Cobalt: An Essential Micronutrient for Plant Growth?

Cobalt is a transition metal located in the fourth row of the periodic table and is a neighbor of iron and nickel. It has been considered an essential element for prokaryotes, human beings, and other mammals, but its essentiality for plants remains obscure. In this article, we proposed that cobalt (Co) is a potentially essential micronutrient of plants. Co is essential for the growth of many lower plants, such as marine algal species including diatoms, chrysophytes, and dinoflagellates, as well as for higher plants in the family Fabaceae or Leguminosae. The essentiality to leguminous plants is attributed to its role in nitrogen (N) fixation by symbiotic microbes, primarily rhizobia. Co is an integral component of cobalamin or vitamin B12, which is required by several enzymes involved in N2 fixation. In addition to symbiosis, a group of N2 fixing bacteria known as diazotrophs is able to situate in plant tissue as endophytes or closely associated with roots of plants including economically important crops, such as barley, corn, rice, sugarcane, and wheat. Their action in N2 fixation provides crops with the macronutrient of N. Co is a component of several enzymes and proteins, participating in plant metabolism. Plants may exhibit Co deficiency if there is a severe limitation in Co supply. Conversely, Co is toxic to plants at higher concentrations. High levels of Co result in pale-colored leaves, discolored veins, and the loss of leaves and can also cause iron deficiency in plants. It is anticipated that with the advance of omics, Co as a constitute of enzymes and proteins and its specific role in plant metabolism will be exclusively revealed. The confirmation of Co as an essential micronutrient will enrich our understanding of plant mineral nutrition and improve our practice in crop production.


INTRODUCTION
Cobalt is an essential nutrient for prokaryotes, human beings, and other mammals but has not been considered an essential micronutrient for plants. Instead, this element, along with other elements, such as aluminum (Al), selenium (Se), silicon (Si), sodium (Na), and titanium (Ti), has been considered as a beneficial element for plant growth (Pilon-Smits et al., 2009;Lyu et al., 2017). An element that can improve plant health status at low concentrations but has toxic effects at high concentrations is known as a beneficial element (Pais, 1992). For an element to be considered essential, it must be required by plants to complete its life cycle, must not be replaceable by other elements, and must directly participate in plant metabolism (Arnon and Stout, 1939). It has been well-documented that there are 92 naturally occurring elements on the earth, wherein 82 of which have been found in plants (Reimann et al., 2001). Plants are able to absorb elements from soils either actively or passively due to their sessile nature. The occurrence of an element in plants, particularly in shoots, must have a purpose. Active transport of an element from roots to shoots may indicate a certain role it plays in plants. As stated in the study by Bertrand (1912), potentially, every element has a biological function that can be assessed properly against a background of a deficiency state, and every element is toxic when present at high enough concentrations, which is known as Bertrand's rule of metal necessity.
Significant progress has been made in plant mineral nutrition since the publication of Bertrand's rule (Bertrand, 1912) and the essentiality concept (Arnon and Stout, 1939). Among the beneficial elements, cobalt (Co) could potentially be an essential plant micronutrient. Co is a core element of cobalamin (vitamin B 12 and its derivatives) and a cofactor of a wider range of enzymes and a component of different proteins in prokaryotes and animals (Maret and Vallee, 1993;Kobayashi and Shimizu, 1999;Harrop and Mascharak, 2013;Odaka and Kobayashi, 2013). Co-containing enzymes and proteins in plants require further investigation and clarification. Rhizobia and other nitrogen (N)-fixation bacteria require Co and cobalamin for fixing atmosphere dinitrogen (N 2 ) into ammonia (NH 3 ), providing plants with the essential macronutrient of N. Co plays a vital role in interaction with iron (Fe), nickel (Ni), and zinc (Zn) in maintaining cellular homeostasis. Similar to other essential micronutrients, plants respond to Co concentrations in soil: at low concentrations, it promotes plant growth but causes phytotoxicity at higher concentrations. However, it is different from other beneficial elements, as plants do exhibit Co deficiency when grown in soils with limited supply.
The objective of this article was to concisely review the importance of Co as a plant micronutrient including its role in N fixation, the occurrence of coenzyme or proteins, and its effects on plant growth as well as Co deficiency and toxicity. We intended that this review could raise an awareness that Co is a potentially essential micronutrient of plants, and further research is needed to confirm this proposition.

COBALT AND NITROGEN-FIXATION IN PLANTS
Cobalt was isolated by Brandt in 1735 and recognized as a new element by Bergman in 1780 (Lindsay and Kerr, 2011). The importance of Co to living things was realized in the 1930s during the investigation of ruminant livestock nutrition in Australia (Underwood and Filmer, 1935). Co was discovered to be essential for animals as it is a component of cobalamin. Five scientists were awarded Nobel Prizes for the investigation of cobalamin (Carpenter, 2004).

Cobalt Is a Core Element of Cobalamin
Cobalamin is a large molecule (C 63 H 88 O 14 N 14 PCo) comprised of a modified tetrapyrrole ring known as corrin with Co 3+ in the center (Osman et al., 2021). Co is not inter-exchangeable with other metals in the cobalamin and cannot be released from the ring unless the ring is broken (Yamada, 2013), implying the significance of Co to cobalamin. There are two biologically active forms of cobalamin, namely, methylcobalamin and adenosylcobalamin in ruminants (Gonzalez-Montana et al., 2020). In human beings, Co is a cofactor of two enzymes, namely, ethylmalonyl-CoA mutase (MCM) and methionine synthase. MCM catalyzes the reversible isomerisation of lmethylmalonyl-CoA to succinyl-CoA. A deficiency of MCM causes an inherited metabolism disorder commonly known as methylmalonic aciduria. Methionine synthase utilizes cobalamin as a cofactor to produce methionine from homocysteine ( Table 1). Reduced activity of this enzyme leads to megaloblastic anemia (Tjong et al., 2020). Ruminant animals produce vitamin B 12 if there is an appropriate supply of Co in their diet. It was reported that 3 to 13% of the Co was incorporated into cobalamin by bacteria in the ruminant animals (Huwait et al., 2015).

Cobalamin Biosynthesis in Bacteria and Archaea
The natural forms of vitamin B 12 are 1,5deoxyadenosylcobalamin, hydroxycobalamin, and methylcobalamin (Nohwar et al., 2020). They are synthesized by a selected subset of bacteria and archaea (Heal et al., 2017;Guo and Chen, 2018), which include Bacillus, Escherichia, Fervidobacterium, Kosmotoga, Lactobacillus, Mesotoga, Nitrosopumilus, Petrotoga, Propionibacterium, Proteobacteria, Pseudomonas, Rhodobacter, Rhizobium, Salmonella, Sinorhizobium, Thermosipho, and Thermotoga (Doxey et al., 2015;Fang et al., 2017). Cyanocobalamin is not a natural form but commercially synthesized B 12 . The production of vitamin B 12 by these microbes involves about 30 enzymatic steps through either aerobic or anaerobic pathways. In addition to being essential for fat and carbohydrate metabolism and synthesis of DNA, vitamin B 12 is a cofactor of many enzymes. There are more than 20 cobalamin-dependent enzymes in those prokaryotes including diol dehydratase, ethanolamine ammonialyase, glutamate, and methylmalonyl-CoA mutase, methionine synthase, and ribonucleotide reductase (Marsh, 1999) (Table 1). These enzymes catalyze a series of transmethylation and rearrangement reactions (Rodionov et al., 2003). Thus, Co is essential for those archaea and bacteria.

Cobalt Plays an Important Role in Biological Nitrogen Fixation
Biological N fixation is a process of converting N 2 from the atmosphere into plant-usable form, primarily NH 3. Biological N fixation (BNF) is carried out by a group of prokaryotes known as diazotrophs, which are listed in Table 2, including bacteria, mainly Rhizobium, Frankia, Azotobacter, Mycobaterium, Azospirillum, and Bacillus; Archaea, such as Methanococcales, Methanobacteriatles, and Methanomicrobiales, and cyanobacteria, like Anabaena, Nostoc, Toypothrix, and Anabaenopsis (Soumare et al., 2020). N 2 -fixing organisms are also classified into three categories: symbiotic, endophytic, and associated groups (Figure 1). Such classifications may not be accurate as some of them, such as those from Acetobacter and Azospirillum, could be associated, as well as endophytic bacteria.

Cobalt Is Essential for Symbiotic Bacteria in N Fixation
There are two major symbioses between N 2 -fixing bacteria and higher plants, one is rhizobia with leguminous plants and the other is Frankia with actinorhizal plants (Wall, 2000). The former involves more than 1,700 plant species in the family Fabaceae, which includes some economically important crops, such as alfalfa, beans, peas, and soybeans. More than 220 species are actinorhizal plants, which are mainly trees and shrubs forming symbiotic relationships with Frankia.
Rhizobia are gram-negative bacteria encompassing Rhizobium, Azorhizobium, Sinorhizobium, Bradyrhizobium, and Mesorhizobium (Table 2, Figure 1A). Co was identified to be essential for Rhizobium in the 1950s and 1960s (Ahmed and Evans, 1960;Reisenauer, 1960). Rhizobium uses nitrogenase to catalyze the conversion of N 2 to NH 3 , which can be readily absorbed and assimilated by plants. Three enzymes, namely, methionine synthase, methyl malonyl-CoA mutase, and ribonucleotide reductase in Rhizobium and Bradyrhizobium species, are known to be cobalamin-dependent and significantly affect nodulation and N fixation. Early studies showed that four soybean seedlings inoculated with rhizobia supplemented with 1 µg/L Co were healthy and produced 25.3 g of dry weight. On the contrary, four rhizobia-inoculated seedlings devoid of Co encountered N-deficiency symptoms and produced 16.6 g of dry weight, a 34.4% reduction in biomass due to the absence of Co (Ahmed and Evans, 1959). A close relationship was established amongst Co supply, cobalamin content in Rhizobium, leghemoglobin formation, N fixation, and plant growth (Kliewer and Evans, 1963a,b). The deficiency in Co significantly affects methionine synthase by reducing methionine synthesis, which subsequently decreases protein synthesis and produces smaller-sized bacteroids (bacteria in the nodules capable of N fixation) (Marschner, 2011). Methyl malonyl-CoA mutase catalyzes the production of leghemoglobin. If Co becomes limited, leghemoglobin synthesis is directly affected, resulting in reduced N fixation and ultimately a shortage of N supply. This is because leghemoglobin can protect nitrogenase from oxygen by limiting its supply (Hopkins, 1995). Ribonucleotide reductase is a cobalamin-dependent enzyme that catalyzes the reduction of ribonucleotides to deoxyribonucleotides, which is a rate-limiting step in DNA synthesis (Kolberg et al., 2004).
The genus Frankia is composed of gram-positive and gramvariable actinomycetes (Wall, 2000). It infects plants through root hairs and produces nodules in the pericycle. Frankia in nodules develops vesicles in which nitrogenase is suited (Huss-Danell, 1997). Co is needed for the synthesis of cobalamin which is in turn needed for N fixation. Actinomyceters are known as active producers of cobalamin (Hewitt and Bond, 1966). N fixation by actinorhizal plants appears to be comparable to the magnitude as that of the legumes (Wall, 2000).
Other symbioses occur in cyanobacteria with Gunnera and cycads. The genus Nostoc infected specialized gland organs located on the stems of Gunnera, such as G. chilensis and G. magellanica (Johansson and Bergman, 1994). Cyanobacteria also form symbiotic relationships with cycads in a special type of root system called coralloid roots (Chang et al., 2019). It has been welldocumented that cyanobacteria require Co for the biosynthesis of cobalamin (Cavet et al., 2003).

Cobalt and Plant Associated N 2 Fixing Bacteria
Associated N 2 -fixing bacteria include Azotobacter, Azospirillum, Beijerinckia, Burkholderia, Clostridium, Herbaspirillum, Gluconacetobacter, Methanosarcina, and Paenibacillus (Table 2, Figure 1C). These bacteria are associated with the roots of Frontiers in Plant Science | www.frontiersin.org a wide range of plants, including corn, rice, sugarcane, and wheat (Aasfar et al., 2021). Among them, the genus Azotobacter was first reported in 1901 and has been used as a biofertilizer thereafter (Gerlach and Vogel, 1902). Notable species found in soils are A. chroococcum, A. vinelandii, A. beigerinckii, A. armeniacus, A. nigricans, and A. paspali (Das, 2019). The genome of A. vinelandii DJ has been sequenced (Setubal et al., 2009). N fixation in these species is under aerobic conditions, and two-component proteins of Mo-dependent nitrogenase catalyze N 2 into NH 3 . Co and vitamin B 12 were found to be required by A. vinelandii OP. Additionally, 5,6dimethylbenzimidazolylcobamide coenzyme was identified in this species, which might play an important role in N fixation (Nicholas et al., 1962). Furthermore, higher concentrations of Co were needed for A. vinelandii to fix N 2 than was needed for the utilization of ammonium compounds (Evans and Kliewer, 1964). Co at a concentration of 0.1 mg/L was reported to increase N fixation in A. chroococcum in Jensen's medium (Iswaran and Rao Sundara, 1964). Culture of A. chroococcum in half-strength N-free Jensen's broth showed that N fixation was enhanced after supplemented with Co at 12.5 mg/L or 25 mg/L (Orji et al., 2018). Azotobacters were able to biosynthesize a series of vitamins, including B 12 in chemically-defined media and dialyzed soil media (Gonzalez-Lopez et al., 1983;El-Essawy et al., 1984). In addition to A. vinelandii and A. chroococcum, Pseudomonas fluorescens, Bacillus megaterium, Bacillus firmus, and Sinorhizobium meliloti also produce cobalamin (Palacios et al., 2014), and the synthesized cobalamin may implicate the enhanced N fixation in these bacteria.
Azosprillum is another important genus of plant-associated N 2 -fixing bacteria. A. brasilense cultured on medium supplemented with 0.2 mM Co was able to accumulate Co up to 0.1 to 0.6 mg per gram of dry biomass . 57 Co emission Mössbauer spectroscopy (EMS) studies of Co in Azospirillum brasilense Sp245 showed that Co activated glutamine synthetase to have two different Co forms at its active sites. In vitro, biochemical and spectroscopic analyses showed that Co 2+ is among the divalent cations, along with Mg 2+ and Mn 2+ , most effective in supporting the activity of glutamine synthetase at different adenylylation states, a key enzyme of N metabolism .

Nitrogen Fixing Bacteria and Crop Productivity
Nitrogen is an essential macronutrient for plants. The application of synthetic N fertilizers has greatly enhanced crop production but also has caused serious environmental problems, such as groundwater contamination and surface water eutrophication (Hansen et al., 2017). As a result, exploring the potential of BNF becomes increasingly important. The symbiotic relationship between rhizobia and legume crops was considered the most important BNF system and estimated to contribute to 227 to 300 kg N/ha/year (Roughley et al., 1995;Herridge et al., 2008). N 2 fixation by actinorhizal plants was estimated to be 240-350 kg N/ha/year (Wall, 2000).
Nitrogen fixation by plant-associated diazotrophs has been estimated to be 60 kg N/ha/year (Gupta et al., 2006;Reed et al., 2011). Moreover, the abundance of associated diazotrophs, such as Azotobacter species in the soil provides not only N (Din et al., 2019) but also phosphorus and plant growth regulators, which resulted in a yield increase of up to 40% in cereals and pulse crops (Yanni and El-Fattah, 1999;Choudhury and Kennedy, 2004;Kannan and Ponmurugan, 2010;Ritika and Dey, 2014;Wani et al., 2016;Velmourougane et al., 2019). Such beneficial effects have been harnessed ecologically in the engineering of Azotobacter species for fixing plant needed N, while reducing the reliance on synthetic N fertilizers for crop production in an environmentally friendly manner (Wani et al., 2016;Bageshwar et al., 2017;Ke et al., 2021).
Endophytic bacteria also contribute significantly to N input. Azoarcus is an endophytic N 2 -fixing diazotroph, and its action in roots of kallar grass increased hay yield up to 20-40 t/ha/year without N fertilizer application in salinesodic, alkaline soils (Hurek and Reinhold-Hurek, 2003). Gluconoacetobacter diazotrophicus (Acetobacter diazotrophicus) is the main contributor in sugarcane and can fix up to 150 kg N/ha/year (Dobereiner et al., 1993;Muthukumarasamy et al., 2005). Many C-4 energy plants, such as Miscanthus sacchariflorus, Spartina pectinate, and Penisettum purpureum can harbor endophytic bacteria, which support the N requirement of these plants (Kirchhof et al., 1997). Gupta et al. (2012) reported that N derived from the air by endophytic bacteria for rice ranged from 9.2 to 47% depending on bacterial species. These results indicate that endophytic diazotrophs have a great potential to enhance the productivity of non-leguminous crops.
The aforementioned bacteria essentially act as the same as gut bacteria in mammals by living between plant cells as endophytes, close association with roots, or symbiotically and become indispensable for plant growth and development. Microorganisms are associated with all plant organs , but roots have the largest number and greatest range of microbes. Thus, a plant growing under field conditions is a community, not an individual. Such associations are collectively termed "phytomicrobiome." The phytomicrobiome is integral for plant growth and function. Microbes play important roles in plant nutrient acquisition, biotic and abiotic stress management, physiology regulation through microbe-to-plant signals, and growth regulation via the production of phytohormones. The foregoing discussion documents the role of Co plays in N 2 fixing rhizosphere bacteria. If we accept that coevolution exists between microbes and plants and the phytomicrobione in general, Co should be considered as an essential element to plants as it is required by symbiotic, endophytic, and associated bacteria.

Cobalt Proteins in Plants
There are several lines of evidence suggesting that plants may have cobalamin-dependent enzymes and Co-containing proteins: (1) The ancestor of the chloroplast is cyanobacteria (Falcón et al., 2010), and Co is required by this group of bacteria. The speculation is that Co may be needed by plants.
(2) Plants have been documented to utilize cobalamin produced by symbiotic, endophytic, and associated N 2 fixing bacteria. Cobalamin concentrations of 37, 26, and 11 µg/100 g dry weight were detected in Hippophae rhammoides, Elymus, and Inula helenium, respectively (Nakos et al., 2017). There is a possibility that cobalamin-dependent enzymes may occur in plants. Poston (1977) reported the identification of leucine 2,3aminomutase in extracts of bean seedlings. Its activity was stimulated by coenzyme B 12 but inhibited by unknown factors. The inhibition was removed by the addition of B 12 , suggesting the presence of a cobalamin-dependent enzyme in higher plants. Subsequently, two coenzyme B 12 -dependent enzymes: leucine 2,3-aminomutase and methylmalonyl-CoA mutase were reported in potato tubers (Poston, 1978), but methylmalonyl-CoA mutase was found to be a phosphatase (Paizs et al., 2008).
(3) Co is required by lower plants, which is to be discussed in the following section. (4) Plants can take up and transport cobalamin (Mozafar, 1994;Sato et al., 2004). A recent study using fluorescent analogs to follow the uptake and transport of cobalamin showed that Lepidium sativum can absorb cobalamin (Lawrence et al., 2018). Seed priming with cobalamin provided significant protection against the salt stress of common beans (Keshavarz and Moghadam, 2017). The incorporation of Co in plant tissue culture media significantly improves plantlet production (Bartolo and Macey, 1989). (5) Co as a metal cofactor of some additional enzymes and proteins are briefly discussed below ( Table 1).
Carboxypeptidases (CPSs, EC: 3.4.16-3.4.18) are proteases hydrolyzing the C-terminal residues of peptides and proteins and release free amino acids individually. CPSs are divided into serine (EC: 3.4.16), metal (EC: 3.4.17), and cysteine (EC: 3.4.18) and occur in animals, bacteria, fungi, and plants. One Zn atom is essential to the catalytic activity of native carboxypeptidase A. Zn can be removed by dialysis at low pH or with chelating agents at neutral pH, which results in the inactivation of the enzyme. The re-addition of the metal restores the dual activities of carboxypeptidase toward peptides and esters. Co was found to be more active than Zn in the enzyme toward peptides and has nearly the same activity toward esters, indicating that Co in the active site is virtually identical to that of Zn in the native enzyme (Maret and Vallee, 1993).
Methionine aminopeptidase (MAP, EC 3.4.11.18) is widely documented in animals, bacteria, yeast, and plants. It is a Codependent enzyme responsible for the cleavage of the N-terminal methionine from newly translated polypeptide chains. Two classes of MAPs (MAP1 and MAP2) were reported in bacteria, and at least one MAP1 and one MAP2 occur in eukaryotes (Giglione and Meinnel, 2001). In Arabidopsis, there are four MAP1s (MAP1A, MAP1B, MAP1C, and MAP1D) and two MAP2s (MAP2A and MAP2B), along with two class 1 peptide deformylases (PDF1A and PDF1B). The plant MAP proteins show significant similarity to the eubacterial counterparts except for MAP1A and two MAP2s. It has been documented that the substrate specificity of PDFs and both organellar and cytosolic MAPs in plants are similar to that of their bacterial counterparts (Giglione et al., 2000). The MAP from Salmonella typhimurium is stimulated only by Co 2+ , not by Mg 2+ , Mn 2+ , or Zn 2+ and is inhibited by metal ion chelator EDTA. E. coli MAP is a monomeric protein of 29 kDa consisting of 263 residues that possess two Co 2+ ions in its active site (Permyakov, 2021).
Peroxidases are isoenzymes present in all organisms, which catalyze redox reactions that cleave peroxides; specifically, it breaks down hydrogen peroxide. The study of Han et al. (2008) found that Co 2+ at a concentration below 0.1 mM increased horseradish peroxidase activity because Co 2+ binds with some amino acids near or in the active site of the enzyme.
Urease is an enzyme occurring in selected archaea, algae, bacteria, fungi, and plants. It catalyzes the hydrolysis of urea into ammonia and carbamic acid. The active site of urease contains two Ni 2+ atoms that are bridged by a carbamylated lysine residue and a water molecule (Carter et al., 2009). The study of Watanabe et al. (1994) reported that urease activity of cucumber leaves was markedly reduced when Ni concentration became <100 ng/L, but supplementing Co restored urase activity. Additionally, urease was also activated by both Co and manganese (Mn) through in vitro assay (Carter et al., 2009).
Cobalt transporters. Transporters specifically for Co have not been reported. The current understanding is that Co can be transported through Fe transporters (Figure 2). In Arabidopsis thaliana, Co is taken up from the soil into epidermal cells of roots by IRON-REGULATED TRANSPORTER 1 (IRT1), which is commonly known for absorption of Fe (Korshunova et al., 1999). Once Co is absorbed inside cells, Ferroportins, FPN1, and FPN2 are responsible for its further movement. IREG1/FPN1 is localized to the plasma membrane and expressed in the steel, indicating it is responsible for the loading of Fe to xylem, and FPN2 is situated the in vacuolar membrane and involved in buffering Fe concentration in the cytosol (Morrissey et al., 2009). Truncated FPN2 causes an elevated level of Co in shoots, while the loss of FPN1 abolishes Co accumulation in shoots. A double mutant of fpn1 fpn2 is unable to sequester Co in root vacuole and cannot transport Co to shoots. These results suggest that Co is likely absorbed and transported in the same way as Fe in plants (Figure 2). Additionally, an ATP-binding cassette (ABC) transporter from Arabidopsis has also been reported to transport Co, Ni, and Pb (Morel et al., 2009). Co movement in leaves is also associated with Ni, and Ni and Co movement in or out of chloroplasts are through an ABC transporter in the mediation of ionic homeostasis in the chloroplast of rice (Li et al., 2020).

Cobalt Substitution of Other Metals
A characteristic of Co is its ability to substitute for other transition metals in a large number of enzymes. Maret and Vallee (1993) listed 37 Co-substituted metalloproteins, of which 24 are native to Zn, nine to copper (Cu), and four to Fe. These enzymes mainly occur in animals, bacteria, and yeast, while a few are in plants. Such a characteristic is closely related to the properties of Co with other metals. The ionic radius of Co 2+ is 0.76 Å, which is similar to 0.74 Å of Zn 2+ , 0.69 Å for Cu 2+ , and 0.76 Å for Fe 2+ . Additionally, based on the available Protein Data Bank structures with Co 2+ , the study Khrustalev et al. (2019) found that Co 2+ is commonly bound by cation traps. The traps are formed by relatively negatively charged regions of random coil between a β stand and α helix and between two β strands in which His, Asp, and Glu residues are situated. On the other hand, these sites are also occupied by other metals ions, such as Cu 2+ , Mg 2+ , Mn 2+ , and Zn 2+ , which play significant roles as catalysts. As a result, Co 2+ could rather readily substitute for these ions in the active sites of enzymes. Additionally, based on the FIND-SITEmetal, a program for the prediction of the metal-binding site, the study of Brylinski and Skolnick (2011) found that Zn, due to a lower coordination number preference, is typically chelated with Cys and His, and His residues have a strong preference for Co, Cu, Fe, Ni, and Zn atoms. Thus, Co is able to replace Cu, Fe, Ni, and Zn in the active sites of enzymes. For example, Co addition alleviated Zn limitation in production of Thalassiosira weissflogii, which was due to Co substitution of Zn in the main isoform of carbonic anhydrase (Yee and Morel, 1996). Co substitution of Zn was also reported in two northeast Pacific isolates of diatoms Pseudo-nitzschia delicatissima UNC1205 and Thalassiosira spp. UNC1203 (Kellogg et al., 2020). Co 2+ has been used as a spectroscopically active substitute for Zn 2+ in enzymes (Bennett, 2010). Substitution of tetrahedral Zn 2+ by highercoordinate Co 2+ often results in a catalytically active species, sometimes with catalytic properties perhaps unexpectedly similar to those of the native enzyme. In the vast majority of cases, no other transition ion than Co 2+ provides a better substitute for Zn 2+ (Maret and Vallee, 1993;Bennett, 2010). Due to these reasons, Co specific enzymes or proteins have not been conclusively identified. With the advance of omics, functions of a large number of gene sequences have not been assigned. Using the FIND-SITE-metal, a program developed for prediction of the metal-binding site, Brylinski and Skolnick (2011) predicted that about 10,953 putative metal-binding proteins in human proteome were bound with Ca, 10,534 bound with Mg, 8,681 with Zn, 1,863 with Fe, 1,246 with Mn, 652 with Co, 476 with Cu, and 403 with Ni. The predicted binding proteins with Co are greater than Cu and Ni in humans. Based on this assignment in the human proteome, it could be extremely difficult to believe that there are no Co-containing enzymes and proteins in plants.

COBALT IS ESSENTIAL FOR LOWER PLANTS
Lower plants are commonly known as non-vascular plants because they do not have xylem and phloem vascular systems. Non-vascular plants are generally divided into bryophytes and algae.

Bryophytes
Bryophytes are seedless plants including Anthocerotophyta (hornworts), Bryophyta (mosses), and Marchantiophyta (liverworts) (Davies et al., 2020). This group of plants is able to absorb Co from air, soil, and water. In an early geochemical survey performed in Wisconsin and adjacent states and Missouri and Kentucky in the US, the study of Shacklette (1965) documented that the mean concentration of Co in 38 samples of liverworts and mosses was 32 mg/kg, and the concentration in the lower plants was closely related to the amount of the element in the soil, suggesting they act as a bioindicator of Co concentration in the environment (Baker, 1981). Mosses sampled from streams of the Idaho Cobalt Belt (U.S.) showed that Co concentrations in the plants almost perfectly correlated with those in the sediments, and the maximum content of Co (2,000 mg/kg) in moss ash corresponded to the maximum concentration of 320 mg/kg in the sediment (Erdman and Modreski, 1984). Mosses, such as Bryum argenteum and Hypnum cupressiforme were also considered to be bioindicators for monitoring heavy metal contamination in the air (Andić et al., 2015). Interestingly, the accumulation of Co did not cause any physiological damages to plants, but their growth was further enhanced.
The ability to take up Co could be related to the nonvascular nature and unidentified transporter. A radiolabel study showed that the total amount of 60 Co accumulated FIGURE 2 | Schematic illustration of cobalt (Co 2+ ) absorption, transport, and distribution in plants. Co 2+ is absorbed from the soil into epidermal cells of roots by an iron transporter (IRT1). Once Co 2+ is absorbed inside cells, Ferroportins (FPN1 and FPN2) are responsible for its further movement. FPN2 transports Co 2+ into vacuoles, resulting in the sequestration of Co in root cells. FPN1 is to load Co 2+ into the xylem. In the xylem, Co 2+ is complexed with citrate, histidine (His), methionine (Met), or nicotianamine (NA) to be translocated to shoots. Co 2+ is released in leaves and participate in metabolisms, which are often associated with nickel (Ni) and iron (Fe). It is shown here the Ni and Co movement in or out of chloroplasts through ATP-binding cassette (ABC) transporter in the mediation of ionic homeostasis in the chloroplast.
in P. commune and D. scoparium under given conditions were 7.1 and 6.1 mg/kg, respectively. More than 95% of 60 Co in D. scoparium was localized extracellular, while 70% of 60 Co in P. commune was localized extracellular and about 20% localized intracellularly. These results showed that Co was largely adsorbed extracellularly, and there were unidentified transporters regulating the transport of Co into intracellular sites.
The enhanced growth could be in part attributed to the symbiotic relationship with cyanobacteria. Some bryophytes, primarily liverworts, and hornworts can form a symbiosis with cyanobacteria, such as Nostoc spp. After infection, Nostoc underwent some morphological and physiological changes by reducing growth rate and CO 2 fixation but enhancing the fixation of N 2 as well as releasing fixed N compounds to the plants. Cyanobacteria, like rhizobia, require cobalamin as a cofactor for nitrogenase complex to fix N 2 (Böhme, 1998). Thus, cyanobacteria-bryophyte symbioses require Co.

Algae
Algae constitute a polyphyletic group ranging from unicellular microalgae, like chlorella and diatoms to multicellular forms, such as the giant kelp, seaweeds, and charophytes (Barsanti and Gualtieri, 2006). Co is essential to some marine algal species, including charophyte, diatoms, and dinoflagellates (Nagpal, 2004). Green alga Chlorella salina exhibited two phases of uptake of Co 2+ (Garnham et al., 1992). The initial phase was rapid and independent of metabolism, and the second phase was slow and dependent on metabolism. Competition studies showed that the Co 2+ uptake system was different from that for Mg 2+ , Mn 2+ , and Zn 2+ . The greatest amount of Co was associated with the cell wall. Co concentrations in the cytosol were 0.17 mM but 2.89 mM in the vacuole, suggesting that Co transport was well-controlled in C. salina. In the work of Czerpak et al. (1994), they studied the responses of a freshwater green alga Chlorella pyrenoidosa to different levels of Co and found that Co in a range from 5 to 50 mM significantly enhanced the growth of Chlorella pyrenoidosa, including 150-160 and 50-60% increase in fresh and dry weights, respectively. Such increase was related to the increase of chlorophylls a and b by 45-65%, water-soluble proteins by 19-20%, total carotenoids 55-65%, and monosaccharides content 55-60%, when compared with the culture devoid of Co. Although mechanisms behind the stimulating effects have not been elucidated, it is likely due to the biosynthesis of cobalamin that enhanced alga growth. Two cobalamin coenzyme 5 ′ -deoxyadenosylcobalamin and methylcobalamin occurred in green alga C. vulgaris, and the addition of cobalamin significantly stimulated green alga growth (Watanabe et al., 1997). Moreover, C. vulgaris grown in Bold's basal medium supplemented with 2 and 2.5 µM CoCl 2 produced 166.23 and 173.32 µg vitamin B 12 per 100 g dry weight (Jalilian et al., 2019). Additionally, many algal species require different combinations of cobalamin, vitamin B 1 , and B 7 (Croft et al., 2005) as they do not have pathways to synthesize cobalamin or may use alternative cobalamin-independent routes bypassing the need for the vitamin (Cruz-Lopez and Maske, 2016;Yao et al., 2018). As Co is a constituent of cobalamin, Co is required by those algae.
Some algal species, such as those in the genera Coccomyxa and Elliptochloris as well as diatoms form symbiotic relationships with cyanobacteria (Grube et al., 2017). Co is required for the growth of cyanobacteria, such as Anabaenza cylindrica Lemm (Holm- Hansen et al., 1954) and Prochlorococcus (Hawco et al., 2020) as they need it for N fixation in specialized cells called heterocysts. Thus, algal species symbiotic with cyanobacteria require Co for N-fixation.

COBALT IMPROVES THE GROWTH OF HIGHER PLANTS
Cobalt content in the crust of the earth ranges from 15 to 30 mg/kg (Roberts and Gunn, 2014). Co in soils is closely related to the weathering of parental minerals, such as cobaltite, smaltite, and erythrite (Bakkaus et al., 2005) as well as Co pollution (Mahey et al., 2020). Co in the surface soils of the world varies from 4.5 to 12 mg/kg with the highest level occurring in heavy loamy soils and the lowest in organic and light sandy soils (Kabata-Pendias and Mukherjee, 2007). However, Co in reference soil samples was found to differ from 5.5 to 29.9 mg/kg in the United States (U.S.) and 5.5 to 97 mg/kg in Chinese soils (Govindaraju, 1994). Pilon-Smits et al. (2009) suggested that soil Co concentrations generally range from 15 to 25 mg/kg.

Cobalt in Higher Plants
Plants absorb Co. Table 3 lists Co concentrations in over 140 non-hyperaccumulating species ranging from 0.04 to 274 mg/kg. Average concentrations of Co in grasses vary from 60 to 270 µg/kg and in clover differ from 100 to 570 µg/kg across Australia, Finland, Germany, Great Britain, Japan, New Zealand, Poland, Sweden, and the US (Kabata-Pendias and Mukherjee, 2007). Legumes absorb more Co than grasses. Plants that accumulate metals to a level 100-fold higher than those typically recorded in common plants are known as hyperaccumulators (Brooks, 1998).
As discussed above, Co specific transporters have not been reported, and a schematic diagram for Co absorption and translocation is presented in Figure 2. After absorption by roots, Co is either sequestrated in the vacuole of root cells or transported to shoots. Co that is being transported to shoots is chelated with ligands. Co has little affinity with phytochelatins (Chen et al., 1997;Cheng et al., 2005), thus the ligands are not likely Co-S bonds. The study by Collins et al. (2010) reported that Co 2+ was complexed with carboxylic acids, which were transported from roots to shoots in wheat or tomato plants. Other ligands are citrate or malate as well as non-proteinogenic amino acids, such as histidine and nicotianamine (Figure 2). Co has low mobility within the leaf tissue and is largely distributed in the vascular system of tomato and wheat leaves (Collins et al., 2010). Co transport from roots to shoots is well-controlled. Using radiolabeled 57 Co, Page and Feller (2005) studied Co transport in wheat plants and found that 80% of 57 Co remained in roots after 4 days of culture, and 50% was retained in the roots after 50 days; during which, some 57 Co moved to the apical part of the main roots, suggesting that the loading of Co to the xylem is well-controlled, probably by FPN1 in wheat plants. In another study, Collins et al. (2010) reported that tomato and wheat plants grown in a nutrient solution containing 2.94 mg/L Co had 4,423 µg/kg and 9,319 µg/kg of Co in roots, respectively; but shoot concentrations of Co were 1,581 µg/kg and 395 µg/kg, respectively. This means that 35.7% of Co absorbed by tomato and 4.2% of Co absorbed wheat plants were transported from roots to shoots. Furthermore, for the 1,581 µg/kg Co in tomato shoots, 846 µg/kg was in the stem, 492 µg/kg in old leaves, only 243 µg/kg in young leaves, indicating that only 5.5% of absorbed Co is transported to actively growing shoots of tomato plants. These transport patterns are like those of titanium  which are strictly controlled by plants. These findings imply that plants probably have unidentified transporters specifically for the transport of Co. Due to its toxicity at higher concentrations, the rigorous control of the transport and distribution would ensure that only an appropriate amount of Co could be transported to actively growing shoots. On the other hand, why was more Co transported to dicot tomato shoots than monocot wheat shoots? One explanation could be that different plants have different ligands for complexing Co, and Co complexed by ligands in tomato was more mobile than that in wheat. Another explanation could be that tomato plants need more Co to fulfill some unidentified roles in shoots. Further research is needed to verify these propositions.
To maintain ionic homeostasis in shoots, particularly in chloroplasts, plants develop mechanisms to mediate Co in chloroplasts. An ARG1 transporter, belonging to the ATPbinding cassette, was identified in rice (Li et al., 2020), which was able to modulate the levels of Co and Ni in chloroplasts to prevent excessive Co and Ni from competing with metal cofactors in chlorophyll and metal-binding proteins in photosynthesis (Figure 2).

Plant Growth Improvement
Cobalt at low concentrations can also promote the growth of non-leguminous crops ( Table 4). Co applied to a sandy soil at 1 mg/kg enhanced shoot and root dry weights of wheat by 33.7 and 35.8%, respectively compared with the control (Aery and Jagetiy, 2000), and the same Co rate applied to a sandy loam soil increased shoot and root dry weights of wheat by 27.9 and 39.6%, respectively, compared with the control. The yield and essential oil contents of parsley (Petroselinum crispum) increased considerably after the application of Co at 25 mg/kg soil (Helmy and Gad, 2002). Plant height, branch numbers, and fruit numbers as well as anthocyanin and flavonoids contents of Hibiscus sabdariffa significantly increased after application of Co at 20 and 40 mg/kg (Aziz et al., 2007). Application of 50, 100, 150, 200, and 250 mg/kg Co to corn plants showed that the root length, shoot height, and the number of cobs and seeds per plant increased when plants were applied with 50 mg/kg Co, but these parameters decreased with 100 mg/kg Co and above . Co applied at 10 mg/kg significantly enhanced the growth of two onion cultivars, bulb yields, bulb length, and bulb quality, such as nutrient and essential oil contents. Bulb diameter and bulb weights were much higher than the control treatment (Attia et al., 2014), but Co concentrations higher than 10 mg/kg significantly reduced the promotive effects. Explanations for the improved growth of non-leguminous plants vary but can be summarized as follows: (1) enhanced tolerance to abiotic stresses, (2) activation of antioxidative enzymes, (3) substitution of active metals, and (4) hormesis. Application of Co has been reported to alleviate drought, salt, heavy metal stresses, thus plant growth is not adversely affected. Co has been reported to suppress plant uptake of cadmium (Chmielowska-Bak et al., 2014). Co application increased free proline accumulation, which counteracted the salt stress. In general, abiotic stresses often cause plant imbalance between production and accumulation of reactive oxygen species (ROS), including superoxide anion (O − 2 ), hydroxyl radical (OH − ), and hydrogen peroxide (H 2 O 2 ) (Sachdev et al., 2021). ROS can activate the antioxidant system of the plant, thus minimizing the damages (Tewari et al., 2002;Choudhury et al., 2017). The antioxidant system includes enzymatic antioxidants: ascorbate peroxidase, catalase, dehydroascorbate reductase, general peroxidases, glutathione reductase, monodehydroascorbate reductase, and superoxide dismutase as well as non-enzymatic antioxidants, mainly ascorbic acid, α-tocopherol, carotenoids, reduced glutathione, plastoquinone/ubiquinone, and flavonoids (García-Caparrós et al., 2020). The action of the antioxidant system could be the first line of defense against the adverse effects. Therefore, it is not surprising to notice increased activities of ascorbate peroxidase, catalase, peroxidase, and superoxide dismutase . Co applied at appropriate concentrations can activate antioxidative enzymes, thus reducing ROS-caused damage. As discussed previously, Co may substitute other nutrient elements when such nutrients have limited availability. Baxter et al. (2008) showed that when Arabidopsis plants are grown under a low Fe concentration, the shoot concentration of Co increased, which was concomitant with the increased expression of Fe transporter IRT1. Additionally, Co contribution to hormesis has been proposed (Shahid et al., 2020). Due to the limited research on Co to date, these explanations may not be on target and incomplete. Our proposition is that the application of the appropriate amount of Co may stimulate rhizosphere bacteria (either symbiotic, endophytic, or associated ones) to fix N 2 , increase soil N, and enhance plant growth. Concomitantly,

Species (common name) Co application Effects on plants References
Actinidia chinensis Planch. cv. Hayward) (Kiwi) Fruit was treated with 1 mM Co 2+ solution Inhibited ACC activity in ethylene biosynthesis Hyodo and Fukasawa, 1985 Adiantum raddianum C. Presl (Delta maidenhair fern) Cut green (frond) was treated with 1 mM Co(NO 3 ) 2 solution Prolonged vase life of frond from 3 to 8.2 days Fujino and Reid, 1983 Arachis hypogaea L. (Peanut) Seeds were treated with Co(NO 3 ) 2 at 500 mg/kg seed and followed by two foliar sprays of cobalt nitrate at 500 mg/L before and after flowering Significantly increased plant height, leaf number, pod yield, shelling percentage, harvest index, and total dry matter Raj, 1987 Arachis hypogaea L. (Peanut) CoSO 4 was mixed with soil at 0.21 kg/ha Resulted in 10% higher kernel yield compared with control (without Co application) Basu and Bhadoria, 2008 Arachis hypogaea L. (Peanut) Seedlings of groundnut at the third true leaf stage were irrigated once with CoSO 4 at 2, 4, 6, and 8 mg/L, respectively Increased plant height, number of branches and leaf number, leaf area index, root length, shoot and root biomass as well as pods numbers, pods weight, oil yield, total proteins, total carbohydrates, total soluble sugars, and total soluble solids Gad, 2012a Argyranthemum sp.  Ahmed and Evans, 1959 Glycine max (L.) Merr. (Soybean) Plants were grown in soil mixed with finely powdered (CoCl 2 ) at the concentration of 50 mg/kg Increased root and shoot length, leaf area, dry weight, yield, and yield components Jayakumar et al., 2009 Glycine max (L.) Merr. (Soybean) Seeds were sown in soil mixed with finely powdered (CoCl 2 ) at 50 mg/kg Increased yield parameters, leaf area, shoot length, total dry weight as well as total phenol percentage Vijayarengan et al., 2009 (Continued) Frontiers in Plant Science | www.frontiersin.org  Co enzymes may be triggered to conduct proper biochemical and physiological activities, such as carbonate dehydratase may enhance photosynthesis and Co-peroxidase may activate the enzymatic antioxidant system. As a result, healthy growing plants would take up more nutrients from the soil and improve their growth and overall stress tolerance.

Other Performance Enhancement
Cobalt has been shown to have other beneficial effects on plants. Co as a component of preservative solutions can improve the postharvest quality of floriculture crops by prolonging the vase life of cut flowers. Cut fronds of Delta maidenhair fern (Adiantum raddianum) placed in deionized water became wilted in just 3 days because of the vascular blockage at the basal end of the petiole. The wilting, however, could be delayed for up to 8 days by adding 1 mM Co as Co(NO 3 ) 2 to the water (Fujino and Reid, 1983). The delay of senescence is attributed to the antibacterial activity of Co (Van Doorn et al., 1991). Co addition to preservative solutions increased leaf diffusive resistance, reduced xylem blockage, sustained water flow and uptake, and prolonged vase life of cut flowers of Rosa hybrida "Samantha". Reddy (1988) suggested that partial closure of stomata by Co was responsible for reducing the water loss/water uptake ratio, and thereby maintaining a higher water potential in the cut roses. Co was also reported to slow the senescence process in harvested lettuce (Tosh et al., 1979). Co 3+ has been reported to form Co-complexes, which have antiviral activities (Chang et al., 2010). In addition to antibacterial and antiviral activities, Co shows inhibitory activity to 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase. Ethylene is synthesized from amino acid methionine by two key enzymes, ACC synthase, and ACC oxidase. Co can block the conversion of ACC to ethylene by inhibiting ACC oxidase activity in the ethylene biosynthesis pathway (Lau and Yang, 1976;Serek et al., 2006), thus increasing the vase life of cut flowers.

COBALT DEFICIENCY OCCURS IN PLANTS
Cobalt deficiency does occur in plants. Its deficiency symptoms include leaf chlorosis and necrosis, growth retardation, and reduced crop yield, resembling N-deficiency in plants (Liu, 1998). Co deficient legumes have reduced plant size, smaller and paleyellow leaves, and smaller pods compared with non-deficiency plants. Root growth is also affected by exhibiting an overall reduction of root volume and root lengths. Nodule size and numbers are less abundant than the plants without Co deficiency. Co deficiency causes reduced synthesis of methionine, thus limiting protein synthesis and contributing to the smaller-sized bacteroids (Marschner, 2011). Sweet lupin is particularly sensitive to Co deficiency . In field-grown lupins, Co deficiency reduced bacteroid number per gram of nodule (Chatel et al., 1978) and affected nodule development and function at different levels . Co deficiency in legumes can be assessed by analysis of Co contents in shoots. In general, deficient symptom appears when shoot Co falls in a range from 0.04 (Ozanne et al., 1963) to 0.02 mg/kg based on dry weight . To correct Co deficiency in leguminous crops, application of Co in a range of 1.8 to 145.6 g per hectare was reported (Havlin et al., 2013). Cobalt deficiency also occurs in non-leguminous plants. Co deficiency causes growth retardation in rubber trees and tomato plants (Wilson and Nicholas, 1967). Symptoms of Co deficiency in corn and wheat showed leaf chlorosis and reduced growth (Wilson and Nicholas, 1967). Low leaves may become necrotic, root systems are reduced with decreased number of N 2 fixing bacteria. Grasses with low contents of Co can result in Co deficiency of sheep and cattle. For countries, like South Australia, Sierra Leone, Malta, New Zealand, and Finland, where soils have low Co contents (Sillanpaa and Jansson, 1992), application of Co could improve forage grass growth and enrich tissue Co content. Thus, the feeding of ruminants with healthy grass can reduce Co deficiency (Lee, 1951;Dewey et al., 1958). Due to low Co concentrations in plants, Co deficiency in grazing animals may occur, which can be corrected by mixing Co salts with fertilizers or sand carriers to broadcast it over grazed pastures.

COBALT TOXICITY IN PLANTS
Cobalt at high concentrations causes cytotoxicity and phytotoxicity in plants, which is similar to Cu, Ni, and Zn. Cytotoxicity is the inhibition of mitosis and damage of chromosomes, and disruption of the endoplasmic reticulum of root tip cells (Rauser, 1981;Smith and Carson, 1981;Akeel and Jahan, 2020). Phytotoxicity varies depending on plant species and the concentration of Co in plant organs. Leguminous plants generally exhibit chlorosis or pale-white color on young leaves, and tomatoes show either interveinal chlorosis or diffused chlorosis on young leaves (Akeel and Jahan, 2020).
Cobalt toxicity to plants is uncommon in natural soils, but it happens when plants grow in Co contaminated soils. Soil contamination by Co is mainly from mining and smelting activities, disposal of sewage sludge, and the use of chemical fertilizers (Hamilton, 1994). As discussed above, plants can control Co absorption, transport, and distribution. However, when Co in contaminated soils becomes highly available, Co may gain a competitive advantage over Fe, resulting in more Co being absorbed than Fe through IRT1. With increasing concentrations of Co inside cells, FPN2 may not be able to effectively sequester Co into the vacuole, resulting in more Co to transport from roots to shoots. Li et al. (2020) showed that Co concentrations in shoots of barley, oilseed rape (Brassica napus), and tomato were linearly correlated with the soil solution Co. As a result, excessive Co in shoots may initially cause oxidative stress, resulting in increased anti-oxidative enzyme activities (Tewari et al., 2002). As the stress progresses, Co may compete with Fe or Mg in the chloroplast by decreasing chlorophyll content (Lwalaba et al., 2017), which causes Fe deficiency with newly growing leaves to be yellowish in color. As reported by Sree et al. (2015), Co is able to inhibit the activity of enzymes involved in the biosynthesis of chlorophyll intermediates, like 5-aminolevulinic acid and protoporphyrin, which will reduce net photosynthetic activities. Co also adversely affects the translocation of P, S, Cu, Mn, and Zn from roots to shoots (Chatterjee and Chatterjee, 2000). All these factors, acting together, can result in phytotoxicity and significantly reduce plant growth.
Different plants show different abilities to tolerate Co. Oat (Avena sativa) plants were adversely affected when grown in a soil solution containing 0.14 mg/L Co (Anderson et al., 1973). Rice (Oryza sativa) plants would develop toxic symptoms when grown in soils with Co ranging from 25 and 50 mg/kg (Kitagishi and Yamane, 1981). The contents of Co could be used for predicting the development of toxicity (Akeel and Jahan, 2020). Toxic symptoms occurred in bush beans when tissue Co contents ranged from 43 to 142 mg/kg (Wallace et al., 1977); similarly, 6 mg/kg in barley seedlings (Davis et al., 1978), and 19 to 32 mg/kg in Sudan grass (Gough et al., 1979). In general, tissue Co contents between 30 and 40 mg/kg are considered critical levels for the potential development of Co toxicity (Macnicol and Beckett, 1985). However, due to evolutionary adaptation, Co hyperaccumulators do not develop toxic symptoms at this concentration level. Co contents in leaves of Rinorea cf. bengalensis can be 1,200 mg/kg (Paul et al., 2020), and Glochidion cf. sericeum can accumulate 1,500 mg/kg Co (Van der Ent et al., 2018). Co hyperaccumulators are not the focus of this article. The reader is referred to publications by Brooks (1977), Brooks et al. (1977Brooks et al. ( , 1980, Baker (1981Baker ( , 1987, Lange et al. (2017), and Yamaguchi et al. (2019) for more information.

CONCLUSIONS AND FUTURE PERSPECTIVES
Cobalt in soils ranges from 15 to 25 mg/kg, wherein plant roots can absorb Co from soils and transport absorbed Co from roots to shoots in a controlled manner. Co concentrations in shoots vary with plant species but are comparable to those of essential elements of Cu, Ni, and Zn. Co was well-documented as a constituent of cobalamin, which is required by symbiotic, endophytic, and associated bacteria in the fixation of N 2 . Biological N fixation contributed significantly to the production of economically important crops, including beans, soybeans, rice, corn, barley, wheat, and sugarcane. The current view of plant-microbe association as a phytomicrobiome resulted from millions of years of co-evolution. The coevolution between plants and N 2 fixing bacteria should remind us of the critical role Co plays and its potential essentiality to plant growth and development. Additionally, plants must have Co enzymes or proteins that are specifically responsible for Co metabolism. Due to its similar properties to other transition elements, its biological roles in plants have been largely ignored and simply attributed to its ability to substitute for those elements.
Further research is warranted to (1) identify specific roles of Co plays in diazotrophs, with an emphasis on endophytic and associated bacteria, (2) ascertain Co-containing enzymes and proteins that are implicated in metabolisms of both lower and higher plants, (3) determine the interactions of Co with other transition metals in the regulation of enzymatic activities, (4) recognize Co as an essential micronutrient for plant growth, and (5) develop nutrient management programs by incorporating a group of particular N fixing bacteria with the appropriate amount of Co as plant-specific fertilizers for improving crop production. With the advance in omics, these tasks should be accomplished in the near future. The recognition of Co as an essential micronutrient would enrich our understanding of plant mineral nutrition and enhance crop productivity.

AUTHOR CONTRIBUTIONS
XH, XW, and JC wrote the manuscript. JL prepared figures. All authors contributed to the acquisition and interpretation of available literature and the conception of the work, revised the manuscript, and approved this final version.

FUNDING
This study was supported in part by the Students Innovation and Entrepreneurship Training Program at Zhongkai University of Agriculture and Engineering with grant number 202111347007 and the General Project of the Natural Science Foundation of Fujian Province with grant number 2020J01867.