treatments). Since there were no significant differences 

in beta dispersion between parent treatments or lines, these results suggest that differences in 

beta-diversity between parental lines under the drought treatment are primarily due to 

differences in community composition verses variability between offspring within lines. 

These results indicate that the prevalent seed microbiome members found in most 

parental lines in this study were consistently packaged in the seed microbiome of the G0, G1 

and G2 plants, regardless of the abiotic treatment applied to the plant. While there was a stable 

signature of the most common and abundant overlapping seed microbiome taxa, there were 

99 

 
 
also many taxa that were only found in one or a few parental lines, and there were more 

unique taxa identified between the G1 drought stress lines than in the other treatment groups. 

Table 2.1  
Stable ASVs that were detected in all three generations. Nine ASVs are found in all 36 
parental lines, two of which are core seed microbiome taxa, labeled “S”. 13 additional 
ASVs were present in 50% or more of the parental lines. An additional core seed 
microbiome member was found in three parental lines. Five of these ASVs align to bean 
rhizosphere core OTUs at >96% identity, labeled “R”. All ASVs listed are also found in 
Generation 0 seeds.  

ASV 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 

16 

17 

18 

19 

20 

21 

22 

50 

Presence  Core 
Taxa 
- 

All lines 

All lines 

All lines 

All lines 

All lines 

All lines 

All lines 

All lines 

All lines 

35 lines 

35 lines 

35 lines 

35 lines 

34 lines 

34 lines 

32 lines 

28 lines 

28 lines 

27 lines 

24 lines 

23 lines 

18 lines 

3 lines 

- 

S 

- 

S 

- 

- 

R 

- 

R 

R 

R 

- 

- 

- 

- 

- 

- 

- 

R 

- 

- 

S 

Class 

Order 

Family 

Genus 

Bacteroidia 

Bacteroidales 

Porphyromonadaceae 

Porphyromonas 

Gammaproteobacteria  Enterobacterales 

Enterobacteriaceae 

Gammaproteobacteria  Pseudomonadales 

Pseudomonadaceae 

Pseudomonas 

Bacilli 

Bacilli 

Staphylococcales 

Staphylococcaceae 

Staphylococcus 

Bacillales 

Bacillaceae 

Bacillus 

Gammaproteobacteria  Xanthomonadales 

Xanthomonadaceae 

Stenotrophomonas 

Gammaproteobacteria  Enterobacterales 

Enterobacteriaceae 

Escherichia-Shigella 

Gammaproteobacteria  Burkholderiales 

Oxalobacteraceae 

Massilia 

Gammaproteobacteria  Pseudomonadales 

Pseudomonadaceae 

Pseudomonas 

Actinobacteria 

Micrococcales 

Micrococcaceae 

Arthrobacter 

Alphaproteobacteria 

Rhizobiales 

Rhizobiaceae 

Ochrobactrum 

Actinobacteria 

Streptomycetales 

Streptomycetaceae 

Streptomyces 

Bacilli 

Bacillales 

Bacillaceae 

Bacillus 

Bacteroidia 

Bacteroidales 

Bacteroidaceae 

Bacteroides 

Bacilli 

Bacillales 

Bacillaceae 

Bacillus 

Gammaproteobacteria  Xanthomonadales 

Xanthomonadaceae 

Xanthomonas 

Gammaproteobacteria  Enterobacterales 

Enterobacteriaceae 

Actinobacteria 

Streptomycetales 

Streptomycetaceae 

Streptomyces 

Alphaproteobacteria 

Rhizobiales 

Rhizobiaceae 

Ochrobactrum 

Gammaproteobacteria  Burkholderiales 

Comamonadaceae 

Gammaproteobacteria  Enterobacterales 

Enterobacteriaceae 

Escherichia-Shigella 

Gammaproteobacteria  Pseudomonadales 

Pseudomonadaceae 

Pseudomonas 

Gammaproteobacteria  Enterobacterales 

Erwiniaceae 

Pantoea 

100 

 
 
 
 
 
 
 
 
 
Previously identified core bean microbiome taxa are present in all three generations. 

The 70 overlapping ASVs identified in this study were compared to the six core common 

bean seed microbiome ASVs identified by Simonin et al. 202250, and to the 48 core common 

bean rhizosphere OTUs identified by Stopnisek and Shade 202184 (Figure 2.5). The 22 most 

prevalent ASVs identified in the dataset are listed in Table 2.1 with their prevalence across 

lines, identity to core bean microbiome taxa, and most resolved taxonomic identification (Table 

2.1). 

Three core common bean seed ASVs were found in the overlapping G1-G2 dataset, 

specifically from the Pseudomonas, Bacillus, and Pantoea genera (Table 2.1, Figure S2.4)50. Two 

of these ASVs, the Pseudomonas and Bacillus core members, were identified in all parental lines 

the dataset, while the Pantoea core seed microbiome member was found in only three parental 

lines (Table 2.1).  

To compare to the core rhizosphere OTUs, the full seed dataset was BLASTed against 

the core rhizosphere fasta sequences at >96% identity (the OTU identity threshold used for the 

core study), and there were 579 total hits to the core OTUs, primarily in the phyla 

Actinobacteria and Proteobacteria. Further, when the overlapping ASVs from our study were 

compared to the BLAST hits, there were nine taxa aligned with core rhizosphere OTUs at >96% 

identity (Figure 2.5, Table 2.1, Table S2.2)84. These taxa were classified in the families 

Oxalobacteraceae (genus Massilia), Comamonadaceae, Rhizobiaceae (genus Ochrobactrum), 

Streptomycetaceae (genus Streptomyces), Sphingomonadaceae (genus Sphingomonas), 

Methyloligellaceae, Devosiaceae (genus Devosia), and Micrococcaceae in the genus 

Arthrobacter, of which our seed ASV aligned at 98.8% identity to the most abundant core OTU 

101 

 
 
 
in the rhizosphere study (Table 2.1)84. Additionally, three unique Proteobacteria found in 

drought lines, including two in the order Rhizobiales, aligned at 97% identity to core 

rhizosphere OTUs from Stopnisek and Shade 202184 (Table S2.3). Along with the core seed 

microbiome members identified in the dataset, this provides evidence that the bean core 

microbiome is vertically transmitted from parent to offspring via the seed.   

DISCUSSION 

Our multigenerational study aimed to understand the impact of abiotic treatment on 

the seed microbiome in common bean over multiple generations, and the potential vertical 

transmission of seed microbiome members under these conditions. Since the seed microbiome 

can be an important factor in the establishment of the plant microbiome, it is important to 

understand which seed microbiome members are being passed from parent to offspring, and 

what advantage or disadvantage these seed members may bring to the next plant generation.  

Motivated by the work of Bintarti et al., we hypothesized that the seed microbiomes of 

the common bean plants in our study would be significantly altered by the drought or nutrient 

treatment applied44. However, there was no statistically supported impact of the plant 

treatments on the seed microbiome communities in either plant generation. These results 

differ from what was observed by Bintarti et al. However, with the known intrinsic variability of 

the seed microbiome, our experiment was designed to have high replication within the 

treatments for increased statistical power49. An additional improvement to the previous 

experiment was the use of representative agricultural soil for plant growth with naturally 

occurring soil microbiota, as opposed to a greenhouse soil. It is important for greenhouse and 

growth chamber studies to be as realistic as possible to true field systems, and our study design 

102 

 
 
 
 
allowed for this with a highly diverse soil community86. This study also incorporated improved 

sequencing depth of the bacterial community in the seed. While our first hypothesis regarding 

the impact of environmental treatment was not supported, this study sheds light on the 

stability of the microbiome that is packaged in the seeds of common bean under varying 

environmental conditions, and a potential mechanism for core microbiome members to 

consistently associate with plant species despite varying climates and growing regions84,87.  

A significant finding of this study was the strong evidence of vertical transmission from 

specific parental plant lines. There were many ASVs present in the seeds of all three plant 

generations that were highly abundant across the dataset, including the 22 prevalent taxa 

identified in the parental lines. These taxa were consistently identified in parents and offspring 

through the generations, despite the plants being grown in the field in G0, receiving different 

abiotic treatments, growing in soils that were collected from the field at different times in G1 

and G2, and the three G2 offspring receiving different abiotic treatments within the same 

parent line. It has been shown previously that common bean seed microbiomes are closely 

associated with the vascular system of the plant verses floral compartments, so we can be 

confident that these taxa were not deposited from the growth chamber system through flowers 

and were transported to the seed via the parent plant vascular tissues21,29. Additionally, the 

specificity of ASVs assigned at 100% sequence identity, and the thorough use of sequencing 

controls, allows for confidence that the taxa identified are the same across the different 

generations and samples88. Since the specific parent plant was the most significant driver of 

differences in the G2 seed microbiome communities, responsible for 35% of the variation in the 

PERMANOVA analysis, these results provide important insight to the field that parental plant 

103 

 
 
 
line should be considered in seed microbiome experiments, and replication within treatments 

and parent lines is important when designing microbiome studies. This is in line with previous 

studies that also reported differences in seed microbial communities based on parental plant 

line in Setaria and oak19,43. Our second hypothesis that characteristic taxa would be transmitted 

from parent to offspring under specific abiotic treatments was not fully supported. However, 

there is clear evidence that a stable seed microbiome was transmitted in all treatment 

conditions in our study.  

In addition to parent plant line driving variation in the second generation, there were 

also many taxa found in this study that have previously been described as “core” microbiome 

taxa for common bean. Within the most prevalent taxa in our study, there were two core bean 

seed ASVs and five core bean rhizosphere OTUs identified across all three generations50,84. 

These core taxa that have been identified across multiple studies are likely important for plant 

health in common bean, and are thus consistently transmitted to each new plant generation. 

The transmitted taxa in our study were highly prevalent across parent lines and were highly 

abundant in the dataset, comprising 17% to 45% of the ASVs in each parent line. At such high 

prevalence across the dataset despite parental line treatment, these taxa, particularly the 22 

most prevalent ASVs, appear to be very stably transmitted from the parent plants to offspring. 

These prevalent taxa include many genera that are commonly associated with plant 

microbiomes, such as Bacillus, Pseudomonas, Arthrobacter, Massilia, Stenotrophomonas, 

Ochrobactrum, and Xanthomonas, which are often associated with beneficial traits such as 

plant growth promotion, biocontrol of pathogens, and stress resistance34,89–94. These stable 

seed microbiome taxa likely establish mutualistic or commensal associations in the microbiome 

104 

 
 
 
of the germinating seedling, and persist through the plant lifecycle until they are packaged 

within the seed for the next generation. However, an alternative mechanism that cannot be 

ruled out by this study, since we did not track specific bacterial cells from the seeds, could be 

that the parent plants are selectively recruiting these same taxa from the soil in each 

generation, and they then are packaged inside the seed through the vascular system95.  

The prevalent taxa in this study should be investigated further for potential use in plant 

microbiome engineering, as they may provide important benefits for plant health96,97. It has 

been demonstrated that applying beneficial plant microbiome members through seed 

treatments or soil inoculation can confer improved growth and health benefits to plants98,99 As 

the need for sustainable solutions to increase agricultural productivity grows, microbiome 

engineering and breeding plants for improved microbiomes will continue to be an important 

area of research100,101. The stable bean seed microbiome members identified here, as well as 

core beneficial seed microbes identified in other plant species, are an excellent starting point 

for future research into plant microbiome engineering50,102,103. 

While the stable prevalent taxa in this study could be identified in parent lines from all 

treatment groups, there were also unique taxa identified between parent lines, and the 

greatest number of unique taxa were identified in the drought parental lines. These taxa were 

only found in one parental line and had low relative abundance in the dataset overall. This 

increased prevalence of rare taxa, creating more distinct communities between parent lines in 

the drought treatment, may be driving the increased influence of the parent on the seed 

microbiome communities under drought stress. It has been documented that environmental 

disturbance can alter microbial communities, and drought can cause significant perturbation of 

105 

 
 
 
plant microbiomes10,104,105. Additionally, more variable and diverse microbial communities are 

also known to be more resilient to environmental disturbances106,107. The parental plants that 

experienced a stress treatment accumulated more variable communities in their seeds, had an 

overall increase of rare taxa in the parent lines, and included additional taxa that were related 

to core bean microbiome members50,84. This may indicate that stressed plants contain enriched 

beneficial core taxa in their seeds that can persist in the offspring plant, and may be 

accumulating increased diversity that could be beneficial over longer time scales. While we did 

not observe a direct influence of plant treatment on the seed microbiome within generations, it 

is possible that over many more generations of stress, the increased variability and core taxa 

members may lead to significant shifts in microbiome communities. 

In conclusion, this study provides evidence of vertical transmission of stable microbiome 

members in the seeds of common bean. We found that parental plant line was highly 

explanatory in driving the seed microbiome of the second plant generation, and this effect was 

heightened when the parent plant experienced a stress treatment. There were many prevalent 

seed microbiome members stably transmitted through three plant generations despite abiotic 

treatment, and these prevalent taxa included known common bean seed and rhizosphere core 

taxa. The prevalent taxa identified here could be an area of interest for microbiome engineering 

for the sustainable support of agricultural crops, and should be investigated further. Parental 

plant has been demonstrated to be an important factor influencing the seed microbiome and 

should be included as an experimental variable when investigating the plant microbiome. 

Overall, this study contributes important knowledge to the field of seed microbiome research 

106 

 
 
 
to inform future studies, and contributes to a significant knowledge gap in the transmission of 

endophytic seed communities under stress conditions. 

107 

 
 
 
 
 
 
REFERENCES 

1. 

2. 

3. 

4. 

5. 

6. 

7. 

8. 

9. 

Compant, S., Samad, A., Faist, H. & Sessitsch, A. A review on the plant microbiome: 
Ecology, functions, and emerging trends in microbial application. J Adv Res 19, (2019). 

Santos, L. F. & Olivares, F. L. Plant microbiome structure and benefits for sustainable 
agriculture. Curr Plant Biol 26, 100198 (2021). 

Finkel, O. M., Castrillo, G., Herrera Paredes, S., Salas González, I. & Dangl, J. L. 
Understanding and exploiting plant beneficial microbes. Curr Opin Plant Biol 38, 155–
163 (2017). 

Trivedi, P., Leach, J. E., Tringe, S. G., Sa, T. & Singh, B. K. Plant–microbiome interactions: 
from community assembly to plant health. Nat Rev Microbiol 18, 607–621 (2020). 

Acuña, J. J. et al. Diversity, Community Structure, and Potential Functions of Root-
Associated Bacterial Communities of Different Wheat (Triticum aestivum) Cultivars 
under Field Conditions. Agronomy 13, (2023). 

Cordovez, V. et al. Successive plant growth amplifies genotype-specific assembly of the 
tomato rhizosphere microbiome. Science of the Total Environment 772, (2021). 

Faist, H. et al. Potato root-associated microbiomes adapt to combined water and 
nutrient limitation and have a plant genotype-specific role for plant stress mitigation. 
Environ Microbiome 18, (2023). 

Howe, A. et al. Seasonal activities of the phyllosphere microbiome of perennial crops. 
Nat Commun 14, (2023). 

Liu, F. et al. Soil indigenous microbiome and plant genotypes cooperatively modify 
soybean rhizosphere microbiome assembly. BMC Microbiol 19, (2019). 

10. 

Santos-Medellín, C., Edwards, J., Liechty, Z., Nguyen, B. & Sundaresan, V. Drought stress 
results in a compartment-specific restructuring of the rice root-associated microbiomes. 
mBio 8, (2017). 

11. 

Tiziani, R. et al. Drought, heat, and their combination impact the root exudation 
patterns and rhizosphere microbiome in maize roots. Environ Exp Bot 203, (2022). 

12.  Wagner, M. R. et al. Host genotype and age shape the leaf and root microbiomes of a 

wild perennial plant. Nat Commun 7, (2016). 

13. 

Barret, M., Guimbaud, J. F., Darrasse, A. & Jacques, M. A. Plant microbiota affects seed 
transmission of phytopathogenic microorganisms. Mol Plant Pathol 17, 791–795 (2016). 

108 

 
 
14. 

15. 

16. 

17. 

18. 

19. 

20. 

Kim, H., Jeon, J., Lee, K. K. & Lee, Y. H. Longitudinal transmission of bacterial and fungal 
communities from seed to seed in rice. Commun Biol 5, 772 (2022). 

Liu, Y. et al. Investigating the endophytic bacterial diversity and community structures in 
seeds of genetically related maize (Zea mays L.) genotypes. 3 Biotech 10, 1–10 (2020). 

Abdelfattah, A., Tack, A. J. M., Lobato, C., Wassermann, B. & Berg, G. From seed to seed: 
the role of microbial inheritance in the assembly of the plant microbiome. Trends 
Microbiol 31, 346–355 (2023). 

Bergna, A. et al. Tomato seeds preferably transmit plant beneficial endophytes. 
Phytobiomes J 2, 183–193 (2018). 

Cope-Selby, N. et al. Endophytic bacteria in Miscanthus seed: implications for 
germination, vertical inheritance of endophytes, plant evolution and breeding. GCB 
Bioenergy 9, 57–77 (2017). 

Escobar Rodríguez, C., Antonielli, L., Mitter, B., Trognitz, F. & Sessitsch, A. Heritability 
and functional importance of the setaria viridis bacterial seed microbiome. Phytobiomes 
J 4, 40–52 (2020). 

Latz, M. A. C. et al. Succession of the fungal endophytic microbiome of wheat is 
dependent on tissue-specific interactions between host genotype and environment. 
Science of the Total Environment 759, (2021). 

21.  Malinich, E. A. & Bauer, C. E. The plant growth promoting bacterium Azospirillum 

brasilense is vertically transmitted in Phaseolus vulgaris (common bean). Symbiosis 76, 
97–108 (2018). 

22. 

23. 

24. 

25. 

26. 

Vannier, N. et al. A microorganisms’ journey between plant generations. Microbiome 6, 
79 (2018). 

Abdelfattah, A., Wisniewski, M., Schena, L. & Tack, A. J. M. Experimental evidence of 
microbial inheritance in plants and transmission routes from seed to phyllosphere and 
root. Environ Microbiol (2021) doi:10.1111/1462-2920.15392. 

Barret, M. et al. Emergence shapes the structure of the seed microbiota. Appl Environ 
Microbiol 81, 1257–1266 (2015). 

Klaedtke, S. et al. Terroir is a key driver of seed-associated microbial assemblages. 
Environ Microbiol 18, (2016). 

Rezki, S. et al. Assembly of seed-associated microbial communities within and across 
successive plant generations. Plant Soil 422, 67–79 (2018). 

109 

 
 
27. 

28. 

29. 

30. 

31. 

Escobar Rodríguez, C., Mitter, B., Barret, M., Sessitsch, A. & Compant, S. Commentary: 
seed bacterial inhabitants and their routes of colonization. Plant Soil 422, 129–134 
(2018). 

Shade, A., Jacques, M. A. & Barret, M. Ecological patterns of seed microbiome diversity, 
transmission, and assembly. Curr Opin Microbiol 37, 15–22 (2017). 

Chesneau, G. et al. Temporal dynamics of bacterial communities during seed 
development and maturation. FEMS Microbiol Ecol 96, (2020). 

Galambos, N. et al. Beneficial Insects Deliver Plant Growth-Promoting Bacterial 
Endophytes between Tomato Plants. Microorganisms 9, (2021). 

Kong, H. G., Song, G. C. & Ryu, C. M. Inheritance of seed and rhizosphere microbial 
communities through plant–soil feedback and soil memory. Environ Microbiol Rep 11, 
479–486 (2019). 

32.  Mitter, B. et al. A new approach to modify plant microbiomes and traits by introducing 
beneficial bacteria at flowering into progeny seeds. Front Microbiol 8, (2017). 

33. 

34. 

35. 

36. 

37. 

38. 

39. 

Fukami, T. Historical Contingency in Community Assembly: Integrating Niches, Species 
Pools, and Priority Effects. Annu Rev Ecol Evol Syst 46, 1–23 (2015). 

Johnston-Monje, D., Gutiérrez, J. P. & Lopez-Lavalle, L. A. B. Seed-Transmitted Bacteria 
and Fungi Dominate Juvenile Plant Microbiomes. Front Microbiol 12, (2021). 

Torres-Cortés, G. et al. Functional microbial features driving community assembly during 
seed germination and emergence. Front Plant Sci 9, (2018). 

Berg, G. & Raaijmakers, J. M. Saving seed microbiomes. ISME Journal 12, 1167–1170 
(2018). 

Bziuk, N. et al. The treasure inside barley seeds: microbial diversity and plant beneficial 
bacteria. Environ Microbiome 16, 1–21 (2021). 

Shahzad, R., Khan, A. L., Bilal, S., Asaf, S. & Lee, I. J. What is there in seeds? Vertically 
transmitted endophytic resources for sustainable improvement in plant growth. Front 
Plant Sci 9, (2018). 

Truyens, S., Weyens, N., Cuypers, A. & Vangronsveld, J. Bacterial seed endophytes: 
Genera, vertical transmission and interaction with plants. Environ Microbiol Rep 7, 40–
50 (2015). 

40. 

Eyre, A. W., Wang, M., Oh, Y. & Dean, R. A. Identification and characterization of the 
core rice seed microbiome. Phytobiomes J 3, 148–157 (2019). 

110 

 
 
41. 

42. 

43. 

44. 

45. 

Liu, X. et al. Limited Impact of Soil Microorganisms on the Endophytic Bacteria of Tartary 
Buckwheat (Fagopyrum tataricum). Microorganisms 11, (2023). 

Rochefort, A. et al. Influence of environment and host plant genotype on the structure 
and diversity of the brassica napus seed microbiota. Phytobiomes J 3, 326–336 (2019). 

Fort, T. et al. Maternal effects shape the seed mycobiome in Quercus petraea. New 
Phytologist 230, 1594–1608 (2021). 

Bintarti, A. F., Kearns, P. J., Sulesky-Grieb, A. & Shade, A. Abiotic Treatment to Common 
Bean Plants Results in an Altered Endophytic Seed Microbiome. Microbiol Spectr 10, 
(2022). 

Vujanovic, V., Islam, M. N. & Daida, P. Transgenerational role of seed mycobiome – an 
endosymbiotic fungal composition as a prerequisite to stress resilience and adaptive 
phenotypes in Triticum. Sci Rep 9, 18483 (2019). 

46.  Walitang, D. I. et al. The influence of host genotype and salt stress on the seed 

endophytic community of salt-sensitive and salt-tolerant rice cultivars. BMC Plant Biol 
18, (2018). 

47.  Wassermann, B. et al. Low nitrogen fertilization enriches nitrogen-fixing bacteria in the 
Brassica seed microbiome of subsequent generations. Journal of Sustainable Agriculture 
and Environment 2, 87–98 (2023). 

48. 

49. 

50. 

51. 

52. 

Dastogeer, K. M. G., Tumpa, F. H., Sultana, A., Akter, M. A. & Chakraborty, A. Plant 
microbiome–an account of the factors that shape community composition and diversity. 
Curr Plant Biol 23, 100161 (2020). 

Bintarti, A. F., Sulesky-Grieb, A., Stopnisek, N. & Shade, A. Endophytic microbiome 
variation among single plant seeds. Phytobiomes J 1–57 (2021). 

Simonin, M. et al. Seed microbiota revealed by a large-scale meta-analysis including 50 
plant species. New Phytologist 234, 1448–1463 (2022). 

Castro-Guerrero, N., Isidra-Arellano, M., Mendoza-Cozatl, D. & Valdés-López, O. 
Common Bean: A Legume Model onthe Rise for Unraveling Responsesand Adaptations 
to Iron, Zinc, andPhosphate Deficiencies. Front Plant Sci 7, 391–417 (2016). 

Siddiq, M., Uebersax, M. A. & Siddiq, F. Global Production, Trade, Processing and 
Nutritional Profile of Dry Beans and Other Pulses. in Dry Beans and Pulses 1–28 (Wiley, 
2022). doi:10.1002/9781119776802.ch1. 

53. 

Uebersax, M. A. et al. Dry beans (Phaseolus vulgaris L.) as a vital component of 
sustainable agriculture and food security—A review. Legume Science 5, (2023). 

111 

 
 
54. 

55. 

56. 

57. 

58. 

59. 

60. 

61. 

62. 

Chivangulula, F. M., Amraoui, M. & Pereira, M. G. The Drought Regime in Southern 
Africa: A Systematic Review. Climate 11, (2023). 

Furtak, K. & Wolińska, A. The impact of extreme weather events as a consequence of 
climate change on the soil moisture and on the quality of the soil environment and 
agriculture – A review. Catena (Amst) 231, (2023). 

Gourdji, S., Läderach, P., Valle, A. M., Martinez, C. Z. & Lobell, D. B. Historical climate 
trends, deforestation, and maize and bean yields in Nicaragua. Agric For Meteorol 200, 
270–281 (2015). 

Caballero-Mellado, J. & Martinez-Romero, E. Soil Fertilization Limits the Genetic 
Diversity of Rhizobium in Bean Nodules. Symbiosis vol. 26 (1999). 

Gao, Y. & Cabrera Serrenho, A. Greenhouse gas emissions from nitrogen fertilizers could 
be reduced by up to one-fifth of current levels by 2050 with combined interventions. 
Nat Food 4, 170–178 (2023). 

Kelly, J. D. et al. Registration of ‘Red Hawk’ Dark Red Kidney Bean. Crop Sci 38, 280–281 
(1998). 

Colovas, J., Bintarti, A. F., Mechan Llontop, M. E., Grady, K. L. & Shade, A. Do-it-Yourself 
Mock Community Standard for Multi-Step Assessment of Microbiome Protocols. Curr 
Protoc 2, (2022). 

Eisenhofer, R. et al. Contamination in Low Microbial Biomass Microbiome Studies: Issues 
and Recommendations. Trends Microbiol 27, 105–117 (2019). 

Caporaso, J. G. et al. Global patterns of 16S rRNA diversity at a depth of millions of 
sequences per sample. Proc Natl Acad Sci U S A 108, 4516–4522 (2011). 

63.  Walters, W. et al. Improved Bacterial 16S rRNA Gene (V4 and V4-5) and Fungal Internal 
Transcribed Spacer Marker Gene Primers for Microbial Community Surveys. mSystems 
1, (2016). 

64. 

65. 

66. 

Apprill, A., Mcnally, S., Parsons, R. & Weber, L. Minor revision to V4 region SSU rRNA 
806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquatic 
Microbial Ecology 75, 129–137 (2015). 

Caporaso, J. G. et al. Ultra-high-throughput microbial community analysis on the 
Illumina HiSeq and MiSeq platforms. ISME Journal 6, 1621–1624 (2012). 

Parada, A. E., Needham, D. M. & Fuhrman, J. A. Every base matters: Assessing small 
subunit rRNA primers for marine microbiomes with mock communities, time series and 
global field samples. Environ Microbiol 18, 1403–1414 (2016). 

112 

 
 
67. 

68. 

69. 

70. 

Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data 
science using QIIME 2. Nature Biotechnology vol. 37 852–857 Preprint at 
https://doi.org/10.1038/s41587-019-0209-9 (2019). 

Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon 
data. Nat Methods 13, 581–583 (2016). 

Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data 
processing and web-based tools. Nucleic Acids Res 41, (2013). 

R Core Team. R: A Language and Environment for Statistical Computing. Preprint at 
(2023). 

71.  McMurdie, P. J. & Holmes, S. Phyloseq: An R Package for Reproducible Interactive 

Analysis and Graphics of Microbiome Census Data. PLoS One 8, (2013). 

72. 

Davis, N. M., Proctor, Di. M., Holmes, S. P., Relman, D. A. & Callahan, B. J. Simple 
statistical identification and removal of contaminant sequences in marker-gene and 
metagenomics data. Microbiome 6, (2018). 

73. 

Oksanen, J. et al. vegan: Community Ecology Package. R package version 2.5-7. 
https://CRAN.R-project.org/package=vegan (2019). 

74.  Wickham, H. Ggplot2: Elegant Graphics for Data Analysis. (Springer-Verlag, New York, 

2016). 

75. 

Shetty, S. biomeUtils: Utilities for microbiome analytics. R package version 0.022. 
Preprint at https://rivm-iiv-microbiome.github.io/biomeUtils/index.html (2023). 

76.  Martinez Arbizu, P. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis. 

(2017). 

77. 

78. 

79. 

80. 

Barnett, D., Arts, I. & Penders, J. microViz: an R package for microbiome data 
visualization and statistics. J Open Source Softw 6, 3201 (2021). 

Gehlenborg, N. UpSetR: A More Scalable Alternative to Venn and Euler Diagrams for 
Visualizing  Intersecting Sets. R package version 1.4.0. Preprint at https://CRAN.R-
project.org/package=UpSetR (2019). 

Fox, J. & Weisberg, S. An R Companion to Applied Regression. (Sage, Thousand Oaks, 
CA., 2019). 

Kassambara, A. rstatix: Pipe-Friendly Framework for Basic Statistical Tests. R package 
version 0.7.2. Preprint at https://CRAN.R-project.org/package=rstatix (2023). 

113 

 
 
81. 

Pedersen, T. L. patchwork: The Composer of Plots. https://CRAN.R-
project.org/package=patchwork (2023). 

82.  Wickham, H. et al. Welcome to the Tidyverse. J Open Source Softw 4, 1686 (2019). 

83. 

84. 

85. 

86. 

87. 

88. 

89. 

90. 

91. 

92. 

Chen, H. VennDiagram: Generate High-Resolution Venn and Euler Plots. Preprint at 
https://CRAN.R-project.org/package=VennDiagram (2022). 

Stopnisek, N. & Shade, A. Persistent microbiome members in the common bean 
rhizosphere: an integrated analysis of space, time, and plant genotype. ISME Journal 15, 
2708–2722 (2021). 

Altschul, S. F. et al. Gapped BLAST and PSI-BLAST: A New Generation of Protein Database 
Search Programs. Nucleic Acids Research vol. 25 (1997). 

Poorter, H. et al. Pampered inside, pestered outside? Differences and similarities 
between plants growing in controlled conditions and in the field. New Phytologist vol. 
212 838–855 Preprint at https://doi.org/10.1111/nph.14243 (2016). 

Pérez-Jaramillo, J. E. et al. Deciphering rhizosphere microbiome assembly of wild and 
modern common bean (Phaseolus vulgaris) in native and agricultural soils from 
Colombia. Microbiome 7, (2019). 

Callahan, B. J., McMurdie, P. J. & Holmes, S. P. Exact sequence variants should replace 
operational taxonomic units in marker-gene data analysis. ISME Journal 11, 2639–2643 
(2017). 

Aviles-Garcia, M. E., Flores-Cortez, I., Hernández-Soberano, C., Santoyo, G. & Valencia-
Cantero, E. The plant growth-promoting rhizobacterium Arthrobacter agilis UMCV2 
endophytically colonizes Medicago truncatula. Rev Argent Microbiol 48, 342–346 (2016). 

Imran, A. et al. Ochrobactrum sp. Pv2Z2 exhibits multiple traits of plant growth 
promotion, biodegradation and N-acyl-homoserine-lactone quorum sensing. Ann 
Microbiol 64, 1797–1806 (2014). 

Ofek, M., Hadar, Y. & Minz, D. Ecology of root colonizing Massilia (Oxalobacteraceae). 
PLoS One 7, (2012). 

Ulrich, K., Kube, M., Becker, R., Schneck, V. & Ulrich, A. Genomic Analysis of the 
Endophytic Stenotrophomonas Strain 169 Reveals Features Related to Plant-Growth 
Promotion and Stress Tolerance. Front Microbiol 12, (2021). 

93.  Wang, X. et al. Early inoculation of an endophyte alters the assembly of bacterial 

communities across rice plant growth stages. Microbiol Spectr (2023) 
doi:10.1128/spectrum.04978-22. 

114 

 
 
94. 

Li, P. et al. Fossil-fuel-dependent scenarios could lead to a significant decline of global 
plant-beneficial bacteria abundance in soils by 2100. Nat Food 4, 996–1006 (2023). 

95.  Medina-Paz, F., Herrera-Estrella, L. & Heil, M. All Set before Flowering: A 16S Gene 

Amplicon-Based Analysis of the Root Microbiome Recruited by Common Bean 
(Phaseolus vulgaris) in Its Centre of Domestication. Plants 11, (2022). 

96. 

97. 

98. 

Arif, I., Batool, M. & Schenk, P. M. Plant Microbiome Engineering: Expected Benefits for 
Improved Crop Growth and Resilience. Trends Biotechnol 38, 1385–1396 (2020). 

Tosi, M., Mitter, E. K., Gaiero, J. & Dunfield, K. It takes three to tango: The importance of 
microbes, host plant, and soil management to elucidate manipulation strategies for the 
plant microbiome. Can J Microbiol 66, 413–433 (2020). 

Bziuk, N., Maccario, L., Sørensen, S. J., Schikora, A. & Smalla, K. Barley Rhizosphere 
Microbiome Transplantation – A Strategy to Decrease Susceptibility of Barley Grown in 
Soils With Low Microbial Diversity to Powdery Mildew. Front Microbiol 13, (2022). 

99. 

Eltlbany, N. et al. Enhanced tomato plant growth in soil under reduced P supply through 
microbial inoculants and microbiome shifts. FEMS Microbiol Ecol 95, (2019). 

100.  Kusstatscher, P. et al. Microbiome-Assisted Breeding to Understand Cultivar-Dependent 

Assembly in Cucurbita pepo. Front Plant Sci 12, (2021). 

101.  Berg, G., Kusstatscher, P., Abdelfattah, A., Cernava, T. & Smalla, K. Microbiome 
Modulation—Toward a Better Understanding of Plant Microbiome Response to 
Microbial Inoculants. Front Microbiol 12, (2021). 

102.  Wang, M., Eyre, A. W., Thon, M. R., Oh, Y. & Dean, R. A. Dynamic Changes in the 

Microbiome of Rice During Shoot and Root Growth Derived From Seeds. Front Microbiol 
11, (2020). 

103.  Wassermann, B., Adam, E., Cernava, T. & Berg, G. Understanding the indigenous seed 
microbiota to design bacterial seed treatments. in Seed Endophytes: Biology and 
Biotechnology 83–99 (Springer International Publishing, 2019). doi:10.1007/978-3-030-
10504-4_5. 

104. 

Lau, J. A. & Lennon, J. T. Rapid responses of soil microorganisms improve plant fitness in 
novel environments. 109, (2012). 

105.  Veach, A. M. et al. Plant Hosts Modify Belowground Microbial Community Response to 

Extreme Drought. mSystems 5, (2020). 

106. 

Shade, A. et al. Fundamentals of microbial community resistance and resilience. 
Frontiers in Microbiology vol. 3 Preprint at https://doi.org/10.3389/fmicb.2012.00417 
(2012). 

115 

 
 
107.  Upton, R. N., Bach, E. M. & Hofmockel, K. S. Belowground response of prairie restoration 

and resiliency to drought. Agric Ecosyst Environ 266, 122–132 (2018). 

116 

 
 
  
 
 
APPENDIX B: 
CHAPTER 3 SUPPLEMENTARY MATERIALS 

Figure S2.1. Experimental Design. Seed microbiome samples were taken from the G0 seed pool 
and G1 and G2 plants. Treatments were applied in G2 in a full factorial design, where one 
offspring from each G1 parent line was treated with each of the three treatments in G2. 

Figure S2.2. Rarefaction curves of quality-filtered microbiome profiles (host reads removed, 
see methods). Each line represents one seed microbiome sample (pool of 5 seeds from the 
same parent plant). The DNA read range is 456-5788. 

117 

 
 
 
 
 
 
 
 
 
 
Average of 20 G0 Samples

Figure S2.3. Mean relative abundance of full dataset ASVs identified at the genus level across 
all three generations. The bar in the top row is the average of the 20 seed samples from G0. 
Bars in the middle row represent G1 parent samples. Bars in the bottom row represent the 
average of the 3 G2 offspring samples in each parent line. The “< 0.5% abund.” category is 
comprised of 263 genera that are less than 0.5% abundant in the dataset. 

118 

 
 
  
 
 
 
 
Figure S2.4. Number of ASVs found overlapping between G1 and G2 within parent lines and 
shared between parent treatment groups. Stars indicate the presence of core seed 
microbiome taxa. There are 99 total taxa represented, of which 85 can also be found in G0. 

Figure S2.5. Mean relative abundance of 70 prevalent overlapping ASVs across parent lines in 
G2 identified at the genus level. All other ASVs have been removed from the dataset. The 
unlabeled legend color represents ASVs that are unresolved at the genus level. 

119 

 
 
 
 
 
 
 
 
 
 
Figure S2.6. Number of ASVs found overlapping in G1 and G2 in only 1 parental line in each 
treatment group. There are 59 unique taxa represented, 49 of which can be found in G0. Dots 
represent taxa that are not found in the G0 dataset. 

Table S2.1 Supplemental File: Supplement metadata table with plant treatment data and 
extraction batches 

Table S2.2 Supplemental File: List of 70 overlapping taxa with taxonomy  

Table S2.3 Supplemental File: Supplemental table of unique taxa  

120 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Elucidating the recruitment timing and development of the root microbiome over the 
common bean lifecycle 

CHAPTER 4: 

121 

 
 
 
 
 
 
ABSTRACT 

Plant microbiomes play an important role in the health of the plant, particularly in the 

rhizosphere and rhizoplane communities. Feedbacks between the growing plant and microbial 

cells in the soil contribute to the assembly of microbial communities over the life of the plant, 

but it is difficult to separate the influence of time from plant development in these 

environments. We conducted an innovative experiment utilizing zebularine treatment, a DNA 

methylation inhibitor, to induce a growth delay in common bean plants. This allowed us to 

investigate the influence of plant growth stage and time across the common bean life cycle. We 

identified closely aligned bacterial communities in the rhizoplane of common bean treated with 

and without zebularine based on plant growth stage, despite the plants growing at different 

rates. Plant growth stage had a stronger influence on the bacterial communities than time, 

while a large proportion of the community variation was attributed to neutral processes. 

Indicator taxa associated with plant growth stages were identified and found to be under a 

selective pressure in the common bean rhizoplane. Selected ASVs included known beneficial 

plant microbiome taxa such as genera Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, 

Pseudolabrys, Duganella, Paenibacillus, Bacillus, Thermoactinomyces and Gaiella. These taxa 

that were under selection in the plant microbiome environment may provide valuable insight 

into key microbiome members that could be beneficial to plant health. 

122 

 
 
 
 
INTRODUCTION 

The root and rhizosphere microbiome plays a critical role in plant health and supports 

the plant by providing important benefits like growth promotion and resilience to biotic and 

abiotic stress1. Particularly in the closely associated soil around the plant root, bacterial and 

fungal communities can assist with water assimilation and nutrient uptake, can out-compete 

pathogenic species, and provide plant growth promoting chemical feedbacks2–5. In the roots of 

leguminous plants such as common bean (Phaseolus vulgaris L.), roots also form symbiotic 

relationships with rhizobia species within their root nodules, which perform symbiotic nitrogen 

fixation and provide usable nitrogen to the plant6–8. Plants also produce root exudates such as 

phytohormones in the soil that impact microbial communities and can enrich specific bacterial 

and fungal species that are beneficial for the plant9–11.  

The microbial communities in the root and rhizosphere are dynamic and diverse. Many 

factors have been investigated that can impact these communities including plant genotype, 

abiotic stress such as drought, soil properties, temperature, and the presence of plant 

pathogens12–16. Additionally, it has been documented that the root microbiome changes over 

the life cycle of the plant in various plant species such as maize, canola, and the medicinal plant 

Panax notoginseng17–19. Alterations in the rhizosphere microbiome over time can further 

support the health of the plant, such as through increased resistance to pathogens3. The 

microbial communities in the plant root system typically become more taxonomically and 

functionally diverse as the plant grows, and shift in community composition can coincide with 

shifts in plant life stages3,20,21.  

123 

 
 
While there are various factors that can influence root and rhizosphere communities, a 

key factor in plant microbiome assembly is the influence of the plant via interactions such as 

the release of phytohormones and plant-microbe feedbacks22,23. Root exudates can influence 

the plant growth-promoting bacteria in the rhizosphere, contribute to the establishment of 

symbiotic associations, and drive community assembly of the microbiome with plant 

growth10,18,24–27. However, there are also other factors that can influence microbial 

communities in soil, such as soil and environmental conditions and the natural succession of 

microbial species over time 28–31. Because plant growth and time are closely linked, it is difficult 

to study the influence of these two variables separately in the plant microbiome. Studies have 

investigated the temporal variation of plant microbiomes over the growing season or plant life 

cycle, but these studies are unable to determine the proportion of microbial change attributed 

to community progression over time versus the influence of plant development22,32.  

This study aimed to tease apart these two variables, plant growth and time, in order to 

determine the proportion of the community succession that could be attributed to these 

variables, and which variable was the primary driver of community assembly in common bean. 

We used a chemical treatment, zebularine (zeb), to induce a plant growth delay in common 

bean plants, which allowed us to study the microbial communities of the plant root system over 

two sampling schemes, plant growth stage and time since planting. Zebularine is a DNA 

methylation inhibitor that is often used to study epigenetic changes in plants33. However, an 

interesting result of zebularine treatment on Common Bean seeds is a significant growth delay 

once planted. When common bean seeds are treated with zebularine, plant growth progresses 

at a rate 2-3 weeks behind plants that are not treated. We grew groups of plants with and 

124 

 
 
without zebularine treatment and sampled the bacterial communities across two different 

sampling schemes, allowing us to separate plant growth from time.  

We hypothesized that the zebularine plants sampled under the growth stage and time 

series sampling schemes would have distinct microbial communities, and these communities 

would either resemble the untreated plants at the growth stages, or the untreated plants at 

days since planting. This would allow us to determine which variable was driving the 

development of the plant root communities. We sampled both the rhizosphere (the soil around 

the root system) and the rhizoplane (the community closely associated with the surface of plant 

roots) to understand which compartment of the root community is most affected by the growth 

stage and time variables. Additionally, the bacterial community in the rhizoplane was assessed 

with the Sloan neutral model to determine the influence of selective and neutral processes in 

the common bean rhizoplane, and indicator taxa associated with bean growth stages were 

investigated. This study provides important insight into how the root community is influenced 

by the plant over time and the role of assembly processes in the rhizoplane.  

METHODS 

Surface sterilization and Zebularine seed treatment 

Phaseolus vulgaris L. var. “Red Hawk”34 seeds were surface-sterilized in a solution of 

10% bleach and 0.1% Tween20, then treated with a solution of 5 mM zebularine (zeb) dissolved 

in Milli-Q purified water, or a control treatment with Milli-Q water only. For surface-

sterilization, seeds were placed in a sterile glass beaker and covered with approximately 3 cm of 

bleach solution. Seeds were soaked for 10 minutes with agitation at 5 minutes, then the bleach 

solution was drained, seeds were briefly rinsed with sterile DI H2O five times, and then drained 

125 

 
 
 
of excess water. After surface-sterilization, seeds were divided into petri dishes lined with 

sterile filter paper. Approximately 20-25 seeds were placed in each petri dish, allowing each 

seed to sit directly on the filter paper. The filter paper in each petri dish was moistened with 1 

mL sterile DI H2O, then 2 mL of 5 mM zebularine solution or Milli-Q water was dripped over the 

seeds in the zeb treatment or control treatment dishes, respectively. The petri dishes were 

stored in the dark at room temperature for three days, with an additional 2 mL zebularine or 

control water added to the seeds each day. The seeds were then planted on the fourth day.   

Plant growth, harvest, and phenotypic measurements 

Before planting, agricultural field soil was collected from the Michigan State University 

Agronomy Farm from a field that had previously grown common bean (42°42'57.4"N, 

84°27'58.9"W, East Lansing, MI, USA). The soil was a sandy loam with an average pH of 7.2. Soil 

was passed through a 4 mm sieve to remove rocks and plant debris and stored in sealed 

buckets at 4°C prior to use. Immediately before use the soil was mixed with autoclaved coarse 

vermiculite at a 50% v/v ratio. Seeds were then planted in the field soil and vermiculite mixture 

in 2x2 inch seedling pots in a high-light BioChambers FLEX™ LED growth chamber with a 16-

hour day/8-hour night cycle at 26°C and 22°C, respectively, and 50% relative humidity. Seeds 

were lightly covered with soil and pots were bottom watered by filling a tray beneath the pots 

with water and allowing the soil to absorb water until thoroughly moist. Pots were watered 

with half-strength Hoagland’s solution once a week and DI H2O every two or three days as 

needed to maintain soil moisture35. As seedlings grew, they were transferred to larger pots, 

either 4x4 inch pots for plants that would be harvested at earlier time points, and 1-gallon pots 

for plants that would be harvested at later time points. 

126 

 
 
 
Five plants from both the control and zebularine treatments were randomly selected for 

destructive sampling at each sample point. Plants harvested based on growth stage were 

harvested at the following stages: V1 - first trifoliate leaves expanded, V2 - second trifoliate, V3 

- third trifoliate, V4 - fourth trifoliate, R1 - first open flowers, R4 - approximately half of the seed 

pods fully developed, R7 - seed pods dry and plant senescing36. Plants harvested based on the 

time series were harvested on the following days after planting: 3, 7, 14, 21, 35, 49, 63. The 

time series harvest days roughly aligned with the control treated growth stage sample points. 

See Figure S3.1 for experimental design schematic. Before harvest, pot trays were emptied of 

water to allow the soil to dry for 2-3 days. On harvest day, plant height was measured with a 

ruler from the surface of the soil to the tallest part of the plant, holding the plant upright to its 

tallest point if it had fallen over. For day 3, a sterilized scoopula was used to scoop out the soil 

closest to the roots of the seedling to be collected as rhizosphere. At all other sampling points, 

the plant and soil were turned out of the pot into a clean tray. The root system was gently 

removed from the soil, placed in a sterile Whirl-pakâ bag, and shaken to collect the 

rhizosphere soil. Then the roots were placed on a white background next to a centimeter ruler 

to be photographed for root analysis. Next, the above-ground biomass was cut from the root 

system and placed in a paper envelope, and the root system was placed in another sterile 

Whirl-pakâ bag. The root and rhizosphere samples were flash-frozen in liquid nitrogen and 

then stored at -80°C until DNA extractions were performed. The above-ground biomass was 

weighed in its envelope immediately after harvest for fresh weight, then placed in a 50°C drying 

oven for 2-3 days and weighed again for dry weight.  

127 

 
 
Root photos were analyzed with ImageJ software. For each photo, the scale of the 

image was set in millimeters using the ruler in the image, and the freehand line tool was used 

to trace and measure the five longest roots starting from where the stem met the soil. The 

average root length was calculated for the plant based on the five length measurements. For 

zebularine plants at early time series sampling points, only the main taproot was measured due 

to lack of lateral roots. Traced paths of roots were estimated for mature plants with dense root 

systems. Visible root nodules were counted in each photo with the ImageJ multi-point tool. 

Nodule count is an estimate based on what was visible in the root photo and does not 

necessarily account for nodules in the entire root system, as some nodules may not have been 

visible. 

Bacterial DNA extraction for microbiome analysis 

Rhizosphere and root samples were processed in randomized batches, see 

supplemental metadata table for extraction batch information. Root samples were thawed at 

room temperature, and a 4-5 cm section of the “active” outer zone of the root system (or the 

entire root system for very small plants) was cut with a sterilized razor blade and placed in a 

sterile 50 mL conical tube with enough sterile DI H2O to cover the roots. The roots were gently 

swirled to remove loosely adhered soil, then transferred to a sterile 15 mL conical tube with 

enough sterile Phosphate Buffered Saline (PBS) + 0.05% Tween20 to cover the roots. Very small 

roots were not rinsed with water and were placed directly in the 15 mL conical tube due to low 

surface area with a small amount of adhered soil. Roots were then vortexed at maximum speed 

in the PBS solution for 1-2 minutes until all rhizoplane soil was removed from the surface of the 

roots. Root material was then removed, and the tubes were centrifuged at 4600 x g, 4°C, for 15 

128 

 
 
 
minutes to pellet the rhizoplane material. Following centrifugation, the PBS supernatant was 

removed and pellets were stored at -20°C until DNA extraction. 

DNA extraction was performed with the DNeasy PowerSoil Pro DNA Kit (Qiagen, 

Germantown, MD, USA). Rhizosphere soil was thawed at room temperature and weighed into 

the bead beating tubes for the first step of the DNA extraction protocol. Rhizoplane pellets 

were thawed on ice and then weighed into the bead beating tubes. DNA extraction was 

performed following manufacturer’s instruction with the following modifications: In step one, 

30-50 μg of sample was used in each tube, then 700 μL solution CD1 was used with 100 μL ATL 

buffer (Qiagen, Germantown, MD, USA). The bead beating step was performed for 15 minutes 

on a vortex genie 24-tube adapter at maximum speed. Lastly, 60 μL of final elution buffer C6 

was used, and tubes were incubated at room temperature for 10 minutes before 

centrifugation. Each extraction batch included a negative control of an empty PowerSoil Pro 

tube at the start of the extraction protocol. For the rhizoplane negative controls, solution CD1 

was swirled in a sterile 15 mL conical tube before being added to the PowerSoil Pro tube. A 

positive control of a custom-made mock bacterial community suspended in solution CD1 was 

used in one extraction batch per compartment37. Controls were processed with the extraction 

batch and then sequenced with the true samples for use in downstream DNA sequence 

decontamination. 

Sequencing and sequence processing 

DNA sequencing was performed by the Michigan State University Research Technology 

Support Facility Genomics Core in East Lansing, Michigan, USA, following their standard 

protocol. The V4 region of the 16S rRNA gene was amplified from the rhizosphere and 

129 

 
 
 
rhizoplane extractions using dual indexed, Illumina compatible primers following the protocol 

described in Kozich, JJ, et al. (2013)38. PCR products were batch normalized using an Invitrogen 

SequalPrep DNA Normalization plate and product recovered from the plates pooled. The pool 

was concentrated and cleaned up using a QIAquick Spin column and AMPure XP magnetic 

beads. The pool was QC’d and quantified using a combination of Qubit dsDNA HS, Agilent 4200 

TapeStation HS DNA1000 and Invitrogen Collibri Illumina Library Quantification qPCR 

assays.This pool was loaded onto one (1) MiSeq v2 Standard flow cell and sequencing was 

carried out in a 2x250bp paired end format using a MiSeq v2 500 cycle reagent cartridge. 

Custom sequencing and index primers complementary to the 515f/806r oligomers were added 

to appropriate wells of the reagent cartridge. Base calling was done by Illumina Real Time 

Analysis (RTA) v1.18.54 and output of RTA was demultiplexed and converted to FastQ format 

with Illumina Bcl2fastq v2.20.0. Sequencing was repeatedly performed in three MiSeq runs to 

maximize the total number of DNA reads returned for the dataset. 

After sequencing, fastq files were processed in QIIME2 after primer removal by the 

sequencing center (QIIME2 version: 2022.8.0)39. Sample fastqs were imported to QIIME2 

format, and samples were denoised, truncated and merged using Dada2 and amplicon 

sequence variants (ASVs) were defined at 100% sequence identity40. Customized truncation 

lengths for each MiSeq run were determined using FIGARO41. Run 1 had a forward truncation 

length of 99 and reverse length of 176, run 2 had forward length 124 and reverse length 151, 

and run 3 had forward length 125 and reverse length 150. Once representative sequences and 

ASV tables were created for each run, the runs were merged into one dataset and 16S 

130 

 
 
taxonomy was assigned with the Silva database release 13842. Taxonomy and ASV tables were 

then exported for further analysis in R.  

Sequence Decontamination  

Data analyses were performed in R version 4.3.1 and R Studio version 2023.06.1+52443. 

There were 20.1 million merged DNA reads in the dataset prior to host removal and 

decontamination. ASV, taxonomy, and metadata tables were imported into the phyloseq 

package and host reads classified as chloroplast and mitochondria were removed using the 

subset_taxa() command in the phyloseq package version 1.44.044. Nine percent of the total 

DNA reads were removed as host reads, leaving 18.3 million total bacterial DNA reads. Datasets 

were decontaminated with the decontam package version 1.20.0 at the 0.55 threshold utilizing 

the specific negative controls for each extraction group, and positive extractions with the entire 

dataset45. After decontamination, there were 16.3 million total DNA reads with a range of 

3,635- 151,332 reads per sample in the full dataset. Rarefaction curves were created using the 

rarecurve() command in the vegan package version 2.6-446. Samples were rarefied to the tenth 

lowest number of reads in the dataset at 22,761 using the rarefy_even_depth() command in 

phyloseq. Rarefaction removed nine samples that did not meet the minimum number of reads. 

Analysis 

Plant phenotypic traits were analyzed with the base R stats command t.test() and 

figures were created with the ggplot2 package version 3.4.247. Alpha diversity species richness 

was assessed using estimate_richness() in phyloseq, and figures were created using the 

plot_richness() command from phyloseq with ggplot2. Bray-Curtis distances were calculated 

with vegdist() from the vegan package and used for all analyses of beta diversity, and 

131 

 
 
 
PERMANOVA statistical tests were performed with adonis2() from vegan. Post-hoc analysis on 

the PERMANOVA results was performed with pairwise.adonis2() from pairwiseAdonis version 

0.4.148. Beta diversity ordinations were created with ordinate() from phyloseq with ggplot2. 

Beta dispersion was assessed with the betadisper() and permutest() commands from the vegan 

package. Datasets were divided into groups where applicable using the ps_filter() command in 

the microViz package version 0.10.1049. Procrustes concordance analysis was performed using 

the vegdist(), monoMDS(), procrustes() and protest() commands in vegan. Heatmap was 

created using the ComplexHeatmap package version 2.16.050,51. Indicator taxa were identified 

with the multipatt() command from the indicspecies package version 1.7.14, and the analysis 

was performed on all ASVs that had greater than 100 reads in the dataset52. The Sloan neutral 

model fit was performed on all ASVs with at least 5 reads in the dataset with the fit_sncm() 

command from the reltools package version 0.1.053. 

Data Availability 

Data analysis code and plant phenotypic measurement data can be found at 

(https://github.com/ShadeLab/Rhizosphere_assembly_Common_Bean). Raw DNA sequences 

can be found on the NCBI Sequence Read Archive under BioProject accession number 

PRJNA1066866. Root photographs can be accessed on figshare.com under the following DOIs: 

doi.org/10.6084/m9.figshare.25054997, doi.org/10.6084/m9.figshare.25055087, 

doi.org/10.6084/m9.figshare.25055774, doi.org/10.6084/m9.figshare.25057067. 

132 

 
 
 
 
RESULTS 

Quantifying the Zebularine Growth Delay 

When the common bean plants were grown with and without zebularine treatment, the 

zeb plants harvested based on the growth stage sampling points were at the same stage as 

control plants, but were harvested 7-14 days after the control plants at the same growth stage 

(Figure S3.2, S3.3). Alternatively, at the time series sampling points, the control treated plants 

were at comparable growth stages to the control plants harvested based on growth stage, 

while zebularine (zeb) treated plants had delayed growth compared to the control (Figure S3.2, 

S3.3). Above and below-ground phenotypic traits, including the plant height, fresh and dry 

shoot biomass, root length, and approximate nodule count, were measured at each sampling 

point. While the zeb treated plants did not reach the same mature height as the control plants, 

the above-ground fresh and dry biomass were similar between control and zeb in the growth 

stage series, with no significant difference between control and zeb dry biomass at plant 

maturity (Figure 3.1, Welch Two Sample t-test: Control Vs Zeb R4 t=-0.94992, df=7.8587, p-

value=0.3704; Control vs Zeb R7, t=1.2701, df=5.5536, p-value=0.2547). Alternatively, above-

ground biomass was delayed in zeb plants in the time series and zeb plants had not yet reached 

maturity at the final sampling point on day 63 (Figure 3.1C, Welch Two Sample t-test: Control Vs 

Zeb Day 63, t=7.6828, df=6.8434, p-value=0.00013***).  

133 

 
 
Figure 3.1. Above-ground plant growth measurements by treatment group and sampling series, 
Growth Stage (left) or Time Series (right). A. plant height (cm). B. above-ground fresh biomass 
(g). C. Above-ground dry biomass (g). Plants had not emerged from the soil for above-ground 
biomass collection at the day 3 time point. Asterisks indicate significance between Control and 
Zeb treatment groups at labeled sampling points. Boxplots n=5, black line indicates mean of 
each sample set. Welch Two Sample t-test, ns= not significant, p-values: *<0.05, **<0.01, 
***<0.001, ****<0.0001. 

134 

 
 
 
In below-ground phenotypic measurements, similar results to plant height were 

observed for plant root length, where the zeb root length was shorter than control overall 

(Figure 3.2A, Welch Two Sample t-test: Control Vs Zeb R4, t=2.9755, df=7.6344, p-

value=0.0186*; Control Vs Zeb R7, t=2.5611, df=6.1922, p-value=0.0416*; Control Vs Zeb Day 

63, t=4.6969, df=7.8593, p-value=0.00162**). However, the growth delay of the zeb plants 

compared to control was still pronounced in the time series root length measurements (Figure 

3.2A). Additionally, even though nodule count was estimated from visible nodules in root 

photos, counts followed a similar trajectory between control and zeb over the growth stages 

with no significant difference in mature plants, while nodule development was delayed in the 

zeb time series plants (Figure 3.2B, Welch Two Sample t-test: Control Vs Zeb R4, t=-0.42188, 

df=7.7992, p-value=0.6845; Control Vs Zeb R7, t=-1.217, df=7.4621, p-value=0.2607). These 

results confirm and quantify the observed growth delay of roughly 2 weeks when common 

bean plants are treated with zebularine prior to planting, which allowed us to study the root 

microbiome over two different developmental time series.         

135 

 
 
Figure 3.2. Below-ground plant development measurements by treatment group and sampling 
series, Growth Stage (left) or Time Series (right). A. average root length (mm). B. number of 
visible nodules present in root photo. Asterisks indicate significance between Control and Zeb 
treatment groups at labeled time points. Boxplots n=5, black line indicates mean of each 
sample set. Welch Two Sample t-test, ns= not significant, p-values: *<0.05, **<0.01, ***<0.001, 
****<0.0001. 

Rhizoplane bacterial communities are modulated by plant growth stage 

Microbial communities in the rhizosphere and rhizoplane were analyzed at each 

sampling point with 16S V4 rRNA gene amplicon sequencing. Samples were rarefied to 22,761 

DNA reads and then grouped based on plant compartment for further analysis (Figure S3.4). In 

the rhizosphere, ASV richness ranged from 1,000 to 2,000 ASVs per sample, and there were no 

significant differences between control and zeb in either sampling series (Figure 3.3). ASV 

136 

 
 
 
richness in the rhizoplane samples ranged from 500 to 1,500 ASVs per sample, and there was a 

pattern in richness over the plant life cycle. In the growth stage samples in both control and zeb 

treated plants, ASV richness was highest at the late vegetative and early reproductive stages 

(V4 and R1) and lowest when the plants reached senescence (Figure 3.4A). The control plants 

sampled in the time series followed a similar pattern, but the zeb treated plants in the time 

series had lower richness until day 35, and higher richness at later time points (Figure 3.4B). 

There were significant differences between control and zeb rhizoplane richness at the R4 

growth stage, and the day 14 and day 35 time points (Welch Two Sample t-test, R4: t=2.6229 

df=7.9952 p=0.03053*; Day14: t=2.8744 df=5.8271 p=0.02921*; Day35: t=-3.101 df=4.8484 

p=0.02795*) (Figure 3.4).   

137 

 
 
 
Figure 3.3. Alpha diversity in rhizosphere samples. There are no significant differences in ASV 
richness between control and zeb in either the A. growth stage series or B. time series. Boxplots 
n=5, black line indicates the mean of each sample group. 

138 

 
 
 
Figure 3.4. Alpha diversity in the rhizoplane. Asterisks indicate a sampling point where there is a 
significant difference in ASV richness between control and zeb samples within the A. growth 
stage series or B. time series. Boxplots n=5, black line indicates the mean of each sample group. 
Asterisks indicate significance between Control and Zeb treatment groups at labeled time 
points (Welch Two Sample t-test, R4: t=2.6229 df=7.9952 p=0.03053; Day14: t=2.8744 
df=5.8271 p=0.02921; Day35: t=-3.101 df=4.8484 p=0.02795). Both growth stage groups and 
the control-time plants have a similar pattern of alpha diversity over the sampling points, with 
their highest alpha diversity at the late vegetative and early reproductive stages. The zeb-time 
plants do not match this pattern. 

Bray-Curtis distance was used to compare the composition of the bacterial communities 

across sampling points. When the two control treated datasets were compared (growth stage 

and time), there were no significant differences based on the sampling series. However, when 

zeb treated growth and time datasets were compared, there were significant differences 

between the sampling series in both the rhizosphere and rhizoplane, further confirming that 

139 

 
 
 
the sampling series are not aligned in zeb treated plants, and there are differences in the 

microbial communities as a result (PERMANOVA, Rhizosphere Zeb Growth vs Zeb Time: 

r2=0.01917, F=1.2897, p=0.038*; Rhizoplane Zeb Growth vs Zeb Time: r2=0.03051 F=2.1087 

p=0.0188*) (Data not shown, statistical results can be found in Table S3.1). 

To determine which control sampling series the zeb microbial communities align with 

more closely, the Bray-Curtis distance was compared between the control and zeb treatment in 

the two sampling series. In all comparisons there were significant differences in the microbial 

communities based on which treatment the plants received, suggesting some influence of the 

zebularine treatment on the resulting root communities regardless of sampling time. However, 

these differences based on treatment accounted for a small amount of the variation in each 

dataset, and samples in the growth stage series were clustered more closely at each growth 

stage between control and zeb verses the samples in the time series (PERMANOVA Control vs. 

Zeb treatment. Rhizosphere - growth: r2=0.03035 F=2.2850 p=0.0001*, Rhizosphere - time: 

r2=0.02404 F=1.7894 p=0.0004*, Rhizoplane - growth: r2=0.03567 F=2.9122 p=0.0015*, 

Rhizoplane - time: r2=0.03382 F=3.2968 p=0.0011*) (Figure 3.5, Table S3.1). When communities 

were compared between control and zeb at the specific sampling points, both rhizosphere 

sampling series were similar at the earliest sampling points, but became significantly different 

from each other after the second sampling point regardless of series (Table S3.1). In the 

rhizoplane, the communities sampled based on the time series were significantly different 

between control and zeb at all time points, while the growth stage sampling points were not 

significantly different at the V3, V4 and R7 stages (Table S3.1). These results suggest that there 

is less direct influence of the plant on the rhizosphere communities in this study, while the 

140 

 
 
rhizoplane is significantly influenced by the plant growth stages, and the communities in the 

zebularine treated plants are more closely aligned to the control plants by growth stage verses 

time. When the variation in the rhizoplane dataset was investigated, 20% of the overall 

variance was explained by plant growth stage, 10% was explained by time (days since planting), 

and 2% of the variance was explained by the control or zeb treatment that the plant received 

(Table 3.1). There was also an interaction observed between the influence of Treatment and 

Growth Stage that accounted for 6% of the variation in the rhizoplane samples (Table 3.1). 

Since the greater proportion of variation is explained by the growth stage, this supports the 

observations that the rhizoplane bacterial communities are modulated by plant growth. 

Additionally, when the beta-dispersion of the Bray-Curtis distances were compared, there were 

no significant differences in distance to spatial median between the zeb and control samples in 

any sampling series, or between any sampling points, suggesting that the variance of the 

bacterial communities between samples was not affected by zeb treatment or sampling point 

(Figure S3.5).  

Table 3.1  
PERMANOVA of variance in the rhizoplane dataset. All samples were assigned a growth stage 
and age time point (days since planting) based on harvest date. The greatest variation in the 
rhizoplane bacterial communities (R2 value) is explained by the growth stage, followed by the 
time point. 

Treatment 
(Control vs Zeb) 
Age (Days) 
Growth Stage 
Treatment:Age 
Treatment:Growth 
Stage 

R2 
0.02418 

0.10648 
0.20613 
0.00872 
0.06569 

F 
4.3122 

18.9896 
2.2975 
1.5558 
1.4644 

141 

p-value 
0.001 

0.001 
0.001 
0.077 
0.002 

Significance 
*** 

*** 
*** 

** 

 
 
 
 
 
Figure 3.5. Beta diversity PCoA of the Bray-Curtis distances of the rhizosphere (A, B) and 
rhizoplane (C,D) microbial communities harvested at the Growth Stage (left) and Time Series 
(right) sampling points. Sampling point is indicated by symbol size, Treatment Groups are 
statistically different in all sampling series (PERMANOVA, A: r2=0.03035 F=2.2850 p=0.0001*, B: 
r2=0.02404 F=1.7894 p=0.0004*, C: r2=0.03567 F=2.9122 p=0.0015*, D: r2=0.03382 F=3.2968 
p=0.0011*). When sampling points are compared between Control and Zeb within each 
ordination, the rhizosphere growth stage and time series samples are statistically different from 
each other at all but the earliest sample points at the V1 and V2 stages, and day 3 and day 7, 
respectively. In the rhizoplane, the growth stage samples are not statistically different at the 
V3, V4 and R7 stages, while all samples are statistically different at each time point in the time 
series (see supplementary table for sampling point statistics). 

142 

 
 
 
 
   
 
Rhizoplane communities are aligned in the growth stage series and contain selected taxa 

To further investigate the similarities between the control and zeb samples in each 

sampling series, a Procrustes concordance analysis was performed to compare the structure of 

the PCoA ordinations. The Procrustes analysis revealed concordance between the rhizoplane 

control and zeb ordinations in the growth stage series, further confirming that the bacterial 

communities in these samples are aligned over the lifecycle of the plant, while all other 

ordinations were not concordant (Rhizoplane Growth: Procrustes Sum of Squares = 0.8591, 

Correlation in a symmetric Procrustes rotation = 0.3754, Significance value = 0.024*. Other 

significance values: Rhizosphere Growth = 0.332, Rhizosphere Time = 0.498, Rhizoplane Time = 

0.989) (Figure 3.6). Additionally, when the 57 most abundant taxa in the ordination are plotted 

on a heatmap across all rhizoplane samples, the control and zeb samples mostly cluster 

together at the sampling points based on growth stage, while in the time series the control and 

zeb cluster together at only the earliest time points and appear to separate from each other at 

later time points (Figure 3.7A). This also supports the finding that there is greater similarity 

between treatments based on growth stage.  

143 

 
 
 
Figure 3.6. Procrustes concordance analysis between Control and Zeb samples in Bray-Curtis 
distance PCoA ordinations. A. Rhizosphere Growth Stage, B. Rhizosphere Time Series, C. 
Rhizoplane Growth Stage, D. Rhizoplane Time Series. Blue arrows indicate distances between 
correlating sample points between Control and Zeb. Only the Rhizoplane Growth Stage (C) 
control and zeb ordinations are concordant with each other (C: Procrustes Sum of Squares = 
0.8591, Correlation in a symmetric Procrustes rotation = 0.3754, Significance value = 0.024*. 
Other significance values: A = 0.332, B = 0.498, C = 0.989).  

In order to further investigate the assembly processes in the rhizoplane of common 

bean, the Sloan neutral model was used to investigate taxa that may be assembled by selection 

in the plant environment54,55. The Sloan neutral model was fit to the entire rhizoplane growth 

stage dataset, including both the control and zeb treatment groups. Of the ASVs in the 

144 

 
 
 
 
 
rhizoplane dataset, 13.6% were above prediction in the model, suggesting selection by the 

plant environment over the life cycle of the plant (Figure 3.7B). Alternatively, 4.3% were below 

prediction, suggesting that they are dispersal limited, while 82.1% of the taxa fell within the 

neutral model, which suggests they are assembled by neutral or stochastic processes (Figure 

3.7B).  

Additionally, because the growth stages had the greatest influence on the assembly of 

the rhizoplane bacterial communities in this study, indicator taxa were identified across the 

rhizoplane growth dataset based on growth stage. There were 204 ASVs identified as indicator 

taxa associated with the plant growth stages, 60 and 63 of which were associated with the late 

vegetative and early reproductive growth stages (V4 and R1), 33 were associated with the V1 

stage, 26 were associated with R7, and few taxa were identified as indicators of the V2, V3 and 

R4 stages. When these taxa were investigated in the Sloan neutral model, 75 indicator taxa 

were above prediction (Figure 3.7B). Most of the indicator taxa that were selected by the plant 

are associated with the V1, V4 and R1 growth stages, while there were very few selected 

indicator taxa from other growth stages (Figure 3.7C). These selected indicator taxa were 

diverse and included ASVs from 29 bacterial classes, with over 40% of the selected taxa 

identified in the classes Alphaproteobacteria, Thermoleophilia, Bacilli and 

Gammaproteobacteria (Figure 3.7C). The selected ASVs included taxa from genera 

Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, Pseudolabrys, Duganella, Paenibacillus, 

Bacillus, Thermoactinomyces and Gaiella, and families Hyphomonadaceae, 

Sphingomonadaceae, and Nitrosomonadaceaeand and many others that are known to be 

commonly associated with plant microbiomes (See Table S3.2 for full taxonomy of selected 

145 

 
 
indicator taxa). Overall, these results confirm the significant influence of the plant growth stage 

on the resulting rhizoplane community, and provide evidence that the microbial communities 

are under selection by the plant.       

146 

 
 
 
A

Growth Stage

Growth Stage

Time Series

Time Series

Treatment
Sampling_point

Treatment
Sampling_point

ASV reads
per sample
2000
1500
1000
500
0

ASV reads
per sample
2000
1500
1000
500
0

ASV Taxonomy

Treatment
Control
Zeb

Treatment
Control
Zeb

Sampling Point

Sampling Point

Day3
Day7
Day14
Day21
Day35
Day49
Day63
V1
V2
V3
V4
R1
R4
R7

ASV Taxonomy

Day3
Day7
Day14
Day21
Day35
Day49
Day63
V1
V2
V3
V4
R1
R4
R7

g__Nitrososphaeraceae
g__Nitrososphaeraceae
g__Bdellovibrio
g__Bdellovibrio
g__Pseudomonas
g__Pseudomonas
g__Pseudolabrys
g__Pseudolabrys
g__Devosia
g__Devosia
g__Ensifer
g__Ensifer
g__Dyadobacter
g__Dyadobacter
g__Hydrogenophaga
g__Hydrogenophaga
g__Allo−Neo−Para−Rhizobium
g__Allo−Neo−Para−Rhizobium
g__Novosphingobium
g__Novosphingobium
f__Comamonadaceae
f__Comamonadaceae
f__Methyloligellaceae
f__Methyloligellaceae
g__Luteolibacter
g__Luteolibacter
g__Bosea
g__Bosea
g__Polaromonas
g__Polaromonas
o__Bacillales
o__Bacillales
g__Phenylobacterium
g__Phenylobacterium
g__Methylophilus
g__Methylophilus
g__Bacillus
g__Bacillus
f__Micrococcaceae
f__Micrococcaceae
g__Massilia
g__Massilia
g__Flavobacterium
g__Flavobacterium
g__Paenarthrobacter
g__Paenarthrobacter
f__Oxalobacteraceae
f__Oxalobacteraceae
g__Paenibacillus
g__Paenibacillus
g__Tumebacillus
g__Tumebacillus
g__Rhizobacter
g__Rhizobacter
f__Spirosomaceae
f__Spirosomaceae
f__Rhizobiaceae
f__Rhizobiaceae
g__Bradyrhizobium
g__Bradyrhizobium
g__Methylotenera
g__Methylotenera
g__Mesorhizobium
g__Mesorhizobium
g__Duganella
g__Duganella

V
S
A

V
S
A

4
6
_
P
R

0
3
1
_
P
R

5
6
_
P
R

3
4
_
P
R

4
7
_
P
R

4
6
6
_
0
P
1
R
_
P
R

0
3
9
1
2
_
1
P
_
R
P
R

8
2
1
_
P
R

5
6
7
_
8
P
_
R
P
R

3
4
0
_
1
P
_
R
P
R

4
7
7
_
7
P
_
R
P
R

8
2
6
1
3
_
_
P
P
R
R

6
0
4
1
2
_
_
P
P
R
R

9
2
5
1
5
_
_
P
P
R
R

7
8
8
_
_
P
P
R
R

0
1
6
_
7
P
_
R
P
R

7
7
8
_
5
P
_
R
P
R

6
3
9
_
0
P
1
R
_
P
R

4
2
8
_
0
P
1
R
_
P
R

5
5
9
_
3
P
1
R
_
P
R

8
_
9
P
8
R
_
P
R

6
7
7
_
5
P
_
R
P
R

8
5
8
_
1
P
1
R
_
P
R

9
0
8
1
3
_
1
P
_
R
P
R

8
0
6
1
5
_
_
P
P
R
R

9
3
8
1
8
_
_
P
P
R
R

9
8
7
_
0
P
1
R
_
P
R

7
5
2
_
4
P
1
R
_
P
R

8
1
5
1
3
_
_
P
P
R
R

8
3
5
1
1
_
1
P
_
R
P
R

6
5
4
_
1
P
1
R
_
P
R

8
8
1
_
1
P
_
R
P
R

7
0
2
1
1
_
_
P
P
R
R

2
4
1
1
_
_
P
P
R
R

5
3
7
_
2
P
_
R
P
R

5
1
5
1
4
_
_
P
P
R
R

4
1
6
1
4
_
_
P
P
R
R

1
1
9
_
1
P
1
R
_
P
R

2
1
0
_
4
P
1
R
_
P
R

1
_
6
P
8
R
_
P
R

7
2
8
_
6
P
_
R
P
R

5
4
6
_
2
P
_
R
P
R

6
4
4
_
4
P
_
R
P
R

9
1
7
1
9
_
_
P
P
R
R

0
4
3
1
_
_
P
P
R
R

6
8
7
_
4
P
_
R
P
R

8
6
4
_
1
P
_
R
P
R

6
2
3
_
1
P
_
R
P
R

4
4
2
_
2
P
_
R
P
R

7
9
8
_
2
P
_
R
P
R

3
_
4
P
5
R
_
P
R

7
4
3
_
5
P
_
R
P
R

4
1
8
_
9
P
_
R
P
R

3
1
2
_
_
P
P
R
R

8
2
9
_
_
P
P
R
R

2
2
5
_
1
P
_
R
P
R

4
5
4
_
3
P
_
R
P
R

3
5
1
_
4
P
1
R
_
P
R

5
1
_
P
R

8
9
3
_
2
P
_
R
P
R

2
_
5
P
2
R
_
P
R

5
7
_
P
R

9
_
P
R

4
3
_
P
R

5
2
_
P
R

3
2
_
P
R

1
4
1
_
P
R

7
1
1
_
P
R

6
1
1
_
P
R

9
5
_
P
R

5
8
_
P
R

1
3
1
_
P
R

5
7
1
_
7
P
_
R
P
R

1
3
2
1
7
_
_
P
P
R
R

5
8
0
_
5
P
_
R
P
R

7
1
5
1
3
_
1
P
_
R
P
R

6
1
9
1
4
_
_
P
P
R
R

9
5
6
_
3
P
1
R
_
P
R

0
5
_
P
R

2
7
_
P
R

6
_
P
R

1
7
_
P
R

0
7
_
P
R

2
3
_
P
R

1
5
_
P
R

5
3
1
_
P
R

2
1
1
_
P
R

9
4
5
_
4
P
1
R
_
P
R

6
3
0
1
1
_
1
P
_
R
P
R

2
3
8
_
1
P
_
R
P
R

0
7
3
_
0
P
1
R
_
P
R

6
_
2
P
8
R
_
P
R

1
5
9
_
3
P
_
R
P
R

2
1
0
1
8
_
_
P
P
R
R

5
4
2
1
0
_
1
P
_
R
P
R

0
1
1
1
6
_
_
P
P
R
R

8
1
0
_
4
P
_
R
P
R

3
0
6
1
1
_
_
P
P
R
R

2
8
9
_
9
P
_
R
P
R

9
3
6
_
9
P
_
R
P
R

0
8
7
_
_
P
P
R
R

2
0
3
1
1
_
1
P
_
R
P
R

1
6
1
_
0
P
1
R
_
P
R

0
4
4
_
2
P
1
R
_
P
R

6
1
3
_
3
P
1
R
_
P
R

9
9
3
_
9
P
_
R
P
R

6
9
1
_
4
P
_
R
P
R

7
_
8
P
4
R
_
P
R

3
1
3
1
2
_
1
P
_
R
P
R

1
0
6
1
4
_
1
P
_
R
P
R

4
2
2
1
6
_
_
P
P
R
R

3
3
1
1
1
_
1
P
_
R
P
R

3
9
3
_
4
P
1
R
_
P
R

1
4
1
_
2
P
1
R
_
P
R

8
4
7
_
3
P
1
R
_
P
R

3
2
5
1
_
_
P
P
R
R

6
4
9
1
7
_
_
P
P
R
R

2
6
4
_
_
P
P
R
R

1
1
0
1
2
_
1
P
_
R
P
R

3
4
4
1
3
_
1
P
_
R
P
R

1
2
3
1
8
_
_
P
P
R
R

7
3
8
1
7
_
_
P
P
R
R

5
_
2
P
3
R
1
_
P
R

9
7
4
_
0
P
1
R
_
P
R

4
_
5
P
2
R
1
_
P
R

0
2
0
1
3
_
_
P
P
R
R

4
3
9
1
2
_
_
P
P
R
R

3
8
3
_
3
P
_
R
P
R

8
7
7
_
3
P
_
R
P
R

2
3
9
1
1
_
_
P
P
R
R

4
0
0
1
2
_
_
P
P
R
R

5
2
0
1
6
_
_
P
P
R
R

0
3
2
_
2
P
1
R
_
P
R

9
2
3
_
7
P
_
R
P
R

3
3
2
_
5
P
_
R
P
R

7
3
9
_
6
P
_
R
P
R

9
1
4
_
9
P
_
R
P
R

0
2
1
_
3
P
_
R
P
R

0
6
8
_
3
P
_
R
P
R

2
5
_
P
R

3
7
5
_
9
P
_
R
P
R

2
2
0
1
0
_
1
P
_
R
P
R

4
4
1
_
P
R

9
6
_
P
R

7
1
_
P
R

0
0
1
_
P
R

5
9
_
P
R

4
4
1
_
P
R

7
1
_
P
R

4
1
8
9
3
3
x
_
_
_
P
P
P
a
R
R
R
T
_
V
S
A

x
a
T
_
V
S
A

C

30

20

Growth Stage
t
n
Indicator Taxa
u
o
C
a
x
a
T
r
o

V1

V2

V3

t
a
c
d
n

i

I

d
e
t
c
e
e
S

l

V4

R1

R4

R7

10

0

Growth Stage
Indicator Taxa

V1

V2

V3

V4

R1

R4

R7

−5

−4

−3

−2

−1

V1

V2

Indicator Taxa
Class

c__Actinobacteria

c__Bdellovibrionia

c__Gammaproteobacteria

c__NB1−j

c__SJA−28

c__Alphaproteobacteria

c__Blastocatellia

c__Gemmatimonadetes

c__Nitrososphaeria

c__Subgroup_5

c__Anaerolineae

c__Chlamydiae

c__Holophagae

c__Nitrospiria

c__Sumerlaeia

c__Babeliae

c__Bacilli

c__Chloroflexia

c__Ktedonobacteria

c__Oligoflexia

c__Thermoleophilia

c__Entotheonellia

c__MB−A2−108

c__Planctomycetes

c__Vicinamibacteria

c__Bacteroidia

c__Fimbriimonadia

c__Methylomirabilia

c__Polyangia

R4

R7

V3

V4
Indicator Taxa Growth Stage

R1

B

1.00

0.75

1.00

y
c
n
a
p
u
0.75
c
c
O

0.50

0.25

30

0.50

y
c
n
a
p
u
c
c
O

0.00

0.25

−6

t
n
u
o
C
a
x
a
T
r
o
t
a
c
d
n
I

i

−6

0.00

d
e
t
c
e
e
S

l

10

20

log10(mean relative abundance)

Indicator Taxa
Class

c__Actinobacteria

c__Bdellovibrionia

c__Gammaproteobacteria

c__NB1−j

c__SJA−28

c__Alphaproteobacteria

c__Blastocatellia

c__Gemmatimonadetes

c__Nitrososphaeria

c__Subgroup_5

c__Anaerolineae

c__Chlamydiae

c__Holophagae

c__Nitrospiria

c__Sumerlaeia

−5

−4

−3

−2

log10(mean relative abundance)

−1

c__Babeliae

c__Bacilli

c__Chloroflexia

c__Ktedonobacteria

c__Oligoflexia

c__Thermoleophilia

c__Entotheonellia

c__MB−A2−108

c__Planctomycetes

c__Vicinamibacteria

c__Bacteroidia

c__Fimbriimonadia

c__Methylomirabilia

c__Polyangia

Figure 3.7. Analysis of bacterial communities in the rhizoplane. A. Heatmap of the 57 most 
abundant ASVs (rows) in the rhizoplane samples (columns) across the Growth Stage and Time 
Series sampling schemes. Treatment bar indicates Control or Zeb treated samples. Sampling 
point bar indicates which sampling point the sample was taken at. Colored column indicates the 
most resolved taxonomic identity of the ASV. B. Abundance-occupancy distribution of ASVs in 
the rhizoplane samples. N=9,569 ASVs. Solid and dashed lines represent the fit of the neutral  

V4
Indicator Taxa Growth Stage

R1

R4

R7

V2

V3

V1

0

147 

 
 
 
 
 
 
 
 
 
Figure 3.7 (cont’d)  
model and 95% confidence interval. Colored points represent ASVs identified as indicator taxa 
in the rhizoplane at each plant growth stage. C. Indicator taxa ASVs that fall above the 
prediction of the neutral model, grouped by which plant growth stage they are associated with. 
Colors indicate the taxonomic identification for the ASV at the Class level. 

DISCUSSION 

The goal of this study was to elucidate the proportion of the influence of plant growth 

stage verses time in driving the assembly of the root microbiome over the common bean life 

cycle. There has been a great deal of work done on the assembly of the rhizoplane and 

rhizosphere over the life cycle of the plant, and many studies have documented the influence of 

plant phytohormones and chemical feedbacks on the development of the rhizosphere22,23,28,56. 

However, it has been difficult to assess the influence of plant growth stage and growing time 

separately in the root microbiome because these variables are tightly linked. This study utilized 

a growth delay caused by zebularine to study the development of the bean microbiome on two 

different sampling series in order to separate the variables of growth and time.  

We found that the rhizoplane bacterial community is highly responsive to plant growth 

stage regardless of plant age in the soil, while the rhizosphere soils are less significantly 

impacted by the plant which has been observed in previous studies27,57. Rhizoplane 

communities in the zebularine treated plants were closely aligned to the control plants based 

on growth stage, despite being harvested weeks apart at times. The beta-diversity of the 

rhizoplane communities was similar between control and zeb treated plants across the growth 

stage sampling points, and the rhizoplane growth stage ordinations were concordant in their 

structures. The influence of plant growth stage on the rhizoplane communities in our study 

accounted for 20% of the variation in the communities, while time accounted for 10%. A 

148 

 
 
 
considerable amount of the variation in the communities is unexplained by the variables in this 

study, which aligns with our observation that a large proportion of taxa in the rhizoplane were 

within the prediction of the neutral model and assembled by stochastic processes. Stochastic 

assembly processes have also been observed in other studies investigating the development of 

the plant microbiome, such as in canola, poplar and soybean(bell 2022, dove 2021, 

moroenyane 2021)18,26,58. We also identified many indicator taxa linked to growth stage in the 

common bean rhizoplane that are under selection by the plant, and found that the bacterial 

communities in the late vegetative and early reproductive stages were associated with the 

highest number of indicator taxa.  

Understanding the influence of the plant on the microbiome is not only important from 

an ecological standpoint, but also has significant implications for the growing field of 

microbiome engineering in agricultural crops59,60. Developing methods to identify plant growth-

promoting bacteria that could be applied to plants in the field is a rapidly growing area of 

research. It can be difficult to isolate beneficial bacteria from plant microbiomes, and bacteria 

used in agricultural applications are not always isolated from the crop they are being applied 

to61–63. Our study proposes a potential framework for identifying bacterial targets for microbial 

applications utilizing the Sloan neutral model. Identifying indicator taxa that are under selection 

by the plant in various species may be a promising starting point in developing crop-specific 

microbial treatments for use in the field 64–67. Additionally, since a majority of the indicator taxa 

under selection in this study were from the V1, V4 and R1 growth stages, this may be a 

potential area for further studies to investigate if there is a greater selective pressure by the 

plant at these growth stages. When biological products are being developed for use in the field, 

149 

 
 
researchers should consider the timing of the application of these microbial species, as the 

growth stage of the plant may be an important factor in maximizing the effect of the product, 

but more work is needed in this area. 

In addition to the timing of microbial applications being a potential consideration for 

plant microbiome applications, the bacterial species being applied is also important for 

effective microbial products. The indicator ASVs identified that were under selection in the 

rhizoplane were highly diverse, but contained many taxa that are known to have plant growth-

promoting properties or are commonly associated with plant microbiomes, including many 

ASVs from the classes Alphaproteobacteria, Thermoleophilia, Bacilli and 

Gammaproteobacteria68–75. These taxa that are being selected at in the rhizoplane environment 

should be investigated for potential isolation and use in microbial inocula in common bean, and 

similar experiments investigating selected taxa could be used to identify important microbiome 

members in other crops as well. 

In summary, this study confirms the significant effect of plant growth on the 

development and assembly of the rhizoplane microbiome in common bean, and reveals the 

selective influence of plant growth stage over the plant life cycle. While time does have an 

independent influence on the bacterial communities in the rhizoplane in common bean, a 

greater proportion of the rhizoplane variation was influenced by plant growth stage. 

Additionally, a large proportion of the variation in the communities was unexplained by the 

variables investigated, and the majority of the ASVs were within the prediction of the Sloan 

neutral model. Indicator taxa associated with the late vegetative and early reproductive stages 

were under selection by the plant and should be investigated for use to improve crop yields in 

150 

 
 
 
 
the field. Our results suggest that plant growth stage may be an important factor to consider 

when developing products for microbiome modification, and further work should investigate 

the selective influence of the plant over time. Overall, this works sheds light on the importance 

of plant growth stage on the development of the rhizoplane microbiome, and provides a 

framework for further identification of plant growth-promoting bacteria for agricultural 

applications.  

151 

 
 
 
 
 
REFERENCES 

1. 

2. 

3. 

4. 

5. 

6. 

7. 

8. 

9. 

10. 

11. 

12. 

Compant, S., Samad, A., Faist, H. & Sessitsch, A. A review on the plant microbiome: 
Ecology, functions, and emerging trends in microbial application. J Adv Res 19, (2019). 

Trivedi, P., Leach, J. E., Tringe, S. G., Sa, T. & Singh, B. K. Plant–microbiome interactions: 
from community assembly to plant health. Nat Rev Microbiol 18, 607–621 (2020). 

Hu, J. et al. Rhizosphere microbiome functional diversity and pathogen invasion 
resistance build up during plant development. Environ Microbiol 22, 5005–5018 (2020). 

Vurukonda, S. S. K. P., Vardharajula, S., Shrivastava, M. & SkZ, A. Enhancement of 
drought stress tolerance in crops by plant growth promoting rhizobacteria. Microbiol 
Res 184, 13–24 (2016). 

Goyal, R. K. & Habtewold, J. Z. Evaluation of Legume–Rhizobial Symbiotic Interactions 
Beyond Nitrogen Fixation That Help the Host Survival and Diversification in Hostile 
Environments. Microorganisms 11, 1454 (2023). 

Mendoza-Suárez, M. A. et al. Optimizing Rhizobium-legume symbioses by simultaneous 
measurement of rhizobial competitiveness and N 2 fixation in nodules. PNAS 117, 9822–
9831 (2020). 

Mora, Y. et al. Nitrogen-fixing rhizobial strains isolated from common bean seeds: 
Phylogeny, physiology, and genome analysis. Appl Environ Microbiol 80, 5644–5654 
(2014). 

Yadegari, M., Asadi Rahmani, H., Noormohammadi, G. & Ayneband, A. Plant growth 
promoting rhizobacteria increase growth, yield and nitrogen fixation in Phaseolus 
vulgaris. J Plant Nutr 33, 1733–1743 (2010). 

Bai, B. et al. The root microbiome: Community assembly and its contributions to plant 
fitness. J Integr Plant Biol 64, 230–243 (2022). 

Feng, H. et al. Listening to plant’s Esperanto via root exudates: reprogramming the 
functional expression of plant growth-promoting rhizobacteria. New Phytologist 239, 
2307–2319 (2023). 

Zhalnina, K. et al. Dynamic root exudate chemistry and microbial substrate preferences 
drive patterns in rhizosphere microbial community assembly. Nat Microbiol 3, 470–480 
(2018). 

Fitzpatrick, C. R. et al. Assembly and ecological function of the root microbiome across 
angiosperm plant species. PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES 115, 
E1157–E1165 (2018). 

152 

 
 
13. 

14. 

15. 

16. 

17. 

18. 

Xiong, C. et al. Host selection shapes crop microbiome assembly and network 
complexity. New Phytologist 229, 1091–1104 (2021). 

Zhang, Y. et al. Huanglongbing impairs the rhizosphere-to-rhizoplane enrichment 
process of the citrus root-associated microbiome. Microbiome 5, 97 (2017). 

Cui, Y. et al. Diversity patterns of the rhizosphere and bulk soil microbial communities 
along an altitudinal gradient in an alpine ecosystem of the eastern Tibetan Plateau. 
Geoderma 338, 118–127 (2019). 

Cordovez, V. et al. Successive plant growth amplifies genotype-specific assembly of the 
tomato rhizosphere microbiome. Science of the Total Environment 772, (2021). 

Cavaglieri, L., Orlando, J. & Etcheverry, M. Rhizosphere microbial community structure 
at different maize plant growth stages and root locations. Microbiol Res 164, 391–399 
(2009). 

Bell, J. K., Mamet, S. D., Helgason, B. & Siciliano, S. D. Brassica napus Bacterial Assembly 
Processes Vary with Plant Compartment and Growth Stage but Not between Lines. Appl 
Environ Microbiol 88, (2022). 

19.  Wei, G. et al. Temporal Dynamics of Rhizosphere Communities Across the Life Cycle of 

Panax notoginseng. Front Microbiol 13, (2022). 

20. 

21. 

22. 

23. 

24. 

Gu, Y. et al. Small changes in rhizosphere microbiome composition predict disease 
outcomes earlier than pathogen density variations. ISME Journal 16, 2448–2456 (2022). 

Liu, D. & Howell, K. Community succession of the grapevine fungal microbiome in the 
annual growth cycle. Environ Microbiol 23, 1842–1857 (2021). 

Steinauer, K. et al. Root exudates and rhizosphere microbiomes jointly determine 
temporal shifts in plant-soil feedbacks. Plant Cell Environ 46, 1885–1899 (2023). 

Hu, L. et al. Root exudate metabolites drive plant-soil feedbacks on growth and defense 
by shaping the rhizosphere microbiota. Nat Commun 9, (2018). 

Ratu, S. T. N. et al. Rhizobia use a pathogenic-like effector to hijack leguminous 
nodulation signalling. Sci Rep 11, (2021). 

25.  Moroenyane, I., Tremblay, J. & Yergeau, É. Temporal and spatial interactions modulate 

the soybean microbiome. FEMS Microbiol Ecol 97, (2021). 

26. 

Dove, N. C. et al. Assembly of the Populus Microbiome Is Temporally Dynamic and 
Determined by Selective and Stochastic Factors. mSphere 6, e01316-20 (2021). 

153 

 
 
27. 

28. 

29. 

Xiong, C. et al. Plant developmental stage drives the differentiation in ecological role of 
the maize microbiome. Microbiome 9, (2021). 

Vetterlein, D. et al. Rhizosphere Spatiotemporal Organization–A Key to Rhizosphere 
Functions. Frontiers in Agronomy 2, (2020). 

Barnett, S. E., Youngblut, N. D. & Buckley, D. H. Soil characteristics and land-use drive 
bacterial community assembly patterns. FEMS Microbiol Ecol 96, (2020). 

30.  Malik, A. A. & Bouskill, N. J. Drought impacts on microbial trait distribution and feedback 

to soil carbon cycling. Funct Ecol 36, 1442–1456 (2022). 

31. 

32. 

33. 

34. 

35. 

36. 

37. 

38. 

39. 

Diaz-Garza, A. M. et al. Temporal Dynamics of Rhizobacteria Found in Pequin Pepper, 
Soybean, and Orange Trees Growing in a Semi-arid Ecosystem. Front Sustain Food Syst 4, 
(2020). 

Fox, A., Widmer, F. & Lüscher, A. Soil microbial community structures are shaped by 
agricultural systems revealing little temporal variation. Environ Res 214, (2022). 

Griffin, P. T., Niederhuth, C. E. & Schmitz, R. J. A comparative analysis of 5-azacytidine-
and zebularine-induced DNA demethylation. G3: Genes, Genomes, Genetics 6, 2773–
2780 (2016). 

Kelly, J. D. et al. Registration of ‘Red Hawk’ Dark Red Kidney Bean. Crop Sci 38, 280–281 
(1998). 

Hoagland, D. R. & Arnon, D. I. The Water-Culture Method for Growing Plants without 
Soil. vol. C347 rev 1950 (College of Agriculture, University of California, Berkeley, 
California, 1950). 

Schwartz, H. F. & Langham, M. A. C. Common Bean Growth Stages. Bean IPM 
https://beanipm.pbgworks.org/common-bean (2023). 

Colovas, J., Bintarti, A. F., Mechan Llontop, M. E., Grady, K. L. & Shade, A. Do-it-Yourself 
Mock Community Standard for Multi-Step Assessment of Microbiome Protocols. Curr 
Protoc 2, (2022). 

Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K. & Schloss, P. D. Development 
of a dual-index sequencing strategy and curation pipeline for analyzing amplicon 
sequence data on the miseq illumina sequencing platform. Appl Environ Microbiol 79, 
5112–5120 (2013). 

Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data 
science using QIIME 2. Nature Biotechnology vol. 37 852–857 Preprint at 
https://doi.org/10.1038/s41587-019-0209-9 (2019). 

154 

 
 
40. 

Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon 
data. Nat Methods 13, 581–583 (2016). 

41.  Weinstein, M. M., Prem, A., Jin, M., Tang, S. & Bhasin, J. M. FIGARO: An efficient and 

objective tool for optimizing microbiome rRNA gene trimming parameters. bioRxiv 
(2019) doi:10.1101/610394. 

42. 

43. 

Quast, C. et al. The SILVA ribosomal RNA gene database project: Improved data 
processing and web-based tools. Nucleic Acids Res 41, (2013). 

R Core Team. R: A Language and Environment for Statistical Computing. Preprint at 
(2023). 

44.  McMurdie, P. J. & Holmes, S. Phyloseq: An R Package for Reproducible Interactive 

Analysis and Graphics of Microbiome Census Data. PLoS One 8, (2013). 

45. 

Davis, N. M., Proctor, Di. M., Holmes, S. P., Relman, D. A. & Callahan, B. J. Simple 
statistical identification and removal of contaminant sequences in marker-gene and 
metagenomics data. Microbiome 6, (2018). 

46. 

Oksanen, J. et al. vegan: Community Ecology Package. R package version 2.5-7. 
https://CRAN.R-project.org/package=vegan (2019). 

47.  Wickham, H. Ggplot2: Elegant Graphics for Data Analysis. (Springer-Verlag, New York, 

2016). 

48.  Martinez Arbizu, P. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis. 

(2017). 

49. 

50. 

Barnett, D., Arts, I. & Penders, J. microViz: an R package for microbiome data 
visualization and statistics. J Open Source Softw 6, 3201 (2021). 

Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in 
multidimensional genomic data. Bioinformatics 32, 2847–2849 (2016). 

51. 

Gu, Z. Complex heatmap visualization. iMeta 1, (2022). 

52. 

53. 

54. 

De Cáceres, M., Jansen, F. & Dell, N. indicspecies: Relationship Between Species and 
Groups of Sites. https://cran.r-project.org/web/packages/indicspecies/index.html 
(2023). 

Sprockett, D. reltools: Microbiome Amplicon Analysis and Visualization. R package 
version 0.1.0. https://rdrr.io/github/DanielSprockett/reltools/ (2024). 

Burns, A. R. et al. Contribution of neutral processes to the assembly of gut microbial 
communities in the zebrafish over host development. ISME Journal 10, 655–664 (2016). 

155 

 
 
55. 

56. 

57. 

Shade, A. & Stopnisek, N. Abundance-occupancy distributions to prioritize plant core 
microbiome membership. Curr Opin Microbiol 49, 50–58 (2019). 

Bowsher, A. W., Kearns, P. J. & Shade, A. 16S rRNA/rRNA Gene Ratios and Cell Activity 
Staining Reveal Consistent Patterns of Microbial Activity in Plant-Associated Soil. 
mSystems 4, (2019). 

Stopnisek, N. & Shade, A. Persistent microbiome members in the common bean 
rhizosphere: an integrated analysis of space, time, and plant genotype. ISME Journal 15, 
2708–2722 (2021). 

58.  Moroenyane, I., Mendes, & L., Tremblay, J., Tripathi, B. & Yergeau, & É. Plant 

Compartments and Developmental Stages Modulate the Balance between Niche-Based 
and Neutral Processes in Soybean Microbiome. Microb Ecol 82, 416–428 (2021). 

59. 

Arif, I., Batool, M. & Schenk, P. M. Plant Microbiome Engineering: Expected Benefits for 
Improved Crop Growth and Resilience. Trends Biotechnol 38, 1385–1396 (2020). 

60.  Mburu, S. W., Koskey, G., Njeru, E. M. & Maingi, J. M. Revitalization of bacterial 

endophytes and rhizobacteria for nutrients bioavailability in degraded soils to promote 
crop production. AIMS Agriculture and Food 6, 496–524 (2021). 

61. 

62. 

63. 

64. 

Eltlbany, N. et al. Enhanced tomato plant growth in soil under reduced P supply through 
microbial inoculants and microbiome shifts. FEMS Microbiol Ecol 95, (2019). 

Talaat, N. B. Effective microorganisms: An innovative tool for inducing common bean 
(Phaseolus vulgaris L.) salt-tolerance by regulating photosynthetic rate and endogenous 
phytohormones production. Sci Hortic 250, 254–265 (2019). 

Yurgel, S. N., Ajeethan, N. & Smertenko, A. Response of Plant-Associated Microbiome to 
Plant Root Colonization by Exogenous Bacterial Endophyte in Perennial Crops. Front 
Microbiol 13, (2022). 

Bziuk, N., Maccario, L., Sørensen, S. J., Schikora, A. & Smalla, K. Barley Rhizosphere 
Microbiome Transplantation – A Strategy to Decrease Susceptibility of Barley Grown in 
Soils With Low Microbial Diversity to Powdery Mildew. Front Microbiol 13, (2022). 

65. 

Timmusk, S., Behers, L., Muthoni, J., Muraya, A. & Aronsson, A. C. Perspectives and 
challenges of microbial application for crop improvement. Front Plant Sci 8, (2017). 

66.  Wang, H. et al. A promising product: Abscisic acid-producing bacterial agents for 

restricting cadmium enrichment in field vegetable crops. Food Chem X 19, (2023). 

67.  Mitter, B. et al. A new approach to modify plant microbiomes and traits by introducing 
beneficial bacteria at flowering into progeny seeds. Front Microbiol 8, (2017). 

156 

 
 
68. 

69. 

70. 

Boss, B. L., Wanees, A. E., Zaslow, S. J., Normile, T. G. & Izquierdo, J. A. Comparative 
genomics of the plant-growth promoting bacterium Sphingobium sp. strain AEW4 
isolated from the rhizosphere of the beachgrass Ammophila breviligulata. BMC 
Genomics 23, (2022). 

Haack, F. S. et al. Molecular keys to the Janthinobacterium and Duganella spp. 
Interaction with the plant pathogen Fusarium graminearum. Front Microbiol 7, (2016). 

Li, M. et al. Composition and function of rhizosphere microbiome of Panax notoginseng 
with discrepant yields. Chinese Medicine (United Kingdom) 15, (2020). 

71.  Moore, E. R. et al. Microbial Drivers of Plant Performance during Drought Depend upon 
Community Composition and the Greater Soil Environment. Microbiol Spectr 11, (2023). 

72. 

73. 

74. 

75. 

Tan, L. et al. Responses of microbial communities and interaction networks to different 
management practices in tea plantation soils. Sustainability 11, (2019). 

Tsotetsi, T., Nephali, L., Malebe, M. & Tugizimana, F. Bacillus for Plant Growth 
Promotion and Stress Resilience: What Have We Learned? Plants 11, (2022). 

Yuan, X. et al. Effects of nitrogen reduction combined with bio-organic fertilizer on soil 
bacterial community diversity of red raspberry orchard. PLoS One 18, (2023). 

Li, P. et al. Fossil-fuel-dependent scenarios could lead to a significant decline of global 
plant-beneficial bacteria abundance in soils by 2100. Nat Food 4, 996–1006 (2023). 

157 

 
 
  
 
 
APPENDIX C: 
CHAPTER 4 SUPPLEMENTARY MATERIALS 

Figure S3.1. Experimental Design. Seeds were treated with and without 5mM zebularine, then 
grown over two sampling schemes based on plant growth stage or days since planting. 

158 

 
 
 
 
 
 
  
 
 
Growth Stage

Time Series

V4

Day 35

Control

Zeb

Figure S3.2. Comparison of plant above-ground growth in control (top) and zebularine treated 
(bottom) plants in the Growth Stage series at the V4 stage, and in the Time Series 35 days after 
planting. 

159 

 
 
 
 
 
Growth Stage

Time Series

Control

Zeb

Control

Zeb

V2

V4

R4

Day 3

Day 21

Day 49

Figure S3.3. Comparison of plant root growth in control and zebularine treated plants in the 
Growth Stage series (left) at the V2, V4 and R4 stages, and in the Time Series (right) 3, 21 and 
49 days after planting. 

160 

 
 
 
 
2000

i

s
e
c
e
p
S

1000

0

0

50000

100000

150000

Sample

Figure S3.4. Rarefaction curves of combined rhizosphere and rhizoplane datasets after 
sequence decontamination. Samples were rarefied to 22,761 reads (red, vertical line). 9 
samples were removed that did not meet the minimum DNA read threshold.  

161 

 
 
 
 
 
 
Figure S3.5. Beta-dispersion, the distance to the spatial median of sample groups, in Bray-Curtis 
PCoA ordinations. A. Rhizosphere Growth Stage, B. Rhizosphere Time Series, C. Rhizoplane 
Growth Stage, D. Rhizoplane Time Series. There are no significant differences in dispersion 
between the Control and Zeb treatments at any sampling point in either root compartment. 

162 

 
 
 
 
 
 
 
Table S3.1  
Bray-Curtis PERMANOVA and pairwise results 
F 

SumOfSqs 

R2 

Df 

Pr(>F) 

PERMANOVA: Rhizosphere, Control Growth vs Control Time 

group 

1 

0.158 

0.01909 

1.2846 

0.0697 

PERMANOVA: Rhizosphere, Zeb Growth vs Zeb Time 

group 

1 

0.1491 

0.01917 

1.2897 

0.038  * 

PERMANOVA: Rhizoplane, Control Growth vs Control Time 

group 

1 

0.02354 

1.4944 

0.1142 

PERMANOVA: Rhizoplane, Zeb Growth vs Zeb Time 

group 

1 

0.383 

0.03051 

2.1087 

0.0188  * 

PERMANOVA: Rhizoplane, Growth Stage series, Control vs Zeb 

treatment 

stage 

treatment:stage 

1 

6 

6 

0.4388 

2.7742 

1.2538 

0.03567 

0.22551 

0.10192 

2.9122 

3.0686 

1.3868 

0.0015  ** 

0.0001  *** 

0.0128  * 

Pairwise test: Rhizoplane, Control vs. Zeb at growth stages 

Zeb_growth_V1_vs_Control_growth_V1 

group_stage 

1 

0.20092 

0.18031 

1.7598 

0.0161  * 

Zeb_growth_V2_vs_Control_growth_V2 

group_stage 

1 

0.25737 

0.17024 

1.6413 

0.0172  * 

Control_growth_V3_vs_Zeb_growth_V3 

group_stage 

1 

0.1877 

0.17949 

1.3125 

0.3031 

Control_growth_V4_vs_Zeb_growth_V4 

group_stage 

1 

0.18781 

0.12141 

1.1055 

0.3406 

Control_growth_R1_vs_Zeb_growth_R1 

group_stage 

1 

0.19062 

0.198 

1.7282 

0.0401  * 

Zeb_growth_R4_vs_Control_growth_R4 

group_stage 

1 

0.34333 

0.19559 

1.9452 

0.0159  * 

163 

 
 
  
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table S3.1 (cont’d) 
Zeb_growth_R7_vs_Control_growth_R7 

group_stage 

1 

0.32125 

0.2024 

1.7763 

0.0726 

PERMANOVA: Rhizoplane, Time series, Control vs Zeb 

treatment 

stage 

treatment:stage 

1 

6 

6 

0.4435 

3.7329 

1.8065 

0.03382 

0.28468 

0.13777 

3.2968 

4.6248 

2.2382 

0.0011  ** 

0.0001  *** 

0.0001  *** 

Pairwise test: Rhizoplane, Control vs. Zeb at days since planting 

Zeb_time_day3_vs_Control_time_day3 

group_stage 

1 

0.20158 

0.15675 

1.4871 

0.0148  * 

Zeb_time_day7_vs_Control_time_day7 

group_stage 

1 

0.24231 

0.17966 

1.7521 

0.0391  * 

Control_time_day14_vs_Zeb_time_day14 

group_stage 

1 

0.24532 

0.24769 

2.3047 

0.0162  * 

Zeb_time_day21_vs_Control_time_day21 

group_stage 

1 

0.21636 

0.17412 

1.4758 

0.0318  * 

Control_time_day35_vs_Zeb_time_day35 

group_stage 

1 

0.43207 

0.35432 

3.8413 

0.015  * 

Control_time_day49_vs_Zeb_time_day49 

group_stage 

1 

0.57954 

0.32921 

3.9261 

0.0085  ** 

Zeb_time_day63_vs_Control_time_day63 

group_stage 

1 

0.32498 

0.21318 

2.1675 

0.0081  ** 

PERMANOVA: Rhizosphere, Growth Stage series, Control vs Zeb 

treatment 

stage 

treatment:stage 

1 

6 

6 

0.2474 

1.1804 

0.9847 

0.03035 

0.14481 

0.1208 

2.285 

1.8168 

1.5157 

1.00E-04  *** 

1.00E-04  *** 

1.00E-04  *** 

Pairwise test: Rhizosphere, Control vs. Zeb at growth stages 

Zeb_growth_V1_vs_Control_growth_V1 

group_stage 

1 

0.13643 

0.16989 

1.2279 

0.1115 

164 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
Table S3.1 (cont’d) 
Control_growth_V2_vs_Zeb_growth_V2 

group_stage 

1 

0.12958 

0.13639 

1.1055 

0.2226 

Control_growth_V3_vs_Zeb_growth_V3 

group_stage 

1 

0.20019 

0.18365 

1.7998 

0.0092  ** 

Control_growth_V4_vs_Zeb_growth_V4 

group_stage 

1 

0.16125 

0.16639 

1.5968 

0.0077  ** 

Control_growth_R1_vs_Zeb_growth_R1 

group_stage 

1 

0.16426 

0.17631 

1.7124 

0.0067  ** 

Zeb_growth_R4_vs_Control_growth_R4 

group_stage 

1 

0.17196 

0.15186 

1.4324 

0.018  * 

Zeb_growth_R7_vs_Control_growth_R7 

group_stage 

1 

0.26716 

0.24428 

2.5859 

0.0084  ** 

PERMANOVA: Rhizosphere, Time series, Control vs Zeb 

treatment 

stage 

treatment:stage 

1 

6 

6 

0.1932 

1.095 

0.8122 

0.02404 

0.13619 

0.10102 

1.7894 

1.6899 

1.2535 

4.00E-04  *** 

1.00E-04  *** 

4.00E-04  *** 

Pairwise test: Rhizosphere, Control vs. Zeb at days since planting 

Zeb_time_day3_vs_Control_time_day3 

group_stage 

1 

0.10228 

0.11854 

1.0758 

0.324 

Zeb_time_day7_vs_Control_time_day7 

group_stage 

1 

0.08906 

0.10415 

0.93 

0.7113 

Control_time_day14_vs_Zeb_time_day14 

group_stage 

1 

0.12559 

0.13481 

1.2466 

0.0298  * 

Control_time_day21_vs_Zeb_time_day21 

group_stage 

1 

0.15609 

0.14827 

1.3926 

0.0069  ** 

Control_time_day35_vs_Zeb_time_day35 

group_stage 

1 

0.16751 

0.16108 

1.5361 

0.0083  ** 

165 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
 
 
 
 
 
Table S3.1 (cont’d) 
Control_time_day49_vs_Zeb_time_day49 

group_stage 

1 

0.16252 

0.1627 

1.3602 

0.0423  * 

Control_time_day63_vs_Zeb_time_day63 

group_stage 

1 

0.19904 

0.16579 

1.5899 

0.0086  ** 

Table S3.2 Supplemental File: Table of selected indicator taxa  

Table S3.3 Supplemental File: Metadata table of DNA extractions 

166 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
CHAPTER 5: 

Conclusions and future directions 

167 

 
 
 
 
 
 
 
 
This work contributes important knowledge to the area of plant microbiome research, 

particularly in understanding the microbiome of common bean plants and how they are 

impacted by environmental stress. Changing climate conditions will continue to pose a threat to 

agricultural crops through impacts such as warming, drought, and extreme weather events, and 

improving the resilience of plants to these impacts is becoming increasingly important1,2. This is 

especially true for staple food crops like common bean, as many people depend on beans for 

their livelihood and survival in areas that are at particular risk for climate change impacts3,4. 

Plant-microbe interactions are also at risk of the negative effects of climate change, and it is 

important to understand how these host-microbe associations are changing to allow 

researchers to better support crops into the future5–7. 

Throughout this research, we examined the impacts of repeated drought exposure on 

common been root and rhizosphere communities, the vertical transmission of seed 

microbiomes under abiotic treatment, and the development of the rhizosphere communities 

over the bean life cycle. In chapter 2, the Red Hawk plants grown in different soils and the 

different plant genotypes were found to have significantly different root and rhizosphere 

microbiomes in alpha and beta diversity measures. The drought stress treatments had little 

effect on the bacterial communities of the Red Hawk plants, but Flavert plant microbiomes 

were significantly impacted by the drought and showed evidence of legacy effects of 

multigenerational exposure. These results suggest that plant genotype may play a role in 

responses to environmental stress, and the microbiome of the Red Hawk variety may be stable 

under drought stress. Microbiome modification to support Red Hawk plants under drought may 

not be effective, and other solutions to climate challenges should be investigated.   

168 

 
 
 
In chapter 3, we investigated the vertical transmission of bacterial taxa through the Red 

Hawk bean seeds over two generations of abiotic treatment with drought stress or increased 

nutrient concentration. The abiotic treatments had little effect on the seed microbiome, with 

the only evidence of a treatment effect in the beta diversity of second-generation seeds after 

the Generation 1 parent had been exposed to drought stress.  However, a stable signature of 

the seed microbiome was observed over multiple generations, with 22 bacterial ASVs identified 

as highly prevalent in the seeds of all three generations, despite differences in soil and 

treatment conditions. The transmitted seed endophytes in our study also contained known 

bacterial taxa that are part of the common bean core microbiome. In summary, we identified a 

stable and consistent seed microbiome in Red Hawk seeds with strong evidence of vertical 

transmission, and these seed taxa should be investigated further to understand their role in 

plant health.  

Lastly, in chapter 4, we explored the separate influences of plant growth stage and time 

on the assembly of the rhizosphere and rhizoplane microbiome in common bean Red Hawk 

plants. Through an innovative experimental design using zebularine treatment to cause a 

growth delay in the Red Hawk plants, we discovered that plant growth stage is the most 

significant driver of the rhizoplane community assembly. The control and zebularine treated 

plants had similar rhizoplane bacterial communities despite being harvested on different dates, 

while the rhizosphere soils were less directly affected by the plant. Plant growth stage 

accounted for 20% of the variation in the rhizoplane communities, while time accounted for 

10%. A large proportion of the bacterial community variation was unexplained in the dataset, 

and investigating the communities with the Sloan neutral model8 revealed that much of the 

169 

 
 
bacterial communities are controlled by neutral or stochastic processes in this study. Indicator 

taxa associated with the plant across growth stages were identified and were found to be under 

selection by the plant environment in the rhizoplane. These ASVs should be investigated further 

as potential targets for beneficial bacterial applications in common bean. 

Future work based on this research could include many different avenues of 

investigation. Each of these experiments identified potentially beneficial plant microbiome 

members that should be investigated further and uncovered potential frameworks for further 

identification of important microbiome members for use in agricultural applications. One area 

that could be investigated is the consistently transmitted ASVs identified in the seed 

microbiome study. These taxa are consistently associated with the seeds over multiple 

generations despite differences in soil and abiotic treatment. The taxa identified in the drought 

and rhizosphere assembly studies should be compared to the prevalent ASVs in the seed study. 

This may elucidate the role of these ASVs in other plant microbiome compartments and their 

potential benefits to the plant under varying environmental conditions. Further, the ASVs 

identified in this work, both in the seed microbiome and the indicator taxa that were under 

selection in the rhizoplane, could be targeted for isolation from common bean seeds in order to 

assess their potential benefits in vivo9. Isolating plant microbiome members has proved 

challenging in past work, but having specific targets for isolation may help create new methods 

to isolate these taxa from plant tissues10–12. Once bacteria are isolated from the plant, this 

opens many doors for plant growth studies to elucidate which isolates are most beneficial and 

what their mechanisms are in supporting the plant. There has been work done utilizing 

microbial inoculants to improve plant performance in controlled growth studies with promising 

170 

 
 
results and it would be particularly interesting to apply the prevalent seed ASVs and rhizoplane 

indicator taxa to plants to study their effects on plant growth13–15. 

Additionally, this work provides a framework for understanding how the pant root 

microbiome is influenced by the plant over time. Analyzing plant microbiomes over the life 

cycle of the plant has elucidated a clear impact of growth stage on the development of the 

rhizoplane in common bean. Utilizing the neutral model to assess the selection of indicator taxa 

by the plant environment has also further advanced our knowledge of the assembly processes 

in the rhizoplane ecosystem. Future studies could investigate the selective influence of the 

plant over time at each growth stage, verses broadly as we investigated here, which could 

provide insight into potential microbial applications in the field. Knowing when beneficial 

microbes are most important for the plant over its growth stages, and how susceptible the root 

microbiome is to microbial invasion at these stages, are important knowledge gaps that should 

be investigated to advance the field of microbiome engineering16–19. This experimental design 

could be applied in various agricultural crops in order to identify key bacterial taxa and growth 

stages in the plant microbiome. 16–19 

Lastly, it is important to consider the applicability of these small-scale studies to plants 

growing in the field. Varying soil types and geographic locations, environmental and disease 

conditions, and management practices can all have an impact on the plant microbiome. It is 

important when studying the plant microbiome to consider these variables that could affect the 

plant in real-world conditions20 . Our drought study partially addresses these challenges by 

studying the effects of treatment in different plant genotypes and soils. As researchers continue 

to develop microbial strategies for supporting crops in the field, it will become increasingly 

171 

 
 
 
important to understand how different plant varieties respond to microbial treatments and 

how environmental factors can alter the effectiveness of these products. Additionally, it is 

important to move research from controlled conditions to the field at larger scales with greater 

replication, in order to understand if the same patterns are evident in both environments. 

Overall, this work provides important knowledge to the field of plant microbiome 

research, particularly in the potential of the plant microbiome to support plants under stress 

conditions, the stable transmission of seed microbiome taxa, and the assembly of the 

rhizoplane over the life of the plant. Understanding this basic science of the common bean 

microbiome is important to move the field forward and will help researchers find sustainable 

solutions for the future of agriculture. The need to support agricultural crops under changing 

climate conditions will become increasingly urgent as the global population continues to rise, 

and plant microbiome engineering may hold the key to improving the resilience of agricultural 

systems21. This research on the common bean microbiome will provide tools and targets for 

future studies in beneficial bacteria for sustainable agricultural solutions. 

172 

 
 
 
 
 
REFERENCES 

1. 

2. 

3. 

4. 

5. 

6. 

7. 

8. 

9. 

10. 

11. 

Gourdji, S., Läderach, P., Valle, A. M., Martinez, C. Z. & Lobell, D. B. Historical climate 
trends, deforestation, and maize and bean yields in Nicaragua. Agric For Meteorol 200, 
270–281 (2015). 

Maity, A. et al. Climate Change Impacts on Seed Production and Quality: Current 
Knowledge, Implications, and Mitigation Strategies. Seed Science and Technology 51, 
65–96 (2023). 

Uebersax, M. A. et al. Dry beans (Phaseolus vulgaris L.) as a vital component of 
sustainable agriculture and food security—A review. Legume Science 5, (2023). 

Shrestha, R. P. & Nepal, N. An assessment by subsistence farmers of the risks to food 
security attributable to climate change in Makwanpur, Nepal. Food Secur 8, 415–425 
(2016). 

Cheng, Y. T., Zhang, L. & He, S. Y. Plant-Microbe Interactions Facing Environmental 
Challenge. Cell Host Microbe 26, 183–192 (2019). 

Jansson, J. K. & Hofmockel, K. S. Soil microbiomes and climate change. Nat Rev Microbiol 
18, 35–46 (2020). 

Li, P. et al. Fossil-fuel-dependent scenarios could lead to a significant decline of global 
plant-beneficial bacteria abundance in soils by 2100. Nat Food 4, 996–1006 (2023). 

Burns, A. R. et al. Contribution of neutral processes to the assembly of gut microbial 
communities in the zebrafish over host development. ISME Journal 10, 655–664 (2016). 

Ulrich, K., Kube, M., Becker, R., Schneck, V. & Ulrich, A. Genomic Analysis of the 
Endophytic Stenotrophomonas Strain 169 Reveals Features Related to Plant-Growth 
Promotion and Stress Tolerance. Front Microbiol 12, (2021). 

Bergna, A. et al. Tomato seeds preferably transmit plant beneficial endophytes. 
Phytobiomes J 2, 183–193 (2018). 

Torres-Cortés, G. et al. Functional microbial features driving community assembly during 
seed germination and emergence. Front Plant Sci 9, (2018). 

12.  Mukherjee, A., Singh, B. K. & Verma, J. P. Harnessing chickpea (Cicer arietinum L.) seed 
endophytes for enhancing plant growth attributes and bio-controlling against Fusarium 
sp. Microbiol Res 237, (2020). 

13. 

Eltlbany, N. et al. Enhanced tomato plant growth in soil under reduced P supply through 
microbial inoculants and microbiome shifts. FEMS Microbiol Ecol 95, (2019). 

173 

 
 
14. 

15. 

16. 

Elkoca, E., Turan, M. & Donmez, M. F. Effects of single, dual and triple inoculations with 
bacillus subtilis, bacillus megaterium and rhizobium leguminosarum bv. phaseoli on 
nodulation, nutrient uptake, yield and yield parameters of common bean (phaseolus 
vulgaris l. cv. ’Elkoca-05’). J Plant Nutr 33, 2104–2119 (2010). 

Igiehon, O. N. & Babalola, O. O. Rhizobium and Mycorrhizal Fungal Species Improved 
Soybean Yield Under Drought Stress Conditions. Curr Microbiol 78, 1615–1627 (2021). 

O’Callaghan, M. Microbial inoculation of seed for improved crop performance: issues 
and opportunities. Appl Microbiol Biotechnol 100, 5729–5746 (2016). 

17.  Mitter, E. K., Tosi, M., Obregón, D., Dunfield, K. E. & Germida, J. J. Rethinking Crop 

Nutrition in Times of Modern Microbiology: Innovative Biofertilizer Technologies. Front 
Sustain Food Syst 5, (2021). 

18. 

19. 

20. 

Timmusk, S., Behers, L., Muthoni, J., Muraya, A. & Aronsson, A. C. Perspectives and 
challenges of microbial application for crop improvement. Front Plant Sci 8, (2017). 

Berg, G., Kusstatscher, P., Abdelfattah, A., Cernava, T. & Smalla, K. Microbiome 
Modulation—Toward a Better Understanding of Plant Microbiome Response to 
Microbial Inoculants. Front Microbiol 12, (2021). 

Poorter, H. et al. Pampered inside, pestered outside? Differences and similarities 
between plants growing in controlled conditions and in the field. New Phytologist vol. 
212 838–855 Preprint at https://doi.org/10.1111/nph.14243 (2016). 

21. 

Arif, I., Batool, M. & Schenk, P. M. Plant Microbiome Engineering: Expected Benefits for 
Improved Crop Growth and Resilience. Trends Biotechnol 38, 1385–1396 (2020). 

174