Rhizobia are a group of bacteria that increase soil nitrogen content through symbiosis with legume plants. The soil and symbiotic host are potentially stressful environments, and the soil will likely become even more stressful as the climate changes. Many rhizobia within the Bradyrhizobium clade, like Bradyrhizobium diazoefficiens, possess the genetic capacity to synthesize hopanoids, steroid-like lipids similar in structure and function to cholesterol. Hopanoids are known to protect against stresses relevant to the niche of B. diazoefficiens. Paradoxically, mutants unable to synthesize the extended class of hopanoids participate in symbioses with success similar to that of the wild type, despite being delayed in root nodule initiation. Here, we show that in B. diazoefficiens, the growth defects of extended-hopanoid-deficient mutants can be at least partially compensated for by the physicochemical environment, specifically, by optimal osmotic and divalent cation concentrations. Through biophysical measurements of lipid packing and membrane permeability, we show that extended hopanoids confer robustness to environmental variability. These results help explain the discrepancy between previous in-culture and in planta results and indicate that hopanoids may provide a greater fitness advantage to rhizobia in the variable soil environment than the more controlled environments within root nodules. To improve the legume-rhizobium symbiosis through either bioengineering or strain selection, it will be important to consider the full life cycle of rhizobia, from soil to symbiosis.
IMPORTANCE Rhizobia, such as B. diazoefficiens, play an important role in the nitrogen cycle by making nitrogen gas bioavailable through symbiosis with legume plants. As climate change threatens soil health, this symbiosis has received increased attention as a more sustainable source of soil nitrogen than the energy-intensive Haber-Bosch process. Efforts to use rhizobia as biofertilizers have been effective; however, long-term integration of rhizobia into the soil community has been less successful. This work represents a small step toward improving the legume-rhizobium symbiosis by identifying a cellular component—hopanoid lipids—that confers robustness to environmental stresses rhizobia are likely to encounter in soil microenvironments as sporadic desiccation and flooding events become more common.


The soil is a precious ecosystem. The health of the soil—measured by organic matter and nutrient content, moisture retention, and the microbial community—predicts how well plants will grow (1). While practices to maintain healthy soil have been known for centuries, as agriculture faces the threat of climate change, these sustainable land management practices have garnered new attention as potential climate change mitigation strategies (2, 3). Crop rotation is an ancient land management strategy that restores nutrients to the soil. Legumes, such as soybean, peanut, and alfalfa, are an important component of crop rotations because they increase soil nitrogen content, reducing reliance on synthesized nitrogen fertilizers (4). However, legumes cannot do this alone: they rely on symbiosis with a group of polyphyletic soil bacteria called rhizobia (5, 6). This symbiosis is a very close interaction, with the bacteria living intracellularly in specialized de novo organs called root nodules. Low pH, low oxygen, and elevated osmolarity are maintained within the nodule environment that the bacteria experience between the plant-derived membrane and the bacteria, called the symbiosome space (79). This environment favors bacterial conversion of nitrogen gas to bioavailable ammonia which eventually is exchanged for reduced carbon in the form of dicarboxylic acids.
One way to improve legume use in crop rotations as a sustainable nitrogen fertilizer is to improve the efficiency of the legume-rhizobium symbiosis. Rhizobial strains with greater symbiotic efficiency, as measured by nitrogen fixation rates and legume growth, have been isolated and applied to the soil or to legume seeds. This strategy has been used successfully at scale with legume crops such as soybean in Brazil (1012). However, rhizobia often fail to stably integrate into the soil community, so these inoculations must be repeated each year (1316). Beyond the symbiosis itself, rhizobia can have positive effects on the plant when living in the soil, such as relieving salinity stress and increasing water and nutrient uptake (1719). These positive effects, which can be observed yearly as the seasons change, will become even more important as the climate changes, leading to drastic changes in precipitation and thus soil water potential and osmotic strength. To successfully use legumes in crop rotations to increase soil nitrogen content as the climate changes, rhizobia must be successful in both the symbiosis and surviving the soil environment. Notably, these bacteria are abundant in a variety of soil types (20).
As recently discussed, possession of an adaptable outer membrane may provide an important fitness advantage in soils (9). Rhizobia produce modified lipid A, a major component of the lipopolysaccharides (LPS) that make up the outer leaflet of the outer membrane (21, 22). These modifications trend toward increasing the hydrophobicity and other cohesive interactions that lead to a more robust outer membrane (23, 24). Additionally, a subset of rhizobia, mostly within the Bradyrhizobium clade, make hopanoids, a class of sterol-like lipids, which maintain resistance to environmental stressors such as pH, temperature, and antibiotics and foster successful symbioses (25, 26). Some members of Bradyrhizobium also attach hopanoids to lipid A, which appears to act as a hydrophobic hook into the inner leaflet of the outer membrane (2729). This hopanoid attached to lipid A (HoLA) is synthesized from extended hopanoids, a subclass of hopanoids with an added hydrophilic tail (Fig. 1) (25).
FIG 1 Hopanoid biosynthesis and HoLA structure. The biosynthesis of hopanoids is shown on the left, from squalene to the unextended hopanoids (magenta) to the extended hopanoids (green). All of these hopanoids may be methylated at the C-2 position by the hopanoid methylase HpnP. These hopanoids are transferred to the outer membrane by the HpnN hopanoid transporter. Extended hopanoids are attached to the very-long-chain fatty acid (yellow) on lipid A to create HoLA, which also resides in the outer membrane.
Intriguingly, generating a mutant that is unable to make any type of hopanoid by removing the first committed step in hopanoid biosynthesis (Δshc) has evaded realization in Bradyrhizobium diazoefficiens, which suggested an essential role for hopanoids in this strain (25). Removing the ability to synthesize C35 or “extended” hopanoids (ΔhpnH), however, was achieved, and it has a large effect on the fitness of B. diazoefficiens in culture (25, 30). The ΔhpnH mutant manifests growth defects at high osmolarity and is unable to grow under low-pH or microaerobic conditions—all conditions thought to characterize the symbiosome space. While the ΔhpnH mutant exhibits defects in planta, especially in root nodule initiation, the nitrogen fixation rate in symbiosis with the tropical legume Aeschynomene afraspera when normalized for nodule dry weight is not significantly different between the ΔhpnH mutant and the wild type (WT) and the majority of ΔhpnH-infected nodules grow at rates comparable to those for the WT (30). Given the growth defects of the ΔhpnH mutant observed in the presence of environmental stresses expected within the symbiosome space, these results were surprising. In this study, we investigated this paradox by exploring the nuanced interplay between the lack of extended hopanoids and an environmentally relevant concentration range of osmolytes and cations.


Hopanoids are conditionally essential in B. diazoefficiens.

Due to the previous difficulties isolating an shc deletion strain in B. diazoefficiens, we constructed a conditional squalene-hopene cyclase (SHC) expression strain in a Δshc background. The conditional expression system employed a cumate repressor system that was modified for B. diazoefficiens, where cumate relieves repression of transcription of the gene of interest (see Fig. S1A in the supplemental material). We tested the growth of this strain and a mutant deficient in producing extended (C35) hopanoids, the ΔhpnH strain, on plates made up of different media (peptone-salts-yeast extract medium with 0.1% arabinose [PSY] and arabinose-gluconate medium [AG]) commonly used to culture rhizobia (Fig. S1B). As expected, the cumate conditional SHC expression strain was unable to grow or form single colonies on PSY medium without cumate present. But to our surprise, when the strain was grown on AG medium plates, growth was restored in the densest part of the streak plate in the absence of cumate; similarly, the ΔhpnH strain grew much better on AG than on PSY. This unexpected qualitative finding led us to hypothesize that hopanoids might not be as important under the AG condition.
While there are many altered components between the two media (Table 1), two aspects that stood out to us were the differences in osmolarity and the differences in divalent cation concentrations. Osmolarity is thought to be elevated in the symbiosome space and can span a wide range in the soil environment (9). Divalent cations are known to stabilize the outer membrane through interactions with LPS, especially calcium (23, 31). The medium osmolarity was 16 mM less and the divalent cation concentration was almost doubled in PSY compared to the AG medium (Table 1), suggesting that hopanoids are necessary to withstand certain levels of osmolytes and/or ionic strength in B. diazoefficiens. Despite this realization, we continued to struggle with obtaining an unmarked, in-frame shc deletion mutant in AG medium, ironically due to the fact that the sucrose selection method we were using to generate the mutant strain (32) likely exposed it to an osmotic stress beyond the threshold it could tolerate. However, we were able to complement the hpnH mutant by putting the hpnH gene under the control of the PaphII promoter integrated at the scoI locus (Fig. S1B), as has been done to constitutively express other genes in B. diazoefficiens previously (33).
TABLE 1 Differences between PSY and AG media
MediumBufferDivalent cationsCarbon sourcesOsmolarity (mOsM)pH
PSY4.3 mM phosphate45 μM CaCl2, 400 μM MgSO4Arabinose, yeast extract, peptone53 ± 0.067
AG5.5 mM HEPES, 5.6 mM MES90 μM CaCl2, 730 μM MgSO4Arabinose, yeast extract, sodium gluconate69 ± 1.26.6
Accordingly, to test how the medium composition affects B. diazoefficiens strains lacking extended hopanoids and what this reveals about the ecophysiological role of extended hopanoids more generally, we proceeded to focus on the ΔhpnH strain for the remainder of our experiments.

Extended hopanoids protect B. diazoefficiens in stationary phase and at low pH.

We began by carefully examining the growth dynamics of our mutant strain in the different media. In PSY, the ΔhpnH strain has a pronounced defect in both exponential and stationary phases (25). Even when ΔhpnH and WT cultures were sampled at the same optical density at 600 nm (OD600) in exponential phase, the ΔhpnH strain had drastically lower viability than the WT. We reasoned that this might be due to initial inoculum viability differences from “overnight” cultures. To test and preempt this, we subcultured twice from an initial turbid culture inoculated with colonies from a fresh plate. Using this technique, we observed that the ΔhpnH mutant strain had only a very slight growth defect compared to the WT and the ΔhpnH complement in the pH 6.6 AG medium (Fig. 2A). To confirm that the similarity in OD600 is due to a comparable number of viable cells, we measured CFU and OD600 after 24 h (mid to late exponential phase) and 72 h (stationary phase) of growth. The OD600 measurements were similar for the two strains at each time point and the CFU were similar at 24 h, but at 72 h, the CFU for the ΔhpnH strain were significantly lower than for the WT and the ΔhpnH complement (Fig. 2B and C). This result supports a broad stationary-phase defect across different media and that our culturing method successfully removed differences in inocula that influenced previous experiments. Using this approach, we tested the ΔhpnH strain’s growth in the pH 5 AG medium (Fig. 2C). The pH within the symbiosome space and in the legume rhizosphere is known to be acidic (7, 34), and previously, the ΔhpnH strain was shown to be unable to grow at pH 6 in PSY medium (25). In pH 5 AG medium, the ΔhpnH strain was able to grow but with a significant defect in growth rate and stationary phase compared to those of the WT and the ΔhpnH complement. Upon inspection, clumping of the ΔhpnH strain was observed in the wells of the plate, perhaps indicating cell death and increased biofilm formation.
FIG 2 B. diazoefficiens ΔhpnH strain is sensitive to low pH and stationary phase. (A and D) Growth of the WT, the ΔhpnH complement, and the ΔhpnH strain in AG medium at pH 6.6 and pH 5 was monitored at an optical density of 600 nm (OD600). Each curve represents the average of three biological replicates. (B and C) CFU per milliliter and OD600 were measured for the WT, the ΔhpnH complement, and the ΔhpnH strain grown in AG medium at pH 6.6 during exponential phase (24 h) and stationary phase (72 h). Error bars (standard deviation) are included (A to D), but some are smaller than the point markers. Statistical significance was determined by unpaired two-tailed t tests.

Physicochemical medium conditions affect the growth of B. diazoefficiens ΔhpnH.

Having discovered that growth defects are condition dependent for the hopanoid-deficient Δshc strain, we decided to check whether this was also true for the growth defect of the ΔhpnH strain in pH 5 AG medium. First, we tested the effects of osmolarity by adding the nonionic osmolyte inositol (Fig. 3). In our study, the WT was unable to grow on inositol alone, indicating that it cannot be used for energy generation (see Materials and Methods). Furthermore, inositol has been used previously to modify osmolarity with B. diazoefficiens (25). The growth rate and lag time parameters were estimated using a Gompertz model. As the concentration of inositol was increased from 25 mM to 400 mM, the growth rate decreased for the WT and the ΔhpnH complement (Fig. 3B). The behaviors of the WT and the ΔhpnH complement strains were not identical, perhaps due to the use of a high-expression promoter instead of the native promoter. Nonetheless, the strains followed similar trends in growth rate and lag across all conditions tested, as shown in Fig. 3B and 4B. In contrast, the growth rate for the ΔhpnH strain increased to a maximum growth rate with 100 mM inositol added, before decreasing. The ΔhpnH strain never reached the same growth rate as the WT, but the differences in the growth rate response over this osmolarity range illustrates that the strains experience these conditions very differently and that a “Goldilocks” osmotic zone exists for the ΔhpnH mutant where its growth is enhanced. A similar trend was observed with lag time, where the ΔhpnH strain exhibited a minimum lag time with 100 mM inositol added. The lag time of the WT and the ΔhpnH complement remained relatively constant up to 100 mM inositol added and then steadily increased. These experiments were also completed with sorbitol as the osmolyte for the WT and the ΔhpnH strain. Sorbitol has been used as an osmolyte previously with this strain (35), and B. diazoefficiens was unable to grow on sorbitol alone in our study (see Materials and Methods). The results with sorbitol were similar to those with inositol, but with more variation, especially in stationary phase, likely indicative of B. diazoefficiens’s ability to catabolize sorbitol (Fig. S2).
FIG 3 B. diazoefficiens ΔhpnH strain growth is sensitive to the concentration of inositol. (A) Growth of the WT, the ΔhpnH complement, and the ΔhpnH strain in AG medium at pH 5 with increasing concentrations of inositol was monitored at OD600. The grayscale of the symbols is darker with increasing concentrations of inositol. (B) Growth rate (μ) and lag time were quantified by fitting a single Gompertz curve to each growth curve from panel A. The results are plotted according to increasing concentrations of inositol. Arrows point to the concentration of inositol (100 mM) where the growth of the ΔhpnH strain was optimized. Error bars (standard deviation) are included, but some are smaller than the point markers.
FIG 4 B. diazoefficiens ΔhpnH strain growth is sensitive to the concentration of calcium. (A) Growth of the WT, the ΔhpnH complement, and the ΔhpnH strain in AG medium at pH 5 with different concentrations of divalent cations was monitored at OD600. The grayscale of the symbols is darker with increasing concentrations of Ca2+ ions. (B) Growth rate and lag time were quantified by fitting a single Gompertz curve to each growth curve from panel A. The results are plotted according to increasing concentration of Ca2+ ions with the low divalent cation condition included at y = 0. All growth curves and quantifications represent the averages of three biological replicates. Error bars (standard deviation) are included, but some are smaller than the point markers.
Next, we tested the effects of divalent cations on growth. Because the AG medium already has almost double the divalent cation concentration as the PSY medium, we created a “low-divalent-cation” pH 5 AG medium using the PSY concentrations of divalent cations, specifically, magnesium and calcium (45 μM Ca2+ and 400 μM Mg2+). We then increased the calcium ion concentration in this low-divalent-cation AG medium. The growth rate and lag time for the WT and the ΔhpnH complement were unaffected by these changing conditions (Fig. 4). However, the ΔhpnH strain grew at a lower rate and with a longer lag time in the low-divalent-cation AG medium, indicating that 45 μM may be particularly stressful for the ΔhpnH strain. The growth rate recovered with 100 μM additional calcium and remained the same. The lag time decreased as well when the calcium concentration was increased. When 500 μM additional calcium was added to the low-divalent-cation medium, the growth curve for the ΔhpnH strain became almost indistinguishable from that with the pH 5 AG medium. These experiments were also completed with magnesium as the divalent cation for the WT and the ΔhpnH strain, with similar but less pronounced results; growth rate remained the same, while lag decreased but only slightly (Fig. S3).
Together, these results indicate that the lack of extended hopanoids can be compensated for by changing the physicochemical properties of the growth medium, specifically, the osmolarity and divalent cation concentration. Unlike the pattern seen with osmolytes for the ΔhpnH strain, where an intermediate concentration minimized lag time and increased growth rate, increasing concentrations of divalent cations increasingly shrank the lag time yet did not appreciably affect the growth rate.

Extended hopanoids are required to regulate the membrane properties of B. diazoefficiens.

In order to explain the growth phenotypes of the ΔhpnH strain, we hypothesized that extended hopanoids play a role in protecting the outer membrane against physicochemical perturbations by maintaining membrane properties such as lipid packing, which refers to the lateral density of lipids within the bilayer. Because lipid packing is correlated with viscosity and bilayer stability, it provides a robust general indicator of changes in membrane biophysical properties (36, 37). Indeed, the unextended hopanoid diplopterol is known to modulate changes in lipid A packing in culture that occur in response to decreased pH (38). To probe the underlying mechanism behind the physicochemical compensation for loss of extended hopanoids, we used the lipophilic dye di-4-ANEPPDHQ (di-4) (Fig. 5A). The general polarization (GP) of di-4 reflects lipid packing, with higher GPs indicative of increased packing. Because of its size (molecular weight [MW], 665.55) and polarity, di-4 should preferentially label the outer leaflet of the outer membrane, and it has been routinely used to monitor changes in surface membrane lipid packing (39). However, given the interplay between membrane stability and permeability, the assumption that di-4 selectively labels the outer leaflet of the outer membrane may not hold if membrane permeability is increased sufficiently to allow di-4 to cross the outer membrane. Outer membranes contain saturated lipid A and have been shown to have lipid packing similar to that of liquid-ordered-phase membranes (40), whereas inner membranes comprise more disordered phospholipids, which should have lower lipid packing similar to that in a liquid disordered phase. Thus, a large decrease in di-4 GP could be interpreted as either a decrease in outer membrane lipid packing or a large increase in outer membrane permeability allowing di-4 to label the inner membrane. Both results would indicate a large change in the mechanical properties of the outer membrane.
FIG 5 B. diazoefficiens ΔhpnH strain is deficient in its ability to regulate its membrane properties. (A) Lipid packing measured by di-4 general polarization (GP) index for the WT, the ΔhpnH complement, and the ΔhpnH strain grown in AG medium at pH 6.6, pH 5, and pH 5 with 100 mM inositol added. Individual measurements are shown as black circles. ΔGP for each strain compared to the AG pH 5 condition is shown. Error bars (standard deviation) are included. (B) Relative cell envelope permeability measured by a fluorescein diacetate diffusivity assay of the WT, the ΔhpnH complement, and the ΔhpnH (white) strain grown in AG medium at pH 6.6, pH 5, and pH 5 with 100 mM inositol added. Individual measurements are shown. Error bars (standard deviation) are included. Statistical significance was determined by unpaired two-tailed t tests.
To evaluate the role of extended hopanoids in outer membrane acclimation to pH and osmotic strength, we compared di-4 GP in cells grown at pHs 6.6 and 5 and in cells grown at pH 5 in the presence and absence of inositol (Fig. 5). The di-4 GP values for the WT and the ΔhpnH complement strains were almost identical across the three conditions (pH 6.6 AG medium, pH 5 AG medium, and pH 5 AG medium plus 100 mM inositol), which indicates that the outer membrane of our complement strain responded very similarly to the WT. The ΔhpnH strain had lower GP values than the WT and the ΔhpnH complement when grown in pH 6.6 AG medium. Additionally, all three strains had greater GP values when grown in pH 5 AG medium, which is consistent with observations that low pH increases the packing and order of LPS (38, 41). However, the difference is only significant for the ΔhpnH strain. Interestingly, the WT and the ΔhpnH complement showed a small decrease in GP values in AG pH 5 medium with 100 mM inositol added, while the ΔhpnH strain GP values continued to increase.
To better interpret these results, we determined the ΔGP values, comparing the change in GP compared to the standard pH 6.6 AG medium condition (Fig. 5). The WT and the ΔhpnH complement underwent very small changes compared to the ΔhpnH strain. In addition to exhibiting comparably more GP variability with changing osmolarity, the ΔhpnH strain showed a relatively large negative shift of around 0.3 GP units at pH 6.6. Such a large negative shift in GP suggested a considerable change in mechanical properties of the outer membrane at higher pH. To determine whether this might be the result of compromised outer membrane integrity, we examined whether cell envelope permeability of the ΔhpnH strain at pH 6.6 was higher than for other conditions and strains. We estimated changes in membrane permeability using an assay based on relative changes in fluorescein diacetate (FDA) diffusivity (42). FDA is nonfluorescent and rapidly diffuses into the cell, where it is hydrolyzed. Fluorescein, which is fluorescent, is produced from the hydrolysis of FDA. Additionally, because fluorescein is charged, it cannot diffuse rapidly out of the cell. Therefore, the rate of increase in fluorescein fluorescence can provide an estimate of the relative permeability of FDA across the cell envelopes of different strains or across various growth conditions. We observed that the ΔhpnH strain had 2.3-fold-higher permeability than the WT at pH 6.6 and 8.3- and 6.6-fold-higher permeability than it showed at pH 5 with and without inositol, respectively (Fig. 5B). The large negative shift in di-4 GP values at pH 6.6 therefore is consistent with both reduced lipid packing and compromised membrane integrity. Overall, these results show that extended hopanoids play an important role in guarding B. diazoefficiens membrane properties against physicochemical perturbations.


By making a serendipitous observation of different growth phenotypes of B. diazoefficiens hopanoid mutants on two types of media, we discovered that the physicochemical environment strongly impacts growth of hopanoid-deficient B. diazoefficiens strains, including the ΔhpnH strain, an extended-hopanoid-deficient mutant. We confirmed that the ΔhpnH strain undergoes significant death in stationary phase but that in contrast to previous results (25), it can grow at pH 5 in a medium with higher osmolarity and divalent ion concentration. We identified environmentally relevant conditions that partially compensate for the B. diazoefficiens ΔhpnH mutant growth defect at low pH: intermediate osmolarity and elevated divalent cation concentrations. Finally, using biophysical techniques, we discovered that extended hopanoids are important for modulating lipid packing of the outer membrane.
When we first discovered that our conditional Δshc mutant could partially grow on solid AG medium without cumate-induced hopanoid production but not on PSY medium, it caused us to reexamine our earlier results with the ΔhpnH strain in PSY (25). As previously shown, the ΔhpnH strain has a stationary-phase defect and increased lag time when grown in PSY. When additional stress was added to the PSY condition, such as increased temperature, lowered pH, or microoxia, the ΔhpnH strain was unable to grow at all. It is possible that due to its stationary-phase defect under these conditions, the ΔhpnH strain inoculum in these experiments may have had fewer viable cells than the WT, despite similar optical density measurements, contributing to the severity of these phenotypes. With this in mind, our revised inoculation protocol enabled us to see that the ΔhpnH strain grows very similarly to WT in AG medium and can even grow at pH 5. Notably, the ΔhpnH strain experienced significant death in stationary phase compared to the WT in AG medium, despite the optical density measurements remaining constant. These results highlight the importance of not relying on optical density to assess bacterial viability, as they can be decoupled, a known but often overlooked phenomenon.
As we tried to understand why the two media affected the growth of our hopanoid-deficient mutants differently, we reflected on their compositions. Differences in pH, divalent cations, and osmolarity stood out to us. Previous research on Δshc mutants of other closely related bacteria have shown that hopanoids are important under both acidic and basic conditions. Specifically, a Δshc mutant of Rhodopseudomonas palustris failed to grow as it made the medium more basic through amino acid metabolism of the complex medium (43, 44). Since both PSY and AG media are complex, B. diazoefficiens increases the pH of these media as well. However, the AG medium also contains a higher buffering capacity at a lower pH than PSY (6.6 versus 7), thus likely extending the time before the pH is increased substantially. Therefore, the conditional Δshc mutant’s lack of growth on PSY medium and partial growth on AG medium is perhaps not surprising and illustrates how important medium composition may be when isolating and growing a mutant deficient in production of hopanoids or another membrane component.
The higher concentration of divalent cations, specifically, magnesium and calcium, in the AG medium than in PSY was interesting because of the role these cations play in outer membrane cohesion. Specifically, magnesium and calcium intercalate within the LPS layer, shielding negatively charged residues, resulting in a more ordered and robust outer membrane in the face of physicochemical stressors (23, 31). Our two hopanoid-deficient mutants cannot make HoLA, a component of LPS that contributes to membrane ordering (27, 45). Recent work in Bradyrhizobium BTai1 has revealed that calcium ions increase membrane bilayer thickness, an indication of increased membrane order, in membrane vesicles containing LPS without a hopanoid attached (45); this phenomenon suggests a mechanism whereby calcium may be able to compensate for lack of hopanoids. Our work shows that the lag time of the ΔhpnH strain in pH 5 AG medium can be reduced by increasing concentrations of divalent cations, supporting this hypothesis. Calcium has a greater effect than magnesium, likely reflecting the fact that calcium ions more strongly increase lipid bilayer rigidity through dehydration effects than do magnesium ions (46). The WT was unaffected by these changes in divalent ion concentration, perhaps due to the presence of HoLA, unlike other Bradyrhizobium strains (47). Interestingly, while calcium is maintained at low concentrations (0.1 μM) in the cytosol of plant cells, calcium has been shown to localize to the symbiosome space (48); sufficient calcium is needed to support bacterial nitrogen fixation (48, 49). That calcium is present and functionally important within the acidic symbiosome space (7) helps contextualize why the ΔhpnH mutant functions reasonably well in planta: elevated calcium levels may help compensate for the loss of extended hopanoids within this environment.
We were surprised when we found that the AG medium has a higher osmolarity than the PSY medium, since our previous studies with PSY medium indicated that hopanoids can be protective against hyperosmotic stress (25). Yet, as previously noted, these media are compositionally different in more than one way. It thus appeared possible that at a lower pH, the relationship between osmolarity and hopanoids might be more nuanced. We hypothesized that hypoosmolarity might also be stressful for hopanoid-deficient mutants that have less robust membranes, and our findings bore this out. When we added inositol to the pH 5 AG medium, the ΔhpnH strain grew better, decreasing the lag time and increasing growth rate with up to 100 mM inositol added. Comparatively, the WT grew more poorly upon even the smallest addition of inositol (25 mM). This result illustrates that the low osmolarity of the medium is particularly stressful to the ΔhpnH strain. Interestingly, inositol was found in the symbiosome space (50). While more research is needed to understand the physical and chemical properties of the internal nodule environment, one study found that the symbiosome space contains approximately 180 mM low-molecular-weight compounds (50), notably similar to the maximally restorative osmolarity in our experiments (100 mM inositol in pH 5 AG medium). It is thus possible that the symbiosome space microenvironment allows the ΔhpnH strain to survive and fix nitrogen over time, despite its obvious growth defects at low pH. Intriguingly, a recent study has shown that trehalose, a similar sugar-compatible solute produced by B. diazoefficiens, is necessary for an efficient symbiosis (30, 35). Beyond the potential for osmotic rescue of the ΔhpnH strain, it is also possible that the membrane of the ΔhpnH strain is complemented within the root nodule, perhaps by taking up plant sterols.
Interestingly, while both osmolarity and divalent cation concentrations affect the growth of the ΔhpnH strain at low pH, the effects are different. Specifically, divalent cation concentration had the greatest effect on lag time, while the added osmolytes affected both lag time and growth rate. These differences suggest that different mechanisms underpin the mutant’s response, despite both having the potential to rigidify the outer membrane. We hypothesize that these differences may arise due to the inositol primarily addressing the root cause of the stress, hypoosmolarity, while the increase in divalent cations protects against the effects of hypoosmolarity.
The sensitivity of hopanoid-deficient strains to specific external conditions suggested that extended hopanoids might help cells respond to environmental changes. In contrast to diplopterol, a shorter hopanoid that contains a hydrophilic group, which has been shown to rigidify the membrane while keeping lipids from entering a gel phase and retaining lateral lipid diffusivity (38, 51), extended hopanoids had only been shown to rigidify the membrane (25, 52, 53). Our biophysical experiments confirm that extended hopanoids are necessary for membrane rigidification but also reveal that the lack of extended hopanoids causes greater problems with membrane stability. The ΔhpnH strain displays much greater variability in lipid packing between conditions than the WT, as evidenced by the larger ΔGP values for the mutant. This result suggests that the WT can adjust its lipid packing to maintain a relatively constant membrane fluidity and is intrinsically more mechanically stable. In contrast, the ΔhpnH strain struggles to adjust its lipid packing in response to changes in osmolarity, and membrane integrity is compromised by changes in pH. The lipid packing of the ΔhpnH strain is primarily affected by the external environment. In the case of the pH 5 AG medium with 100 mM inositol added, increased osmolarity, inositol, and sugars more generally are known to rigidify membranes (5457), explaining the increased GP values for the ΔhpnH strain. On the other hand, the WT can adjust its membrane to counteract environmentally triggered membrane rigidification, thus leading to slightly lower GP values. Overall, these results indicate that extended hopanoids play an important role in B. diazoefficiens adjustment to and fortification against the external environment.
In conclusion, the lack of hopanoids, specifically, extended hopanoids—which are required for HoLA biosynthesis—makes B. diazoefficiens particularly sensitive to environmental conditions in ways that are relevant to its life cycle. That the lack of extended hopanoids can be partially compensated for by a moderately high osmotic level helps to resolve the paradox of why the ΔhpnH mutant can be symbiotically successful if given sufficient time to develop within root nodules. Yet the mutant’s sensitivity to hypoosmotic conditions suggests that hopanoids may provide a fitness advantage to rhizobia in waterlogged soils, where osmolytes and divalent cations are diluted. Together, our findings emphasize the importance of considering the full ecophysiological picture when attempting to understand the selective benefits of a given molecular component to an organism.


Bacterial strains, culture media, and chemicals.

All strains used in this study are described in Table S1 in the supplemental material. All strains were grown aerobically with shaking at 250 rpm. Escherichia coli strains were grown in lysogeny broth (LB) at 37°C (58). B. diazoefficiens strains were grown at 30°C in either peptone-salts-yeast extract medium with 0.1% arabinose (PSY) (59, 60) or arabinose-gluconate medium (AG) (described in the supplemental material) (61, 62). The pH of the AG medium was adjusted to pH 6.6 using an NaOH solution or to pH 5 using an HCl solution. For pH 5 AG medium, the HEPES buffer component was replaced with 5.5 mM 2-(N-morpholino) ethanesulfonic acid (MES) buffer for a total of 11 mM MES. A low-divalent-cation pH 5 AG medium was made containing 45 μM CaCl2 and 400 μM MgSO4. Inositol, sorbitol, CaCl2, or MgCl2 was added to the appropriate base medium. For induction of the cumate-inducible promoter, cumate was added to liquid or solid medium for a final concentration of 25 μM from a 400× stock solution in ethanol (6366). Agar plates were made containing 1.5% (wt/vol) agar. Antibiotics were used for selection at the following concentrations (in micrograms per milliliter): spectinomycin (Sp), 100; kanamycin (Km), 100; and tetracycline (Tc), 20 for liquid cultures of B. diazoefficiens and 50 for plates for B. diazoefficiens and for E. coli. The following chemicals were purchased from Sigma-Aldrich unless otherwise noted: glycerol (VWR), HEPES (Gold BioTechnology), sodium chloride yeast extract, magnesium sulfate, magnesium chloride, sorbitol, sodium hydroxide (Fisher Scientific), arabinose (Chem-Impex International, Inc.), sodium sulfate (Mallinckrodt Chemical), peptone, and agar (Becton, Dickinson).

DNA methods, plasmid construction, and transformation.

All plasmid constructions and primers used in this study are described in Table S1 in the supplemental material. Standard methods were used for plasmid DNA isolation and manipulation in E. coli (67). The strong constitutive promoter Prrn-mut2 (68) was annealed from oligonucleotides Prrn-mut2_f and Prrn-mut2_r and cloned into HpaI/BsrGI-digested pQH2, resulting in plasmid pQH2-Prrn-mut2. The resulting inducible system (containing Pbla-mut1T-cymR* and Prrn-mut2 flanked by cuO) was subsequently excised with SpeI and PciI and ligated into pRJPaph-lacZYA prepared with SpeI and NcoI, resulting in plasmid pRJPcu1-lacZYA. The PCR product of the shc gene was cloned into the pRJPcu1 plasmid to obtain an expression plasmid. The pRJPcu-shc plasmid was mobilized into the WT, followed by the pGK302, the markerless deletion vector to delete shc (blr3004) (69). Plasmids were mobilized by conjugation from E. coli S17-1 into B. diazoefficiens strains as previously described, with the following modifications (70). The pRJPcu-shc plasmid was stably integrated as a single copy into the scoI downstream region of B. diazoefficiens (leaving scoI unaffected) as described previously (33).

Induction conditions and reporter activity measurements.

Cultures were grown in PSY to mid-exponential phase and induced with 25 μM (final concentration) cumate or pure ethanol (the solvent for cumate) for controls. For quantitative LacZ assays, cells were centrifuged (5,000 × g) and washed twice. β-Galactosidase assay was done as previously described (58). One biological replicate was defined as an independent culture; each replicate was assayed in technical duplicates of which the arithmetic mean was used for final data plotting.

Streaking strains.

Liquid cultures were grown from plates to early stationary phase (OD, 0.9 to 1.2 as determined by a Beckman Coulter UV-visible [UV-Vis] spectrophotometer) in AG medium. The cultures were spun down and resuspended to an OD600 of 0.5 in fresh medium. Ten microliters of culture was spotted onto each plate and spread using a sterilized spreader.

Osmometer measurements.

The osmolarity of PSY and AG media were measured using a Wescor Vapro 5520 vapor pressure osmometer. Before use, the osmometer was calibrated using 100 mM, 300 mM, and 1,000 mM OptiMole standards from ELITechGroup.

B. diazoefficiens pregrowth.

Five milliliters of fresh medium was inoculated with multiple colonies per strain picked using a sterile stick. After 2 to 3 days, when the cultures were turbid (OD600, >1), these cultures were subcultured into 5 mL of fresh medium and allowed to grow to mid to late exponential phase (OD600, 0.5 to 1). A sample from this mid- to late-exponential-phase culture was then inoculated into fresh medium. After this culture reached mid to late exponential phase, it was used as an inoculum for experiments (unless noted otherwise).

Testing growth on inositol and sorbitol.

Modified AG medium was prepared by removing the yeast extract, arabinose, and sodium gluconate with 51 mM carbon composed of either inositol or sorbitol. WT B. diazoefficiens was grown on AG plates, and then 5 mL of fresh AG medium was inoculated and multiple colonies per strain were picked using a sterile stick. After 2 to 3 days, when the culture was turbid, a sample was spun down and resuspended in AG media without any carbon source. This sample was used to inoculate multiple tubes of modified AG media. These cultures were monitored for 1 week without observing any growth. This experiment was completed twice.

Growth curves.

The growth curve assays were performed in 96-well tissue culture plates (Genesee Scientific) using a Spark 10M multimode microplate reader (Tecan, Grödig, Austria). Wells were topped with 50 μL of autoclaved mineral oil. Optical density absorbance was taken at 600 nm at 30-min increments and 30°C with continuous linear shaking.

Growth curve parameter estimation.

To estimate the maximum specific growth rate (μm) and lag time (λ) for the growth curves, the data from each growth curve were fitted using an R application that relies on nonlinear least squares to fit nonlinear models to the following Gompertz curve equation (71, 72):
OD600=A×exp{exp[μmA(λ − t) + 1]} + C
where A is the final OD600, t is the time in hours, and C is an adjustment for initial OD600. For some of the growth curves, any apparent death phase was removed for analysis because the application was not made to handle the death phase. The following start parameters were used: A = 0.86, lag = 8, μm = 0.11, and C = 0.065. The bounds were kept at the preset parameters. The R application can be found at the following GitHub repository: https://github.com/scott-saunders/growth_curve_fitting. The specific version of the growth curve fitting R application used was retrieved on 24 April 2022 and can be found here: https://github.com/scott-saunders/growth_curve_fitting/blob/master/growth_curve_fitting_ver0.2.Rmd. The following is a direct link to the application that can be run locally: https://scott-h-saunders.shinyapps.io/gompertz_fitting_0v2/#section-parameter-estimates.

Viable-cell plate counts.

Viable-cell plate counts were performed by serially diluting samples in fresh AG medium. Dilutions spanning 6 orders of magnitude were plated on AG agar plates as 10-μL droplets. Plates were incubated at 30°C. Colonies were counted after 4 days for the WT and ΔhpnH complement and after 5 days for the ΔhpnH strain.

Lipid packing.

Three biological replicates were grown at 30°C and harvested at mid-exponential phase (OD600 = 0.5). Cells were kept at 30°C throughout the procedure. Harvested cells were washed twice by pelleting at a relative centrifugal force (rcf) of 5,000 for 7 min and resuspension in fresh medium at an OD600 of 0.2. Cells were then transferred to black-bottom 96-well plates and stained with 80 nM di-4-ANEPPDHQ (Thermo Fisher; D36802). All spectroscopic measurements were carried out using a SPARK 20 plate reader (Tecan, Grödig, Austria) equipped with a thermostat capable of maintaining the temperature with an accuracy of ±1°C. Samples were incubated for 30 min at 150 rpm. Fluorescence emission was measured in the “top reading mode” of the setup in the epi-configuration using a 50/50 mirror and two monochromators (for selecting excitation and emission wavelengths). The sample was excited with a xenon flash lamp with the excitation monochromator set to 485 nm (20-nm bandwidth). The fluorescence emission was measured at wavelengths of 540 nm and 670 nm (20-nm bandwidth each).
General polarization (GP) was calculated from di-4 emission using the following formula:
GP=I540 − I670I540 + I670
where I is the fluorescence emission intensity at a given wavelength after subtraction of the signal measured for the blank suspension.

FDA permeability assay.

To assess relative changes in permeability, we employed an assay based on the hydrolysis of fluorescein diacetate (FDA) to fluorescein after it diffuses into the cell (42, 73). Three biological replicates of each strain were grown until early exponential phase (OD600, 0.5). Harvested cells were washed twice by pelleting at an rcf of 5,000 for 7 min and resuspension in fresh medium at an OD600 of 0.2. Cells were then transferred to black-bottom 96-well plates and FDA (Sigma F7378) was added directly to each well to a final concentration of 20 μM. All spectroscopic measurements were carried out using a SPARK 20 plate reader (Tecan, Grödig, Austria) equipped with a thermostat capable of maintaining the temperature with the accuracy of ±1°C. Immediately following the addition of FDA, samples were measured every 3 min for 39 min. Excitation was set to 485 nm (20-nm bandwidth), and emission intensity was measured at 525 nm (20-nm bandwidth). The relative permeability of FDA hydrolysis was estimated as the slope of fluorescence versus time for 39 min.

Statistical tests.

To determine if observed differences between conditions or strains were statistically significant, we performed unpaired two-tailed t tests. Statistical significances are indicated in figures as follows: ns, not significant (P ≥ 0.05); *, P ≤ 0.05; **, P ≤ 0.01; and ***, P ≤ 0.001.


We thank members of the Newman lab for their helpful comments and insights, especially Brittany Belin and all past members of Team Hopanoid. We are grateful to Hans Martin-Fischer for his encouragement and support of our work.
This research was enabled by an NSF graduate research fellowship foundation (E.T.), NASA (NNX16AL96G to D.K.N.), a German Federal Ministry of Education and Research BMBF grant (to J.S.; project 03Z22EN12), and a VW Foundation “Life” grant (to J.S.; project 93090).

Supplemental Material

File (jb.00442-21-s0001.pdf)
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.


Doran JW, Zeiss MR. 2000. Soil health and sustainability: managing the biotic component of soil quality. Appl Soil Ecol 15:3–11.
Norris CE, Congreves KA. 2018. Alternative management practices improve soil health indices in intensive vegetable cropping systems: a review. Front Environ Sci.
Tilman D, Cassman KG, Matson PA, Naylor R, Polasky S. 2002. Agricultural sustainability and intensive production practices. Nature 418:671–677.
Foyer CH, Nguyen H, Lam H-M. 2019. Legumes—the art and science of environmentally sustainable agriculture. Plant Cell Environ 42:1–5.
Oldroyd GED, Murray JD, Poole PS, Downie JA. 2011. The rules of engagement in the legume-rhizobial symbiosis. Annu Rev Genet 45:119–144.
Gibson KE, Kobayashi H, Walker GC. 2008. Molecular determinants of a symbiotic chronic infection. Annu Rev Genet 42:413–441.
Pierre O, Engler G, Hopkins J, Brau F, Boncompagni E, Hérouart D. 2013. Peribacteroid space acidification: a marker of mature bacteroid functioning in Medicago truncatula nodules. Plant Cell Environ 36:2059–2070.
Hunt S, Layzell DB. 1993. Gas exchange of legume nodules and the regulation of nitrogenase activity. Annu Rev Plant Physiol Plant Mol Biol 44:483–511.
Tookmanian EM, Belin BJ, Sáenz JP, Newman DK. 2021. The role of hopanoids in fortifying rhizobia against a changing climate. Environ Microbiol 23:2906–2918.
Reckling M, Hecker J-M, Bergkvist G, Watson CA, Zander P, Schläfke N, Stoddard FL, Eory V, Topp CFE, Maire J, Bachinger J. 2016. A cropping system assessment framework—evaluating effects of introducing legumes into crop rotations. Eur J Agron 76:186–197.
Bullock DG. 1992. Crop rotation. CRC Crit Rev Plant Sci 11:309–326.
Loureiro MDF, Kaschuk G, Alberton O, Hungria M. 2007. Soybean [Glycine max (L.) Merrill] rhizobial diversity in Brazilian oxisols under various soil, cropping, and inoculation managements. Biol Fertil Soils 43:665–674.
Zhang NN, Sun YM, Li L, Wang ET, Chen WX, Yuan HL. 2010. Effects of intercropping and Rhizobium inoculation on yield and rhizosphere bacterial community of faba bean (Vicia faba L.). Biol Fertil Soils 46:625–639.
Roughley RJ, Gemell LG, Thompson JA, Brockwell J. 1993. The number of Bradyrhizobium sp. (Lupinus) applied to seed and its effect on rhizosphere colonization, nodulation and yield of lupin. Soil Biol Biochem 25:1453–1458.
Corich V, Giacomini A, Vendramin E, Vian P, Carlot M, Concheri G, Polone E, Casella S, Nuti MP, Squartini A. 2007. Long term evaluation of field-released genetically modified rhizobia. Environ Biosafety Res 6:167–181.
O’Callaghan M. 2016. Microbial inoculation of seed for improved crop performance: issues and opportunities. Appl Microbiol Biotechnol 100:5729–5746.
Karmakar K, Rana A, Rajwar A, Sahgal M, Johri BN. 2015. Legume-rhizobia symbiosis under stress, p 241–258. In Arora NK (ed), Plant microbes symbiosis: applied facets. Springer India, New Delhi, India.
Ilangumaran G, Smith DL. 2017. Plant growth promoting rhizobacteria in amelioration of salinity stress: a systems biology perspective. Front Plant Sci 8:1768.
Zahran HH. 1999. Rhizobium-legume symbiosis and nitrogen fixation under severe conditions and in an arid climate. Microbiol Mol Biol Rev 63:968–989.
Delgado-Baquerizo M, Oliverio AM, Brewer TE, Benavent-González A, Eldridge DJ, Bardgett RD, Maestre FT, Singh BK, Fierer N. 2018. A global atlas of the dominant bacteria found in soil. Science 359:320–325.
Choma A, Komaniecka I, Zebracki K. 2017. Structure, biosynthesis and function of unusual lipids A from nodule-inducing and N2-fixing bacteria. Biochim Biophys Acta Mol Cell Biol Lipids 1862:196–209.
Serrato RV. 2014. Lipopolysaccharides in diazotrophic bacteria. Front Cell Infect Microbiol 4:119.
Nikaido H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593–656.
Komaniecka I, Zamłyńska K, Zan R, Staszczak M, Pawelec J, Seta I, Choma A. 2016. Rhizobium strains differ considerably in outer membrane permeability and polymyxin B resistance. Acta Biochim Pol 63:517–525.
Kulkarni G, Busset N, Molinaro A, Gargani D, Chaintreuil C, Silipo A, Giraud E, Newman DK. 2015. Specific hopanoid classes differentially affect free-living and symbiotic states of Bradyrhizobium diazoefficiens. mBio 6:e01251-15.
Belin BJ, Busset N, Giraud E, Molinaro A, Silipo A, Newman DK. 2018. Hopanoid lipids: from membranes to plant-bacteria interactions. Nat Rev Microbiol 16:304–315.
Silipo A, Vitiello G, Gully D, Sturiale L, Chaintreuil C, Fardoux J, Gargani D, Lee H-I, Kulkarni G, Busset N, Marchetti R, Palmigiano A, Moll H, Engel R, Lanzetta R, Paduano L, Parrilli M, Chang W-S, Holst O, Newman DK, Garozzo D, D’Errico G, Giraud E, Molinaro A. 2014. Covalently linked hopanoid-lipid A improves outer-membrane resistance of a Bradyrhizobium symbiont of legumes. Nat Commun 5:5106.
Busset N, Di Lorenzo F, Palmigiano A, Sturiale L, Gressent F, Fardoux J, Gully D, Chaintreuil C, Molinaro A, Silipo A, Giraud E. 2017. The very long chain fatty acid (C26:25OH) linked to the lipid A is important for the fitness of the photosynthetic Bradyrhizobium strain ORS278 and the establishment of a successful symbiosis with Aeschynomene legumes. Front Microbiol 8:1821.
Komaniecka I, Choma A, Mazur A, Duda KA, Lindner B, Schwudke D, Holst O. 2014. Occurrence of an unusual hopanoid-containing lipid A among lipopolysaccharides from Bradyrhizobium species. J Biol Chem 289:35644–35655.
Belin BJ, Tookmanian EM, de Anda J, Wong GCL, Newman DK. 2019. Extended hopanoid loss reduces bacterial motility and surface attachment and leads to heterogeneity in root nodule growth kinetics in a Bradyrhizobium-Aeschynomene symbiosis. Mol Plant Microbe Interact 32:1415–1428.
Clifton LA, Skoda MWA, Le Brun AP, Ciesielski F, Kuzmenko I, Holt SA, Lakey JH. 2015. Effect of divalent cation removal on the structure of gram-negative bacterial outer membrane models. Langmuir 31:404–412.
Schäfer A, Tauch A, Jäger W, Kalinowski J, Thierbach G, Pühler A. 1994. Small mobilizable multi-purpose cloning vectors derived from the Escherichia coli plasmids pK18 and pK19: selection of defined deletions in the chromosome of Corynebacterium glutamicum. Gene 145:69–73.
Ledermann R, Bartsch I, Remus-Emsermann MN, Vorholt JA, Fischer H-M. 2015. Stable fluorescent and enzymatic tagging of Bradyrhizobium diazoefficiens to analyze host-plant infection and colonization. Mol Plant Microbe Interact 28:959–967.
Bolan NS, Hedley MJ, White RE. 1991. Processes of soil acidification during nitrogen cycling with emphasis on legume based pastures. Plant Soil 134:53–63.
Ledermann R, Emmenegger B, Couzigou J-M, Zamboni N, Kiefer P, Vorholt JA, Fischer H-M. 2021. Bradyrhizobium diazoefficiens requires chemical chaperones to cope with osmotic stress during soybean infection. mBio 12:e00390-21.
Ma Y, Benda A, Kwiatek J, Owen DM, Gaus K. 2018. Time-resolved laurdan fluorescence reveals insights into membrane viscosity and hydration levels. Biophys J 115:1498–1508.
Steinkühler J, Sezgin E, Urbančič I, Eggeling C, Dimova R. 2019. Mechanical properties of plasma membrane vesicles correlate with lipid order, viscosity and cell density. Commun Biol 2:337.
Sáenz JP, Sezgin E, Schwille P, Simons K. 2012. Functional convergence of hopanoids and sterols in membrane ordering. Proc Natl Acad Sci USA 109:14236–14240.
Zgurskaya HI, Löpez CA, Gnanakaran S. 2015. Permeability barrier of Gram-negative cell envelopes and approaches to bypass it. ACS Infect Dis 1:512–522.
Sáenz JP, Grosser D, Bradley AS, Lagny TJ, Lavrynenko O, Broda M, Simons K. 2015. Hopanoids as functional analogues of cholesterol in bacterial membranes. Proc Natl Acad Sci USA 112:11971–11976.
Brandenburg K, Seydel U. 1990. Investigation into the fluidity of lipopolysaccharide and free lipid A membrane systems by Fourier-transform infrared spectroscopy and differential scanning calorimetry. Eur J Biochem 191:229–236.
Levental KR, Malmberg E, Symons JL, Fan Y-Y, Chapkin RS, Ernst R, Levental I. 2020. Lipidomic and biophysical homeostasis of mammalian membranes counteracts dietary lipid perturbations to maintain cellular fitness. Nat Commun 11:1339.
Welander PV, Hunter RC, Zhang L, Sessions AL, Summons RE, Newman DK. 2009. Hopanoids play a role in membrane integrity and pH homeostasis in Rhodopseudomonas palustris TIE-1. J Bacteriol 191:6145–6156.
Farrell MJ, Finkel SE. 2003. The growth advantage in stationary-phase phenotype conferred by rpoS mutations is dependent on the pH and nutrient environment. J Bacteriol 185:7044–7052.
Vitiello G, Oliva R, Petraccone L, Vecchio PD, Heenan RK, Molinaro A, Silipo A, D’Errico G, Paduano L. 2021. Covalently bonded hopanoid-lipid A from Bradyrhizobium: the role of unusual molecular structure and calcium ions in regulating the lipid bilayers organization. J Colloid Interface Sci 594:891–901.
Papahadjopoulos D, Portis A, Pangborn W. 1978. Calcium-induced lipid phase transitions and membrane fusion. Ann N Y Acad Sci 308:50–66.
Macció D, Fabra A, Castro S. 2002. Acidity and calcium interaction affect the growth of Bradyrhizobium sp. and the attachment to peanut roots. Soil Biol Biochem 34:201–208.
Izmailov SF. 2003. Calcium-based interactions of symbiotic partners in legumes: role of peribacteroid membrane. Russ J Plant Physiol 50:553–566.
Andreev IM, Andreeva IN, Dubrovo PN, Krylova VV, Kozharinova GM, Izmailov SF. 2001. Calcium status of yellow lupin symbiosomes as a potential regulator of their nitrogenase activity: the role of the peribacteroid membrane. Russ J Plant Physiol 48:308–317.
Tejima K, Arima Y, Yokoyama T, Sekimoto H. 2003. Composition of amino acids, organic acids, and sugars in the peribacteroid space of soybean root nodules. Soil Sci Plant Nutr 49:239–247.
Mangiarotti A, Genovese DM, Naumann CA, Monti MR, Wilke N. 2019. Hopanoids, like sterols, modulate dynamics, compaction, phase segregation and permeability of membranes. Biochim Biophys Acta Biomembr 1861:183060.
Kannenberg E, Blume A, McElhaney RN, Poralla K. 1983. Monolayer and calorimetric studies of phosphatidylcholines containing branched-chain fatty acids and of their interactions with cholesterol and with a bacterial hopanoid in model membranes. Biochim Biophys Acta Biomembr 733:111–116.
Chen Z, Sato Y, Nakazawa I, Suzuki Y. 1995. Interactions between bacteriohopane-32,33,34,35-tetrol and liposomal membranes composed of dipalmitoylphosphatidylcholine. Biol Pharm Bull 18:477–480.
Crowe LM, Mouradian R, Crowe JH, Jackson SA, Womersley C. 1984. Effects of carbohydrates on membrane stability at low water activities. Biochim Biophys Acta 769:141–150.
Los DA, Murata N. 2004. Membrane fluidity and its roles in the perception of environmental signals. Biochim Biophys Acta 1666:142–157.
Andersen HD, Wang C, Arleth L, Peters GH, Westh P. 2011. Reconciliation of opposing views on membrane-sugar interactions. Proc Natl Acad Sci USA 108:1874–1878.
Yamazaki M, Ohnishi S, Ito T. 1989. Osmoelastic coupling in biological structures: decrease in membrane fluidity and osmophobic association of phospholipid vesicles in response to osmotic stress. Biochemistry 28:3710–3715.
Miller JH. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Mesa S, Hauser F, Friberg M, Malaguti E, Fischer H-M, Hennecke H. 2008. Comprehensive assessment of the regulons controlled by the FixLJ-FixK2-FixK1 cascade in Bradyrhizobium japonicum. J Bacteriol 190:6568–6579.
Regensburger B, Hennecke H. 1983. RNA polymerase from Rhizobium japonicum. Arch Microbiol 135:103–109.
Sadowsky MJ, Tully RE, Cregan PB, Keyser HH. 1987. Genetic diversity in Bradyrhizobium japonicum serogroup 123 and its relation to genotype-specific nodulation of soybean. Appl Environ Microbiol 53:2624–2630.
Cole MA, Elkan GH. 1973. Transmissible resistance to penicillin G, neomycin, and chloramphenicol in Rhizobium japonicum. Antimicrob Agents Chemother 4:248–253.
Kaczmarczyk A, Vorholt JA, Francez-Charlot A. 2013. Cumate-inducible gene expression system for sphingomonads and other Alphaproteobacteria. Appl Environ Microbiol 79:6795–6802.
Eaton RW. 1997. p-Cymene catabolic pathway in Pseudomonas putida F1: cloning and characterization of DNA encoding conversion of p-cymene to p-cumate. J Bacteriol 179:3171–3180.
Eaton RW. 1996. p-Cumate catabolic pathway in Pseudomonas putida Fl: cloning and characterization of DNA carrying the cmt operon. J Bacteriol 178:1351–1362.
Ledermann R. 2017. Role of general stress response in trehalose biosynthesis for functional rhizobia-legume symbiosis. Doctoral dissertation. ETH Zurich, Zurich, Switzerland.
Gibson DG, Young L, Chuang R-Y, Venter JC, Hutchison CA, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343–345.
Beck C, Marty R, Kläusli S, Hennecke H, Göttfert M. 1997. Dissection of the transcription machinery for housekeeping genes of Bradyrhizobium japonicum. J Bacteriol 179:364–369.
Masloboeva N, Reutimann L, Stiefel P, Follador R, Leimer N, Hennecke H, Mesa S, Fischer H-M. 2012. Reactive oxygen species-inducible ECF σ factors of Bradyrhizobium japonicum. PLoS One 7:e43421.
Hahn M, Meyer L, Studer D, Regensburger B, Hennecke H. 1984. Insertion and deletion mutations within the nif region of Rhizobium japonicum. Plant Mol Biol 3:159–168.
Tjørve KMC, Tjørve E. 2017. The use of Gompertz models in growth analyses, and new Gompertz-model approach: an addition to the Unified-Richards family. PLoS One 12:e0178691.
Zwietering MH, Jongenburger I, Rombouts FM, van ’t Riet K. 1990. Modeling of the bacterial growth curve. Appl Environ Microbiol 56:1875–1881.
Rizk S, Henke P, Santana-Molina C, Martens G, Gnädig M, Nguyen NA, Devos DP, Neumann-Schaal M, Saenz JP. 2021. Functional diversity of isoprenoid lipids in Methylobacterium extorquens PA1. Mol Microbiol 116:1064–1078.

Information & Contributors


Published In

cover image Journal of Bacteriology
Journal of Bacteriology
Volume 204Number 719 July 2022
eLocator: e00442-21
Editor: Anke Becker, Philipps University Marburg
PubMed: 35657706


Received: 30 August 2021
Accepted: 3 May 2022
Published online: 3 June 2022


Request permissions for this article.


  1. bradyrhizobia
  2. climate change
  3. hopanoids
  4. osmotic stress
  5. robustness
  6. soil microbiology



Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California, USA
Lisa Junghans
B CUBE Center for Molecular Bioengineering, Technische Universität Dresden, Dresden, Germany
Gargi Kulkarni
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California, USA
Institute of Microbiology, ETH Zurich, Zurich, Switzerland
B CUBE Center for Molecular Bioengineering, Technische Universität Dresden, Dresden, Germany
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California, USA
Division of Geological and Planetary Science, California Institute of Technology, Pasadena, California, USA


Anke Becker
Philipps University Marburg


The authors declare no conflict of interest.

Metrics & Citations



  • For recently published articles, the TOTAL download count will appear as zero until a new month starts.
  • There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.
  • Citation counts come from the Crossref Cited by service.


If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. For an editable text file, please select Medlars format which will download as a .txt file. Simply select your manager software from the list below and click Download.

View Options

Figures and Media






Share the article link

Share with email

Email a colleague

Share on social media

American Society for Microbiology ("ASM") is committed to maintaining your confidence and trust with respect to the information we collect from you on websites owned and operated by ASM ("ASM Web Sites") and other sources. This Privacy Policy sets forth the information we collect about you, how we use this information and the choices you have about how we use such information.
FIND OUT MORE about the privacy policy