Hopanoid lipids may facilitate aerobic nitrogen fixation in the ocean

Abstract

Cyanobacterial diazotrophs are considered to be the most important source of fixed N₂ in the open ocean. Biological N₂ fixation is catalyzed by the extremely O₂-sensitive nitrogenase enzyme. In cyanobacteria without specialized N₂-fixing cells (heterocysts), mechanisms such as decoupling photosynthesis from N₂ fixation in space or time are involved in protecting nitrogenase from the intracellular O₂ evolved by photosynthesis. However, it is not known how cyanobacterial cells limit O₂ diffusion across their membranes to protect nitrogenase in ambient O₂-saturated surface ocean waters. Here, we explored all known genomes of the major marine cyanobacterial lineages for the presence of hopanoid synthesis genes, since hopanoids are a class of lipids that might act as an O₂ diffusion barrier. We found that, whereas all non−heterocyst-forming cyanobacterial diazotrophs had hopanoid synthesis genes, none of the marine Synechococcus, Prochlorococcus (non−N₂-fixing), and marine heterocyst-forming (N₂-fixing) cyanobacteria did. Finally, we conclude that hopanoid-enriched membranes are a conserved trait in non−heterocyst-forming cyanobacterial diazotrophs that might lower the permeability to extracellular O₂. This membrane property coupled with high respiration rates to decrease intracellular O₂ concentration may therefore explain how non−heterocyst-forming cyanobacterial diazotrophs can fix N₂ in the fully oxic surface ocean.

Marine cyanobacterial diazotrophs, i.e., those capable of reducing dissolved dinitrogen gas (N₂) into ammonia through N₂ fixation, are key suppliers of bioavailable N, a limiting nutrient for primary production in the ocean (1). Biological N₂ fixation is solely performed by the O₂-sensitive nitrogenase enzyme (2), and understanding how low intracellular O₂ concentrations are maintained in fully oxic open waters is a long-standing question that has attracted much interest (3–6).

Although, a priori, it would seem that N₂ fixation is incompatible with the O₂-evolving photosynthetic lifestyle of cyanobacteria, it is known that these microorganisms have evolved a variety of strategies to protect nitrogenase from O₂ inactivation. For example, some filamentous cyanobacteria, including the symbionts of marine diatoms, form specialized cells called heterocysts (7). A microaerobic environment is created inside heterocysts by inactivating oxygenic photosynthesis, by maintaining or enhancing respiration, and by the formation of an extra glycolipid cell envelope outside the cell wall (8). In contrast, non−heterocyst-forming cyanobacteria such as the filamentous Trichodesmium or the free-living unicellular Crocosphaera must separate photosynthesis and N₂ fixation either spatially or temporally to avoid exposing nitrogenase to the O₂ that they produce during the light hours (9, 10). In the unicellular cyanobacterial symbiont UCYN-A, all of the genes for the synthesis of the O₂-evolving photosystem II (PSII) apparatus have been lost and so UCYN-A doesn’t generate O₂ (11). None of the aforementioned strategies, however, can protect nitrogenase of non−heterocyst-forming cyanobacterial diazotrophs from the O₂ that diffuses across cell membranes from the environment (including host photosynthesis in the case of UCYN-A). Mechanisms such as respiration, the Mehler reaction, and/or other O₂ scavenging strategies have been proposed as potential ways to overcome this problem (12), but whether these mechanisms are sufficient to lower the O₂ concentration in the inner cell while N₂ fixation takes place remains unknown.

We have discovered a consistent pattern of distribution of hopanoid synthesis genes among marine cyanobacteria that suggests that they may play an important role in marine N₂ fixation. Hopanoids are a class of membrane lipids that have been shown to confer special properties to cell membranes (13). Hopanoids can intercalate into lipid bilayers of membranes due to their planar and hydrophobic structure and might decrease their permeability to O₂ (14). Approximately 10% of bacteria, including plant-associated diazotrophs, have the gene for the synthesis of hopanoids (the squalene−hopene cyclase gene shc) (13). Interestingly, the only direct evidence showing that hopanoids facilitate N₂ fixation comes from studies of the terrestrial N₂-fixing heterotrophic bacteria Frankia. In Frankia sp., hopanoids might serve as an O₂ diffusion barrier in their N₂-fixing vesicles (15), with the thickness of the vesicle envelope directly correlated to the external O₂ concentration (16). However, this linkage was later questioned based on the observation of high proportions of hopanoids in membranes regardless of the N status in Frankia sp. (17).

We compiled data on hopanoid production and mined the publicly available genomes of marine cyanobacteria to provide an exploration of the presence of hopanoid biosynthetic and modification genes across all of the major marine cyanobacterial lineages, including both diazotrophs and non-diazotrophs (Fig. 1). We found that the shc gene for synthesizing hopanoids was consistently present in all of the non−heterocyst-forming cyanobacterial diazotrophs, including unicellular cyanobacterial symbionts with extremely reduced genomes such as UCYN-A. In contrast, none of the non-diazotrophic marine Synechococcus and Prochlorococcus, which are the dominant cyanobacteria in the ocean (18, 19), nor the heterocyst-forming marine cyanobacteria Calothrix rhizosoleniae SC01 and Richelia intracellularis HH01 had the shc gene in their genomes. The same pattern was observed for almost all of the hopanoid modification genes except for the hpnK and hpnP genes (Fig. 1). These observations suggest that, whereas the capacity of allocating hopanoids into cell membranes may be universal across all marine non−heterocyst-forming diazotrophic cyanobacteria, it is absent from all marine Synechococcus and Prochlorococcus (which do not fix N₂) and from heterocyst-forming marine cyanobacteria (which already protect nitrogenase from O₂ by heterocysts). Furthermore, in Crocosphaera and Cyanothece, the transcription of the shc gene peaks right before the nitrogenase-encoding gene (nifH) starts increasing its expression level (data collected from ref. 20), and simultaneous expression of both markers has also been detected in UCYN-A (21). These patterns are further supported by previous observations of hopanoid production in the cyanobacterium Crocosphaera watsonii WH8501 in the context of N₂ fixation (22, 23). However, the role of hopanoids in N₂ fixation was discarded because C. watsonii WH8501 showed constant levels of hopanoids regardless of light−dark periods or the availability of fixed N (23).

Fig. 1.

Hopanids in marine cyanobacteria. (Upper) A schematic representation of the hopanoid biosynthesis and modification pathways, including enzymes and products. (Lower) Summary of the presence/absence of the genes involved in the synthesis and modification of hopanoids across a selection of the major marine cyanobacterial lineages. All of the available marine cyanobacterial genomes in NCBI (May 2019) were screened for this analysis, yet only 21 are shown, for simplification. Asterisk (*), experimentally tested in refs. 20 and 21 (n.t., not tested). Enzymes participating in hopanoid pathways: squalene−hopene cyclase (SHC), hopanoid biosynthesis-associated radical SAM protein (HpnH), hopanoid-associated phosphorylase (HpnG), hopanoid biosynthesis-associated glycosyltransferase protein (HpnI), hopanoid biosynthesis-associated protein (HpnK), hopanoid biosynthesis-associated radical SAM protein (HpnJ), aminotransferase (HpnO), hopanoid 2-methyltransferase (HpnP), and hopanoid C3 methylase (HpnR); 3-methylhopanoid production has never been found in marine cyanobacteria (28); hpnO was absent in all of the screened strains. Dashed arrows indicate that enzymes driving intermediate steps are unknown. See ref. 13 for further details on hopanoid biosynthesis.

We thus propose that the presence of hopanoids in the whole-cell membrane is a conserved trait in marine non−heterocyst-forming cyanobacterial diazotrophs that might confer protection to nitrogenase by reducing the rate of diffusion of extracellular O₂ into the cell. In parallel, as shown for Cyanothece (24), increases in respiration rates can presumably lower the intracellular O₂ concentration to levels suitable for nitrogenase activity while fulfilling the adenosine 5′-triphosphate (ATP) demand required for N₂ fixation. Although the constant levels of hopanoids to total lipids has previously been argued to discount a role of hopanoids in marine N₂ fixation (23), we believe that hopanoids reduce O₂ membrane permeability that limits the diffusion rate and facilitates respiratory protection of nitrogenase. It is also possible that hopanoids can form rafts, i.e., membrane microdomains with high hopanoid content that promote dynamic changes in membrane permeability based on redistributions of hopanoid molecules in the membrane (13). Hopanoid rafts have been detected in C. watsonii (25), which suggests that Crocosphaera might have such dynamic changes in membrane permeability.

Since members of non−heterocyst-forming freshwater cyanobacteria (e.g., Aphanothece, Pleurocapsa, endosymbionts of the diatoms Rhopalodia gibberula and Epithemia turgida) and noncyanobacterial diazotrophs (e.g., Azotobacter) also have the shc gene, we believe that our hypothesis, which provides a mechanism that restricts O₂ diffusion analogous to the heterocyst, may provide an important research direction for future studies devoted to understanding N₂ fixation in different environments (marine, freshwater, terrestrial) as well as other O₂-sensitive processes (e.g., methanogenesis) when happening in well-oxygenated environments (26, 27).