Hopanoid lipids promote soybean–Bradyrhizobium symbiosis

ABSTRACT The symbioses between leguminous plants and nitrogen-fixing bacteria known as rhizobia are well known for promoting plant growth and sustainably increasing soil nitrogen. Recent evidence indicates that hopanoids, a family of steroid-like lipids, promote Bradyrhizobium symbioses with tropical legumes. To characterize hopanoids in Bradyrhizobium symbiosis with soybean, we validated a recently published cumate-inducible hopanoid mutant of Bradyrhizobium diazoefficiens USDA110, Pcu-shc::∆shc. GC-MS analysis showed that this strain does not produce hopanoids without cumate induction, and under this condition, is impaired in growth in rich medium and under osmotic, temperature, and pH stress. In planta, Pcu-shc::∆shc is an inefficient soybean symbiont with significantly lower rates of nitrogen fixation and low survival within the host tissue. RNA-seq revealed that hopanoid loss reduces the expression of flagellar motility and chemotaxis-related genes, further confirmed by swim plate assays, and enhances the expression of genes related to nitrogen metabolism and protein secretion. These results suggest that hopanoids provide a significant fitness advantage to B. diazoefficiens in legume hosts and provide a foundation for future mechanistic studies of hopanoid function in protein secretion and motility. 
IMPORTANCE
 A major problem for global sustainability is feeding our exponentially growing human population while available arable land decreases. Harnessing the power of plant-beneficial microbes is a potential solution, including increasing our reliance on the symbioses of leguminous plants and nitrogen-fixing rhizobia. This study examines the role of hopanoid lipids in the symbiosis between Bradyrhizobium diazoefficiens USDA110, an important commercial inoculant strain, and its economically significant host soybean. Our research extends our knowledge of the functions of bacterial lipids in symbiosis to an agricultural context, which may one day help improve the practical applications of plant-beneficial microbes in agriculture.

N itrogen-fixing soil bacteria known as rhizobia are major contributors to the global nitrogen cycle and engage in economically and environmentally significant symbioses with legume plants (1).These symbioses help reduce the amount of nitrogen fertilizers required in commercial agriculture, ultimately lowering agricultural greenhouse gas emissions (2).Understanding the rhizobial genes that promote these symbioses has the potential to enhance agricultural sustainability.
In the symbiotic context, prior work on hopanoids has focused on the interactions between Bradyrhizobia and the Aeschynomene genus of semi-aquatic plants, a broadly distributed forage legume in the developing world.In the photosynthetic Bradyrhi zobium BTAi1, a deletion strain of the shc gene encoding squalene-hopene cyclase, the enzyme that catalyzes the first committed step of hopanoid biosynthesis (Fig. 1A), was found to have lower survival and aberrant bacteroid morphology in its native Aeschy nomene evenia host (5).In Bradyrhizobium diazoefficiens USDA110, a native soybean (Glycine max) symbiont with a broad range of compatible hosts, only mutant strains that lack specific hopanoid subclasses have been examined.A B. diazoefficiens strain lacking the hopanoid 2-methylase HpnP has no phenotype in symbiosis with Aeschynomene afraspera or soybean, whereas loss of the enzyme HpnH that generates hopanoids with extended hydrocarbon tails inhibits symbiosis with A. afraspera but not soybean (3).Further analysis found that hpnH deletion slows the development of the symbiosis-spe cific root nodule organ in A. afraspera, in part due to the strain's reduced motility and poor in planta survival (4).
It is not known whether the ∆hpnP and ∆hpnH phenotypes are similar to wild type (WT) in soybean because hopanoids are not relevant in this host, or whether the knockout of only certain hopanoid subclasses is insufficient.The roles of hopanoids in Bradyrhizobium-soybean symbiosis have been difficult to study, as the shc gene required for hopanoid synthesis appears to be essential in the main soybean symbiont, Bradyrhi zobium diazoefficiens USDA110 (3,29).Answering this question is an important step in determining the relevance of hopanoids for agriculture, given the global dominance of soybeans as cash crops.
In this study, we examine the effects of complete hopanoid loss on the symbioses between B. diazoefficiens and soybean using a recently published cumate-inducible shc knockout strain (24).We find that loss of shc expression is sufficient to deplete mem brane hopanoids and results in similar, though more severe, stress sensitivity phenotypes as observed for other hopanoid mutants.Hopanoid depletion also strongly inhibits B. diazoefficiens-soybean symbiosis over an extended period, with reduced nodule biomass and low in planta bacterial survival.In A. afraspera, complete hopanoid loss abolishes B. diazoefficiens symbiosis.RNA-seq analyses suggest that these symbiotic phenotypes may result from upregulation of protein secretion and nitrogen assimilation and the disrup tion of LPS biosynthesis and flagellar motility, the latter of which is confirmed by loss of swimming motility in the shc depletion conditions.Taken together, this work validates a new tool for exploring hopanoid function in B. diazoefficiens, expands the relevance of hopanoids in symbiosis to a major crop plant, and suggests potential molecular mecha nisms for hopanoids in legume-rhizobia symbiosis.

Validation of a cumate-inducible shc mutant strain of B. diazoefficiens
To characterize hopanoid functions in B. diazoefficiens, we used a cumate-inducible shc mutant strain Pcu-shc::∆shc (abbreviated as Pcu-shc hereafter) that was constructed previously by colleagues (24) (Fig. S1).We first sought to validate this strain by quantify ing the dose effect of cumate on shc gene expression and bacterial growth.We per formed qRT-PCR (∆∆Ct method with an internal control of 16S rRNA) to measure the relative expression (RE) of shc in WT and Pcu-shc strains after treatment with 5-50 µM cumate or an ethanol (EtOH) vehicle control (Fig. 1B).We found that the expression of shc is insensitive to cumate in WT, whereas the RE of shc in Pcu-shc was approximately 10fold lower than WT in the absence of cumate.The mutant strain had 5-10-fold higher shc expression than WT after cumate treatment, though we saw minimal changes in shc expression within the range of cumate concentrations tested.This result confirmed transcriptional induction of shc in Pcu-shc by cumate.In parallel, we performed growth curves of the two strains in rich medium (AG) on a plate reader for samples treated with either 25 µM cumate (+C) or EtOH control (-C).WT grown with cumate had similar growth dynamics to WT without cumate, suggesting that at 25 µM concentration, cumate does not affect WT growth.In contrast, Pcu-shc without cumate had pronounced growth defects in both exponential and stationary phases compared to WT without cumate, and growth of Pcu-shc was restored to WT levels with 25 µM cumate supplementation (Fig. 1C).We also found that high doses of cumate inhibited bacterial growth (Fig. S2), and therefore chose 25 µM cumate for all following assays, consistent with previous studies (24).
To determine if the low levels of shc expression in Pcu-shc without cumate are sufficient to deplete hopanoids from the membrane, we used GC-MS (gas chromatog raphy-mass spectrometry) protocols adapted from previous studies (27) to extract and analyze hopanoids from B. diazoefficiens membrane.In WT, the hopanoid precursor squalene (eluted at ~14.25 min), diploptene (at ~16.96 min), diplopterol (at ~17.97 min), and tetrahymanol (at ~18.85 min) were identified regardless of cumate addition (Fig. 1D, left panels).Pcu-shc did not make any detectable hopanoids without cumate treatment but did produce squalene, whereas cumate restored the production of all types of hopanoids.Curiously, squalene levels were also depleted by cumate addition (Fig. 1D, right panels), likely due to SHC overexpression to levels that fully exhausted the squalene substrate.Taken together, our assays demonstrate that cumate depletion in Pcu-shc results in complete hopanoid loss and growth inhibition, and that 25 µM cumate is sufficient to restore WT lipid composition and growth.

Hopanoids affect stress tolerance in free-living cultures
Previous work in B. diazoefficiens demonstrated that specific hopanoid subclasses, namely the 2-methyl hopanoids and extended hopanoids, are required for survival of membrane-targeting stresses in culture (3,4,27); however, whether stress survival is further inhibited by total hopanoid loss has not been investigated in this organism.To address this question, we performed growth curves of Pcu-shc and WT strains under three abiotic stresses-low pH, high temperature, and high osmotic pressurecommonly encountered by rhizobia within plant nodules (30).Under low pH conditions (pH = 5.0 compared to the optimal pH = 6.6),WT growth was unaffected irrespective of cumate addition (Fig. 2A), whereas Pcu-shc failed to grow without cumate supplemen tation.Similarly, Pcu-shc cannot survive without cumate under high temperatures (37°C compared to 30°C) (Fig. 2B), but both low pH and high-temperature growth could be partially restored by cumate in this strain.Under osmotic stress caused by 5% sucrose, Pcu-shc still grew without cumate, albeit slowly, and in this case growth inhibition was fully relieved by cumate (Fig. 2C).Dual temperature and osmotic stress roughly phenocopied temperature stress alone, with a slight growth elevation (Fig. 2D).

Hopanoids participate in soybean-B.diazoefficiens symbiosis
Given that hopanoids are essential for B. diazoefficiens survival of nodule-relevant stresses in culture, we hypothesized that they may participate in the organism's native symbiosis with soybean.To test this hypothesis, we inoculated WT or Pcu-shc onto soybeans grown in carbon-and nitrogen-free hydroponic growth medium, with or without 25 µM cumate.At 27 days post-inoculation (dpi), soybeans inoculated with Pcushc displayed similar nutrient starvation symptoms as non-inoculated (NI) control plants, including uniform foliage chlorosis and significantly stunted growth (Fig. 3A and B).Additionally, Pcu-shc-inoculated soybeans produced a strikingly lower number of nodules (~76% less) with reduced nodule dry mass (~72% lower) compared to plants inoculated with WT (Fig. 3C through E).As expected from the low nodule biomass, rates of nitrogen fixation per plant, quantified by the Acetylene Reduction Assay (ARA), were reduced by a commensurate amount (~65%) in the Pcu-shc-inoculated soybeans compared to WT-inoculated plants (Fig. 3F).However, normalizations of acetylene reduction by nodule number and dry mass were not significantly different between WTand Pcu-shc-inoculated strains (Fig. 3G and H), indicating that the reduction of nitrogen fixation in Pcu-shc-inoculated soybean was due to the decreased nodule number, rather than a core nitrogen fixation defect.An independent biological replicate of these assays was performed that yielded consistent results (Fig. S3A through H).
We also examined soybean symbiosis phenotypes at 45 dpi, to determine whether the hopanoid defect could be compensated over time, as shown for extended hopanoid mutants in A. afraspera (4).In line with the results at 27 dpi, soybeans inoculated with Pcu-shc grew much slower with shorter shoots, lower nitrogen fixation, fewer nodules, and lower nodule dry weight than WT-inoculated plants at 45 dpi (Fig. S4), demonstrat ing that hopanoid phenotypes could not be compensated over this time period.
Interestingly, none of the Pcu-shc phenotypes could be restored to WT in plants supplemented with 25 µM cumate (Fig. 3).We suspect that cumate was not absorbed into nodule tissues or was metabolized by soybean hosts, though it is also possible that loss of squalene in the Pcu-shc plus cumate condition plays a role (Fig. 1E).Alternatively, some of the symbiotic phenotypes might result from other elements on the pRJPcu-shc vector that is integrated into the Pcu-shc genome, such as the plasmid mobilization or tetracycline resistance genes.We think this interpretation is unlikely, as integration of plasmids with the same backbone has been used to generate many tagged strains of B. diazoefficiens, and none of these strains had such severe defects in symbiosis as Pcu-shc (31). of nitrogenase activity per plant by the Acetylene Reduction Assay (ARA).The area of the ethylene peak from each acetylene-treated plant, which is positively correlated with nitrogen fixation, was recorded and normalized by reaction time (F), nodule number (G), and nodule weight (H).The median and the quantiles were calculated using GraphPad Prism 9.0.Levels of significance were evaluated using t test and annotated as follows: P > 0.05, non-significant (ns); P < 0.01, two asterisks; P value < 0.001, three asterisks; P value < 0.0001, four asterisks.

Hopanoid mutant-infected soybean nodules are disorganized with low symbiont loads
To assess nodule and bacterial morphologies in symbiosis, we prepared semi-thin (100 µm) sections of ~20 nodules pooled from 5 to 10 plants per treatment.Under stereo microscopy, most Pcu-shc-infected nodules (95%) had dark pigments indicative of starch granules surrounding the infection zone (Fig. 4E through H).Interestingly, some Pcu-shc nodules (~18%) displayed segmented or "patchy" infection zones with starch granules located at the edge of patches (Fig. 4E, F and H).In contrast, all WT-infected nodules had an evenly distributed infection zone pattern across the nodule (Fig. 4A and C), even at lower infected cell density (Fig. 4B and D), and the frequency of nodules with starch granules was low (9%).We note that these starch granules are not strictly a sign of poor nitrogen fixation; though nearly every Pcu-shc infected nodule displayed this phenotype, our previous assay showed that acetylene reduction per nodule dry mass between WT and Pcu-shc was not significantly different (Fig. 3H).Instead, the starch granules are an indication of an imbalance between the photosynthesis of plants and the ability of bacteroids to metabolize them.
To examine if loss of hopanoids affects bacteroid viability and density within nodules, we stained a similar number of nodule cross-sections as above (~20 nodules from five plants) with propidium iodide (PI; a membrane-impermeable DNA dye and "dead" cell marker), SYTO9 (a membrane-permeable DNA dye and "live" cell marker), and Calcofluor white (a plant cell wall stain).High-resolution confocal imaging revealed that all WT-Cinfected nodule cells contained a high density of symbionts predominantly stained with SYTO 9 (Fig. 4I through L; Fig. S5).SYTO9 and PI staining of Pcu-shc-C-infected nodules yielded more variable results.A portion of Pcu-shc-C-infected nodules (~30%; 7 out of 22) contained no or very few symbionts that were predominantly PI-stained (Fig. 4M through P; Fig. S6), while others contained a patchy distribution of predominantly SYTO9-stained bacteria.We used Fiji (ImageJ) to quantify these phenotypes by measuring (i) the area of the nodule infection zone (IZ) containing bacteroids (Fig. 4Q); (ii and iii) the intensity of the SYTO9-stained "live" cells (Fig. 4R) or PI-stained "dead" cells (Fig. 4S) in the IZ; and (iv) the combined PI and SYTO9 intensity (Fig. 4T) in the IZ.This analysis revealed that a larger fraction (~50%) of the IZ in WT-C-infected nodules contained bacteroids compared to Pcu-shc-C-infected nodules (~33%) (Fig. 4Q).WT-C-infected nodules also had a significantly higher number of SYTO 9-stained "live" cells, but had no difference in the number of PI-stained "dead" cells, compared to Pcu-shc-C-infected nodules (Fig. 4R through T).Our microscopy analyses additionally demonstrated that in largely empty Pcu-shc-C-inoculated nodules, plasma membranes of soybean nodule cells were separated from their cell walls, suggesting cell death (Fig. 4N; Fig. S6).Considering that ARA per nodule and per nodule dry mass were equivalent between WT-C and Pcu-shc-C infected nodules, we assume that the low nitrogen fixation of empty or under-occupied Pcu-shc-C nodules is counteracted by a subset of WT-like nodules with high occupancy (Fig. S7).This heterogeneity of hopanoid mutant developmental phenotypes may also explain the high variability in ARA per nodule and per nodule dry mass (Fig. 3G and H).
Cumate supplementation had minimal effects on both WT-and Pcu-shc-infected nodule phenotypes.There were no significant differences between WT-C and WT+C nodules in any of the measurements.Although the bacteroid-occupied area of the IZ of Pcu-shc+C-infected nodules was restored to the WT+C level, neither the intensity of "live" bacteroid cells nor total bacteroids were fully restored (Fig. 4Q and T).This suggests that Pcu-shc+C can infect the plants normally but cannot survive in the harsh nodule environment, consistent with its inability to enhance plant growth or nitrogen fixation relative to Pcu-shc-C.An independent biological replicate of these assays yielded consistent results (Fig. S3I through L).Sections were stained with Calcofluor (cyan, plant cell wall), SYTO 9 (yellow, "live" bacteria), and propidium iodide (magenta, membrane-compromised "dead" bacteria and plant nuclei).(Q-T) Fraction and intensity of nodule infection zones (IZ) containing "live" (SYTO 9-stained) and "dead" (PI-stained) bacteroids (Q) and intensity of "live" (R), "dead" (S), and total bacterial cells (T).a.u.= arbitrary units of fluorescence.Levels of significance are calculated and annotated as in Fig. 3.

Total hopanoid loss prevents B. diazoefficiens symbiosis with A. afraspera
We also examined the symbiotic phenotypes of Pcu-shc without cumate in A. afraspera, in which extended hopanoid and 2-methylated hopanoid mutant phenotypes were previously examined (4).In this host, the plant shoot height and the number and mass of nodules infected by Pcu-shc-C were substantially reduced at 28 dpi (Fig. S8A through  D), and all of these mutant-infected nodules were much smaller than WT-C (Fig. S8E and  F).Although only a few nodules were present per plant, nodule sections stained as above suggest all nodules were empty or contained a low density of undifferentiated bacteria (Fig. S8G).

Transcriptomic studies suggest that hopanoids suppress the expression of flagellar motility-related genes and enhance expression of genes involved in protein secretion
To understand the molecular mechanisms underlying the phenotypes of Pcu-shc under hopanoid-free conditions, whole-genome transcriptomics of free-living WT and Pcu-shc with and without cumate were investigated.Normalized read counts of WT and Pcu-shc per treatment from two biological replicates with three technical replicates each (with the exception of WT without cumate, which has only five total replicates, as one sample contained poor-quality RNA) are provided in Table S2.We obtained 25-50 million reads per sample with 200×-400× coverage, and 96% of the reads could be aligned to the B. diazoefficiens USDA110 genome.In all samples, a large majority of the genes (98%) had non-zero counts.As expected from our qRT-PCR assays (Fig. 1B), the coverage and average of read counts of shc in Pcu-shc-C (160 ± 6) were ~65-fold lower than WT-C (10,580 ± 460), whereas they were ~7-fold higher in Pcu-shc+C (80,841 ± 3,043) compared to WT+C (10,064 ± 345) (Fig. 5A; Table S2).Also, as expected from our growth curves and symbiosis phenotypes, the Pearson correlation coefficients between Pcu-shc-C replicates with all other samples were lower (~0.94, in blue) than the coefficients of correlation between all other samples (~0.99 in orange) (Fig. S9).
We performed differentially expressed gene (DEG) analysis on transcripts from Pcushc-C compared to WT-C (Fig. 5D; Table S3), where DEGs were defined by a log 2 fold change ≥ 1 or ≤−1 and a false discovery rate-adjusted P value ≤ 0.05.We excluded DEGs from analysis that were also upregulated or downregulated in Pcu-shc+C relative to WT+C (Fig. 5E; Table S4), as these DEGs are likely to result from the presence of pRJPcushc backbone elements other than Pcu-shc.We note that the effect of the pRJPcu-shc backbone appears to be minimal, as only 56 genes (<1% of the genome) (Fig. 5B and C, numbers in circles) were differentially expressed in both Pcu-shc conditions (±C) relative to similarly treated WT controls, whereas 782 genes (~9% of the genome) were affected only in Pcu-shc-C (Fig. 5B and C, numbers in squares).The effect of cumate on gene expression can be inferred by comparing DEGs between the WT+C versus WT-C and Pcu-shc+C versus Pcu-shc-C data sets (Fig. S10; Tables S7 and 8).Only 15 DEGs are shared for these data sets (Fig. 5B and C, numbers in triangles); however, there are 95 DEGs that are upregulated in WT after cumate addition, but are not upregulated by cumate in Pcu-shc, suggesting that cumate effects might differ between these strains.Unfortunately, it is not possible to separate cumate effects that are unique to Pcu-shc from the effects of shc gene overexpression.
The majority of unique DEGs in Pcu-shc-C compared to WT-C were uncharacterized genes.A heatmap of the top 20 upregualted and and downregulated DEGs is included in Fig. S11.To assess the accuracy of these DEGs, we selected 12 genes (6 upregulated genes, 5 downregulated gene, and 1 unaffected gene) and quantified their transcript levels by qRT-PCR.We found all the relative expressions of those genes in Pcu-shc-C compared to WT-C were consistent with the RNA-seq data (Table S9), confirming the fidelity of our RNA-seq data for further analysis.
Among downregulated DEGs, many genes are involved in the regulation of flagella (e.g., fla and flaA), flagellar assembly (e.g., flaF and flaE), and chemotaxis (e.g., cheR), which are highlighted in the volcano plot (Fig. 5D).GO term enrichment analysis also confirmed that flagellum organization and flagellum-dependent motility were the top GO terms associated with downregulated DEGs (Fig. 5F).Genes involved in  S9).(E and F) GO term analysis of downregulated (F) and upregulated (E) in biological process categories.lipopolysaccharide (LPS) biosynthesis and symbiosis, such as rafD and rafE (32), also are among the downregulated DEGs (Fig. 5D).This may reflect the role of hopanoids as a structural component of the lipid A moiety of LPS in Bradyrhizobium (5).
GO term enrichment analysis indicated upregulated genes are most strongly associated with nitrogen metabolic processes (Fig. 5G).These include the ammonium transporter amtB and each gene in the operon blr2803-2809, which is involved in NO 3-assimilation.Proteins encoded include: an ABC-type NO 3− transport system NrtABC (Blr2803-05), a major facilitator superfamily (MFS)-type NO 3−/ NO 2− transporter (Blr2806), bacterial hemoglobin (Blr2807), a FAD-dependent NAD(P)H oxidoreductase (Blr2808), and the catalytic subunit of the assimilatory NO 3− reductase NasA (Blr2809).The most upregulated gene in Pcu-shc-C, uppS, is an isoprenyl transferase involved in terpenoid biosynthesis, upstream of hopanoid biosynthesis.Genes involved in secretion systems, such as hlyD and bll6293 of the type I secretion system (T1SS) and an operon composed of 25 type III secretion system (T3SS) genes (e.g., rhcJ, rhcN, rhcQ, nolB, rhcS, rhcR, etc.), are the second major category of upregulated DEGs in Pcu-shc-C (Fig. 5D and G).Recent research revealed that mutants in secretion systems in plant-beneficial bacteria modulate prokaryotic and eukaryotic interactions in the rhizosphere (33).

Hopanoid depletion prevents swimming motility
To determine whether the reduced expression of flagellar motility genes was sufficient to inhibit Pcu-shc-C motility, we performed a swimming assay in semisolid agar plates.We found that WT swimming was unaffected by cumate (Fig. 6A), with a motility halo at 12 dpi that was twofold larger than at 2 dpi.The Pcu-shc-C inoculum did not appear to swim between 2 and 12 dpi, whereas the cumate-complemented strain expanded similarly to WT (Fig. 6B).

DISCUSSION
Relatively few studies explore the biological function of bacterial lipids on plant-microbe interactions.A difficulty of such work is the construction and validation of lipid biosynthesis mutants.This validation typically requires lipid extraction and quantification by mass spectrometry, a technically challenging approach that is less widely available than methods for tracking protein depletion.Here, we provide a thorough validation of an inducible total hopanoid mutant, previously generated by colleagues through a clever genetic approach to bypass the difficulty of selecting for clean shc knockouts (24).We demonstrate that the strain is suitable for cumate dose-dependent manipulation of not only shc expression but also membrane hopanoid content in a pure culture setting.

Consequences of hopanoid loss in stress survival
In the absence of hopanoids, B. diazoefficiens stress sensitivity is markedly more severe than reported in other hopanoid mutants.Under pH stress at pH 4.5-5.0,Δshc strains of Rhodopseudomonas palustris TIE-1 (27) and Burkholderia cenocepacia K56-2 (10) have delayed growth but are still viable, whereas the equivalent total hopanoid knockout of B. diazoefficiens cannot grow.This may suggest that hopanoids are more central to pH stress resistance in our organism, consistent with previous work on extended hopanoid mutants (ΔhpnH) of B. diazoefficiens (3,24).However, we cannot discount the effects of distinct medium compositions used to cultivate the different species.
In the context of temperature stress, the ΔhpnH mutant of B. diazoefficiens also cannot grow at 37°C (3) suggesting extended hopanoids are the primary contributors to this phenotype in the Pcu-shc strain.This ΔhpnH strain has more mild growth inhibition under osmotic stress (24) than Pcu-shc-C, and the 2-methylated hopanoid mutant ΔhpnP has no osmotic stress phenotype (3), indicating that the short hopanoid class is the most relevant to osmotic stress survival.Overall, the stress sensitivity phenotypes of Pcu-shc-C under our culturing conditions are more severe than previously published mutants.The strong Pcu-shc-C phenotypes, combined with its inducibility, demonstrate that it is a strong model for future molecular studies of hopanoid-mediated stress resistance.

B. BTAi1 Δshc in A. evenia, B. diazoefficiens ΔhpnH in A. afraspera, and B. diazoefficiens
Pcu-shc in soybean all had a roughly twofold lower rate of acetylene reduction per plant than the WT strain at 2-4 weeks post-inoculation, with a subset of nodules exhibiting low symbiont load, disorganization of the central infection zone, and signs of early nodule senescence or nodule cell death (3)(4)(5), What differs between the three models is which hopanoids must be depleted to yield this common phenotype.In A. afraspera hosts, total hopanoid loss prevents symbiosis entirely, but the extended hopanoid knockout ΔhpnH yields the common phenotype, suggesting that short hopanoids are most important in the B. diazoefficiens-A.afraspera interaction.In soybean and A. evenia, the common phenotype is recapitulated in a total hopanoid knockout, though we note that ΔhpnH and ΔhpnP were not tested in A. evenia.
Why total hopanoid loss elicits a stronger phenotype in the B. diazoefficiens-A.afraspera symbiosis is not clear.It is possible that the types of stresses associated with this host are different from those of A. evenia and soybean, but we find this conclusion unlikely.Instead, we believe that the stronger phenotype of B. diazoefficiens Pcu-shc in A. afraspera reflects the fact that these two species do not interact in nature.Though B. diazoefficiens can engage in this symbiosis, it is less efficient than the native symbiont B. sp.ORS285 (34), suggesting that B. diazoefficiens has not been adapted to survive in A. afraspera tissue.Regardless, the similarity of the total hopanoid knockout pheno types of the native B. BTAi1-A.evenia and B. diazoefficiens-soybean pairings confirm that hopanoids are indeed important for natural Bradyrhizobium-legume symbiosis.All studies also agree that 2-methyl hopanoids produced by HpnP are not major contrib utors to symbiosis, and instead that short and extended hopanoids are the more significant lipids in this context.

Molecular mechanisms of hopanoid function
Prior analysis of the B. diazoefficiens-A.afraspera symbiosis proposed that the reduced nitrogen fixation of hopanoid mutants is a product of at least two independent defects: delayed initial infection of plant roots, perhaps relating to reduced strain motility, and slower root nodule growth, perhaps due to lower levels of symbiont proliferation and survival in planta (4).This manuscript supports a similar model in soybean.We observed both reduced flagellar assembly gene expression and loss of swimming motility in the Pcu-shc-C condition, consistent with the reduced swimming motility of B. diazoef ficiens ΔhpnH (4) and B. cenocepacia Δshc (10), as well as lower stress resistance in culture and bacteroid densities in nodules.Our transcriptomics work introduces several new potential mechanisms by which hopanoid depletion inhibits symbiosis: excessive nitrogen assimilation, altered secretion of symbiotically relevant effector proteins, and loss of chemosensation and taxis.
How hopanoids affect these processes at the molecular level is not clear.Some of the observed phenotypes may arise from activation of general stress response (GSR) mechanisms in hopanoid mutants.In R. palustris TIE-1, the expression of 2-methyl hopanoids is regulated by the extracellular function (ECF) sigma factor, which mediates GSR in this organism (35).Though it has not been tested directly, it is possible that the levels of some or all hopanoids are also linked to GSR activation in B. diazoefficiens, or to the activation of membrane-specific stress response pathways.Another hypothesis is that hopanoids affect membrane-based processes through their role in biophysically reinforcing or compartmentalizing cell membranes, reviewed in (36).Bacterial flagellar motility, chemotaxis, protein secretion, and the nitrate assimilatory pathway all involve large membrane-associated protein complexes that may rely on proper membrane mechanics to function.There is some evidence for this hypothesis in the case of protein secretion: the Sec translocon ATPases of Escherichia coli (37) and Streptococcus pneumo niae (38) are sensitive to membrane lipid composition and biophysics.This possibility will be an exciting topic of future investigation.

Challenges for future tool development
As we have demonstrated, the inducible mutant strategy used to develop the Pcu-shc strain ( 24) is a useful workaround for generating mutants of B. diazoefficiens genes that are essential or otherwise difficult to delete cleanly.However, it has limited use in planta, as we could not rescue shc depletion phenotypes in the host.Further tool development to identify tissue-permeable inducer molecules, and the identification of promoters that rescue Bradyrhizobium gene expression at specific stages of nodule development as have been done for Sinorhizobium meliloti (39) would significantly enhance our ability to understand bacterial gene function in Bradyrhizobium-legume symbioses.

Bacterial strains and cultivation
Wild-type Bradyrhizobium diazoefficiens strain USDA110 spc4 (WT B.diazoefficiens USDA110) originally was obtained as a gift from Dr Hans-Martin Fischer (ETH Zurich) to Dr. Dianne Newman (Caltech), who then gifted the strain to Dr. Belin.Hopanoid mutant strain Pcu-shc::∆shc (DKN1784) (24) was provided as a gift from Dr Dianne Newman (Caltech).Briefly, to construct Pcu-shc::∆shc in reference 24, the Newman group generated a plasmid (pRJPcu-shc) in which a copy of the shc gene (blr3004) was introduced into the multiple cloning site on a cumate-inducible promoter (Pcu).This vector then was integrated into the symbiotically silent blr1132' locus in the B. diazoefficiens genome.Finally, the endogenous shc gene (blr3004) was deleted in the presence of cumate (under which the copy of shc on the pRJPcu-shc backbone is expressed) to produce strain Pcu-shc::∆shc.
For all experiments, cells from 10% glycerol stocks stored at −80°C were streaked on agar plates made with rich AG medium (4.6 mM sodium gluconate, 6.6 mM arabinose, 1 g/L yeast extract, 6 mM NH 4 Cl, 5.6 mM MES, 5 mM HEPES, 1 mM Na 2 HPO 4 , 1.76 mM Na 2 SO 4 , 88 µM CaCl 2 , 25 µM FeCl 3 , and 0.73 mM MgSO 4 , pH 6.6).Plates were grown aerobically at 30°C until single colonies appeared, typically for 3-5 days, and colonies were selected for inoculation into 5 mL liquid-rich AG.Liquid cultures were grown to OD 600 = 0.5-0.8under aerobic conditions in Eppendorf Innova incubating shakers at 30°C and 250 rpm.Culture tube OD 600 values were measured using a Thermo Scientific GENESYS 30 Visible Spectrophotometer with a test tube adapter module.

qRT-PCR
qRT-PCR was performed as previously described with minor modifications (40).Briefly, total RNA from WT or Pcu-shc::shc strains at early exponential phase (OD 600 = 0.5-0.8)was extracted using an RNeasy Mini Kit (Qiagen), according to the manufacturer's recommended protocol.The genomic DNA was removed using on-column DNase-I digestion (Qiagen) for 15 min.RNA concentration and purity were determined using a Nanodrop one (Thermo Scientific).Total RNA (500 ng) was reverse transcribed using random primers (Promega) and AMV Reverse Transcriptase (Promega), according to the manufacturer's protocol.
Quantitative PCR (qPCR) with Fast SYBR Green PCR Master Mix (Thermo Fisher) was performed using the CFX96 Real-Time System (Bio-Rad) in Optical 96-Well Fast Plate (Stellar Scientific).Each reaction contained 5 ng cDNA and 250 nM gene-specific primers.The relative expression of the target genes was normalized to the 16S ribosomal RNA housekeeping gene (Ensembl transcript ID: EBT00050756951, gene name: rrn16S; NCBI gene ID: AAV28_03955) based on the mean of cycle threshold (Ct) values and ΔΔCt method.All values are the means ± standard errors of three replicates.The primers used for qRT-PCR are listed in Table S1.

Growth curves
Three individual colonies from WT and Pcu-shc::∆shc strains on agar plates were picked up and cultivated in 5 mL AG media, as noted above.Once the bacterial strains reached the early exponential phase (OD 600 = 0.5-0.8), the culture OD 600 values were adjusted to the same value in the fresh AG medium.Cultures were then diluted 1:100 into 200 µL fresh AG medium in individual wells of a 96-well plate.For growth curves of WT and Pcu-shc::∆shc under variable cumate concentrations, AG media was supplemented with either 0.1% EtOH only (for minus cumate samples) or 0.1% EtOH with 5, 10, 25, 50, or 100 µM cumate (for plus cumate samples).For growth of the strains under stress, AG medium was supplemented with 5% sucrose or adjusted to pH 5.0.Each well was overlaid gently with 30 µL of mineral oil to prevent evaporation.
The plate was placed into a microplate reader BioTek EPOCH 2 set to 30°C (most conditions) or 37°C (heat stress conditions) with continuous shaking.The OD 600 was measured every 2 h for 96 h.Data were plotted in Prism GraphPad.

Hopanoid extraction and quantification by GC-MS
Hopanoid lipids were purified and quantified as previously described with minor modifications (27,41).Briefly, 5 mL cultures of WT and Pcu-shc::∆shc were harvested in early exponential phase by spinning down in a swinging bucket centrifuge at 3,250 × g for 30 min.Cell pellets were resuspended in 50 µL of water and transferred into 9 mm Clear Glass Screw Thread Vials (1.7 mL Tapered Base, Thermo Scientific SKU C40009) with Vial Caps (Thermo Scientific, SKU 60180-516).MeOH (125 µL) and CH 2 Cl 2 (62.5 µL) were added in sequence in a 2:1 ratio.The vials then were capped and sonicated in a Branson M2800H Ultrasonic Bath (40 kHz, Power 110 W) for 30 min at room temperature.Total lipids were extracted by the addition of ~200 µL CH 2 Cl 2 and mixed by pipetting.After the sample settled for 2-4 h at room temperature, permitting the organic and aqueous layer to separate, the lower organic layer that contained the hopanoids was collected, dried in a 60°C oven, and acetylated in 100 µL of 1:1 pyridine:acetic anhydride (Ac 2 O) for 30 min at 60°C.Acetylated hopanoids were analyzed on a Thermo Scientific TRACE 1300 gas chromatograph coupled with an ISQ 7000 single quadrupole mass spectrometer.Protocols were modified from those previously described (27,28).About 1 µL of acetylated hopanoids was injected into the gas chromatograph and run through a Restek Rxi-XLB column (30 m × 0.25 mm × 0.10 µm, Restek 13708) with an injector temperature of 325°C and in split mode (4:1), with helium as the carrier at constant flow of 1.5 mL/min.The GC oven temperature was programmed as follows: 100°C for 2 min, then ramped at 15°C/min to 320°C, followed by a 5 min hold at 320°C.The mass spectrometer was operated in full scan mode over 50-750 amu at 70 eV in electron impact (EI) ionization mode.The MS transfer line temperature was held at 325°C and the ion source temperature at 250°C.Hop-22 (30)-ene (Sigma-Aldrich, PubChem Substance ID: 32974796), also known as the short hopanoid diploptene, was used as standard.
Data collected from GC-MS were processed by the Chromeleon 7.3 Chromatogra phy Data System (CDS) Software.Compounds were identified by comparison with the standard in terms of retention times, NIST compound library hits identified in Chrome leon, and mass spectra of molecular ions, as described previously (3,5,27,28,42,43).

Soybean and Aeschynomene cultivation and inoculation
Aeschynomene afraspera (A.afraspera) seeds were produced in-house and cultivated as previously described (4).Soybean (Glycine max) seeds (non-GMO forage variety) were purchased from the Hancock Seed Company.Seeds were sterilized by rinsing in 95% EtOH, incubating in 5% bleach for 5 min on a nutating shaker, and washing five times in ultrapure water.Using sterile metal forceps, cleaned with ethanol periodically in a biosafety cabinet, sterilized seeds were placed on freshly poured 1% water/agar plates.The plates were sealed with parafilm and covered in aluminum foil to protect from light.Seeds were germinated in the dark at 30°C for 72 h.
For planting, seven to nine seedlings for each treatment (WT-C, WT+C, Pcu-shc-C, Pcu-shc+C, and non-inoculated control) were guided into sterile 400 mL borosilicate glass beakers (soybean assays) or 250 mm × 25 mm/100 mL large tubes (A.afraspera assays) through perforated aluminum foil caps.Beakers or tubes were filled with autoclaved buffered nodulation medium (BNM) modified from the previous study (44) with adjustment to pH 6.5 with ~1 mL of 10M KOH per liter.The BNM medium contained 2 mM Cacl-2H 2 O, 2 mM MES buffer, Nod major salts (0. ).An additional 500 µM KNO 3 was added to provide the minimum nitrogen required for soybean growth before nodule formation.Seedlings were grown for 5-7 days in Conviron GEN2000 chambers with light intensities of 600 µmol/m 2 /s at 28°C and 80% RH for a 16:8 h day:night cycle.
WT and Pcu-shc::∆shc cultures were grown to OD 600 = 0.5-0.8 in AG media with cumate and then subcultured with or without cumate, as described above, to OD 600 = 0.5-0.8.Bacterial cultures were pelleted at 3,520 × g for 30 min at room temperature and resuspended in rich AG to a final OD 600 = 1.0.After adjusting the culture OD 600 , 4 mL or 1 mL of culture suspension was added to each soybean or A. afraspera plant, respectively, along with 400 µL of 25 mM (1,000×) cumate, or 400 µL of EtOH only.Plants inoculated with different strain/cumate combinations were cultivated in separate bins to prevent cross-contamination.After inoculation, plants were maintained in GEN2000 chambers until harvesting for nodule analysis at 27-28 dpi or 40-45 dpi under the same conditions described above.In longer experiments (exceeding roughly 30 dpi), plants were replenished with BNM as required.

Acetylene reduction assays and plant phenotyping
At ~27 or 45 dpi, soybeans were cut at the hypocotyl with sterile razor blades, and the aboveground tissues were used for measuring the shoot height while the roots were placed inside 250 mL Balch-type serum vials containing ~10 mL of ultrapure water.Vials were sealed using bromobutyl rubber stoppers and aluminum crimp caps, and 10 mL of headspace was removed and replaced with 10 mL acetylene gas (Airgas USA) using a sterile syringe with a 16G hypodermic needle.The ARA reaction start time was recorded immediately after acetylene injection.Vials were then inverted to verify a gas-tight seal was maintained, based on the absence of visible gas bubbles in the water.Vials containing air/water only or non-inoculated plants were also treated with acetylene as controls.
Acetylene-treated plants were maintained overnight in GEN2000 chambers and serially removed for headspace analysis by GC-MS.Gastight syringes with remova ble 22s G/2 inch/point style #2 needles (Hamilton Company) were used to inject 100 µL headspace samples into a TRACE 1300/ISQ 7000 GC-MS equipped with a TracePLOT TG-BOND Q+ column (30 m × 0.32 mm × 10 µm, Thermo Scientific, Cat.no.26005-6030) with an injector temperature of 100°C in splitless mode, with helium as the carrier at constant flow of 1.5 mL/min.The GC oven was held at a constant 60°C.The mass spectrometer was operated in 20-100 amu at 70 eV in EI ionization mode.The MS transfer line temperature was held at 250°C and the ion source temperature at 250°C.
The compound peak areas and ARA reaction stop time were recorded.Acetylene and ethylene were identified by searching the NIST compound library.The ARA reaction duration was hand calculated.The areas of the ethylene peak were calculated and the acetylene reduction rates were calculated as follows: ARA/hour/plant = area of ethylene peak reaction duration (hours) After measuring acetylene reduction, plants were removed from the bottles and the nodules were counted and carefully harvested using a razor blade.The harvested nodules were transferred into pre-weighed Eppendorf tubes and were photographed and then dried at a temperature of 60°C for at least 48 h before weighing to calculate their dry mass.For reporting acetylene reduction per nodule and per milligram of nodule dry weight, the formulas below were used:

Nodule staining and microscopy
Root nodule semi-thin sections (100 µm thickness) were collected using a 7000smz-2 vibratome (Campden Instruments).Brightfield images of nodule sections were taken using a Zeiss Stemi 508 stereo microscope with an Axiocam 305 color camera.For fluorescence images, nodule sections were immediately stained with 5 µM SYTO 9 and 30 µM PI using a LIVE/DEAD BacLight Bacterial Viability Kit (Thermo Fisher Scientific Cat.no.L13152), according to the manufacturer's protocol.After LIVE/DEAD staining, the cells were fixed in PBS containing 3.7% paraformaldehyde (Sigma-Aldrich, SKU F8775) for 30 min at room temperature.Fixed sections were washed 3× in PBS and transferred to a solution of PBS with 0.01% Calcofluor stain (Fluorescence brightener 28 disodium salt solution; Sigma-Aldrich, SKU 910090).Samples were stained at room temperature for 30 min or overnight at and then washed 3× in PBS.Samples were then mounted onto slides in one-well, 20 mm diameter, 0.12 mm deep Secure-Seal spacers (Thermo Fisher Scientific Cat.no.S24736) in Fluoromount-G mounting medium (Thermo Fisher Scientific, Cat.no.E141473) and were sealed with 20 mm square #1.5 coverslips and Seche Vite fast dry nail polish (Amazon).The slides were cured at 4°C overnight prior to imaging.
High-resolution images were taken using a ZEISS LSM 980 confocal microscope with an Airyscan 2 super-resolution module using 10× (0.45 NA) or 63× (1.4 NA) objectives.The images were acquired by brightfield or in fluorescence mode with the following filters for each dye: Calcofluor-405 nm laser excitation and 422-477 nm emission; SYTO 9-488 nm laser excitation and 495-550 nm emission ("LIVE" stain); PI-561 nm laser excitation with 607-735 nm emission ("DEAD" stain).All fluorescence images were acquired using GaAsP-PMT detectors.Quantitative analysis of nodule images was performed using custom ImageJ macros.

RNA-seq library construction and analysis
Three individual colonies of WT and Pcu-shc::shc grown with and without 25 µM cumate were cultured to early exponential phase and total RNAs were extracted by Qiagen RNeasy Mini Kit as described previously (45).RNA quantification was achieved using a Nanodrop one (Thermo Scientific) and RNA quality was monitored with an Agilent 2100 Bioanalyzer using an RNA 6000 Pico Kit (Agilent Technologies).Ribosomal RNA (rRNA) was depleted and cDNA libraries were prepared from 100 ng total RNA by Stranded Total RNA Prep Ligation with Ribo-Zero Plus (Illumina).The quantity and quality of the prepared library were analyzed by the Qubit dsDNA BR Assay Kit and the Agilent 2100 Bioanalyzer using a DNA 1000 Kit, respectively.A library of samples (2 ng/sample) under each condition (30 Mb read/each sample) was loaded on the Illumina Nextseq 500 system with a NextSeq 500/550 High Output Reagent CartriDEG v2 (Illumina) for single-end reads at the Carnegie Embryology Department's in-house sequencing facility.Two independent biological replicates were performed.
Raw data from the two independent experiments were analyzed together by the nf-core/rnaseq pipeline built using Nextflow (46).Based on the gene anno tation of Bradyrhizobium diazoefficiens USDA 110 in EnsemblBacteria (https://bac teria.ensembl.org/index.html),reads were mapped to the genome resulting in a compressed binary version of the Sequence Alignment Map (BAM files) and WIG files for reads visualization.DEG analysis of Pcu-shc::shc compared to WT was performed using DESeq2 in RStudio (47,48).To make accurate comparisons of gene expression between samples, raw counts were normalized by the DESeq2's size factor and batch factor of each sample to correct for variability of the sequencing depth and batch effects from independent experiments.The ratios of normalized counts of each gene in Pcu-shc::shc and WT were computed as a fold change (fc).Genes with differences in expression (false discovery rate [FDR] < 0.05; log 2 fc was 1 or −1) were further used to construct the volcano plots by GraphPad Prism 9.0.The Pierson correlation and heatmap of the top 20 upregulated and downregulated DEGs were constructed by ImageGP fc was 1 or −1) were further used to construct the volcano plots by GraphPad Prism 9.0.The Pierson correlation and heatmap of the top 20 upregulated and downregulated DEGs were constructed by ImageGP (49) (https://www.bic.ac.cn/ImageGP/).Venn diagrams were constructed by ImageGP.The gene ontology (GO) functional enrichment analysis of the upregulated gene sets and downregulated gene sets was performed using PANTHER (50, 51) (http://pantherdb.org/).

Swimming motility assays
Swimming motility assays were performed as previously described with some modifications (52).Three single colonies of each bacterial strain were grown to turbidity (OD 600 = 1.15-1.3) in 5 mL of AG media, then diluted to an OD 600 of 0.02 in 10 mL of fresh AG containing 0.5 g/L of arabinose and gluconate without yeast extract.Cultures were grown to the early exponential phase (OD 600 = 0.25-0.4)and then diluted to an OD 600 of 0.05 in fresh starvation media (AG without yeast extract but with 0.1 g/L arabinose and gluconate).For each bacterial strain, 5 µL of the adjusted cultures were dropped onto the surface of swimming plates (0.3% agarose starvation media) at sites equally distant from each other and from the sides of plates.After inoculation, the plates were incubated in a humidity-controlled (80% relative humidity) environmental chamber at 30°C for 12 days, with daily images taken using a Panasonic Lumix GX85 Mirrorless Camera.Colony expansion was determined by measuring the lengths and widths of the visible bacterial motility halos and calculating the average of the length and width for each colony.

Statistical analyses
Mean values and standard errors (SEM) or median and quantiles with at least three technical replicates were calculated using GraphPad Prism 9.0.Levels of significance were evaluated using either t test or one-way ANOVA, followed by Tukey's multiple comparison tests by GraphPad Prism 9.0.P values larger than 0.05 were considered non-significant (ns).

FIG 1
FIG 1 Validation of a cumate-inducible shc strain Pcu-shc::shc (Pcu-shc) of B. diazoefficiens USDA 110.(A) The first committed step of hopanoid biosynthesis.Squalene-hopene cyclase encoded by shc cyclizes linear squalene (IV) into the C 30 short hopanoids, diploptene (I), and/or diplopterol (II).The tetrahymanol synthase encoded by ths (blr0371) converts diploptene to tetrahymanol (III) (28).(B) qRT-PCR of shc in WT and Pcu-shc in early exponential phase (OD 600 = 0.5-0.8)with varied cumate concentrations.The internal control gene was 16S rRNA and relative expression (RE) of shc was calculated by ∆∆Ct method.(C) Four-day growth curves of WT and Pcu-shc with and without 25 µM cumate in AG media at pH 6.6, 30°C.WT-Cumate and Pcu-shc-Cumate (abbr.WT-C and Pcu-shc-C) denote strains supplemented with an ethanol solvent control or ethanol plus cumate.(D) Total ion chromatograms of lipids extracted from WT and Pcu-shc at early exponential phase by gas chromatography-mass spectrometry (GC-MS).Roman numerals represent the same compounds as in panel A.

FIG 2
FIG 2 Cumate treatment rescues general stress sensitivity of Pcu-shc.Four-day growth curves of WT and Pcu-shc with and without cumate in AG medium under (A) low pH stress at pH 5.0, (B) high-temperature stress at 37°C, (C) high osmotic stress of 5% sucrose, and (D) dual stresses from high temperature and high osmolality.

FIG 3
FIG 3 Pcu-shc is an inefficient soybean symbiont at 27 days post-inoculation (dpi).(A) Comparison of growth of soybeans inoculated with WT and Pcu-shc with and without cumate, respectively (N = 9 plants each).Non-inoculated (NI) soybeans (N = 7 plants) were grown as controls.(B) Median and quantiles of shoot height (cm) per plant in panel A. (C) Images of nodules in 1.5 mL Eppendorf tubes collected from soybean in panel A. Each tube contains all nodules from a single plant.(D and E) Median and quantiles of nodule number and nodule weight (mg) per plant for the treatments of panel A. (F-H) GC-MS quantification

FIG 4
FIG 4 Pcu-shc forms some empty nodules on soybean at 27 days post-inoculation (dpi).(A-H) Representative brightfield images of nodule cross-sections for WT-C (A and B), WT+C (C and ), Pcu-shc-C (E and F), and Pcu-shc+C (G and H).Scale bars on each image represent 1 mm.Regions with orange dashed lines are "patchy" infection zones.Red arrows point to starch granules.(I-P) Representative confocal images of nodule cross-sections infected by WT-C (I and J), WT+C (K and L), Pcu-shc-C (M and N), and Pcu-shc+C (O and P).Sections were stained with Calcofluor (cyan, plant cell wall), SYTO 9 (yellow, "live" bacteria), and propidium

FIG 5
FIG 5 Pcu-shc transcriptomics suggests defects in motility and chemotaxis and enhanced nitrogen metabolism and protein secretion.(A) Read coverage of the shc genomic region in WT and Pcu-shc with and without cumate.(B and C) Venn diagrams of upregulated and downregulated differential expression genes.(D and E) Volcano plots of all genes based on read fold change (fc) and false discovery rate (FDR) of Pcu-shc-C compared to WT-C and Pcu-shc+C compared to Pcu-shc-C.The red dots represent upregulated genes (log 2 fc ≥ 1, FDR < 0.05), the blue dots represent downregulated genes (log 2 fc ≤ −1, FDR < 0.05), and the gray dots represent non-differentially expressed genes (−1 < log 2 fc < 1, FDR ≥ 0.05) and false-positive DEGs due to cumate and construct effects.The genes with asterisks are confirmed by qRT-PCR in TableS9).(E and F) GO term analysis of downregulated (F) and upregulated (E) in biological process categories.

FIG 6
FIG 6 Pcu-shc displays compromised flagellar motility.Swimming motility of WT and Pcu-shc with and without cumate in starvation medium agarose plates (AG without yeast but with 0.1 g/L arabinose and gluconate).(A) WT and Pcu-shc inocula grown with and without cumate on starvation medium plates from 2 to 12 dpi.(B) The mean ± standard errors (SEM) of diameters of colonies from the four treatments in (A).N = 3 technical replicates for each treatment.