Crocosphaera watsonii – A widespread nitrogen‐fixing unicellular marine cyanobacterium

Crocosphaera watsonii is a unicellular N2‐fixing (diazotrophic) cyanobacterium observed in tropical and subtropical oligotrophic oceans. As a diazotroph, it can be a source of bioavailable nitrogen (N) to the microbial community in N‐limited environments, and this may fuel primary production in the regions where it occurs. Crocosphaera watsonii has been the subject of intense study, both in culture and in field populations. Here, we summarize the current understanding of the phylogenetic and physiological diversity of C. watsonii, its distribution, and its ecological niche. Analysis of the relationships among the individual Crocosphaera species and related free‐living and symbiotic lineages of diazotrophs based on the nifH gene have shown that the C. watsonii group holds a basal position and that its sequence is more similar to Rippkaea and Zehria than to other Crocosphaera species. This finding warrants further scrutiny to determine if the placement is related to a horizontal gene transfer event. Here, the nifH UCYN‐B gene copy number from a recent synthesis effort was used as a proxy for relative C. watsonii abundance to examine patterns of C. watsonii distribution as a function of environmental factors, like iron and phosphorus concentration, and complimented with a synthesis of C. watsonii physiology. Furthermore, we have summarized the current knowledge of C. watsonii with regards to N2 fixation, photosynthesis, and quantitative modeling of physiology. Because N availability can limit primary production, C. watsonii is widely recognized for its importance to carbon and N cycling in ocean ecosystems, and we conclude this review by highlighting important topics for further research on this important species.

Marine N 2 -fixation studies historically focused on Trichodesmium, and this genus has been the subject of a number of intensive reviews of its physiological features, distribution, and physiology (Capone et al., 1997;Kranz et al., 2011;McKinna, 2015).The symbiotic unicellular diazotroph Candidatus Atelocyanobacterium thalassae (UCYN-A) was discovered relatively recently (Zehr et al., 1998) and has been the subject of interest and extensive research efforts (Farnelid et al., 2016;Zehr et al., 2016).The function and the distribution of marine diazotrophs, including those of Crocosphaera watsonii (UCYN-B), have been summarized previously (Luo et al., 2012;Masuda et al., 2022;Shao et al., 2023;Tang & Cassar, 2019;Thompson & Zehr, 2013;Zehr & Capone, 2020).In this review, we have focused on C. watsonii, summarizing the phylogenetic clustering, physiological diversity, ecological niche, and regulation of photosynthesis and N 2 fixation of C. watsonii.We conclude by summarizing the knowledge obtained by quantitative modeling of C. watsonii physiology.

PHYLOGENETIC CLUSTERING
The ability to fix N 2 occurs across cyanobacterial lineages, and the complete nif gene operon has been detected in more than one-third of sequenced cyanobacterial genomes (Latysheva et al., 2012;Watanabe & Horiike, 2021).Among marine plankton, diazotrophs are represented by several phylogenetically separated lineages (see Appendix S1 in the Supporting Information; Figure 2a).Crocosphaera spp.(UCYN-B and -C) and Candidatus Atelocyanobacterium thalassae (UCYN-A) form a unique lineage of unicellular representatives of cyanobacterial diazotrophic marine plankton (Figure 2a) that belongs to a broader cluster of single-celled N 2 -fixing cyanobacteria inhabiting various ecosystems, including terrestrial, epilithic, soil, and freshwater habitats (genera Rippkaea, Aphanothece, Gloeothece; Mareš et al., 2019).The systematics of unicellular cyanobacteria has been a challenge due to morphological convergence and researchers using diverse taxonomic concepts over time, including traditional botanical (Geitler, 1932), bacteriological (Rippka et al., 1979), and more recently phylogenetic (Johansen & Casamatta, 2005)  morphological observations, transmission electron microscopy (TEM), phylogenetic analysis (of the 16S rRNA gene and two additional housekeeping loci), ITS rRNA region secondary structure prediction, and assessment of the source habitat type (Mareš et al., 2019).The most striking result from the analysis of rRNA operon data by Mareš et al. (2019) was that several strains previously referred to as Cyanothece sp. or Gloeocapsa sp.turned out to be members of the genus Crocosphaera.Consequently, the genus Crocosphaera currently contains not only C. watsonii (UCYN-B) but also species corresponding to the UCYN-C clade: C. subtropica (previously Cyanothece sp.ATCC51142), C. chwakensis (previously Cyanothece sp.CCY0110), and other species still waiting to be described (Figure 2b).Recently, a novel Crocosphaera species, Crocosphaera waterburyi, has been proposed based on metagenomic analysis from the North Pacific Ocean (Cleveland et al., 2023), and so the genus continues to expand.Furthermore, a morphologically and ecologically convergent group of diazotrophic Crocosphaera-like isolates has formed a separate genus-level lineage described as Zehria (Mareš et al., 2019).
A closer look at the whole lineage (Figure 2b) reveals that widespread non-photosynthetic diazotrophic unicellular endosymbionts of eukaryotic algae have formed sister clades to Crocosphaera.One of them, UCYN-A or Candidatus Atelocyanobacterium thalassae, contained endosymbionts of marine Prymnesiophytes (Hagino et al., 2013;Thompson et al., 2012); another more distantly related lineage contained the endosymbionts (also called "spheroid bodies") of rhopalodiacean diatoms (Nakayama & Inagaki, 2017).In both cases, the symbionts had abandoned oxygenic photosynthesis; they function almost as organelles, providing fixed N 2 to their hosts in exchange for fixed carbon (Prechtl et al., 2004;Thompson et al., 2012).Intriguingly, the marine diatom Climacodium frauenfeldianum has been shown to be capable of hosting Crocosphaera as a photosynthetic diazotrophic endosymbiont (Caputo et al., 2019;Carpenter & Janson, 2000).
The clustering and relationships among the individual Crocosphaera-like free-living and symbiotic lineages seem to be identical based on the 16S rRNA gene and a robust set of 120 universal bacterial housekeeping proteins (Parks et al., 2021) (Figure 2b), although the whole genome is missing for Zehria.Although the corresponding clades can be identified in the nifH tree (Figure 2b), the C. watsonii (UCYN-B) group holds a basal position, and its nifH gene sequences are more similar to Rippkaea and Zehria (89.2%-90.1% pairwise identity) than to other Crocosphaera species (80.5%-85.0%).Such a phylogenetic pattern possibly indicates a horizontal gene transfer (HGT) event leading to the replacement of the C. watsonii nifH gene by the gene from an unknown but related marine unicellular diazotrophic cyanobacterium.Alternatively, the 16S rRNA gene of Zehria could have been subject to HGT.Based on previous reports, HGT in the N 2 fixation gene cluster is possible, although uncommon, in cyanobacteria (Bolhuis et al., 2010;Latysheva et al., 2012).The nifH gene itself seems to exhibit HGT more frequently than other nif genes, and its phylogeny is often incongruent with the 16S rRNA gene phylogeny in N 2 -fixing bacteria in general (Gaby & Buckley, 2014).Given the scarcity of reported HGT events in cyanobacterial rRNA genes (Johansen et al., 2017), we suggest that an HGT event in the nifH gene of C. watsonii might be more likely than in the 16S rRNA gene of Zehria.Future studies that include whole genome sequencing of Zehria are needed to validate this hypothesis.
Studies suggest low genomic diversity of Crocosphaera watsonii (Webb et al., 2009;Zehr et al., 2007).For example, analysis of a set of five functional genes showed greater than 99% nucleotide identity among seven C. watsonii strains (Zehr et al., 2007).In addition, sequences of the internal transcribed spacer region between the 16S and 23S rRNA genes (ITS rRNA region) from phenotypically distant cultures were highly similar (Webb et al., 2009).Cultivated and in situ C. watsonii genomes have shown an unusually high number of transposases (Bench et al., 2011(Bench et al., , 2013;;Hewson et al., 2009;Mes & Doeleman, 2006;Zehr et al., 2007), suggesting a mechanism for adaptation to changing environmental conditions, which could underpin the variety of phenotypic differences observed among strains (Bench et al., 2011).Whole genome sequences of six C. watsonii strains have shown that C. watsonii WH8501 has a larger genome (6.2 Mb vs. 4.5-5.8Mb), a higher number of transposase genes (1211 genes vs. 165-223 genes), and a higher number of strain-specific transposase genes (71 vs. 4-19) than the five other strains analyzed (Bench et al., 2013).These results suggested that the C. watsonii WH8501 strain is an outlier relative to the others examined and may not be the most representative strain of the natural C. watsonii population.However, this warrants further study across a broader set of both isolates and field samples.
The comprehensive dataset of UCYN-B nifH gene copy number, updated after the study by Tang and Cassar (2019), and by Shao et al. (2023) revealed nonhomogeneous distributions (Figure 3a).We note the relative increase in UCYN-B nifH gene copy number in the western South Pacific, around Fiji, in the northeastern Pacific, near the Hawaii islands, in the tropical North Atlantic, in the western subtropical South Atlantic, in the Philippine Sea, and in the eastern north Indian Ocean (Figure 3a; Shao et al., 2023 and references therein).The maximum copy number (7.9 × 10 6 nifH gene copies • L −1 ) was reported from 37 m depth off Fiji (15° S 175° E; Moisander et al., 2010;Figure 3a).Diazotroph distributions can be influenced by many factors including nutrient supply and supply ratios (Shiozaki, Kodama, & Furuya, 2014;Subramaniam et al., 2008), which may be the case for some of the areas of increased UCYN-B nifH gene copy number.For example, comparatively low UCYN-B nifH gene copy number in the Indian Ocean may be attributed to higher upward nitrate fluxes to the surface because of the shallow nitracline (Sato et al., 2022;Shiozaki, Ijichi, et al., 2014).A negative relationship between a shallow nitracline and UCYN-B nifH gene copy number was reported from the northern South China Sea (Shiozaki, Ijichi, et al., 2014).Taken together, identifying the factors that determine the distribution of UCYN-B nifH gene copy number is an ongoing area of study.
The vertical distribution or UCYN-B nifH gene copy number was restricted to above 200 m of depth, with >10 5 nifH gene copies • L −1 observed above 150 m (Figure 4a).Temperature seems to be an important niche trait for diazotrophs, and the UCYN-B nifH gene copy number was observed to increase with increasing water temperature in regional and global studies (Moisander et al., 2010;Tang & Cassar, 2019).Although the UCYN-B nifH gene copy number was typically highest between 25 and 30°C, the UCYN-B nifH gene has been identified in regions as cold as 10.1°C (Goebel et al., 2010;Shao et al., 2023).Further, the UCYN-B nifH gene copy number at >25°C was not always high (>10 5 nifH gene copies • L −1 ), highlighting that temperature is not the only explanatory factor influencing UCYN-B nifH gene copy number (Figure 4b).Laboratory experiments using clonal cultures have reported the optimum temperatures of eight Crocosphaera watsonii strains (WH8501, WH8504, WH0002, WH0003, WH0005, WH0006, WH0401, and WH0402) to be between 16 and 36°C (Fu et al., 2014;Qu et al., 2022;Webb et al., 2009); this broad range supports the variation seen in UCYN-B nifH gene copies as a function of temperature (Figure 4b).Despite the breadth of the temperature range, UCYN-B nifH gene copies were detected at higher temperatures than Trichodesmium nifH gene copies by Tang and Cassar (2019), and this is consistent with culture data where growth rates of C. watsonii strains WH0005, WH0003, WH0401 and WH0402 with growth optimum temperature at 28-30°C were ~25% less than those of Trichodesmium at 24°C and ~50% higher than Trichodesmium at 28°C (Fu et al., 2014).

PHYSIOLOG ICAL ECOLOGY
The classical view on marine N 2 fixation assumes that the preferred ecological niches of diazotrophs are limited mainly to surface waters of oligotrophic areas, characterized by low DIN, saturated dissolved oxygen, and high light intensities (Zehr & Capone, 2020).This view is confirmed by patterns in UCYN-B nifH gene copy number where the copy number was typically highest (>10 5 nifH gene copies • L −1 ) in the surface under low nitrate and phosphorus concentrations, that is, <0.02 μM of nitrate (t-test; t = −12.115,df = 405, p < 0.001) and <0.14 μM phosphate (t-test; t = −17.306,df = 253, p < 0.001; Figure 4a,c,d).However, copies of UCYN-B nifH were detected in the subsurface and at nutrient concentrations above the 0.02 μM nitrate and 0.14 μM phosphate concentrations referenced above (Figure 4a,c,d).Culture studies have shown N 2 fixation is inhibited by increasing the concentration of added NH + 4 or NO − 3 (Dekaezemacker & Bonnet, 2011;Knapp et al., 2012), and the extent of inhibition of fixed N (both NH + 4 and NO − 3 ) on N 2 fixation is more pronounced under low light intensity (25 μmol photons • m −2 • s −1 ) compared to high light intensity (175 μmol photons • m −2 • s −1 ; Garcia & Hutchins, 2014).These results suggest that Crocosphaera watsonii would likely use N 2 in low N surface waters and use combined N in the subsurface, but patterns of N uptake and N 2 fixation are an active area of ongoing study.3 (Dekaezemacker & Bonnet, 2011;Großkopf & Laroche, 2012;Masuda et al., 2013;Rabouille et al., 2021).Although C. watsonii is known as freeliving, self-aggregates of multiple cells have been reported both from field observations (Foster et al., 2013) and clonal cultures (Mohr et al., 2013).Aggregation of cells may be related to extracellular polysaccharide (EPS) production (Foster et al., 2013).Observations of concurrent combined N assimilation and N 2 fixation may be related to a division of labor observed among C. watsonii sub-populations in self-aggregates in which some cells in the population fix N 2 and release NH + 4 while other cells take up NH + 4 (Masuda et al., 2020).Cellular heterogeneity in N 2 fixation activity was observed from in situ Crocosphaera-like cells (Foster et al., 2013) and C. watsonii WH8501 and PS0609 isolates (Masuda et al., 2020;Mohr et al., 2013) using a combination of stable isotope labeling and NanoSIMS analysis.Interestingly about 30% of cells in the culture study of Masuda et al. (2020) did not perform N 2 fixation, but these non-N 2 -fixing cells did perform photosynthesis (Masuda et al., 2020).Genomic information supports both NH + 4 and NO − 3 assimilation (Shi et al., 2010).In sum, these observations suggest C. watsonii can utilize both N 2 and combined N as nitrogen sources.
The excretion of NH + 4 is estimated at a range of up to 26%-66% of fixed N 2 (Dron, Rabouille, Claquin, Le Roy, et al., 2012;Masuda et al., 2013).This is a potentially substantial source of so-called new N into low N systems.It had previously been shown that NH + 4 release was particularly active for an isolate with larger cell size (Crocosphaera watsonii WH0003) compared to a smallcelled isolate (C.watsonii WH8501; Dekaezemacker & Bonnet, 2011), and elevated amounts of EPS have been reported from larger cells (>4 μm; Sohm et al., 2011;Webb et al., 2009), which could influence patterns in self-aggregation.Such self-aggregation may influence broader aspects of or C cycling by sinking fixed C and N.For example, C. watsonii was observed to be one of the sinking particles during an elevated carbon flux event (Poff et al., 2021), and self-aggregation may play a role in this carbon flux.Multiple mechanisms for acclimation to low dissolved inorganic phosphate (DIP) environments have been identified in C. watsonii.For example, C. watsonii can grow on dissolved organic phosphorus (DOP) instead of DIP as a sole phosphorus source including both phosphomonoesters and phosphodiesters (Yamaguchi et al., 2020), and its genome (e.g.strain WH8501) contains genes that encode for the alkaline phosphatase enzymes needed to hydrolyze dissolved organic phosphoesters (Bench et al., 2013;Dyhrman & Haley, 2006).When extracellular P is in excess, C. watsonii uses the constitutively expressed phosphate inorganic transport (PIT) system to transport P into the cell.Under P-limiting conditions, a high-affinity phosphate-specific transport (PST) system uses ATP-mediated transport, and the high-affinity phosphate-binding protein (PstS) is upregulated to maximize phosphate uptake to meet cellular P demand (Dyhrman & Haley, 2006;Pereira et al., 2016).Given its regulation pattern, expression of the pstS gene may be a good molecular marker of C. watsonii P stress (Pereira et al., 2016(Pereira et al., , 2019)).Notably, some C. watsonii strains have up to seven copies of the pstS gene (Bench et al., 2013).In low DIP environments, toxic arsenate could be transported into cells through the PST system, but C. watsonii has an arsenate detoxification pathway in its genome (Dyhrman & Haley, 2011).Unlike other cyanobacteria (i.e., Prochlorococcus, Synechococcus, and Trichodesmium), C. watsonii does not substitute sulpholipids for phospholipids, nor utilize phosphonates in low P environments (Dyhrman & Haley, 2006;Pereira et al., 2016;Van Mooy et al., 2009).Recently, a study of the interaction between temperature and P limitation showed that at elevated temperatures, genes for sulfolipid biosynthesis (sgdB) were upregulated, suggesting that C. watsonii may substitute sulpholipids for phospholipids under certain conditions (Deng et al., 2021).This potential P sparing mechanism may, in part, underpin the similar low-P and P-replete growth rates observed at high temperatures by Deng et al. (2021) and warrants further study for how this might relate to C. watsonii fitness with future global warming.
The distribution of diazotrophs is often constrained by the availability of iron (Mills et al., 2004;Monteiro et al., 2010;Moore et al., 2009).In the case of Crocosphaera watsonii WH8501, Fe limitation led to a significant decrease in growth rate; C, N, and Chl a contents per cell; and N 2 -and CO 2 -fixation rates (Jacq et al., 2014).Under Fe stress, C. watsonii induces the IdiA protein as a cellular Fe scavenging mechanism (Webb et al., 2001).In addition, Fe limitation is known to reduce cell size, and this effect has been documented for C. watsonii strains WH8501 and WH0003 (Garcia et al., 2015;Jacq et al., 2014).Two phases of response have been observed depending on the degree of Fe limitation: (i) Under a moderate Fe limitation, the biovolume of C. watsonii was strongly reduced, but metabolic activity remained (Jacq et al., 2014), and (ii) with increasing Fe deprivation, biovolume remained unchanged but the growth and N 2 -and CO 2 -fixation rates decreased (Jacq et al., 2014).Advantages of decreased cell size in a low nutrient environment have been proposed, such as a high cell surface area-to-volume ratio and thin diffusion boundary layer, as well as reduced material and energetic investment for cell division (Garcia et al., 2015).Interestingly, under P-deficient conditions, C. watsonii WH0003 grew and fixed N 2 faster when Fe was deficient than when Fe was replete, where decreases in cell size were only achieved when Fe and P were co-limiting (Garcia et al., 2015).In addition, Fe and P colimitation appears to support higher growth when compared to Fe or P limitation alone (Yang et al., 2022).This finding suggests the complex influence both Fe and P availability may have on C. watsonii growth and N 2 fixation.
The Fe half-saturation constant for growth of Crocosphaera watsonii was twice as low as that of Trichodesmium (Bucciarelli et al., 2013;Jacq et al., 2014), indicating C. watsonii may have a competitive advantage over Trichodesmium in low Fe environments.One study on the interactive effect of Fe and temperature showed that at optimum temperature, nitrogen-specific iron-use efficiency was higher under Fedepleted conditions than Fe replete conditions (Yang et al., 2021).Results of proteomic analysis have suggested that internal Fe recycling between photosynthesis and N 2 fixation occurs in C. watsonii (Saito et al., 2011).The Fe availability is also important in the response of C. watsonii to future increased CO 2 levels: Growth and N 2 fixation are-predicted to increase with future increasing CO 2 concentration (Gradoville et al., 2014) but only if Fe is not limiting (Fu et al., 2008).Much remains unknown regarding the physiological mechanisms that help C. watsonii cope with Fe and P limitation.

TEMPOR AL SEPAR ATION OF N 2 FIX ATION AND PHOTOSYNTHESIS
A crucial aspect of Crocosphaera watsonii biology is the temporal separation of N 2 fixation and photosynthesis (Mohr et al., 2010;Shi et al., 2010;Tuit et al., 2004;Waterbury & Rippka, 1989;Wilson et al., 2017).Oxygenic photosynthesis during the day generates cellular energy and reductant power, and C. watsonii restrains the peak of N 2 fixation to the night period when there is no light-dependent photosynthesis (Dron, Rabouille, Claquin, Le Roy, et al., 2012;Masuda et al., 2018;Shi et al., 2010).The temporal segregation helps to solve the problem of nitrogenase being inactivated by oxygen, a byproduct of oxygenic photosynthesis (Fay, 1992;Gallon, 1992).genome expression analysis on C. watsonii grown under a 12:12 h light:dark cycle using a microarray identified that 47.4% of genes had a diel expression pattern (Shi et al., 2010).This diel expression pattern was also observed in Crocosphaera field populations, underpinning diel changes in N 2 fixation (Wilson et al., 2017).During periods of light, C. watsonii fixes inorganic carbon (Masuda et al., 2018;Mohr et al., 2010) and accumulates storage carbohydrates (Dron, Rabouille, Claquin, Chang, et al., 2012).A part of photosynthetically fixed carbon is exuded outside of cells as EPS and transparent exopolymeric particles (TEP; Rabouille et al., 2017;Sohm et al., 2011).The internally stored carbohydrates are then degraded and used for respiration at the end of the light phase or at the beginning of the dark phase (Dron, Rabouille, Claquin, Chang, et al., 2012) to decrease cellular O 2 concentrations, supply ATP, and reduce the power required to fuel N 2 fixation (Großkopf & Laroche, 2012).Without the presence of additional stressors, N 2 fixation starts at the end of the day, peaking in the middle of the night (Dron, Rabouille, Claquin, Le Roy, et al., 2012;Mohr et al., 2010).This is also the time of DNA synthesis (Dron, Rabouille, Claquin, Le Roy, et al., 2012).Cell division in C. watsonii occurs in the middle of the light period (Dron et al., 2013;Wilson et al., 2017) in parallel to peaks of carbon fixation and growth.The timing of these metabolic processes in cyanobacteria is often assumed to be regulated by the circadian clock based on the circadian-related kai genes (Shi et al., 2010), which are proposed to vary with light-driven changes in energy metabolism (Rust et al., 2011).However, in the case of C. watsonii WH8501, cyclic patterns of gene expression related to N 2 fixation (nifH), photosynthesis (psaC, psbA, and psbO), and circadian clock dynamics (kaiA and kaiC) continues in constant light (Mohr et al., 2013;Pennebaker et al., 2010).

PHOTOSYNTHESIS
The photosynthetic apparatus of Crocosphaera watsonii is comparable to other unicellular cyanobacteria (Zehr et al., 2007), such as the model Synechocystis sp.PCC 6803.There is strong diel periodicity both in C. watsonii's structure and function.Like in other cyanobacteria, light is absorbed by pigments of photosynthetic antennae that involve membrane proteins containing chlorophyll a and the extramembrane phycobilisomes (Ting et al., 2002).The dominant phycobilins of C. watsonii are the short-wavelength-absorbing phycoerythrin and phycourobilin (Webb et al., 2009).This is in contrast to UCYN-C (C.subtropica; Mareš et al., 2019) which predominantly contains the longer-wavelength-absorbing phycocyanin.According to genome analysis, C. watsonii strains WH8501 and WH0401 lack the Orange Carotenoid Protein (OCP) that is required for the proper function of the energy dissipative photoprotection mechanism modulating the flow of energy from phycobilisomes to reaction centers in most of the sequenced cyanobacteria (Bao et al., 2017;Kerfeld et al., 2017).The OCP photoprotective mechanism is also absent in other ecologically important marine cyanobacteria from oligotrophic regions (Prochlorococcus, Synechococcus, Trichodesmium;Bao et al., 2017;Kelman et al., 2009), suggesting that the homeostasis of energy flow under conditions of severe nutrient limitations is achieved by regulation at other levels, e.g., by investing the excess energy into exporting carbon as EPS or TEP.We note that C. watsonii WH0003 does possess OCP (Kerfeld et al., 2017), so there is variability between strains in terms of photoprotection mechanisms.
Another mechanism known for optimizing the energy flow in cyanobacteria under changing light conditions is the so-called state transition (Kirilovsky et al., 2014), a process that modulates the coupling and distribution of absorbed energy flow from phycobilisomes to photosystems I and II (PSI and PSII).The presence and diel rhythms of state transitions have been reported for UCYN-C (Crocosphaera subtropica, formerly Cyanothece; Meunier et al., 1997).We have observed a similar diel rhythm for state transitions in C. watsonii (S. Rana et al., unpublished data).Recently, another photoprotective, energy-quenching mechanism mediated by the IsiA protein has been proposed to operate in cyanobacteria, not only under limited Fe concentrations but also under high light (Chen et al., 2017(Chen et al., , 2021)).Crocosphaera watsonii contains the isiA gene.Whether this photoprotective mechanism operates in C. watsonii under the low Fe and high light conditions that exist in oligotrophic areas remains to be elucidated.
The direct effect of illumination on decomposition and lack of recovery of PSII in Crocosphaera watsonii has been confirmed by biochemical protein analysis together with variable chlorophyll fluorescence measurements (Masuda et al., 2018;Rabouille & Claquin, 2016).The decline of PSII activity and decreased PSII abundance in the dark phase has been explained by a combination of monomerization of PSII and gradual disassembly of a large portion of PSII core complexes (Masuda et al., 2018).Specifically, PSII reaction center D1 protein (PsbA) abundance decreased in the dark, and the non-functional version of D1 (due to the absence of key amino acid residues essential for binding the oxygenevolving CaMn 4 O 5 clusters; the so-called "rogue" D1 or rD1; Murray, 2012;Sheridan et al., 2020;Wegener et al., 2015) was detected only in a small percentage of PSII during the dark period (Masuda et al., 2018).The role of rD1 remains unknown.Suppressed N 2 fixation activity under a subjective dark period (i.e., a time when cells were kept in a continous light condition and not returned to the expected dark phase; Mohr et al., 2013) together with a decreasing trend of N 2 fixation activity at the beginning of the light period (Compaoré & Stal, 2010;Shi et al., 2010) suggests that nitrogenase synthesis regulated by the circadian clock and is directly entrained by the energy metabolisms, as Rust et al. (2011) posited in relation to Synechococcus elongatus, as well as by inactivation of the nitrogenase complex by O 2 generated during photosynthesis.

QUANTITATIVE MODELING
There are a growing number of studies that use quantitative models to evaluate Crocosphaera watsonii biogeography and physiology.This research had its start with the numerical modeling of diazotroph biogeography as a function of light and resource limitation (Monteiro et al., 2010(Monteiro et al., , 2011;;Saito et al., 2011).For example, Saito et al. ( 2011) used a numerical model and proteomics experimental data from C. watsonii, to argue that modulation of iron-metalloenzyme inventory allowed Crocosphaera to inhabit a broader range of regions low in Fe, predicting a wide distribution in the subtropical gyres (Saito et al., 2011).Further use of numerical models in subsequent studies have been expanded to look at broader patterns of diversity and biogeography in diazotrophs in general and with Crocosphaera-like analogs (Dutkiewicz et al., 2015(Dutkiewicz et al., , 2020)).
Separately, more detailed physiological models have been developed (Grimaud et al., 2014;Inomura, Deutsch, et al., 2019;Inomura, Masuda, & Gauglitz, 2019;Nicholson et al., 2018), which Inomura et al. (2020) have termed "coarse-grained models" (Inomura et al., 2020).These types of models started with the representation of C and N fluxes (Grimaud et al., 2014), and recently, O 2 and iron fluxes were included (Inomura, Deutsch, et al., 2019;Nicholson et al., 2018).Models focused on day-night cycles have shown that the many metabolic pathways in Crocosphaera watsonii can be explained by a cellular clock (Grimaud et al., 2014).In particular, the metabolisms of N 2 fixation, respiration, and photosynthesis are strongly coupled with the time-dependent allocation of intracellular Fe (Inomura, Deutsch, et al., 2019; Figure 5).Also, the temperature dependencies of respiration may explain why C. watsonii is constrained to warm regions (Church et al., 2009;Fu et al., 2014;Moisander et al., 2010): Lower temperature may increase the intracellular O 2 (Inomura, Deutsch, et al., 2019), which may prevent N 2 fixation.This trait differs from what is observed in UCYN-A: UCYN-A may tolerate low temperatures (Harding et al., 2018;Shiozaki et al., 2020;Shiozaki, Fujiwara, et al., 2018).Another coarse-grained model resolved broad-brush macromolecular allocation (Inomura, Masuda, & Gauglitz, 2019) based on the chemostat culture study of C. watsonii by Masuda et al. (2013).Inomura, Masuda, & Gauglitz (2019) showed the that by fixing nitrogen, C. watsonii can increase their population sizes and expand their niche despite the presence of ammonium.A recent model of O 2 diffusion coupled with simple metabolic fluxes suggested that a strong barrier against oxygen exists in C. watsonii (Inomura et al., 2017;Inomura, Deutsch, et al., 2019).This barrier is consistent with the idea that hopanoidenriched membranes, which are a found in nonheterocyst-forming cyanobacterial diazotrophs like C. watsonii, might limit O 2 permeability and thus help to maintain N 2 fixation (Cornejo-Castillo & Zehr, 2019;Sáenz, 2010).Last, a recent model resolves a heterogeneous metabolic population of C. watsonii, suggesting that heterogeneous N 2 fixation may reduce the requirement of the population, extending its viable depth range (Masuda et al., 2020).

CONCLUSIONS
The availability of the environmental nifH sequences shed light on the previously unappreciated small unicellular cyanobacterial nitrogen fixer Crocosphaera watsonii (Zehr et al., 2001).Extensive field observations, recently compiled by Shao et al. (2023), have corroborated the wide distribution of UCYN-B nifH genes.In combination with chemical and physical parameters, we now know that UCYN-B nifH gene copies are highest in warm (greater than ~25°C), low nutrient surface waters; however, its distribution is not limited to these regions, and the regulation mechanisms of its distribution are not yet well known.Measuring the physiological responses of C. watsonii to chemical and/or physical changes in laboratory experiments, combined with theoretical modeling, is a promising new area of study.Thus, with the increasing amount of both field observations and controlling parameters, knowledge of the global distribution pattern of C. watsonii and its drivers is expected to increase rapidly in the near future.
Over the past decade, multiple approaches have been applied and extensive effort has been put toward estimating the contribution of Crocosphaera watsonii to global N 2 fixation, which has led to remarkable discoveries.In the future, it will be important to combine in situ observations, culture experiments, and modeling to better estimate the contribution of C. watsonii to global C and N cycling.For example, current UCYN-B nifH gene copy abundances do not directly represent cell abundances (Gradoville et al., 2022;Sargent et al., 2016;White et al., 2018), the ratios of nifH copies per cell abundances are likely different among different diazotroph groups, and not all the cells fix N 2 , even though all cells fix C (Foster et al., 2013;Masuda et al., 2020;Mohr et al., 2013).Even though UCYN-B nifH gene copy measurements via qPCR are based on the detection of genomic DNA, the significant diel changes in DNA topology (Pennebaker et al., 2010), in gene expression, and in metabolic activities (Dugenne et al., 2020;Shi et al., 2010) indicate that the time of the day should be taken into account in future analyses of C. watsonii distribution.Furthermore, intriguing but not yet fully understood phenomena have been discovered, such as (1) intracellular recycling of Fe (Saito et al., 2011), photosystem II inactivation in the dark (Masuda et al., 2018;Rabouille & Claquin, 2016), (2) the division of metabolic activities in the cells of selfaggregates (Masuda et al., 2020), and (3) increased growth under co-limitation compared to single nutrient limitation (Yang et al., 2022).Although most of the studies have focused on the response of C. watsonii to physical and chemical conditions, we do not fully understand the effects of co-limitation (e.g., temperature, light, and nutrient source or multiple nutrient sources) nor know the full extent of the organisms that are grazing C. watsonii (Dugenne et al., 2020;Wilson et al., 2017).
Even though most of the culture studies have been based on the Crocosphaera watsonii WH8501 strain, there is intra-specific C. watsonii diversity (Bench et al., 2013).Together with continuous efforts to isolate new strains, the comparison of physiological studies based on multiple clonal cultures may be important.Furthermore, future models of oceanic nutrient cycling should consider the phylogenetic and physiological diversity of unicellular marine diazotrophs within the Crocosphaera genus as well as other related freeliving and symbiotic lineages.In sum, the drastic diurnal variation in metabolisms, including the circadian clock, the relationship between photosynthesis and N 2 fixation, and intercellular heterogeneity in metabolic activity, make C. watsonii not only an important model for learning about diazotrophs but also a model organism for studying cyanobacteria in general.

DATA AVAI L ABI LI T Y STATEM ENT
The data supporting this study are available on request from the corresponding author (TM).
approaches.Recently, the taxonomy of unicellular diazotrophs was revised based on modern comprehensive techniques, which involved F I G U R E 2 Phylogeny of Crocosphaera and related organisms.(a) Phylogenomic tree (120 conserved proteins) of cyanobacteria, showing the clustering of planktonic marine diazotrophs in multiple separated derived lineages.The Crocosphaera (UCYN-B and C) and UCYN-A lineages host unicellular diazotrophic cyanobacteria, while other planktonic marine cyanobacteria are either filamentous diazotrophs or unicellular non-diazotrophs.(b) Comparison of phylogenetic clustering within the UCYN-A, B, and C lineages based on 16S rRNA gene, multilocus (120 proteins), and nifH trees.The UCYN A-C clades and the clade of related diatom endosymbionts are detectable in all three trees; however, the topology of the nifH tree is incongruent with the conserved loci, indicating possible horizontal gene transfer among related lineages.The methods for phylogenetic inference are provided as Appendix S1. [Color figure can be viewed at wileyonlinelibrary.com]

F
I G U R E 3 A global map of UCYN-B nifH copies plotted from Shao et al. (2023).(a) The geographic distribution of the maximum observed UCYN-B nifH copies (>10 copies • L −1 ) at a location.The plots were compiled from 3416 published observations of UCYN-B nifH gene copy number from 0 to 200 m depth (Shao et al., 2023 and references therein).The colors in the circles indicate the maximum abundance on a logarithmic scale, the black dots indicate a result reported as below the detection limit by Shao et al. (2023).The background color is the nitrate concentration with contour lines indicating the surface nitrate concentrations obtained from the World Ocean Atlas 2018.Red lines indicate the 20°C isotherm of the sea surface temperature.(b) Latitudinal distribution of the maximum UCYN-B nifH copies (>10 copies • L −1 ), shown with violin plots using data from Shao et al. (2023).The colored dots show the temperature of the seawater at the sampling station.The black dots indicate the absence of temperature data.[Color figure can be viewed at wileyonlinelibrary.com]Both NH + 4 and NO − 3 assimilation can save energy (ATP) relative to N 2 fixation, but Crocosphaera watsonii keeps fixing N 2 in the presence of NH + 4 and NO −

F
Relationship of UCYN-B nifH gene copies • L −1 and physical and chemical parameters based on the data set used for Figure 3 (Shao et al., 2023 and references therein).UCYN-B nifH gene copies • L −1 against (a) depth, (b) temperature, (c) nitrate concentrations, and (d) phosphorus concentrations.

F
I G U R E 5 An example of a recent quantitative model of Crocosphaera watsonii (Cell Flux Model of Crocosphaera: Inomura, Deutsch, et al., 2019).(a) A schematic of the model.(b-e) Model results of diurnal patterns are compared with data.Model results and data are represented by curves and circles, respectively.(b-d) N 2 fixation, respiration, and photosynthesis under two O 2 conditions.Data are from Großkopf and LaRoche (2012).(e) Fe allocation within the cell.Gray shading indicates the dark period.Fe allocation data are from Saito et al. (2011).[Color figure can be viewed at wileyonlinelibrary.com]