About

Devanshi Khokhani is an Assistant Professor in the Department of Plant Pathology at the University of Minnesota. Her research focuses on exploring beneficial and pathogenic plant-microbe interactions to provide sustainable agricultural solutions. She is particularly interested in understanding the virulence mechanisms of bacterial plant pathogens and multitrophic interactions involving nitrogen-fixing bacteria, mycorrhizal fungi, and crop plants. Her work aims to deepen the understanding of these complex interactions to support sustainable crop production and plant health.

Research topics

• Botany
• Cell biology
• Genetics
• Biology
• Biochemistry

Selected publications

• First report of bacterial blight of <i>Camelina sativa</i> caused by <i>Xanthomonas campestris</i> in Minnesota

  Plant Disease · 2026-05-20

  articleSenior author

  Camelina sativa is a short-season Eurasian annual cultivated in Minnesota, USA, as a winter intermediate oilseed crop for sustainable aviation fuel. In June and July of 2025, bacterial blight was observed in breeding trials grown at four MN locations: St. Paul (STP), Rosemont (RMT), Becker (BEC), and Grand Rapids (GRA). The symptoms were initially observed as water-soaked and dark spots on the lower leaves, then turned black with a yellow margin. As the disease progressed, dark lesions began to appear on the stems and seed pods. The severely infected pods were deformed, shriveled, and died prematurely (Fig. S1A, B). Disease incidence ranged from 0 to 30% across the locations. Symptomatic samples from each site were tested positive on the Xanthomonas ImmunoStrip® test (Agdia, Inc), and yellow, viscous colonies with entire margins were isolated on yeast-dextrose-carbonate medium (YDC) from the supernatant of surface sterilized leaf and stem homogenate (Basnet et al. 2024) (Fig. S1.E). A minimum of two isolates were collected from individual plants at each location (2 from STP, 3 from RMT, 4 from BEC, 3 from GRA). To confirm Koch’s postulates, 3-week-old vernalized C. sativa ‘Joelle’ seedlings, grown in a growth chamber (16-h photoperiod, 21°C day/18°C night; n = 9 plants per isolate), were inoculated by piercing the leaf midvein with sterile toothpicks dipped into bacterial suspension (OD600 = 0.1, ~1x108 CFUs/mL) or sterile phosphate buffer saline (mock-inoculated control). Plants were bagged for 48 h to maintain humidity. Two isolates per location and two assay replicates were used for infection assays. At 3 dpi, water-soaked lesions developed in the infected plants, while controls remained healthy (Fig. S1D). By 4 dpi, chlorosis and necrosis appeared; controls showed no symptoms. At 5 dpi, the pathogen was re-isolated. DNA was extracted using DNeasy UltraClean Microbial Kit (Qiagen Inc., Germantown, MD), and a section of the 16S rRNA subunit was sequenced for preliminary genus confirmation using the universal primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-TACGGYTACCTTGTTACGACTT-3′). All the sequenced isolates shared &gt;94% identity with Xanthomonas when compared using NCBI BLASTn search. Hybrid Oxford Nanopore Technologies/Illumina sequencing was performed on one isolate per origin (STP XUMN-3, RMT XUMN-4, BEC XUMN-5, GRA: XUMN-6); all four had CheckM values ≥99.6% complete with 0.0% contamination. FastANI analysis (Jain et al., 2018) using complete whole genome sequences of representative Xanthomonas campestris (Xc) strains (Dubrow et al. 2022) and contigs from UW-Madison isolates sequenced as part of their recent first report (Basnet et al. 2024) revealed that these isolates are closely related to those reported to have caused bacterial blight in field pennycress in 2024 in Minnesota (Chiu et al., 2025), as FastANI estimates between them are &gt;98.3%. Additionally, FastANI estimates show &gt;99.0% similarity among camelina isolates and &lt;98.5% similarity to pennycress isolates. While this pattern may suggest a clonal relationship among camelina isolates, this interpretation is based solely on ANI and should be treated as tentative. Camelina has attracted strong interest in the Upper Midwest and the EU for its role in decarbonization and ecosystem services (Berti et al. 2025; Buffi et al. 2026), highlighting the need to monitor emerging pathogens, such as X. campestris, to anticipate economic risks and inform disease management strategies.

  PublisherDOI

• Ecogenomic Diversity of <i>Clavibacter nebraskensis</i> in North America

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-02-12

  articleOpen access

  Abstract Goss’s wilt and leaf blight of maize is caused by Clavibacter nebraskensis and has reemerged as an important disease in North America. Despite its epidemiological relevance, this species remains poorly characterized in terms of population structure, functional diversity, and ecological differentiation, particularly among strains reported from Mexico. In this study, long-read whole-genome sequencing and phenotypic assays were used to characterize genomic diversity, virulence, and fitness-associated traits in C. nebraskensis . We generated 24 long-read genomes, including 20 contemporary Mexican isolates and four historical United States strains collected between 1969 and 1996, and compared them with publicly available genomes from North America and South Africa. Phylogenomic analyses confirmed that all strains cluster within the C. nebraskensis clade, and gene accumulation curves supported a closed pangenome with accessory gene variation linked to geographic origin and isolation period. Functional assays showed strain-level variation in virulence, enzymatic activity, bacteriocin antagonism, polysaccharide production, biofilm formation, and pigmentation. Cellulolytic activity was associated with disease severity, whereas pigment-related traits were linked to thiamine metabolism. Overall, these results indicate that C. nebraskensis comprises ecologically heterogeneous populations, structured around alternative survival and competition strategies. Integrating genome-wide comparisons with functional characterization of fitness-related traits provides a framework for understanding the biological factors underlying Goss’s wilt. dynamics

  PublisherDOI

• Comparative Analysis of Soil Microbial Communities in High-Tunnel and Field Agricultural Systems

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-02-07

  articleSenior authorCorresponding

  Abstract High tunnels and open-field systems differ markedly in soil physicochemical properties, yet their effects on belowground microbiomes remain poorly understood. We characterized bacterial and fungal communities in paired high-tunnel and adjacent field soils from 100 small-scale vegetable farms across Minnesota, integrating amplicon sequencing of 16S rRNA and ITS2 regions with soil nutrient data, arbuscular mycorrhizal fungi (AMF) spore counts, and microbial co-occurrence networks. High-tunnel soils had higher pH, organic matter, and multiple macronutrients (notably P, K, and N forms) and lower bulk density than fields, reflecting intensive organic amendments and reduced leaching. Despite these differences, bacterial and fungal alpha diversity did not differ between environments, whereas beta diversity analyses revealed strong shifts in community composition. High tunnels were enriched in salt- and stress-tolerant bacterial phyla (Firmicutes, Deinococcota, Patescibacteria, Halanaerobiaeota, Halobacterota) and saprotrophic fungal groups (Mortierellomycota, Ascomycota, Basidiomycota, Mucoromycota), while several oligotrophic or symbiotic taxa, including Acidobacteriota and Glomeromycota, declined. Glomeromycota relative abundance was negatively correlated with high soil phosphorus, whereas AMF spore densities did not decline, suggesting suppression of active mycorrhizal symbioses rather than propagule loss under high-nutrient conditions. Co-occurrence network analyses showed that bacterial and fungal networks in high tunnels were less dense, more modular, and exhibited higher ratios of positive to negative associations than field networks, consistent with stress-induced shifts toward more facilitative interactions. Collectively, our results indicate that high-tunnel production homogenizes soil microbiomes and selects for stress- and high-nutrient-adapted taxa, with potential consequences for nutrient cycling, AMF function, and long-term agroecosystem outcomes.

  PublisherDOI

• Managing soil sustainably on small-scale vegetable farms: Lessons learned from high tunnel and open field vegetable production

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-01-28 · 1 citations

  articleOpen access

  ABSTRACT Small-scale vegetable farms are increasingly important to local food systems, but the soils on these farms are not well understood, particularly in high tunnel production environments. Therefore, this study aimed to 1. Compare soil nutrients and soil health metrics in high tunnels and nearby open fields. 2. Document soil nutrient accumulation on diversified vegetable farms and assess loss potential. 3. Explore the impacts of specific management practices (input use, cover crops, tillage, and soil testing) and farm demographics on a variety of soil health and soil nutrient metrics. Just under half of the high tunnels in this study had soluble salt accumulation, which was associated with higher soil nitrate concentrations. The pH of many high tunnel soils was above the optimal range for crop production, which was correlated with irrigation water alkalinity. Some high tunnel soils had rapid water infiltration rates, with implications for irrigation water management. Both high tunnel and open field soil were rich in nutrients compared with other Minnesota farms. Preliminary assessments suggested risks to surface and groundwater from nutrient runoff and leaching. While farmer experience and more years in vegetable production were negatively associated with soil health metrics, management practices including reduced tillage, organic management, and application of plant-based compost were positively associated with soil health. Cation exchange capacity and permanganate oxidizable carbon did not provide significantly more insight than simply measuring organic matter. Arbuscular mycorrhizal fungal spore counts were inconclusive, but aggregate stability and bulk density were responsive to farmer reported soil management activities. Core ideas High tunnel soil tends to be rich in nutrients and organic matter. They also accumulate soluble salts, likely from excess inputs Irrigation water routinely tested high in pH and alkalinity. These factors may explain high soil pH in high tunnels. Small-scale vegetable farms often have high concentrations of soil nutrients in both high tunnels and open fields, with potential to cause environmental contamination through leaching and runoff Vegetable production may be inherently hard on soil health, but conservation practices including reduced tillage, organic management, and use of plant-based composts can improve soil health in these production systems

  PublisherOA PDFDOI

• Comparative genomics of strains of varying virulence of the Gram-positive bacterial pathogen of maize, Clavibacter nebraskensis

  Research Square · 2026-01-26

  preprintOpen accessSenior author
  PublisherDOI

• Refining the root-associated microbial consortia for enhanced biocontrol of the root-rot pathogen of corn

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-09-27

  preprintOpen accessSenior authorCorresponding

  Abstract Microbial consortia play a crucial role in plant protection by suppressing soil-borne pathogens. A previously studied root-associated microbial consortium consisting of seven bacterial strains (C7) demonstrated biocontrol activity against seedling blight in corn caused by Fusarium verticillioides . To enhance its biocontrol potential, we incorporated a free-living bacterial strain (S8) exhibiting biocontrol activity, forming a modified community (C8). We evaluated the biocontrol efficacy of S8, C7, and C8 against four major corn pathogens: Pythium torulosum, Fusarium graminearum, Fusarium subglutinans , and Rhizoctonia solani . Plate assays revealed that S8 and C8 exhibited the highest inhibition against P. torulosum (&gt;65% growth inhibition) but were less effective against Fusarium species (25–30%), while none of the communities restricted R. solani growth. In pot assays under growth chamber conditions S8 alone exhibited superior pathogen suppression compared to C7 and C8. However, integrating S8 into C7 did not enhance overall biocontrol efficacy. Community analysis via 16S amplicon sequencing revealed no significant shifts in C7 community strain abundance upon S8 introduction, suggesting a lack of establishment of S8 into the C7 community. Given that some individual strains exhibited stronger pathogen suppression than C7 and had variable effects on root biomass and plant height, we designed three sub-communities (SC1, SC2, SC3) based on the highest inhibitory activity. Plant assays demonstrated that inoculation with SC1 and SC2 restored plant height and root biomass, indicating that biocontrol efficacy is primarily driven by specific strain combinations rather than the broader community. Our findings underscore the importance of refining microbial consortia to maximize synergistic interactions and minimize antagonism, advancing sustainable disease management in agriculture.

  PublisherOA PDFDOI

• First Report of Bacterial Blight of Field Pennycress Caused by <i>Xanthomonas campestris</i> in Minnesota

  Plant Disease · 2025-06-05

  articleOpen accessSenior author

  Field pennycress (Thlaspi arvense) is an annual Eurasian winter/spring plant domesticated as an oilseed cover crop. In June 2024, water-soaked, diffuse black spots 1-3 mm in diameter with larger chlorotic halos were observed on pennycress leaves and seed pods in the Southern Research and Outreach Center in Waseca, MN plots. Significant bacterial streaming was observed under the microscope from symptomatic seed pods, and the tissue tested positive on an agdia immunostrip for Xanthomonas spp. Similar symptoms and water-soaked black lesions on seedling leaves were observed in pennycress research plots in St. Paul, MN, from August to October 2024, with symptoms affecting 70-80% of plants in September. Symptomatic tissue samples were surface sterilized with 0.75% NaOCl for 30s, rinsed in sterile Milli-Q water, macerated, and extracted in sterile 0.85% NaCl for 15 min, then streaked onto agar media. Seed pod extracts were streaked onto King’s B medium (King et al. 1954; Basnet et al. 2024), while leaf extracts were streaked onto Wilbrink’s medium (Sands et al. 1986) supplemented with 75 mg/l cycloheximide. After incubation for 72 h at 26°C, yellow, viscous colonies with entire margins were restreaked onto yeast dextrose calcium carbonate medium (YDC; Wilson et al. 1967) to ensure pure culture. To confirm Koch’s postulates, vernalized T. arvense seedlings were grown in a growth chamber (14 h photoperiod, 20°C day, 18°C night,) and inoculated at 5 weeks old with one of two isolates (XUMN-1 - Waseca seed pod origin, n=15 or XUMN-2 - St Paul leaf origin, n=15) by piercing the midvein with a sterile toothpick dipped into bacterial suspension (OD600 = 0.12; 1x108 CFU/mL). Plastic bags were fitted over plants for 48 h postinoculation, and environmental conditions were altered to promote bacterial growth (16 h photoperiod, constant 26°C, 250 umol/(m2)/s, 60% humidity; Basnet et al. 2024). Chlorosis and necrosis around inoculation sites were visible beginning at 4 days post inoculation (dpi), with symptoms developing into necrotic lesions and chlorosis by 7 dpi, which were absent in the mock treatment. Isolates were recovered using methods described above and DNA was extracted with Qiagen’s DNeasy UltraClean Microbial Kit. The sequenced 16s rRNA subunits (primers 27F: 5’-AGAGTTTGATCMTGGCTCAG-3’; 1492R: 5’-TACGGYTACCTTGTTACGACTT-3’) of all isolates’ DNA extracts shared >99% identity with Xanthomonas spp. based on an NCBI BLASTn search. Hybrid Illumina/Oxford Nanopore whole genome sequencing of the isolates was performed to confirm isolate species identity (CheckM value of our submitted genome was 99.64% complete with 0% contamination). FastANI analysis (Jain et al. 2018) with complete whole genome sequences of representative Xanthomonas campestris (Xc) strains (Dubrow et al. 2022) and contigs from UW-Madison isolates sequenced as part of their recent first report (Basnet et al. 2024) revealed that both isolates from this study shared >98.4% estimated identity with Xc pv. barbarae strain 3054 (RefSeq GCF_028750135.1), Xc pv. incanae strain CFBP1606, and Xc pv. barbarae strain 10-16. Additionally, these isolates are likely clonal with one another and with the UW-Madison greenhouse-originated isolates, as FastANI estimates between all five strains are >99.8%. After confirmation of the clonal nature of this study’s isolates, one strain was submitted to NCBI (NCBI BioProject ID PRJNA1209959). Pennycress is a high value cover crop and oilseed crop. Therefore, Xc epidemics can have economic and environmental impacts.

  PublisherOA PDFDOI

• First Report of Bacterial Wilt of Ginger Caused by <i>Ralstonia pseudosolanacearum</i> in the Continental United States

  Plant Disease · 2025-01-16 · 5 citations

  articleOpen accessSenior author

  Ginger (Zingiber officinale) is an herbaceous perennial in the Zingiberaceae family grown primarily in tropical to subtropical biomes as a culinary spice, a traditional medicine, and a landscaping plant. While ginger grows at soil temperatures above 20°C, several farmers in the upper Midwestern US farmers grows short-season ginger in high tunnels. In 2023 and 2024, growers in southeastern Minnesota reported a new disease of ginger. USDA traced the origin of the ginger rhizomes imported by these growers to Peru (Soto-Heredia et al., 2024). Plants exhibited wilting, chlorosis, pseudostem collapse, and plant death. Rhizomes were squishy with a fetid, rotting odor. Disease incidence ranged from scattered plants to 50% of the crop. Symptomatic samples tested positive for Ralstonia solanacearum species complex (RSSC) using AgDia ImmunoStrips. Rhizome tissue extracts plated onto CPG + 1% tetrazolium chloride agar and mSMSA agar produced irregular, round-to-fluidal white colonies with pink centers after 48 h of incubation at 28°C (Kelman 1954; Engelbrecht 1994). Diagnostic RSSC phylotype-specific multiplex PCR primers Nmult21:1F, Nmult21:2F, Nmult23:AF, Nmult22:InF, Nmult22:RR, and 759f/760r confirmed isolate identity; two isolates (UMN24-1, UMN24-2) yielded bands at 280 bp and 144 bp, identifying them as RSSC and phylotype I respectively (Fegan and Prior 2005; Opina et al. 1997). Additionally, 16S rRNA PCR of these isolates was sequenced using primers 27F (5'-AGAGTTTGATCMTGGCTCAG-3')/1492R (5'-TACGGYTACCTTGTTACGACTT-3') and shared >99% identity (isolate UMN24-1: 99.00% 27F, 99.53% 1492R; UMN24-2: 99.32% 1492R) with R. pseudosolanacearum (Rps) strain Gj707 (GenBank CP104497.1). Eight ginger plants in individual pots were grown in a Conviron growth chamber (28°C, 250 μmol/m2/s, 16 h photoperiod) and inoculated at the six-leaf stage (6 weeks old) by pipetting 109 CFU of the isolate, UMN24-1, suspended in sterile water into a 30°angled 1 cm incision on the pseudostem. We scored disease incidence based on wilted leaves per total leaves on each plant for 14 days post inoculation (dpi). While control plants remained healthy, inoculated plants showed symptoms (wilting, vascular discoloration, chlorosis) at 4 dpi, were 100% wilted by 6 dpi, and all pseudostems completely collapsed by 11 dpi (Fig. S1). Bacteria recovered from the inoculated plants matched the original isolate in appearance and positively identified as RSSC phylotype I using the PCR analyses described above. To confirm taxonomic identification of UMN24-1, the complete genome of this isolate was sequenced using hybrid Illumina and Oxford Nanopore sequencing (NCBI BioProject ID PRJNA1178296, release date upon publication) (Fig. S2). Average nucleotide identity (ANI) analysis of UMN24-1 using KBase's FastANI app reported 99.8% identity shared with Japanese RSSC phylotype I strain MAFF311693 (GenBank GCF 015698385.1). We also sequenced UMN24-2, and its phylogenetic analysis revealed that the two isolates are clonal (Fig. S2). Hence, we did not submit its sequence to the NCBI. This marks the first documented case of Rps in ginger in the United States. Similar to bacterial wilt in ginger closely related crop like turmeric has been repeatedly reported in northern Europe in recent years (EPPO 2024). While RSSC phylotype I is geographically limited by its minimal cold tolerance, the emergence of a pathogen with a potentially broad host range has worrying implications in a changing climate.

  PublisherOA PDFDOI

• Comparative genomics of the Gram-positive bacterial pathogen of maize, Clavibacter nebraskensis, reveals genomic variation between strains of varying aggressiveness

  Research Square · 2025-10-21

  preprintOpen accessSenior author
  PublisherOA PDFDOI

• Virulence is not directly related to strain success <i>in planta</i> in <i>Clavibacter nebraskensis</i>

  mSystems · 2024-11-29 · 2 citations

  articleOpen accessSenior author

  ABSTRACT Goss’s wilt and leaf blight of maize is an economically important disease caused by the Gram-positive bacterium, Clavibacter nebraskensis ( Cn ). Little is known about the ecology and pathogenesis of this bacterium. Here, we used phenotypic assays and a high-throughput whole-genome sequencing approach to explore among-strain variation in virulence and multistrain reproductive success in planta . Our survey of 41 strains revealed that more recently sampled strains tended to have higher virulence than strains sampled before 2010 and tended to be more genetically divergent from the reference strain, isolated in 1971. More detailed assays with a representative sample of 13 of these strains revealed that host genotype (resistant or susceptible) did not strongly affect strain success and that strain success in planta in multi-strain communities was not closely associated with virulence in single-strain assays. Two weakly virulent strains, CIC354 and CIC370, had the greatest reproductive success, whereas the most highly virulent strains did not significantly change in frequency in any host genotype. A genomic analysis revealed candidate genes, including putative virulence factors (i.e., a secreted cellulase), responsible for among-strain variation in reproductive success. IMPORTANCE Non-pathogenic strains of many bacterial pathogens are reported to coexist with pathogenic strains in symptomatic plants. To understand the ecology and pathogenesis of the pathogen population, it is essential to study strain dynamics in the context of the host. We created a community of 13 strains exhibiting diverse virulence phenotypes and used this community to infect the host plant. We compared the strain frequency of these strains before and after the host infection. Contrary to our hypothesis of highly virulent strains being selected by the susceptible host, we found that weakly virulent strains were selected by both resistant and susceptible host lines. We identified several genes associated with strain frequency shifts suggesting their role in strain colonization, virulence, and fitness.

  PublisherDOI

Frequent coauthors

• Caitilyn Allen

  University of Wisconsin–Madison

  9 shared

• Jean‐Michel Ané

  University of Wisconsin–Madison

  8 shared

• Kevin Garcia

  University of Wisconsin–Madison

  8 shared

• Tiffany M. Lowe‐Power

  6 shared

• Cristobal Carrera Carriel

  University of Wisconsin–Madison

  6 shared

• Quanita J. Choudhury

  University of Georgia

  6 shared

• Matthew Moscou

  Twin Cities Orthopedics

  6 shared

• Kevin R. Cope

  Oak Ridge National Laboratory

  6 shared