Regulation and distinct physiological roles of manganese in bacteria

ABSTRACT

Manganese (Mn²⁺) is an essential trace element within organisms spanning the entire tree of life. In this review, we provide an overview of Mn²⁺ transport and the regulation of its homeostasis in bacteria, with a focus on its functions beyond being a cofactor for enzymes. Crucial differences in Mn²⁺ homeostasis exist between bacterial species that can be characterized to have an iron- or manganese-centric metabolism. Highly iron-centric species require minimal Mn²⁺ and mostly use it as a mechanism to cope with oxidative stress. As a consequence, tight regulation of Mn²⁺ uptake is required, while organisms that use both Fe²⁺ and Mn²⁺ need other layers of regulation for maintaining homeostasis. We will focus in detail on manganese-centric bacterial species, in particular lactobacilli, that require little to no Fe²⁺ and use Mn²⁺ for a wider variety of functions. These organisms can accumulate extraordinarily high amounts of Mn²⁺ intracellularly, enabling the nonenzymatic use of Mn²⁺ for decomposition of reactive oxygen species while simultaneously functioning as a mechanism of competitive exclusion. We further discuss how Mn²⁺ accumulation can provide both beneficial and pathogenic bacteria with advantages in thriving in their niches.

This review provides an overview of Mn²⁺ transport and the regulation of its homeostasis in bacteria, with a focus on lactobacilli, and describes functions of Mn²⁺ beyond being a co-factor for enzymes, such as oxidative stress and competitive exclusion.

INTRODUCTION

The divalent cation of manganese (Mn²⁺) has a variety of important roles in bacterial cells (Fig. 1). Depending on the species and conditions, it can function as a cofactor for diverse enzymes involved in central carbon metabolism (Holland and Pritchard 1975; Chander, Setlow and Setlow 1998), nucleotide metabolism (Martin and Imlay 2011), translation (Stetter and Zillig 1974) and signaling (Missiakas and Raina 1997). Moreover, it is known to associate intracellularly with nucleic acids, proteins and metabolites (Chandrangsu, Rensing and Helmann 2017; Waters 2020) and therefore also plays a role beyond catalytic functions. While the unconventional role of helping in radiation resistance of extremophiles had been known for some time (Daly et al. 2004), recently more roles of Mn²⁺ have been discovered, showing the diverse functions of this trace metal in (i) lactobacilli that scavenge it to gain a competitive growth advantage over yeast and mold in dairy (Siedler et al. 2020), and (ii) highly specialized newly isolated bacteria in which it serves as an independent energy source (Yu and Leadbetter 2020) (Fig. 1).

Figure 1.

Functions of Mn²⁺ in bacteria. From left to right, the organisms become more manganese centric and use Mn²⁺ for an increasing number of cellular functions of which they share the ones depicted as overlapping boxes; the two depicted separately at the very right have very specialized functions for Mn²⁺ that are not shared by other species. The species given in the top row are the most-studied examples, but the traits are not necessarily exclusive to these species.

The most widespread and essential role of Mn²⁺ across bacterial species relates to oxidative stress resistance (Fig. 1). Under normal growth conditions, most bacteria do not have a high requirement for Mn²⁺, as most enzymes use iron (Fe²⁺) and not Mn²⁺ as a cofactor (Jakubovics and Jenkinson 2001; Horsburgh et al. 2002; Helmann 2014). Moreover, Fe²⁺ is an essential cofactor for the respiratory chain and hence required for all respiring organisms (Jakubovics and Jenkinson 2001). The requirement for Mn²⁺ and Fe²⁺ differs depending on the species, and bacteria can be divided into three different groups based on their preference and requirement for Fe²⁺ or Mn²⁺. Escherichia coli is the best-studied example of a bacterial species with a highly so-called iron-centric metabolism, as it requires Fe²⁺ for normal growth and Mn²⁺ only under oxidative stress conditions (Anjem, Varghese and Imlay 2009; Paruthiyil et al. 2019). In contrast, some bacteria have been found to require little to no Fe²⁺ for normal growth and instead display a high requirement for Mn²⁺, and hence are called manganese centric. Such bacteria are found in the genus Borrelia, the causative agent of Lyme disease, but are especially abundant in species formerly belonging to the recently reclassified Lactobacillus genus, particularly those of the current Lactiplantibacillus and Lacticaseibacillus genera (Helmann 2014; Chandrangsu, Rensing and Helmann 2017). For simplicity, the term lactobacilli will forwardly in this review refer to species of the former broadly defined Lactobacillus genus. An intermediate group of bacteria including Bacillus subtilis can be considered neither iron- nor manganese centric as they require both Fe²⁺ and Mn²⁺ for normal growth and functioning (Chandrangsu, Rensing and Helmann 2017). In these organisms, most enzymes use Fe²⁺ as a cofactor, but Mn²⁺ is used as well, for example as an essential cofactor for sporulation in B. subtilis (Jakubovics and Jenkinson 2001; Que and Helmann 2002).

While lactobacilli can accumulate and tolerate very high amounts of Mn²⁺, excess accumulation of Mn²⁺ in nonmanganese-centric organisms can easily lead to cytotoxicity - primarily through mismetallation of proteins (Chandrangsu, Rensing and Helmann 2017). Distinct bacteria thus have to tightly regulate Mn²⁺ homeostasis to different extents to ensure sufficient levels for Mn²⁺-dependent stress responses and other functions, yet low enough to prevent cytotoxicity (Waters 2020). They achieve this mostly through the control of importers and exporters, mediated by transcription factors and riboswitches responsive to Mn²⁺ and/or oxidative stress.

Several recent reviews are available on Mn²⁺ transport and homeostasis (Juttukonda and Skaar 2015; Chandrangsu, Rensing and Helmann 2017; Waters 2020). In the current review, we will first provide a detailed description of the regulation of Mn²⁺ transport and homeostasis regulation, highlighting differences between various classes of bacteria based on whether they are iron- or manganese centric. Second, we will discuss the roles of Mn²⁺ in these distinct bacterial classes in light of homeostasis and functional origin, and elaborate on how Mn²⁺ is involved in not only oxidative stress but also in both beneficial and harmful interactions between bacteria and other organisms. Throughout the topics of this review, we particularly focus on the intriguing manganese-centric physiology of lactobacilli.

Mn²⁺ TRANSPORTERS

Two major types of bacterial Mn²⁺ importers are known: the NRAMP (natural resistance-associated macrophage protein) family type transporter MntH, and the ABC (ATP-binding cassette) transporter type MntABC (Fig. 2). The NRAMPs are a family of secondary active transporters involved in metal ion homeostasis (Shi, Zhao and Kong 2014; Jensen and Jensen 2015). The bacterial NRAMP type was first discovered and characterized in Salmonella enterica serovar Typhimurium (S. Typhimurium) and E. coli based on homology with eukaryotic transporters, and named MntH as it is a highly specific Mn²⁺transporter that symports protons (H⁺) (Kehres et al. 2000). For the ABC transporter, the nomenclature varies between organisms, e.g. SitABC, SloABC and MtsCBA (see Table S1, Supporting Information). For reasons of clarity, we refer to this transporter as MntABC throughout this review. In Lactiplantibacillus plantarum, a potential third type of importer, named MntA (not to be confused with MntABC), has been identified (Hao, Reiske and Wilson 1999), but its role in Mn²⁺ import is not fully confirmed (Groot et al. 2005). In this review, it will not be further discussed and, in line with other literature, we are referring to the A-subunit of MntABC whenever MntA/mntA is mentioned.

Figure 2.

Manganese transporters in bacteria. For presence and absence of the different transporters and their regulation, see Table S1 (Supporting Information) and Figs 3–5. Note that the nomenclature for the ABC transporter differs per species. Throughout this review, we will refer to it as MntABC, but other names are SloABC, MtsABC and SitABC (or CBA). Figure based on Bane (2015).

Whereas primary active transporters such as MntABC require the hydrolysis of ATP for function, secondary active transporters like MntH couple the favorable energy of the passage of one molecule to power the transport of another (Bane 2015) (Fig. 2). It is not yet fully understood what the distinct roles of the two importers MntH and MntABC are. In Streptococcus spp., deletion of either mntH or mntABC decreases virulence, albeit to different extents (Eijkelkamp, McDevitt and Kitten 2015; Kajfasz et al. 2020). In several bacteria, including streptococci and B. subtilis, the two transporters were shown to have distinct affinities for Mn²⁺, and to be active at different stages of growth (Que and Helmann 2002; Wang, Tong and Dong 2014). In S. Typhimurium, the affinity for Mn²⁺ was found to be 0.1 µM (1.8 ng/L) for both MntH and MntABC (Kehres et al. 2000, 2002). The cation inhibition profiles were found to be strikingly similar between these two structurally different transporters, with the exception of Zn²⁺, which strongly inhibits MntABC but not MntH (Kehres et al. 2000, 2002). Moreover, MntH in S. Typhimurium was shown to function best under acidic conditions, while MntABC works optimally at slightly alkaline pH and loses functionality in mildly acidic conditions (Kehres et al. 2002). Several lactobacilli contain two or even three MntH transporters as well as the ABC transporter; presumably all these have different affinities and physiological functions (Serata, Yasuda and Sako 2018; Siedler et al. 2020), which will be further discussed later in this review.

For Mn²⁺ uptake, no molecular equivalents to the siderophores such as those used for Fe²⁺ chelation have been described in detail. The peptide antibiotic bacitracin produced by bacilli has been reported to bind Mn²⁺ and increase its import and hence could potentially be considered as a ‘manganiphore’ (Haavik 1979; Archibald 1986), but little follow-up literature is available on this topic. Moreover, bacitracin has been described to bind other metal ions and to have low affinity for Mn²⁺, with no binding at low pH (Ming and Epperson 2002). Therefore, competition for Mn²⁺ appears to occur through differences in transporter affinity and efficiency. This could explain why several manganese-centric Lactobacillus spp. encode for multiple mntH importer genes, as high intracellular Mn²⁺ concentrations are essential in these species (Serata, Yasuda and Sako 2018; Siedler et al. 2020).

While only two types of Mn²⁺ importers are conserved throughout bacteria, up to five different classes of bacterial Mn²⁺ exporters are described in literature (Fig. 2) (Zeinert et al. 2018; Waters 2020). Except for the GTT1-family MneA from Vibrio cholera, we will discuss the regulation of these exporter classes in the next sections. Mn²⁺ export takes place whenever intracellular Mn²⁺ levels are too high and intoxicate cells, leading to a rapid release of Mn²⁺ into the environment after a short period of bacteriostasis (Fisher et al. 1973; Huang et al. 2017). Only recently, the two major Mn²⁺ exporters of B. subtilis, MneP and MneS, belonging to the cation diffusion facilitator family, were identified (Huang et al. 2017). This was later followed by the discovery of two more putative secondary Mn²⁺ exporters YceF and YkoY, belonging to the TerC family (Paruthiyil et al. 2019) (Fig. 2; Table S1, Supporting Information). In E. coli, a putative export pump MntP, belonging to its own family, has been described. Deletion of mntP resulted in elevated intracellular Mn²⁺ levels and increased manganese sensitivity (Waters, Sandoval and Storz 2011). A detailed review on Mn²⁺ transporters has recently been published by others (Waters 2020).

REGULATION OF Mn²⁺ HOMEOSTASIS

Genetic regulatory elements

Regulation of Mn²⁺ homeostasis mostly takes place on the transport level. Even though functions of Mn²⁺ and the level of manganese centricity of a given bacterial species differ, most regulators of Mn²⁺ transporters are shared between the distinct groups (Patzer and Hantke 2001). First and foremost, the metalloprotein transcriptional regulator MntR, for Mn²⁺transport regulator, constitutes the central regulator of Mn²⁺ homeostasis across bacteria, even though MntR homologs display only around 30% similarity between Gram-positive and Gram-negative bacteria (Shi, Zhao and Kong 2014). MntR controls intracellular Mn²⁺ levels by coordinating the transcription of importers, and, depending on the organism, also exporters (Figs 3 –5; Table S1, Supporting Information). MntR forms a homodimer that, through binding of one Mn²⁺ ion per subunit, undergoes a conformational change, which increases the affinity for its DNA binding sites. In B. subtilis, it has been shown that the dynamic response range of MntR spans from 4 to 20 µM (0.22 to 1.1 µg/mL) (Huang et al. 2017). This relatively minor range reflects that tight regulation is required to correctly balance the intracellular concentration of a trace element that is both essential and toxic.

Figure 3.

Manganese transport and regulation in the iron-centric E. coli. The individual transcriptional regulators are depicted at the top, including their mechanisms, e.g. metal binding, as well as in the figure to indicate their binding sites. For MntR, an additional symbol for the DNA binding site (BS-mntR) in light blue has been chosen for clarity. Repression and activation symbols indicate the influence of the individual transcription factors shown above upon Mn²⁺ addition. Genes are depicted in thick arrows and transcription start sites are shown in thin arrows. The base pair (bp) numbers indicate the distance from the gene to the transcription start site. The 2D structure of folded RNA is depicted with and without Mn²⁺ addition. The coding regions of mntR and mntP are shown in green and lilac, respectively, while the upstream region is shown in gray.

Figure 5.

Manganese transport and regulation in the manganese-centric L. plantarum. MntR homodimer is depicted in light blue and a thick light blue line represents the DNA binding site. Genes are depicted in thick arrows and transcription start sites are shown in thin arrows. Repression and activation symbols indicate the influence of the individual transcription factors shown above upon manganese access conditions. The base pair (bp) numbers indicate the distance from the gene to the transcription start site. MntH* represents the Mn²⁺ transporter that has been shown to be responsible for competitive exclusion (Siedler et al. 2020), while MntH# depicts the regulation of the additional one to two mntH genes. No MntR binding site was found in front of mntH# or mntR, but responsiveness upon Mn²⁺ addition or depletion was shown experimentally. Mn²⁺ concentrations from low to very high are depicted below with the indicated repression or activation symbol to depict the regulation. The responsible regulatory element has not been identified yet. Note that mntH1 and mntH2 in the papers of Tong et al. (2017a,b) are two subunits of the ABC transporter based on primer sequence and hence depicted and named here as such.

Besides MntR, three other transcription factors are known to influence Mn²⁺ concentrations in bacterial cells, which will be described in later sections in detail. Fur, for Ferric uptake regulator, like MntR also belongs to the metalloprotein transcriptional regulator family and responds directly to Fe²⁺ ions. Fur represses mostly Fe²⁺ uptake genes under high Fe²⁺ concentrations (Bags and Neilands 1987), and plays a role in Mn²⁺ homeostasis. Additionally, the H₂O₂-sensing transcription factors PerR in Gram-positives and OxyR in Gram-negatives (Patzer and Hantke 2001) induce expression of Mn²⁺ uptake mechanisms either directly or indirectly under oxidative stress (Guedon et al. 2003). PerR of B. subtilis binds Fe²⁺ or Mn²⁺ depending on the respective availability of the two metals, which is a prerequisite for PerR to bind DNA and repress transcription of regulon members. PerR with Fe²⁺ as cofactor is sensitive to metal-catalyzed oxidation by H₂O₂, which causes the release of PerR from its DNA binding boxes enabling expression of downstream genes (Chen, Keramati and Helmann 1995; Herbig and Helmann 2001; Lee and Helmann 2006). OxyR gets activated directly by H₂O₂ through oxidation of a cysteine residue and controls its regulon by direct interaction with RNA polymerase (Imlay 2008; Anjem, Varghese and Imlay 2009; Shi, Zhao and Kong 2014).

Apart from transcription factors, Mn²⁺ can influence the expression of several genes via Mn²⁺-responsive riboswitches. One of these is known as the yybP–ykoY motif, named after the yybP and ykoY genes that it precedes in B. subtilis. yybP–ykoY motifs are widespread in bacteria (Barrick et al. 2004) and shown to bind Mn²⁺ in E. coli, B. subtilis (Dambach et al. 2015) and Lactococcus lactis (Price et al. 2015) in which they mainly regulate Mn²⁺ export. A second Mn²⁺ riboswitch that influences Mn²⁺ import rather than export is found and conserved in several iron-centric Gram-negative enteric bacteria. Both riboswitches will be described in more detail in the respective sections later. The fact that the quantity of Mn²⁺-related proteins are tuned on both transcriptional and translational levels again highlights the essentiality of tightly regulating intracellular Mn²⁺ levels (Dambach et al. 2015). In addition to the regulators introduced earlier, nongenetic factors such as pH and organic acids play a role in regulating Mn²⁺ homeostasis, which will be discussed in a separate section.

The first studies on the function and regulation of Mn²⁺ metabolism focused on E. coli and S. Typhimurium as model organisms. Many subsequent investigations in B. subtilis and to some extent also in streptococci and lactococci have been performed. Remarkably few studies have focused on lactobacilli. This is surprising, as Mn²⁺ plays such a major role in these species that are among the most manganese-centric organisms known, together with Borrelia (Lisher and Giedroc 2013). As described later, oxidative stress responses as well as intracellular Mn²⁺ levels differ in bacteria with an iron- or manganese-centric metabolism, which is reflected in the way Mn²⁺ levels are regulated under such and normal conditions. Due to this, and in combination with the fairly low homology between Gram-positive and Gram-negative MntR, we will further discuss the molecular basis of regulation of Mn²⁺ homeostasis separately for each bacterial group (iron centric, iron- and manganese dependent, and manganese centric).

Iron-centric bacteria (most Gram-negatives)

Escherichia coli solely depends on Mn²⁺ under oxidative stress conditions (Anjem, Varghese and Imlay 2009) and harbors one importer (MntH) and one exporter (MntP). Several layers of regulation are involved to fine-tune the expression of both transporters, which ensures the optimal Mn²⁺ concentration for a given condition (Fig. 3). Under Fe²⁺ and/or Mn²⁺-rich conditions, Mn²⁺ is scarcely imported into the E. coli cell because mntH transcription is repressed both by Fur and MntR (Patzer and Hantke 2001; Anjem, Varghese and Imlay 2009; Shi, Zhao and Kong 2014). Inactivation of Fur disrupts the Fe²⁺-based but not the Mn²⁺-based regulation of mntH, indicating that MntR and Fur function independently (Kehres et al. 2002). Expression of mntH becomes strongly upregulated by the presence of H₂O₂ (Anjem, Varghese and Imlay 2009), but does not respond to paraquat, an inducer of the formation of O₂⁻ radicals (Kehres et al. 2000). Accordingly, the E. coli mntH promoter contains separate binding sites for MntR, Fur and OxyR, but not for the main O₂⁻ sensor SoxS (Fig. 3) (Kehres et al. 2000). It was experimentally shown that the H₂O₂-based OxyR-based regulation is dominant over Fur/Fe²⁺ regulation (Kehres et al. 2000). We speculate that MntR regulation, in its turn, overrules that of OxyR, as MntR binding indicates that a sufficient amount of Mn²⁺ is present in the cells even under oxidative stress conditions.

In E. coli, MntH levels are also affected by a small protein named MntS, the expression of which is MntR controlled (Fig. 3) (Waters, Sandoval and Storz 2011). MntS helps MntR to repress mntH under moderate to high Mn²⁺ conditions, although it currently remains unknown how MntS exerts this effect (Martin et al. 2015).

Another layer of regulation was originally observed in an S. Typhimurium mutant lacking mntR where addition of Mn²⁺ still resulted in slight repression of the generally increased mntH transcription. This led to the discovery of a conserved terminator-like structure in the 5′ Untranslated Region (UTR) of the mntH gene, which functions as a Mn²⁺-responsive riboswitch (Shi, Zhao and Kong 2014). The riboswitch, later on also found in E. coli and other enteric bacteria, constitutes a traditional ‘OFF’ riboswitch with a highly selective affinity for Mn²⁺ (Shi, Zhao and Kong 2014). When Mn²⁺ is bound to the riboswitch, the latter adopts a Rho-independent terminator-like structure that confers premature termination of mntH transcription under high Mn²⁺ conditions, thereby fine-tuning mntH expression (Shi, Zhao and Kong 2014) (Fig. 3; Table S1, Supporting Information).

Escherichia coli MntR and Fur not only function as repressors of Mn²⁺ import but also as activators for expression of Mn²⁺ exporters. Under nontoxic Mn²⁺concentrations, gene expression of the MntP exporter is repressed by H-NS, a general histone-like DNA-binding protein in E. coli and related bacteria (Bertin et al. 2001). However, when intracellular Mn²⁺ levels become excessively high, both MntR and Mn²⁺-mismetallated Fur bind the mntP promoter, thereby alleviating repression by H-NS (Dambach et al. 2015). MntP levels in E. coli are also regulated by the Mn²⁺-responsive yybP–ykoY-type riboswitch (which is different from the one mentioned earlier for mntH) (Waters, Sandoval and Storz 2011; Dambach et al. 2015) (Fig. 3). Moreover, MntS overexpression under high Mn²⁺ conditions resulted in increased Mn²⁺ toxicity, suggesting that MntS could be involved in posttranscriptional regulation of MntP as well (Martin et al. 2015). Since these experiments have been performed under nonphysiological conditions—mntS is normally not expressed under high Mn²⁺ conditions (Waters, Sandoval and Storz 2011)—further experiments are required to elucidate the physiological role of MntS.

Bacteria utilizing both iron and manganese (most Firmicutes)

Many bacteria, including B. subtilis and streptococci among other Firmicutes, require Fe²⁺ along with Mn²⁺ under normal growth conditions. Moreover, in many bacilli, Mn²⁺ is important for sporulation and peptide antibiotic production (Archibald 1986; Jakubovics and Jenkinson 2001). These bacteria have both MntH and MntABC transporters for Mn²⁺ import plus two dedicated exporters MneS and MneP (Fig. 4). B. subtilis is believed to have a third uptake system, since an mntA–mntH double mutant can still grow when only micromolar levels of Mn²⁺ are provided (Que and Helmann 2002).

Figure 4.

Manganese transport and regulation in B. subtilis(A) and Streptococcusoligofermentans(B), which require both Fe²⁺ and Mn²⁺ and hence are neither iron- nor manganese centric. MntR homodimer is depicted in light blue and its regulation, e.g. metal binding, and H₂O₂ availability, are shown at the top. A thick, light blue line represents the DNA binding site. Genes are depicted in thick arrows and transcription start sites are shown in thin arrows. Repression and activation symbols indicate the influence of the individual transcription factors shown above upon Mn²⁺ addition. The base pair (bp) numbers indicate the distance from the gene to the transcription start site. The 2D structure of folded RNA is depicted with and without Mn²⁺ addition.

Unlike E. coli, no Fur-like regulation of Mn²⁺-related genes takes place in B. subtilis (Fig. 4). Rather, Mn²⁺ import and efflux systems are regulated directly by MntR and indirectly by the H₂O₂ response regulator PerR that acts in a complex regulatory circuit involving H₂O₂, Mn²⁺ and Fe²⁺ (Guedon et al. 2003). Bacillus MntR is among the most extensively studied of all MntR variants. Crystal structures of B. halodurans MntR revealed that three instead of the previously assumed two Mn²⁺ molecules are bound in the active dimerized molecule (Lee et al. 2019). B. subtilis mntR deletion mutants constitutively express both mntH and mntABC, which increases sensitivity to Mn²⁺ (as well as to Cd²⁺) and causes Mn²⁺ intoxication (Que and Helmann 2002; Huang et al. 2017). Whereas mntH and mntABC are completely repressed by MntR, mneP and mneS display constitutive expression, which get 5- to 10-fold enhanced by active (Mn²⁺-bound) MntR (Huang et al. 2017). Binding of MntR to the mneP/mneS promoter and subsequent gene activation only occurs at Mn²⁺ levels several folds higher than those required for transcriptional repression of mntABC and mntH (Huang et al. 2017). This is in line with the finding that all three MntR binding boxes in the mneP promoter need to be occupied to induce full expression, which entails that more active (Mn²⁺-bound) MntR molecules are required to activate mneP compared with mntABC and mntH transcription (Huang et al. 2017) (Fig. 4).

Streptococcus oligofermentans also utilizes both MntH and MntABC. Mutant generation showed that MntABC rather than MntH is the dominant Mn²⁺ importer in this species (Wang, Tong and Dong 2014). MntABC facilitates protection against H₂O₂ and O₂⁻, which are naturally abundant in a common habitat of S. oligofermentans, the mammalian oral cavity (Wang, Tong and Dong 2014). MntR, named SloR in Streptococcus, binds to and regulates transcription from the mntABC promoter (Chen et al. 2017). It displays two unique regulatory functions as next to Mn²⁺ it responds to H₂O₂ availability and the redox state of the cell (Chen et al. 2017). MntR in S. oligofermentans was found to contain three cysteine residues: Cys11 is responsible for DNA binding, Cys123 for metal binding and Cys156 is of unknown function and is less conserved than the other two. H₂O₂ was shown to oxidize all three cysteine residues in MntR, leading to deactivation and even degradation of the regulator. This derepresses mntABC transcription and causes an increase in Mn²⁺ uptake, enabling cells to withstand oxidative stress (Chen et al. 2017). Note that this form of regulation is comparable to the one seen in E. coli, but instead of having two separate regulators (OxyR and MntR), both roles are now fulfilled by MntR. Contrary to OxyR, the Gram-positive equivalent PerR does not directly bind the mntH promoter, and we speculate that this is why MntR took up this function in this Gram-positive organism.

Mn²⁺-responsive riboswitches present in Gram-positive bacteria differ mechanistically from those in Gram-negatives. In B. subtilis, the yybP–ykoY riboswitch precedes the genes yybP and ykoY (Dambach et al. 2015). yybP and ykoY encode a predicted membrane protein of unknown function and encode a TerC family membrane protein, respectively. The role of the Mn²⁺ riboswitch-regulated yybP remains elusive, but ykoY was recently proposed to constitute a Mn²⁺ exporter, partially overlapping in function with another newly identified exporter YceF (Paruthiyil et al. 2019) (Fig. 4; Table S1, Supporting Information). YkoY and YceF might form an extra release valve to back up the two main exporters MneP and MneS, or have a role in metalating secreted or membrane proteins (Paruthiyil et al. 2019). Hence, whereas the yybP–ykoY riboswitch in E. coli regulates the main exporter mntP, it seems only of secondary importance in B. subtilis.

In L. lactis and Streptococcus pneumoniae, the yybP–ykoY riboswitch controls similar P-type ATPase Mn²⁺exporters (YoaB in L. lactis, MgtA in S. pneumoniae) (Price et al. 2015; Martin et al. 2019). In S. pneumoniae, the yybP–ykoY riboswitch is located directly upstream of the constitutively expressed Mn²⁺ exporter mntE (Martin et al. 2019). The constitutive expression of mntE might entail that S. pneumoniae maintains relatively low intracellular Mn²⁺ levels. Therefore, these species can tolerate high levels of Mn²⁺ unless mntE is deleted, resulting in toxic Mn²⁺ accumulation but only when grown under high Mn²⁺conditions (Rolerson et al. 2006; Martin et al. 2017). A similar observation was done for Enterococcus feacalis(Lam et al. 2020)). This effect is worse in an mntR-mntE double knockout mutant (Martin et al. 2017). Transcription of Mn²⁺ exporters in S. pneumoniae and B. subtilis only increases under high Mn²⁺ conditions in cells sensitized to Mn²⁺, for example via deletion of mntR. It was therefore suggested that the riboswitch functions as an extra checkpoint to prevent Mn²⁺ toxicity, and, in addition, could be involved in Ca²⁺ efflux (Martin et al. 2019). Moreover, magnesium ions (Mg²⁺) were found to ‘pre-fold’ the conformation and possibly influence the kinetics of the yybP–ykoY riboswitch by priming the affinity of the RNA structure for Mn²⁺ (Sung and Nesbitt 2019).

Altogether, the discovery of different ways of regulating Mn²⁺ transport within Firmicutes hints toward the existence of large interspecies variation in terms of physiological roles that can be attributed to Mn²⁺.

Manganese-centric bacteria (mostly lactobacilli)

Species within the Lactobacillus, Deinococcus and Borrelia genera have been shown to accumulate remarkably high intracellular levels of Mn²⁺, which they use for oxidative stress resistance (Archibald and Fridovich 1981), competitive exclusion (Siedler et al. 2020), radiation resistance (Daly et al. 2004) or circumvention of host Fe²⁺ sequestration (Aguirre et al. 2013) (Fig. 1). Lactiplantibacillus plantarum does not require Fe²⁺ (Archibald 1983) but instead accumulates as much as 30–35 mM intracellular Mn²⁺ (Archibald and Duong 1984). Some L. plantarum strains require at least 2 µM (0.11 mg/L) for growth and show a considerable decrease in biomass and growth rates below this concentration (Hao, Reiske and Wilson 1999; Groot et al. 2005). The importance of Mn²⁺ uptake in lactobacilli is indicated by the presence of a notable number of uptake systems, with up to three mntH genes as well as mntABC (Fig. 5; Table S1, Supporting Information) (Groot et al. 2005; Tong et al. 2017a,b; Serata, Yasuda and Sako 2018; Siedler et al. 2020). Generally, only mntABC and one of the mntH transporters are mainly expressed under Mn²⁺ starvation conditions in L. plantarum (Groot et al. 2005). Please note that different studies use different numberings for mntH orthologs, depending on genome location.

Mn²⁺ uptake and regulation remain to be studied in lactobacilli, despite the importance of this metal ion in the overall physiology of these bacteria. For instance, MntR has not been studied in any Lactobacillus, although binding sites have been identified in silico upstream of mntH1 and mntABC (Novichkov et al. 2013) (Fig. 5). In contrast, PerR and Fur binding sites were not found in the proximity of mntH or mntABC (Novichkov et al. 2013), indicating that the regulation and role of Mn²⁺ in these highly manganese-centric organisms deviate from the previously discussed bacteria. In a L. plantarum mutant lacking mntH2 (mntH1 in the L. paracasei study of Siedler et al. 2020), mntABC expression increased and, vice versa, mntH2 expression increased in an mntABC mutant. These results suggest the existence of cross-regulation of mntABC and mntH2 (Groot et al. 2005), potentially via MntR. Several studies examined gene expression of Mn²⁺ transporters in L. plantarum at different Mn²⁺ concentrations. One group reported on the transcription of five individual mntH genes of one strain, named mntH1–5 (Tong et al. 2017a,b), but based on primer sequence, mntH1 and 2 are instead two subunits of the MntABC transporter. The transcription of all these Mn²⁺ transporter genes was upregulated under Mn²⁺ starvation conditions (0 mg/L) compared with standard growth conditions (16 mg/L or 291 µM), while mntR expression was mildly downregulated (Tong et al. 2017b). When grown with an excess of 960 mg/L (±17.5 mM) Mn²⁺, all five genes were downregulated, while mntR was upregulated; at an even higher concentration of 5760 mg/L (±105 mM) Mn²⁺, mntH4 and mntH5 were upregulated and mntR was even more upregulated (Tong et al. 2017a) (Fig. 5). This could either indicate a dual role of MntH4 and MntH5 as both importer and exporter, depending on the Mn²⁺ concentration, or indicate that some other molecule is transported. The finding that importers are expressed at lower Mn²⁺ concentrations is in line with another study that showed that mntH1 and mntH2 (mntH2 here likely is mntH3 in the papers of Tong et al.) are not expressed at 100–300 µM (5.5–16.5 mg/L), but their transcripts can be detected below 3 µM (0.165 mg/L) (mntH1 and mntH2) and 10 µM (0.55 mg/L) (mntH2) (Groot et al. 2005). No expression of mntH3 was found under the tested conditions, and based on the genetic context of this gene it was suggested to play a role in Fe²⁺ instead of Mn²⁺ transport (Groot et al. 2005).

Whereas the K_m in L. plantarum for Mn²⁺ uptake (0.2 µM; 0.011 mg/L) is similar to that of other organisms such as E. coli and B. subtilis (∼0.1 µM; 0.0055 mg/L), the V_max (rate of uptake) is much higher, namely 23.8 nmol mg⁻¹ of protein min⁻¹ (Archibald and Duong 1984). In this light, it would be highly relevant to compare the K_m value of Bacillus MntR (4–20 µM; 0.22–1.1 mg/L) with that of Lactobacillus MntR variants, which presumably have adopted a different dynamic range due to higher Mn²⁺ requirements. Already in early studies on Mn²⁺ homeostasis in L. plantarum, scientists speculated about the mechanism behind the extremely high V_max of Mn²⁺ uptake. It was suggested that this was either mediated by a specific transporter with extraordinarily high transport activity, or by a high quantity of transporters (Archibald 1986). Recently, it was shown that the main mntH gene involved in Mn²⁺ uptake in both L. paracasei and L. rhamnosus displays extremely high transcription — similar to the level of highly expressed glycolytic genes (Siedler et al. 2020). This provides an indication that the large quantity of MntH transporters present within the cell membrane is responsible for a high V_max, although it cannot be ruled out that the specified Lactobacillus MntH also possesses improved Mn²⁺ uptake activity. However, it does imply that organisms encoding an MntH could have the potential to import high levels of Mn²⁺ by increasing transporter copy numbers.

As MntH-mediated Mn²⁺ transport is proton coupled (Fig. 2), it comes at an indirect energetic cost in relation to proton homeostasis. Leakage of only a fraction of the intracellular Mn²⁺ could therefore be detrimental to growth of organisms with high intracellular Mn²⁺ levels. Indeed, this seems to be avoided as no leakage of intracellular Mn²⁺ was observed when L. plantarum was transferred from a Mn²⁺-rich to a Mn²⁺-depleted medium (Archibald 1986). Interestingly, no reports of Mn²⁺ exporters in lactobacilli are available, so it is currently unknown whether and when they perform active export of Mn²⁺. Further research is needed to determine how both import and export are regulated.

Taken together, while some facilitators of Mn²⁺ homeostasis have been studied in lactobacilli, the underlying regulation remains largely unknown. The elucidation of such mechanisms that drive high intracellular Mn²⁺ levels will be crucial in generating an understanding of the rather particular and intriguing Mn²⁺ homeostasis found in lactobacilli and other manganese-centric bacteria. Most studies in lactobacilli have looked at expression under fairly or extremely high Mn²⁺ levels. While this is certainly relevant in many environments, it would be of equal interest to investigate the low end of Mn²⁺ concentrations as they exist in milk and fermented dairy (Siedler et al. 2020) or the human gut (Juttukonda and Skaar 2015). Also, the importance and level of Mn²⁺ uptake by lactobacilli under oxidative stress conditions remain to be addressed, as all studies so far did not test mutants under oxidative stress conditions.

Nongenetic factors influencing Mn²⁺ transport and intracellular storage

Apart from genetic regulation, other factors are known to affect Mn²⁺ uptake. As stated earlier, the Mn²⁺ importers MntH and MntABC are both influenced by pH. In S. Typhimurium, Mn²⁺ uptake by MntABC was highest at pH 8 and was reduced to 50% of its maximum capacity at around pH 6.7 and to nearly 23% at pH 6.0 (Kehres et al. 2002). The affinity of MntH for Mn²⁺ was independent of pH (with K_m not impacted), but the maximum uptake rate increased 3-fold as the pH decreased from 8.2 to 5.5 (Kehres et al. 2000). It was suggested that the effect of pH is physicochemical and not regulatory. However, more recently, it was found that pH-dependent regulation is exerted through the yybP–ykoY Mn²⁺-responsive riboswitch in E. coli and B. subtilis. The level of yybP–ykoY transcripts increased under alkaline conditions, suggesting that the riboswitch orchestrates regulation by both responding to pH and Mn²⁺ (Dambach et al. 2015).

In both L. plantarum and E. coli, Mn²⁺ uptake is dependent on energy and the proton gradient, as it is inhibited by carbonyl cyanide m-chlorophenyl hydrazine and dinitrophenol, two compounds that uncouple oxidative phosphorylation and dissipate the proton gradient (Silver, Johnseine and King 1970; Archibald and Duong 1984). Since L. plantarum lacks a respiratory chain, it generates a proton gradient via ATP hydrolysis by a proton-pumping ATPase. Treatment of cells with an inhibitor of proton-pumping ATPases (N,N'-dicyclohexylcarbodiimide) abolished Mn²⁺ uptake (Archibald and Duong 1984). In contrast, treatment with sodium arsenate, an inhibitor of substrate-level phosphorylation, blocked import only partially (Archibald and Duong 1984).

In E. coli, uptake improved with increasing temperatures (Silver, Johnseine and King 1970). In L. plantarum, Mn²⁺ uptake increases with temperatures up to 37°C, but already at 4°C significant uptake was seen (Archibald and Duong 1984). The effect of temperature has not been further studied or mentioned.

In Streptococcus, Co²⁺ and Fe²⁺ were found to inhibit Mn²⁺ uptake but in concentrations 100-fold larger than Mn²⁺ (Kehres et al. 2000). The inhibiting effect of Fe²⁺ on both MntABC and MntH was found to be pH dependent, potentially because the transporters obtain an increased affinity for Fe²⁺ at low pH (Kehres et al. 2000). While the inhibition profiles were strikingly similar for all other ions, Zn²⁺ strongly inhibited MntABC but not MntH (Kehres et al. 2002). The intricate link between Mn²⁺ homeostasis and that of other transition metals is emphasized in a recent study, in which it was unveiled that dysregulation of Mg²⁺ export suppressed Mn²⁺ and Co²⁺ toxicity in a B. subtilis mntR–mntH double mutant (Pi, Wendel and Helmann 2020).

As much as 30–35 mM (1.65-1.93 g/L) Mn²⁺ can accumulate intracellularly in L. plantarum. In these cells and other hyper-scavengers, Mn²⁺ is proposed to be associated with granules of polyphosphate (Archibald and Fridovich 1982). Up to 80% was estimated to be stored in an inactive form in this way, possibly representing a reservoir of Mn²⁺ for times of scarcity while preventing toxicity, whereas the remaining free Mn²⁺ is readily available for metabolic activity (Archibald and Duong 1984). When cells are transferred from Mn²⁺-rich to Mn²⁺-limiting conditions, growth rate becomes affected only after several generations, further supporting the idea of Mn²⁺ storage. Rather than changing the total cellular Mn²⁺ content, L. plantarum cells are therefore proposed to ensure homeostasis of the freely available Mn²⁺ portion rather than the total cellular content (Archibald and Duong 1984). In line with this hypothesis, the intracellular Mn²⁺ concentration no longer responds to oxidative stress and does not depend on the extracellular concentration when exceeding 200 µM (11 mg/L). In contrast, phosphate limitation leads to a decrease in intracellular Mn²⁺ even if the extracellular concentrations are >200 µM. As an effect of the tight relationship between Mn²⁺ and polyphosphate granules, phosphate-limited cells are unable to accumulate large amounts of Mn²⁺. Vice versa, polyphosphate granules were not observed in medium with low Mn²⁺ levels (Archibald and Duong 1984). Nevertheless, Mn²⁺ uptake appeared to continue at neutral pH (6.7) regardless of phosphate as long as 20 mM organic acids were present, which seem to make Mn²⁺ more available to the cells (Archibald and Duong 1984). At pH 5.5, cells could only obtain Mn²⁺ efficiently in the presence of citrate or other tricarboxylic acids, while the acids themselves were not taken up (Archibald and Duong 1984).

Across several bacterial species, the ppk gene encoding polyphosphate kinase, which catalyzes polyphosphate synthesis from ATP, seems to play a major role in polyphosphate accumulation capacity and oxidative stress resistance (Gray and Jakob 2015). Surprisingly though, only few studies are available on polyphosphate accumulation in general and in lactobacilli specifically, and its exact roles and mechanisms in stress resistance are only partially understood (Gray and Jakob 2015). In some bacteria such as E. coli and L. casei, polyphosphate has also been shown to play a role in coping with stress induced by salt, heat or low pH, but a similar function could not be identified for L. plantarum (Gray and Jakob 2015; Alcántara et al. 2018). Besides its role in intracellular stress resistance, polyphosphate is an important immunomodulating signal molecule in bacteria–host interactions (a process in which reactive oxygen species [ROS] also play a big role) and has been suggested to play a role in the probiotic effects of lactobacilli (Segawa et al. 2011; Gray and Jakob 2015).

In general, effective intracellular storage of Mn²⁺ is of prime importance in organisms with high intracellular Mn²⁺ levels. Its association with various low molecular weight (LMW) complexes beyond polyphosphate granules has been reported in several organisms (Archibald and Fridovich 1982; Barnese et al. 2012). These include LMW complexes such as inorganic phosphate species, nitrogenous compounds including short peptides, and nucleotides, in addition to carbonate and organic acids (Archibald and Fridovich 1982; Daly et al. 2010; Lisher and Giedroc 2013). More research is needed to identify mechanistic details and species-specific variation in these mechanisms, and it also remains to be identified whether dedicated ‘chaperones’ exist, or unspecific binding is the main mechanism of ‘sequestration’.

ROLES, EFFECTS AND FUNCTIONAL ORIGIN OF Mn²⁺ TRANSPORT AND HOMEOSTASIS IN BACTERIA

Understanding how Mn²⁺ transport and regulation is achieved in iron- or manganese-centric bacteria reveals clues about the primary role and evolution of such systems in these bacterial groups, on which we will elaborate in this section. Mn²⁺ is used for diverse functions in the cells, with increasing complexity dependent on the requirements for this trace metal (Fig. 1).

Mn²⁺ in oxidative stress resistance in iron-centric vs manganese-centric bacteria

In aerobic environments, cells are naturally exposed to small amounts of ROS in the form of superoxide (O₂⁻) and hydrogen peroxide (H₂O₂). These are formed from molecular oxygen extracellularly by environmental redox reactions and/or intracellularly via enzyme autoxidation (Horsburgh et al. 2002; Seaver and Imlay 2004; Imlay 2008). Most aerobic bacteria keep ROS levels within viable limits through the action of superoxide dismutases (SODs), reductases, peroxidases and catalases. SODs scavenge O₂⁻ and convert it into O₂ and H₂O₂, a redox reaction known as superoxide dismutation. In most organisms, H₂O₂ is then degraded by catalases and peroxidases that convert it into water and oxygen. Some anaerobic bacteria depend on reductases instead of SODs for O₂⁻ removal (Horsburgh et al. 2002; Imlay 2008; Anjem, Varghese and Imlay 2009). The continuous removal of ROS from the intracellular environment is crucial because Fe²⁺-loaded enzymes, despite their prevalence in bacteria, are inherently susceptible to Fenton reactions, i.e. oxidation reactions of H₂O₂ with Fe²⁺ resulting in highly reactive hydroxyl radicals that cause oxidation of nearby molecules such as proteins and DNA (Imlay 2008; Anjem, Varghese and Imlay 2009).

ROS levels can rise above those in normal aerobic conditions through a variety of causes. For instance, Lactic Acid Bacteria (LAB) and mammalian macrophages can secrete large amounts of H₂O₂ to fight off competitors or pathogenic bacteria (Imlay 2008). Also, plants and some bacteria can produce redox-cycling compounds that generate both H₂O₂ and O₂⁻ through the oxidation of redox enzymes and concomitant transfer of electrons to O₂ (Imlay 2008). Furthermore, when exposed to light, H₂O₂ can build up in solutions through photochemistry, while an elevation of H₂O₂ might occur whenever anaerobic sediments containing reduced metal and sulfur compounds get in contact with water in which oxygen molecules are dissolved (Imlay 2008). Under conditions with elevated oxidative stress, the regular ROS scavenging enzymes are insufficient to prevent oxidative damage, and additional response systems get activated. Because Mn²⁺, unlike Fe²⁺, does not result in Fenton reactions (Jakubovics and Jenkinson 2001; Anjem, Varghese and Imlay 2009), it lies at the basis of both enzymatic and nonenzymatic oxidative stress responses and thus plays a unique and ubiquitous role in coping with high ROS levels.

A bacterium that employs Mn²⁺ in an enzymatic manner during oxidative stress is for example E. coli, which possesses two intracellular SODs. One uses Mn²⁺ as a cofactor (Mn-SOD), while the other utilizes Fe²⁺ (Fe-SOD). Under oxidative stress conditions, the latter switches from a Fe²⁺ to a Mn²⁺ isozyme to reduce Fenton reaction-induced self-damage (Imlay 2008). A nonenzymatic use of Mn²⁺ for coping with oxidative stress can be found in many LAB, which are often devoid of either SODs or catalases (Chen et al. 2017). For example, the SOD-negative L. casei only survives oxidative stress induced by O₂⁻ generated by paraquat in the presence of Mn²⁺ (Serata, Yasuda and Sako 2018). The exact mechanism by which Mn²⁺ prevents oxidative damage in a non-SOD way is not yet fully understood. The main hypothesis is that Mn²⁺ bound in LMW complexes functions as ROS scavenger, thereby replacing the function of SOD. For example, Mn²⁺ in complex with physiologically relevant anions such as phosphates, carbonates or organic acids is capable of superoxide dismutation (Archibald and Fridovich 1982; Barnese et al. 2012). Depending on the anion associated with Mn²⁺, either catalytic or noncatalytic dismutation can occur. For instance, while Mn-pyrophosphate displays noncatalytic behavior, Mn-orthophosphate was found to act in a catalytic manner (Barnese et al. 2012). Moreover, it has been demonstrated in vitro that Mn-LMW complexes are capable of decomposing H₂O₂ (Stadtman, Berlett and Chock 1990; Liochev and Fridovich 2004), the product of superoxide dismutation, although it is currently unknown how this translates to in vivo activity. Another mechanism has been proposed for E. coli, in which under oxidative stress and high Mn²⁺ uptake conditions, Mn²⁺ outcompetes Fe²⁺ in catalytic sites and thereby prevents site-specific Fenton reactions, but this model has not yet been tested (Imlay 2008). Since lactobacilli accumulate little to no Fe²⁺, it seems likely that such a mechanism is more specific for iron-centric organisms like E. coli, and that the mechanisms of nonenzymatic Mn²⁺-based oxidative stress resistance differ between manganese- and iron-centric organisms.

Since Mn²⁺ does not lead to Fenton reactions, the exchange of Fe²⁺ by Mn²⁺-mediated enzymatic reactions in itself contributes to oxidative stress resistance. To achieve this, the uptake of Fe²⁺ from the external milieu should be minimized and instead large amounts of Mn²⁺ have to be taken up to outcompete any trace amounts of Fe²⁺, as Fe²⁺ typically binds enzymes at such high affinities that it easily replaces other metals, causing mismetallation of Mn²⁺-dependent enzymes (Gerwien et al. 2018). Indeed, lactobacilli are known for their Fe²⁺-independent lifestyle, displaying an absence of a requirement for Fe²⁺ (Bruyneel, vande Woestyne and Verstraete 1989; Weinberg 1997), with Fe²⁺ seemingly completely excluded from some species; down to 2 Fe²⁺ atoms per cell were detected in L. plantarum, compared with 1 × 10⁶ for E. coli (Archibald 1983). High Mn²⁺ and low Fe²⁺ levels are thus linked, and the latter is likely either an effect or a prerequisite of high intracellular Mn²⁺ levels.

When Fe²⁺ is almost completely expelled from cells, aerotolerance does not depend on the presence of ROS-scavenging enzymes (enzymatic response), which could account for the lack of SODs in many lactobacilli. The depletion of Fe²⁺ further enables the high H₂O₂ production observed for some lactobacilli, as Fe²⁺-mediated oxygen radical formation is avoided. However, without Fe²⁺, several cellular functions cannot be carried out. Among those is aerobic respiration that involves Fe²⁺-dependent cytochromes (Jakubovics and Jenkinson 2001). In line with this, most LAB do not contain heme compounds and use fermentation rather than aerobic respiration for energy generation. In contrast, bacteria that rely largely on respiration for energy storage, such as E. coli and B. subtilis, cannot avoid the use of Fe²⁺, and thus have to utilize ROS scavenging and reducing enzymes such as SODs, reductases and peroxidases. Since respiration yields more energy than fermentation, it seems likely that there are additional advantages of the Mn²⁺-centric metabolism of lactobacilli, which will be discussed next.

Influence of Mn²⁺ on cell growth and development

Besides oxidative stress resistance, Mn²⁺ plays an important role in certain aspects of bacterial growth and development (Fig. 1). While its presence is a prerequisite for cell growth in manganese-centric organisms, its function is multifaceted in the group that is neither iron- nor manganese centric where it additionally influences virulence, sporulation and biofilm formation. Interestingly, this group can be further subdivided into (i) species with a minimal requirement for Mn²⁺, for which trace amounts of Mn²⁺ present as ‘contamination’ in Chemically Defined Medium (CDM) are sufficient and (ii) species with a greater demand for this trace metal that require specific addition of Mn²⁺ to CDM, while also still requiring Fe²⁺ (though less than the iron-centric organisms).

The ovoid-shaped lactic acid bacteria such as enterococci, streptococci and lactococci fall into the first subgroup, and Mn²⁺ has a limited effect on their physiology under normal growth conditions: transcriptomic studies in S. pneumoniae grown in Mn²⁺-limiting conditions revealed changes in only five genes, encoding two subunits of MntABC, a cation efflux system protein and virulence-associated proteins PcpA and PrtA (Ogunniyi et al. 2010). Recent studies identified more genes that were differentially regulated in the absence of Mn²⁺ in S. mutansand S. sanguinis including upregulation of Mn²⁺ uptake systems and changes in genes important for Mn²⁺-specific cellular functions by Mn²⁺-dependent enzymes (Kajfasz et al. 2020; Puccio et al. 2020). To date, eight enzymes are known that employ Mn²⁺ as a cofactor in streptococci (Kuipers et al. 2016; Martin et al. 2017), including Mn-SOD discussed earlier, and a Mn²⁺-dependent ribonucleotide reductase RNR NrdEF, required for synthesis of deoxynucleotides from ribonucleotides under aerobic conditions (Makhlynets et al. 2014; Rhodes et al. 2014). Interestingly, the Mn²⁺-dependent group of enzymes comprises a set of Mn²⁺-requiring phosphatases that are crucial for normal cell physiology and development, as well as for virulence. These phosphatases are strongly affected by perturbations in Mn²⁺ homeostasis, which can be achieved by addition of the highly competitive Zn²⁺ transition metal or the metal chelator EDTA, employing mutants devoid of Mn²⁺ import or export, or changing the oxygen availability (Geno et al. 2014; Ong, Walker and McEwan 2015; Martin et al. 2017; Puccio et al. 2020). For instance, a balanced phosphatase activity of PhpP, mediated by Mn²⁺ availability, is crucial for maintaining the phosphorylation status of key cell division proteins in S. pneumoniae (Martin et al. 2017). In mntR/mntE mutant strains, too high Mn²⁺ levels result in hyperactive PhpP phosphatase and aberrant cell division (Martin et al. 2017). On the other hand, a shortage of Mn²⁺ was found to render Mn²⁺-dependent phosphatases inactive. This is likely because the Mn:Zn ratio was altered, leading to the displacement of Mn²⁺ with Zn²⁺ in Mn²⁺-dependent phosphatases, rendering the latter inactive (Martin et al. 2017; Puccio et al. 2020). Zn²⁺ is also known to bind the Mn²⁺ solute-binding protein MntC of MntABC with high affinity, essentially blocking uptake and resulting in depletion of Mn²⁺ and altered Mn:Zn ratios (McDevitt et al. 2011; Eijkelkamp et al. 2014).

Two Mn²⁺-dependent phosphatases have a major role in streptococcal capsular polysaccharide (CPS) biosynthesis, which is crucial for virulence and biofilm formation (Hardy et al. 2001) — namely CpsB and phosphoglucomutase Pgm. Pgm is inhibited by Zn²⁺, resulting in decreased hyaluronic acid capsule production in S. pyogenes (Ray 1967; Ong, Walker and McEwan 2015). CpsB activity increases under increased oxygen availability, i.e. when metabolism switches to a more manganese-centric state (Geno et al. 2014). In S. sanguinis, expression of pgm and cpsB, as well as deoB, encoding a Mn²⁺-dependent phosphopentomutase, was found to be downregulated upon Mn²⁺ depletion, while that of papP, coding for a nucleotide phosphatase that uses Mn²⁺ as a cofactor and is involved in membrane lipid homeostasis, was upregulated (Puccio et al. 2020). CPS production, mediated by CpsB and Pgm, constitutes one of the core virulence factors of streptococci (Hardy et al. 2001; Geno et al. 2014; Ong, Walker and McEwan 2015; Wu et al. 2016), and deletion of papP is known to attenuate virulence (Cron et al. 2011; Bai et al. 2013). Interestingly, S. aureus was recently found to express a second metal-independent Pgm under Mn²⁺ limitation, required for growth under such conditions but also necessary for host infection (Radin et al. 2019). A clear link exists between Mn²⁺ homeostasis and virulence, which will be further discussed in the next section. Interestingly, also the inorganic pyrophosphatase PpaC relies on Mn²⁺ for optimal function, which has been proposed to have a putative role in the formation of LMW Mn²⁺-phosphate complexes involved in oxidative stress tolerance (Martin et al. 2017). Hence, similar to what is observed in the manganese-centric lactobacilli, phosphate and phosphatases are closely linked to Mn²⁺ homeostasis.

Species such as bacilli and staphylococci belong to the second subgroup. Generally, this group of bacteria is more sensitive to Mn²⁺ restriction, despite possessing high-affinity Mn²⁺ uptake systems. Additionally, Mn²⁺ efflux is under control of MntR in B. subtilis, resulting in a highly Mn²⁺-sensitive mntR deletion mutant strain, whereas this is not the case for mntR deletion mutants in streptococci, where the exporters are constitutively expressed (Martin et al. 2019) (Fig. 4). Moreover, in contrast to the relatively few changes that occur in streptococci, microarray data showed that, in B. subtilis, 10% of all of genes were differentially expressed when Mn²⁺ was omitted from the medium (Mhatre et al. 2016).

The dependency of B. subtilis and S. aureus on the Mn²⁺-dependent isoform of Pgm likely explains why these species require more Mn²⁺, as this renders Mn²⁺ essential for growth on glucose (Oh and Freese 1976; Radin et al. 2019). Recent investigations in our laboratory have shown that Listeria spp. have a similar requirement for Mn²⁺ and contain a highly homologous Pgm (van Gijtenbeek et al. manuscript in preparation). The role of the Mn²⁺-dependent Pgm reaches further than its essential roles in glycolysis and CPS synthesis/virulence. Besides being required for vegetative growth of B. subtilis, Mn²⁺ is essential during the early phases of sporulation. This is likely due to Pgm activity, which is also required for sporulation (Oh and Freese 1976; Vasantha and Freese 1979). In addition, it has been suggested that the phosphorylation state of Spo0A, a major regulator during the initiation phase of sporulation, is possibly dependent on Mn²⁺ as well (Mhatre et al. 2016). This seems plausible when looking at the strong link between phosphatases and Mn²⁺ as described in the previous paragraph. Also Spo0A-regulated antimicrobial peptide production was reduced in B. subtilis grown in medium not supplemented with Mn²⁺ (Mhatre et al. 2016). While Mn²⁺ is essential for growth and sporulation in B. subtilis, too high concentrations (>15 µM) resulted in reduced growth rate and sporulation in Bacillus (Vasantha and Freese 1979; Cheung, Vitkovic and Brown 1982). This might be due to displacement of Fe²⁺ by Mn²⁺ when levels of the latter are too high, evidenced by the upregulation of Fur (Guedon et al. 2003) and emphasizing the importance of a strict regulation of Fe²⁺ and Mn²⁺ balance in these organisms.

More recently, it was shown that Mn²⁺ is also required for biofilm formation by B. subtilis (Shemesh and Chaia 2013; Mhatre et al. 2016), as well as L. plantarum (Nozaka et al. 2014). Low Mn²⁺ concentrations result in decreased biofilm formation in both species. This seems another way in which Mn²⁺ is involved in stress responses, as biofilm formation and sporulation occur under stress conditions (i.e. stationary phase, which involves nutrient depletion, oxygen stress and toxin production) (Mhatre et al. 2016). Interestingly, Mn²⁺ seems to have opposing effects on biofilm formation by iron-centric or manganese- and manganese/iron-centric bacteria. In contrast to the more or fully manganese-centric organisms, biofilm formation in E. coli is negatively influenced by the presence of Mn²⁺ (Guo and Lu 2020). Moreover, in uropathogenic E. coli low concentrations of Fe²⁺ and Mn²⁺ lead to the formation of biofilm aggregates, which disperse upon an increase in these metal ions, leading to recurrence of the infection (Rowe, Withers and Swift 2010). This is likely related to the success or failure of mammalian hosts to deplete essential ions as a defense mechanism and subsequent virulence development (Rowe, Withers and Swift 2010). Such competition and virulence mechanisms based on Mn²⁺ (and Fe²⁺) are discussed in the next section.

Nutritional immunity and competitive exclusion: Mn²⁺ in pathogenic and beneficial bacteria as two sides of the same coin

Mn²⁺ and its transport has long been recognized to be involved in streptococcal and enterococcal development toward virulence (Loo et al. 2003; McAllister et al. 2004; Rolerson et al.2006; Abrantes et al. 2013; Kajfasz et al. 2020). Although the existence of a clear relationship between Mn²⁺ transport and virulence development by pathogenic bacteria is acknowledged, the role of Mn²⁺ in beneficial interspecies interactions has remained more elusive (Waters 2020). Recently, it is becoming increasingly clear that Mn²⁺ also plays a crucial role in beneficial interactions that occur within bacterial communities, at the bacteria–eukaryotic host interface or in bacteria–plant symbioses. In this section, we briefly explore both harmful and beneficial relationships that exist, sometimes simultaneously, in various niches between two or more species that have been reported to be affected by Mn²⁺ availability. Within this scope, we will also discuss the beneficial role of the hyper-scavenging of Mn²⁺ exerted by highly manganese-centric lactobacilli.

The oral cavity forms an example of a niche where two neighboring organisms use the same Mn²⁺uptake systems to survive oxidative stress and drive either pathogenesis or prevention thereof. Here, the beneficial commensal S. oligofermentans produces large amounts of H₂O₂ to fight off other microorganisms that compete for nutrients, including pathogens. Since S. oligofermentans cells do not have a catalase, they rely on high Mn²⁺ uptake to survive oxidative stress imposed by their own H₂O₂ production (Wang, Tong and Dong 2014; Chen et al. 2017). Mutant cells devoid of a high-affinity Mn²⁺ transporter showed a 5.7-fold decrease in survival rate, while overexpression of the transporter resulted in a 12-fold increase when challenged with H₂O₂ (Wang, Tong and Dong 2014). Hence, Mn²⁺ import by S. oligofermentans is self-beneficiary but at the same time delivers an advantage to the host. In order to overcome the stress induced by H₂O₂ production by S. oligofermentans, the cariogenic pathogen S. mutans, in its turn, utilizes and requires the same Mn²⁺ importers for its survival, growth and virulence (Paik et al. 2003; Rolerson et al. 2006; Kajfasz et al. 2020). In the same niche, competition for Mn²⁺ to fend off pathogenic microorganisms is also employed by the host itself. The general concept of restricting access to Mn²⁺ and other transition metals such as Zn²⁺ and Fe²⁺ is known as nutritional immunity (Kehl-Fie and Skaar 2010). The specific depletion of Mn²⁺ is realized, among other mechanisms, by the Mn²⁺-chelating protein calprotectin expressed by healthy epithelial cells in the oral cavity (Gómez et al. 1995; Corbin et al. 2008).

In other locations of the mammalian body, calprotectin is believed to be secreted by neutrophils that undergo apoptosis in response to microbial infections, for instance in the inflamed gut and at wound sites (Urban et al. 2009; Diaz-Ochoa et al. 2014). It has been shown to be a critical host element in the defense against a broad range of bacteria (Juttukonda and Skaar 2015). As discussed earlier, nutritional immunity via Mn²⁺ scavenging by mammalian hosts is often paired with production of ROS. It is evident that Mn²⁺ restriction in the presence of ROS leads to a highly toxic environment for bacteria that depend on increasing intracellular Mn²⁺ for oxidative stress survival.

Since manganese-centric lactobacilli are strong competitors for Mn²⁺ and sometimes also produce antibacterial peptides and/or H₂O₂, it is tempting to speculate that these mechanisms, alone or in combination, enable lactobacilli to thrive as beneficial microbiota in the human host. On top of that, the LAB-based beneficial microbiota might help to capture Mn²⁺ in the human gut otherwise available for more harmful bacteria (Knaus et al. 2017). Indeed, the high Mn²⁺ uptake by L. plantarum has been proposed to be an underlying mechanism for its probiotic properties (Tong et al. 2017a). Mn²⁺ restriction and ROS production by the mammalian host can be counteracted by the overexpression of Mn²⁺ importers triggered under Mn²⁺-limiting conditions. Like S. mutans, many pathogens, including various virulent streptococci (Paik et al. 2003; Johnston et al. 2006; Wichgers Schreur et al. 2011) as well as E. coli (Sabri et al. 2008), S. Typhimurium (Diaz-Ochoa et al. 2016), E. faecalis (Colomer-Winter et al. 2018) and S. aureus (Kehl-Fie et al. 2013), rely on MntABC and/or MntH for survival in the host. Another path to cope with detrimental low Mn²⁺ concentrations has recently been postulated for S. aureus and S. pneumoniae that have at least one strictly Mn²⁺-dependent enzyme essential for glycolysis. Under Mn²⁺ limitation, both species shift from glucose utilization to amino acids for energy generation, presumably to reduce the level of Mn²⁺ they require for growth (Ogunniyi et al. 2010; Radin et al. 2016).

Similar to Mn²⁺ scavenging, mammals have developed a variety of mechanisms to scavenge Fe²⁺ (Johnson and Wessling-Resnick 2012). Borrelia burgdorferi, the causative agent of Lyme disease, circumvents this by being fully manganese centric with no requirement for Fe²⁺, with highly effective Mn²⁺ uptake systems (Posey and Gherardini 2000). Also, Helicobacter pylori uses Mn²⁺ instead of Fe²⁺ as a cofactor for several key enzymes to survive in low Fe²⁺ environments (Haley and Gaddy 2015).

Bacterial Mn²⁺ uptake is not only relevant in mammals but has recently also been shown to exert a key function in certain plant symbioses. In young nodules of galegoid-clade legumes, mntH and mntABC were found to be highly expressed in the plant symbiont Rhizobium leguminosarum (Hood et al. 2017). Interestingly, both genes appeared essential for nitrogen fixation, and their expression was only upregulated in galegoid-clade and not in phaseloid-clade legumes. It was suggested that this host effect is related to a difference in Mn²⁺ availability and/or ROS concentration in the different plant species (Hood et al. 2017).

Another niche where Mn²⁺ scavenging provides a growth advantage for beneficial organisms is found in dairy products, where lactobacilli inhibit growth of yeast and mold during milk-to-yogurt fermentations through Mn²⁺ scavenging—a process known as competitive exclusion (Siedler et al. 2020). The analogy between the mechanisms used by lactobacilli for competitive exclusion and mammalian hosts for nutritional immunity is intriguing, as both use homologs of the NRAMP type Mn²⁺ transporter MntH for their defensive Mn²⁺ scavenging.

Functional origin of Mn²⁺ scavenging in manganese-centric lactobacilli

The finding that Mn²⁺ accumulation in lactobacilli has functions beyond basal Mn²⁺ homeostasis and oxidative stress resistance poses an intriguing question with regard to the order at which these functions evolved. One clue might be taken from the analogy between the use of MntH/NRAMP transporters by both lactobacilli and mammals for strategies to outcompete other organisms. It is tempting to speculate that MntH1 in lactobacilli evolved to have a major function in this, as deletion of the L. paracasei mntH1 only abolished inhibition of yeast and mold related to Mn²⁺ starvation but did not affect normal milk acidification, a proxy for growth of L. paracasei (Siedler et al. 2020). Contrary to L. paracasei, deletion of mntH in B. subtilis did result in a clear growth defect, as cells did not reach maximum cell densities when provided with <3 µM (0.165 mg/L) Mn²⁺ (Que and Helmann 2002). Since L. paracasei has three mntH genes, the two other MntH variants, of which at least one is likely to transport Mn²⁺ as evidenced by studies performed in the closely related L. casei, (Serata, Yasuda and Sako 2018), could take over a possible growth-supporting role of MntH1. However, neither compensated for the uptake level required to substantially deplete the environment from Mn²⁺, indicating that MntH1 alone is required and sufficient for competitive exclusion (Siedler et al. 2020). This also further indicates a different role and regulation of Mn²⁺ in highly manganese-centric organisms like LAB and iron/manganese-centric organisms like B. subtilis, and possibly a different functional evolution of the different systems.

Another hint that MntH1 transporters in manganese-centric lactobacilli might have evolved to serve a distinct role is provided by the conditions under which MntR responds to Mn²⁺ in various bacteria. In S. oligofermentans, transcription of Mn²⁺ scavenging systems only becomes derepressed under oxidative stress after three cysteine residues within MntR get oxidized, thereby releasing MntR inhibition (Chen et al. 2017). However, MntR in lactobacilli has only one cysteine residue, indicating a difference in the effect of oxidative stress on MntR and hence Mn²⁺ import.

Why many LAB, unlike most other aerotolerant organisms, use nonenzymatic instead of enzymatic mechanisms to cope with oxidative stress and developed a fully manganese-centric metabolism remains a topic of debate. One hypothesis is that bacteria obtained distinct mechanisms to endure the transformation of the world from anaerobic to aerobic after the first photosystems evolved around 2.4–3.8 billion years ago (Khademian and Imlay 2020). Before that, most enzymatic reactions were optimized to function under anaerobic conditions. Iron was the ion of choice for this, since its ferric form is highly soluble and bioavailable under low oxygen conditions and has a highly biologically relevant reduction potential, enabling a multitude of reactions (Khademian and Imlay 2020). One of the results of the rise in atmospheric oxygen was the drop in solubility and hence bioavailability of ferric iron, resulting in a low-iron environment. Moreover, the iron-dependent enzymes became more prone to oxidative damage (Khademian and Imlay 2020). As recently stated by Khademian and Imlay, Fe²⁺ was so important in the metabolism of bacteria that they generally did not evolve in the direction of an iron-free metabolism, but instead developed mechanisms to cope with this new scenario. These include the use of Mn-SOD and Mn²⁺ import for oxidative stress protection, scavenging of Fe²⁺ through the use of siderophores and substituting Fe²⁺ for Mn²⁺ where possible (Khademian and Imlay 2020). Lactobacillaceae are believed to have diverged from staphylococci and bacilli around 1.8 billion years ago (Battistuzzi, Feijao and Hedges 2004; Duar et al. 2017) and thus after the rise in atmospheric oxygen. Possibly, they evolved away from Fe²⁺ usage and toward a manganese-centric ion homeostasis to resist oxidative stress, which in turn also provided some of them with an additional growth advantage through competitive exclusion. As such, they constitute one of the few examples of bacteria that have been managed to shift their metabolism to a fully Mn²⁺-dependent, Fe²⁺-free one. The intriguing question why apparently only lactobacilli took this solution remains to be answered.

Overall, it is not yet clear in what order the different functions of accumulating extremely high Mn²⁺ levels by certain lactobacilli evolved. Did the accumulation of high Mn²⁺ concentrations result in a competitive advantage after which SOD genes were lost as it became redundant? Or did it originally evolve as a means of dealing with oxidative stress while lacking SOD and did this give the added advantage of competitive exclusion?

CONCLUDING REMARKS AND OUTLOOK

From the current literature, it is clear that Mn²⁺ conveys a variety of long-underestimated roles in bacteria that go beyond being a mere cofactor of enzymes. Overall, the physiological role of Mn²⁺ and the regulation of its homeostasis appear different in bacterial species depending on their level of iron- or manganese centricity. To date, most knowledge on the regulation of its homeostasis and involvement in protein function is derived from iron-centric organisms that require Mn²⁺ in either small amounts or only under certain conditions, such as oxidative stress. However, not much is known about the transporters and the regulation involved in Mn²⁺ homeostasis in manganese-centric bacteria. The high intracellular Mn²⁺ levels in lactobacilli seem to have multifaceted roles beyond oxidative stress resistance, such as competitive exclusion, and enabling both polyphosphate granules and the low Fe²⁺ levels that in turn facilitate H₂O₂ production toward inhibition of other organisms. The negative correlation between Mn²⁺ and Fe²⁺ levels observed in certain lactobacilli could represent an alternative evolutionary strategy for coping with oxidative stress. Other advantages of high intracellular Mn²⁺ levels such as radiation resistance have been identified and further studies could reveal additional beneficial effects in other organisms, while additional species with high Mn²⁺ concentrations also remain to be identified. Moreover, most studies on Mn²⁺ uptake in lactobacilli have evaluated mutant strains only under standard growth conditions. Characterizing mutant strains of highly manganese-centric organisms under various stress conditions is likely to result in the discovery of yet unidentified functions and regulatory mechanisms of Mn²⁺ homeostasis in both beneficial and pathogenic bacteria.

ACKNOWLEDGMENTS

We acknowledge Ana Rute Neves for critical reading of the manuscript.

Conflict of Interest

The authors are employed by Chr. Hansen A/S, a company that develops and commercializes bacterial cultures.