Paramagnetism in Bacillus spores: Opportunities for novel biotechnological applications

Abstract Spores of Bacillus megaterium, Bacillus cereus, and Bacillus subtilis were found to exhibit intrinsic paramagnetic properties as a result of the accumulation of manganese ions. All three Bacillus species displayed strong yet distinctive magnetic properties arising from differences in manganese quantity and valency. Manganese ions were found to accumulate both within the spore core as well as being associated with the surface of the spore. Bacillus megaterium spores accumulated up to 1 wt.% manganese (II) within, with a further 0.6 wt.% adsorbed onto the surface. At room temperature, Bacillus spores possess average magnetic susceptibilities in the range of 10−6 to 10−5. Three spore‐related biotechnological applications—magnetic sensing, magnetic separation and metal ion adsorption—were assessed subsequently, with the latter two considered as having the most potential for development.


| INTRODUCTION
Bacillus cells initiate sporulation as a protective mechanism in response to nutrient starvation. Successful formation of heat resistant, metabolically dormant spores typically requires mineral rich growth medium, including a requirement for manganese ions (Greene & Slepecky, 1972) (Charney, Fisher, & Hegarty, 1951). The importance of manganese for enzymatic activity in biological systems is wellestablished (Culotta, 2000); however, while most microorganisms only require trace amounts of manganese, spores are unusual in that they accumulate a much greater amount than vegetative cells (Curran, Brunstetter, & Myers, 1943). Warth, Ohye, and Murrell (1963) reported that Bacillus spores could contain as much as 1.5 wt.% manganese. Such high levels of uptake suggest that manganese has an important function within spores. Many authors have explored the role of manganese ions in several of the spore's resistance properties (Ghosh et al., 2011;Levinson & Hyatt, 1964), as well as in germination (Levinson & Sevag, 1954;Levinson, Sloan, & Hyatt, 1958). However, only one group has reported the intrinsic magnetic properties conferred by the high manganese content in spores (Melnik et al., 2007).
Manganese is a chemically versatile transition metal, which can exist in several oxidation states ranging from +1 to +7. This allows manganese to display a wide variety of magnetic behaviors, from ferromagnetic manganese oxides (Du, Zhang, Sun, & Yan, 2009) to diamagnetic permanganate compounds (Goldenberg, 1940). The exact chemical state of manganese in the spore is not known, however, evidence suggests that Mn 2+ ions compete with Ca 2+ ions to form a chelate with dipicolinic acid (DPA) inside the spore core (Bailey, Karp, & Sacks, 1965). In this case, bacterial spores are very likely to be paramagnetic at room temperature, exhibiting a small attractive force toward permanent magnets.
The magnetic properties of spores may be of biotechnological value, particularly in sensing, and separation related applications.
Current magnetic methods for the separation of spores require them to be tagged with antibody functionalized magnetic beads (Shields et al., 2012). However, with the exception of magnetotactic bacteria (Blakemore, 1975), most microorganisms tend to be diamagnetic, thus the intrinsic paramagnetism of spores may be sufficient to achieve efficient separation.
In sensory applications, most methods rely on a similar antibody functionalisation and tagging procedure (Gómez de la Torre et al., 2012;Wang et al., 2015) without exploiting the spore's intrinsic paramagnetic properties. Conceivably, potential cross-species variance in the latter may be detectable by a sensitive magnetometer and used to identify dangerous species such as B. anthracis. However, increased understanding of the magnetic properties of spores is required before any applications can be considered. Accordingly, the aim of this work was to fully characterize the magnetic properties of three species of Bacillus spores, determining the type of magnetism that they exhibit, and determining the role of manganese in conferring this magnetism. Possible alterations to the magnetic properties as well as novel biotechnological applications are also explored.

| Bacterial strains and sporulation conditions
Three lab strains of Bacillus were studied in the current work: B. megaterium QM B1551, provided by Prof. P. S. Vary (Illinois); B. cereus 569, provided by Prof. A. Moir (Sheffield, UK), and B. subtilis PS832, provided by Prof. P. Setlow (CT). The sporulation conditions and spore purification procedures followed were those described previously by Nicholson and Setlow (1990), Clements and Moir (1998) and Christie, Ustok, Lu, Packman, and Lowe (2010) for the sporulation of B. subtilis, B. cereus, and B. megaterium, respectively. The latter two species were additionally cultured in media containing manganese concentrations varying between 20 and 1,000 µM. In addition, the uptake of iron and cobalt ions in B. megaterium spores was investigated by culturing spores in media containing 100 µM of FeSO 4 and CoCl 2 , respectively.

| Sample preparation for magnetometer measurements
Clean spore suspensions, containing >95% spores (checked by optical microscopy), were concentrated to a paste-like density by centrifugation.
Between 100 and 150 µl of this paste was pipetted into the cap and body of a size 0 gelatine capsule, pre-frozen to −80°C. This achieves a quick freeze and prevents the gelatine from extracting water from the spore paste. The frozen spore capsule was then lyophilized at −20°C and 250 mTorr pressure for 8 hr. Capsules were sealed when dry and weighed to determine the dry mass of spores within. To minimize any background signal, spore capsules were suspended within a straw supplied by Quantum Design (San Diego, CA) with the aid of two low paramagnetic glass rods. The straw was then sealed with cotton at both ends and Kapton™ tape at the lower end before being placed into the magnetometer for measurement.

| Magnetic property measurements
The magnetic response of spore samples was measured in a superconducting quantum interference device (SQUID) (Quantum Design MPMS 2XL) for both field and temperature dependence. For field dependence, the temperature was set to 5 K and a full hysteresis sweep between 5 T and −5 T was conducted. In temperature dependence, zero field cooling measurements were conducted: the sample was first cooled to 2 K before a 0.1 T magnetic field was applied. The temperature was then increased to 300 K in 2 K intervals.

| Imaging and spectroscopy
B. megaterium spore samples were prepared for scanning electron microscopy (SEM) imaging and energy dispersive X-ray spectroscopy (EDS). Around 1 ml of a high density (OD 600 > 10) spore suspension was deposited on a Leit tab pre-coated with poly-L-lysine to increase spore adhesion. After 5 min, the tabs were gently rinsed with distilled water to remove any unwanted chemicals and subjected to a fast freeze in liquid ethane to preserve the spore's morphological features.
Frozen samples were lyophilized overnight at a fixed pressure of 10 −3 mbar and a temperature programme of initially −90°C for 2 hr then subsequently increasing by 10°per hour up to a final temperature of 30°C. Lyophilized samples were then sputter coated with a thin layer (25-50 nm) of conductive carbon and stored in a dry location, at room temperature, prior to imaging. SEM imaging and EDS were undertaken using a FEI XHR Verios 460 (Hillsboro, OR) mounted with an AMETEK EDAX detector device (Mahwah, NJ).

| Inductively coupled plasma-optical emission spectroscopy (ICP-OES)
The lyophilized spore capsule samples prepared for magnetic measurements were submitted for ICP-OES analysis to determine the manganese content in spores (Medac Ltd., Chobham, UK).

| Spore demineralisation
Spores were demineralized by acid titration as previously described by Marquis and Bender (1985). A solution of 5 wt.% spores was titrated with hydrochloric acid over approximately 4 hr until the pH stabilized at 4.0. To ensure that all minerals were removed, the suspension was washed three times by centrifugation and resuspension in deionized water and the titration process repeated. All acid treated spores were prepared from the same culture batch as their non-treated counterparts and were thoroughly washed in deionized water before conducting experiments.

| Surface adsorption experiments
Approximately 0.2 g of MnCl 2 or FeSO 4 salt crystals were added to a 25 ml suspension of 1 wt.% B. megaterium spores and the solution was mixed well. Once the crystals had dissolved, 1 ml of 1 M sodium hydroxide solution was pipetted into the suspension, forming a colored precipitate. The suspension was then vortexed until no further oxidation occurred and the color of the precipitate was stable. A total of 2 ml of 2 M HCl was subsequently added to the suspension, which was incubated on ice overnight. Spores were separated from inorganic debris by repeated rounds of centrifugation and resuspension in deionized water. Suspensions were cleaned to give over 95% phasebright spores before experiments were conducted.

| Theory
Paramagnetic materials possess at least one unpaired electron that can align with the externally applied magnetic field; however, the extent to which these electrons can align with the magnetic field is low as thermal excitation randomizes the spin orientation for most electrons and only a small proportion will align with the applied field.
In most cases, the mass magnetisation, M, of a paramagnetic material depends linearly on the applied field, H, according to, where χ m is the mass magnetic susceptibility (m 3 kg −1 ). According to Shevkoplyas, Siegel, Westervelt, Prentiss, and Whitesides (2007), the magnetic susceptibility of the material, following conversion from centimetre-gram-second units to SI units, may be expressed as, where χ spore is the dimensionless magnetic susceptibility of the spore and ρ spore is its density in kg m −3 . At low enough temperatures or high enough fields, however, saturation occurs and the field dependence can be expressed using the Langevin function, where M is the mass magnetisation of the sample (A m 2 kg −1 ), N is the number of magnetic entities per unit kg of spores, m is the magnetic moment of each entity in Bohr magnetons, μ B is the Bohr magneton constant (9.274 × 10 −24 J T −1 ), k is the Boltzmann constant (1.381 × 10 −23 m 2 kg s −2 K −1 ), µ 0 is the magnetic permeability of free space (4π × 10 −7 H m −1 ), and H is the applied magnetic field (A m −1 ).
The magnetic moment, m, should indicate the oxidation state of the manganese contained within the spores and the number of magnetic entities, N, can be converted into a weight percentage of manganese, w Mn , according to, The magnetic susceptibility of paramagnetic materials is found to obey Curie's law, where C is the Curie constant given by Cullity (1972) as, Siegel et al. (2006) developed a microfluidic device (Supplementary Figure S1) that was used to separate superparamagnetic beads using a weak electromagnetic field. The theoretical equations later derived by Shevkoplyas et al. (2007) were employed to assess the feasibility of exploiting the magnetic properties of spores for separation. The setup consists of a 40 µm wide channel sandwiched between two electromagnets. The distance between the surface of the channel and the center of the electromagnet is 70 µm. In their analysis, the electromagnet was modeled as a single electrical wire carrying current I. According to Shevkoplyas et al. (2007), the displacement time, t, required for a spore of radius R to move through a liquid of viscosity µ from its initial position, x i , to the final position, x f , is given by

| Statistics
Statistical analysis was performed using Microsoft Excel (Redmond, WA). The method of least squares was used to fit experimental data to the corresponding equations. At each temperature or magnetic field strength, the magnetisation value was evaluated as the average of three scans; the errors associated with each measurement were evaluated by the magnetometer's inbuilt software. These errors were then used to estimate the uncertainties associated with the least squares fitting and hence the error in the fitting parameters.

| Magnetic behavior of B. megaterium spores
It is well-known that bacterial spores contain a wide range of inorganic elements when cultured in media with an adequate metal balance (Murrell, 1969). The EDX spectrum in Figure 1a shows that B. megaterium spores grown in standard supplemented nutrient broth (SNB) possess Ca, P, Mg, K, and Mn as their main inorganic constituents. This result is in agreement with those reported by Murrell (1967) with the exception that the spectrum contains no trace of Fe or Na. These elements are most likely present in quantities below the detection limit for the EDX detector (typically 0.1 wt.%). In B.
subtilis, Granger, Gaidamakova, Matrosova, Daly, and Setlow (2011) used ICP-OES to find the spore's iron abundance to be around 0.005 wt.%, two orders of magnitude lower than the typical amount of manganese present. Additionally, good agreement is found between the relative size of the EDX spectra peaks and the relative proportions of each element determined by Curran, Brunstetter, and Myers (1943) and Murrell (1967) for a variety of Bacillus species and strains.
The larger Mn peaks in Figures 1b and 1c indicate that the developing spores had taken up additional manganese when produced in higher manganese concentration media. Furthermore, as the Mn peak becomes larger the Ca peak decreases while all other peaks remain similar in size. The former observation is in agreement with the hypothesis that calcium and manganese ions compete for sites in the spore, as suggested previously by Slepecky and Foster (1959) and Levinson and Hyatt (1964). When spores were cultured in the presence of other divalent transition metal salts, no uptake was detected by EDX (data not shown). This result suggests that the uptake mechanism is specific to manganese and if spores do accumulate other transition metal ions, the amounts are below the detection limit for EDX analyses. This suggests that the ionic state of the manganese is 2+ (Figgis & Lewis, 1964) and is very likely to be part of a tetrahedral or octahedral complex.

| Magnetic properties of spores of other species and vegetative cells
An observable paramagnetic response in the presence of a magnetic field gradient is not a property exclusive to B. megaterium spores. Melnik et al. (2007)

| Effect of culture medium inorganic salt concentration on spore paramagnetic properties
The concentration of manganese in culture media had a strong impact on the strength of the measured signal as it determines the amount of manganese that the mature spore contains (Supplementary Tables S1 and S2). As the manganese content increased, the corresponding magnetic moments decreased from 5.9 to 5.1 and from 5.1 to 3.8 for B. megaterium and B. cereus spores, respectively. According to Figgis and Lewis (1964), lower magnetic moments indicate that the predominant manganese valency in spores with enhanced manganese content is increased by one. As the drop in magnetic moment occurs gradually with manganese concentration, it indicates that more than one oxidation state of manganese is present in the spore and that the relative proportions of each is determined by the concentration of manganese during spore formation. Given the range of magnetic moments observed, it is likely that B. megaterium spores contain a combination of the 2+ and 3+ oxidation states and B. cereus spores a combination of 3+ and 4+, however, the presence of smaller quantities of other oxidation states cannot be ruled out. Figure 3 illustrates the effect of culture medium manganese concentration on the manganese content of B. megaterium and B. cereus spores. The accumulation of manganese in both acid treated and untreated spores can be adequately described by an exponential decay function of the form where [Mn] is the concentration of manganese in the culture medium, and a 1 and a 2 are constants which represent the saturation manganese weight percent and the culture concentration which yields 63% of the saturation value, respectively. At low manganese concentrations (<100 µM), the quantity of manganese uptake is linear and almost identical for both acid treated and untreated spores. At higher concentrations, the manganese content tends asymptotically toward 1.0 and 1.6 wt.% for acid treated and untreated spores, respectively.
The accuracy of values is confirmed by the agreement with ICP-OES measurements for selected samples, as observed in Figure 3. Although the Langmuir isotherm is equally adequate in modelling the data in Figure 3, the isotherm describes the surface adsorption of materials rather than the combined effects of adsorption and absorption.
Further analysis of the data using the Langmuir isotherm (Supplementary Figure S3) can be found in the discussion.
The consistently lower manganese content in acid-treated spores indicates that manganese is externally bound and removed by this treatment. Furthermore, it suggests that spores accumulate manganese through two mechanisms, giving rise to the two asymptotes observed in Figure 3a. The lower asymptote is likely to be determined by the manganese transport process that is active during sporulation, whereas the higher asymptote is a physical limit imposed by adsorption on to the surface of the spore.  Figure S2). This suggests that developing B. megaterium spores have some form of specificity toward the uptake of manganese ions and neither adsorb nor absorb significant amounts of iron or cobalt. A similar observation for the uptake of iron ions was reported by Granger et al. (2011) in B. subtilis, but this result for cobalt contradicts the findings of Slepecky and Foster (1959). In the latter, B.
megaterium spores grown in media containing 680 µM Co 2+ possessed 0.14 wt.% cobalt. This difference may be caused by the lower concentration of cobalt (150 µM) used in this work.

| Magnetic susceptibility at room temperature
The spore's magnetic susceptibility at 300 K can be determined from the slope of the magnetisation chart. Figure 4 shows the magnetisation behavior of B. megaterium spores cultured in media with three different concentrations of MnCl 2 . All three data sets show linear behavior and the magnetic susceptibility can be calculated from Equation (2) (using SI units) as 1.0 × 10 −6 , 7.2 × 10 −6 , and 1.9 × 10 −5 for 20, 50, and 100 µM MnCl 2 , respectively. As the magnetic moments do not reach saturation at 300 K, these susceptibility values are indicative of spores of the Bacillus genus; therefore, at room temperature, the average magnetic susceptibility of Bacillus spores lies in the range 10 −6 and 10 −5 . The manganese coated spore sample possesses a strong magnetic signal at 5 K, displaying hysteresis and a very large coercive field of approximately 500 Oersted. To illustrate their magnetic strength, even at room temperature, Figure 5c contains an image of these spores aligning with the magnetic field lines between two circular neodymium magnets; a behavior that resembles that of iron filings. The shape of Figure 5c is uncommon as it derives from two contributions-a paramagnetic component from within the spore and a surface contribution that appears to be ferromagnetic. Figure 5d reveals that the surface manganese undergoes a transition at temperatures above 43 K, which corresponds to the Curie temperature of Mn 3 O 4 (Boucher, Buhl, & Perrin, 1971). These results strongly indicate that the manganese shell surrounding the spore exhibits bulk ferromagnetic properties at 5 K; nevertheless, an antiferromagnetic contribution, due to the presence of uncompensated spins on the surface, cannot be entirely excluded (Cooper et al., 2013). The iron coated spore sample demonstrates that the adsorbate is not limited to manganese. These spores bear the distinctive bright yellow color of a hydrated iron oxide-hydroxide (FeO(OH) · nH 2 O).

| Surface adsorption
Figure 5b reveals that the crystals form thin rods and aggregate in a foliated manner, fitting the description for goethite (Cornell & Schwertmann, 2004a). In this case, the iron coated spores would be antiferromagnetic with a Néel temperature of 400 K (Cornell & Schwertmann, 2004b); however, further analyses are required to confirm the identity of the iron compound.

| DISCUSSION
To date, there have only been two accounts reporting the intrinsic magnetic properties of bacterial spores (Melnik et al., 2007;Sun, Zborowski, & Chalmers, 2011). This work provides the first comprehensive characterisation of these properties in Bacillus spores and explores the extent to which they can be altered. In agreement with previous results (Warth et al., 1963), manganese is the only non-diamagnetic element that accumulates to any significant amount in spores and SQUID measurements have confirmed that this manganese is paramagnetic at all temperatures up to 300 K. Similar to Sun et al. (2011), we determined that several oxidation states of manganese (2+, 3+, and 4+) are present in spores; however, our results indicate a lower manganese content, in the range of 0.2 wt.% rather than 10 wt.%. Increasing the concentration of manganese in the culture medium raises the manganese content in spores approximately 10-fold. Nevertheless, the force experienced by these spores in a magnetic field would still be several orders of magnitude smaller than that experienced by ferromagnetic entities. With this in mind, potential applications associated with spore paramagnetism include: (i) Sensing: Manganese was found to be the cause for paramagnetism in all three Bacillus species studied, yet their respective SQUID measurements presented some clear differences.
The chemical state and environment in which the manganese ions exist in the spore differs across the species-under standard conditions, not only do the spores contain different amounts of manganese, their magnetic moments also range between 5.1 and 5.9 µ B . These differences may be useful for the detection of more harmful, pathogenic species such as B. anthracis or Clostridium difficile. There are, however, two major obstacles to overcome. The first concerns sensitivity; the device must be sensitive enough to detect the magnetic signal from, typically, a small quantity of spores.
This obstacle is perhaps the easiest to overcome as there are commercial paramagnetic oxygen sensors that can measure the concentration of oxygen at room temperature. According to Wills and Hector (1924), the magnetic susceptibility of oxygen at room temperature is of the order of 10 −7 , suggesting that these devices would also be capable of detecting the magnetic signal emitted by spores. The second obstacle concerns differentiation; the ability to distinguish specific spore species or strains from a mixture of spores and impurities. Although most impurities will possess a very different magnetic susceptibility to spores, it will be challenging to differentiate between spore species. In particular, the spore's magnetisation does not saturate at room temperature and therefore it will not be possible to obtain magnetisation curves such as those found in Error bars indistinguishable when plotted difficulty of overcoming this challenge, it is unlikely that magnetic fields could be used to discern between different spore species.
(ii) Separation: A microfluidic device designed to expose spores in aqueous suspension to an open gradient permanent magnet has been demonstrated to capture and deposit B. atrophaeus spores onto a glass slide (Melnik et al., 2007). To assess further the theoretical considerations associated with spores' intrinsic paramagnetism for separation applications, a microfluidic device such as that presented by Shevkoplyas et al.(2007) may be considered. To estimate the shortest time required for separation, the highest values for magnetic susceptibility (10 −5 ) and electric current (1 A, maximum current used in their work) will be assumed. Furthermore, spores may be approximated as spheres with radius 0.5 µm. The liquid viscosity is expected to be around 10 −3 Pa s. Using these values, Equation (7) estimates that, under the influence of a weak magnetic field, it would take over 166 hr for the spores to reach the sidewall from the midpoint of the microfluidic channel. At this timescale, separation would not be feasible, however it should be noted that the magnetic field generated in the channel is weak, at approximately 2.5 mT for 1 A current. For comparison, a typical permanent neodymium magnet has a remanence between 0.2 and 1.2 T. As the wire is only a proxy for the electromagnet in the experimental setup, the current passing through it may be increased indefinitely to achieve the desired magnetic field.
Therefore, by increasing the current through the wire to 100 A, a magnetic field of 0.25 T is produced and the separation time would then be reduced to approximately 60 s. This calculation illustrates that while it is possible to exploit a spore's intrinsic magnetic properties for separation, it does require moderately strong magnetic fields (see Supplementary Figure S4 and Melnik et al., 2007). This analysis also disregards the fact that the calculated magnetic susceptibility is an average property and therefore natural variation will mean that longer times will be required to achieve high separation yields.
(iii) Adsorption: Manganese is accumulated both inside the spore and on its surface. while the presence of manganese in the core is well known (Thomas, 1964), adsorption to the spore integuments is less well characterized. Alderton and Snell (1963) demonstrated that bacterial spores behave like ion exchange gels, where following an acid regeneration treatment, the surface becomes free to take up structures that show base exchange behavior. The results in Figure   3a may be interpreted further using an adsorption model.
Subtracting the manganese content of acid-treated spores in Figure   3a from that of untreated spores yields an adsorption curve. Note that the concentration of manganese has been adjusted by approximately 43 µM in order to account for the onset of adsorption. By assuming that manganese adsorbs onto spores following the Langmuir isotherm, an inverse plot may be used to obtain the adsorption parameters q m = 14.4 mg/g and K D = 0.88 mM −1 (Supplementary Figure S3). Based on Figure 3a, the adsorption capacity was higher than expected, however, in the absence of competing ions it does appear to be a reasonable value.
Compared with alternative natural manganese adsorbents (Omri & Benzina, 2012), a maximum adsorption capacity of 14.4 mg Mn 2+ per gram of spore may not be particularly attractive. Equally, spores are not limited to the adsorption of manganese ions. The activated surface of acid treated spores, for example, may find an application in removing heavy metal ions from contaminated waters.
Additionally, this work has shown that spores can be easily functionalized with ferromagnetic nanoparticles, greatly increasing their magnetic susceptibility and enabling their use in a wide array of biotechnological applications for example, drug delivery.

| CONCLUSION
Bacillus spores are paramagnetic due to the high manganese content accumulated within the spore core and that associated with the spore coat (which may be surface bound or distributed throughout the coat).