Manganese is an essential constituent of many metalloenzymes and serves as an enzyme activator. The normal concentrations of
Mn range from 0.072 μM to 0.27 μM in human blood (
Aschner and Aschner, 200) and from
20 μM to 53 μM in human brain (
Bowman and Aschner, 2014). We observed that
THP1 cells, a cell line derived from peripheral blood, started to gain antiviral activity when the medium Mn2+ level reached 2 μM. Thus, the concentrations of Mn2+ required to activate innate immunity appeared to be within the physiological range. We further demonstrated that the
cytosolic Mn levels increased 60 times to 5.8–6.8 μM after virus infection. Critically, Mn2+ supported cGAS to produce cGAMP with low levels of dsDNA while Mg2+ didn’t at all. The elevated cytosolic Mn2+ thus significantly lowered the detection limit of cells to cytosolic dsDNA or DNA viruses. Accordingly, Mn-deficient mice were highly susceptible to DNA virus as
Tmem173−/− mice did. Accumulation of extracellular soluble Mn was also observed. It is conceivable that the extracellular Mn was released by virus-infected cells, both alive and pyroptotic. In fact, the increased Mn contents in bronchoalveolar lavage fluid, white blood cells, and alveolar macrophages in virus-infected mice confirmed the physiological relevance of the Mn extracellular release, suggesting that Mn alerts host both locally at the site of infection and systemically through the circulation. Interestingly, Burleson’s group reported that a single intraperitoneal injection of 10, 20, or 40 μg MnCl2/g body weight caused significantly enhanced NK activity, probably mediated by the production of type I-IFNs (
Smialowicz et al., 1988). Although the amount of Mn used in that work was 10 times higher than what we used (10–40 μg versus 1–2 μg MnCl2/g body weight), we believe that it was most likely through the cGAS-STING activation.
Mn2+ is very similar to Mg2+ in terms of the chemical properties. Previous work showed that Mn2+ is able to replace Mg2+
in vitro in almost half of Mg2+-dependent enzymes, in which the catalytic activity of the enzymes are often maintained (
Wedler, 1993). In this study, however, we found that Mn2+ not only replaces Mg2+ in cGAS activation, but also enhances its enzymatic activity and ligand sensitivity. Similar results have been reported in the insulin receptor associated protein kinase, the kinase activity of which was activated substantially by Mn2+, but not by Mg2+ (
Suzuki et al., 1987,
Wente et al., 1990). It is possible that cGAS binds Mn-ATP much tighter than Mg-ATP, thereby activating cGAS. Alternatively, Mn2+ may induce an activating conformational change of cGAS protein. The crystal structure of cGAS with Mg2+ or Mn2+, probably in the presence of ATP/GTP, will provide insight into the detailed mechanism. Nevertheless, we found that in contrast to Mg2+, cGAS incubated with Mn2+ is inclined to precipitate, suggesting that Mn2+ may induce cGAS protein into a compact conformation which is easier to be activated. Unfortunately, this feature disallowed us to test the interaction of cGAS with ATP, GTP, or dsDNA in the presence of Mn2+. In addition, Mn2+ promoted STING activation through the enhanced cGAMP-STING affinity. This result agreed with previous work showing that Mn2+ promotes the dimerization of c-di-GMP and 3′3’-cGAMP (
Roembke et al., 2014,
Stelitano et al., 2013). Mn is the only essential metal of which the transportation and the subcellular distribution in organelles are not defined (
Horning et al., 2015,
Kwakye et al., 2015). We found that upon DNA virus infection, Mn2+ is liberated from mitochondria and/or Golgi apparatus, the major intracellular reservoirs for Mn2+ storage. In addition, free Mn2+ may also be released from Mn-binding proteins such as metallothioneins and calprotectin, an Mn2+-sequestering protein constituting about 40% of the total cytoplasmic protein in neutrophils (
Brophy and Nolan, 2015), and/or albumin, the major Mn2+-binding protein in plasma (
Foradori et al., 1967,
Rabin et al., 1993). The released Mn2+ from organelles and Mn2+-binding proteins altogether leads to the elevated cytosolic Mn2+ and the activation of cGAS-STING pathway. However, excessive exposure or accumulation of Mn is harmful to the central nervous system due to its preferential Mn uptake by the brains and spinal cords. It is reported that
Manganism occurs in response to acute Mn exposures, while Parkinsonism may result from long-term exposure to low levels of Mn (
Martinez-Finley et al., 2013). In fact, the cellular toxicity of Mn (mainly exerted by Mn2+) has long been recognized and attributed to multiple mechanisms, nevertheless the
molecular basis is still inconclusive, or even contradictory in some cases (
Horning et al., 2015). In particular, elevated type I-IFNs were implicated in the development of Parkinson’s disease. In some extreme cases,
type I-IFN treatment caused severe Parkinsonism that was resolved after interferon withdrawal (
Mizoi et al., 1997,
Sarasombath et al., 2002). We showed that Mn2+ accumulation causes prominent innate immune activation, leading to the production and secretion of cytokines.
Such pathological immune activation in the central nervous system will certainly contribute to the cellular and tissue damage, culminating in Manganism and Parkinsonism. These findings therefore may provide new insights into the molecular basis underlying the development of Manganism, in which Mn-caused toxicity has long been recognized. New therapeutic attempts should focus to prevent this detrimental immune activation in the central nervous system.
