Materials and Methods

Culture of NB from Serum
NB specimens were cultured from FBS or HS as described before [5], [85]. Approval for the use of human samples in this study was obtained from the Institutional Review Board of Chang Gung Memorial Hospital (Gueishan, Taiwan, Republic of China). Written informed consents were signed by the individuals who provided blood samples. Human blood was obtained from healthy volunteers by venipuncture following sterilization of the skin with alcohol. The blood was withdrawn into sterile Vacutainer tubes containing no anticoagulant (Becton, Dickinson & Company, Sparks, MD, USA). Whole blood was centrifuged at 1,500× g for a period of 15 min at room temperature. The supernatant corresponding to HS was retrieved and placed into another tube. The FBS (Biological Industries, Kibbutz Beit Haemek, Israel; PAA Laboratories, Pashing, Austria) and HS used throughout this study were sterilized by filtration through both 0.2-µm and 0.1-µm membranes (Pall Corp., Ann Arbor, MI, USA) prior to use. To culture NB, both sera were diluted into DMEM (Gibco, Carlsbad, CA, USA) to final concentrations ranging from 0.1% to 10%. Culture was performed in 24-well plates with flat-bottom wells and covering lid (Corning, Inc., Corning, NY, USA) using a final volume of 1 ml per well. Culture was also performed in 75-cm2 flasks with 0.2-µm vented caps (Corning) using a final volume of 20 ml per flask. Untreated DMEM was used as a negative control. The culture plates and flasks were incubated at 37°C for several months in the humidified atmosphere of a cell culture incubator. Nanobacterium sp. strain DSM 5820 was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ; Braunschweig, Germany). Culture of “nanons”, which was initially called “Nanobacterium sp. strain Seralab 901045” [46], [86], was kindly provided by Dr. Didier Raoult (Unité des Rickettsies, Centre National de la Recherche Scientifique UMR 6020, Faculté de Médecine, Marseille, France). Both NB strains DSM 5820 and “nanons” were originally isolated from commercially available FBS used for cell culture purposes [86].
To prepare NB samples for electron microscopy and spectroscopy analyses, the cell culture medium of a flask containing a 1-month-old culture of NB was discarded and the NB sample which consisted of a white precipitate adherent to the flask was scraped using a sterile cell scraper (Corning). The precipitate was resuspended in 1 ml of DMEM and centrifuged at 16,000× g for 15 min at room temperature. The pellet was washed twice with DMEM, HEPES buffer (20 mM HEPES, 1 mM CaCl2, 2 mM Na2HPO4, 0.02% sodium azide, and 0.15 M NaCl, pH 7.4), or double-distilled water using the same centrifugation steps. The NB specimen was resuspended in a small volume of double-distilled water and used for the microscopy and spectroscopy analyses.

Photography and Spectrophotometry
Images of the 24-well plates used throughout this study were obtained using a scanner operating in the reflective light mode (Scan Maker 8700, MicroTek, Hsinchu, Taiwan) as described earlier [2]. Spectrophotometry readings of 24-well plates were performed at 650 nm using a Spectra Max M2 spectrophotometer (Molecular Devices, Sunnyvale, CA, USA), essentially as described [2]. Throughout this study, photographic and A650 turbidity readings referred to as “Day 1” were taken within one hour following the preparation of each 24-well plate.

Culture of NB-Like Particles from Protease-Treated Serum and Boiled Serum
Stock solutions of porcine pancreas trypsin (Sigma, St-Louis, MO, USA) or bovine pancreas chymotrypsin (Sigma) were prepared in water at a concentration of 5% (w/v). The protease solutions were sterilized by filtration through a 0.2-µm membrane. The NB-like particles depicted in Fig. 2 were prepared by adding trypsin or chymotrypsin into FBS or HS at a final concentration of 0.5% (v/v), followed by incubation at 37°C for 2 hours. A final volume of 1 ml of serum was used. Following incubation, the solution was diluted to final concentrations ranging from 0.1% to 10% (v/v) in DMEM and the mixture was incubated in cell culture conditions for several weeks. Alternatively, this experiment was repeated by treating FBS or HS with trypsin or chymotrypsin that had been boiled at 95°C for 1 hour. As a negative control, the stock solutions of trypsin or chymotrypsin were diluted into DMEM to final concentrations ranging from 0.01% to 3% (v/v) prior to incubation. The DMEM used for these experiments contained 0.02–0.2% sodium azide in order to prevent contamination.
To culture NB-like particles from boiled serum as shown in Fig. 3, HS was first diluted to 25% (v/v) using double-distilled water. FBS and 25% HS solutions were boiled at 95°C for 10, 30, or 120 min. The boiled sera were then diluted into DMEM to final concentrations ranging from 0.1% to 10% (v/v) and the solutions were incubated in cell culture conditions for several weeks. In some instances, stock solutions of 0.25 M CaCl2 and NaH2PO4 (both at pH 7.4) were successively added at a final concentration of 1 mM prior to incubation. The pH of the stock solution of 0.25 M CaCl2 was adjusted to 7.4 with 1 M HCl or 1 M NaOH whereas the pH of the NaH2PO4 solution was adjusted to 7.4 with 0.25 M Na2HPO4. These solutions were also sterilized by filtration through a 0.2-µm membrane prior to use. Parallel experiments were also conducted by adding aliquots from 0.25 M NaHCO3 pre-adjusted to pH 7.4, to the same final concentrations as those of CaCl2 and Na2HPO4. Results were virtually identical compared to experiments without carbonate. For brevity, only the data for the addition of calcium and phosphate are shown in the present study.

Seeding of NB-Like Particles by Fetuin-A, Albumin, and Serum in Supersaturated Solutions
Stock protein solutions of BSF and HSA were prepared by dissolving the protein in HEPES buffer at a concentration of 10 mg/ml, followed by filtration through 0.2-µm filters. Several lots of proteins were used for the experiments described in this study. For Fig. 4, BSF and HSA were obtained from AppliChem (Boca Raton, FL, USA); for Fig. 5, BSF and HSA were from Sigma; and for Fig. 6, BSF was from Sigma while HSA consisted of a sterile solution used for intravenous injections (Plasbumin®-25; Talecris Biotherapeutics, Inc., Research Triangle Park, NC, USA). The stock solutions of proteins were then diluted into DMEM, individually or in combination, to concentrations varying between 0.7 µg/ml and 40 mg/ml. Other cell culture media obtained from Gibco were used in parallel, including Roswell Park Memorial Institute 1640 or RPMI-1640, F12 medium, medium 199, Glascow minimum essential medium, and Leibovitz L-15 medium. In some experiments, the precipitating reagents CaCl2 and NaH2PO4 were added to final concentrations varying between 0.1 mM to 1 mM. The final solution volume used was 1 ml. The solutions were then incubated in cell culture conditions for several weeks. The concomitant addition of carbonate, as outlined in the previous section, yielded similar results and data for such experimens are not shown here.
To culture NB-like particles from boiled protein solutions as shown in Fig. 7, solutions of BSF purified from FBS (Sigma) or BSA purified from FBS (Sigma) were prepared in HEPES buffer at a final concentration of 25 mg/ml. The protein solutions were filtrated through 0.2-µm membranes prior to use. These protein solutions were boiled at 95°C for 10, 30, or 120 min. The boiled protein solutions were diluted in DMEM at final concentrations ranging from 0.02 mg/ml to 2 mg/ml for boiled BSF and from 0.04 mg/ml to 4 mg/ml for boiled HSA. In some experiments, the precipitating reagents CaCl2 and NaH2PO4 were added successively each at a final concentration of 1 mM to the DMEM solutions containing proteins. The solutions were incubated in cell culture conditions for several weeks.
In order to evaluate the possibility that adsorbed proteins nucleate NB-like particles, BSF (Sigma), BSA (Sigma), or HSA (Talecris) were used in adsorption experiments. Adsorption of the proteins to polystyrene 24-well plates was performed by covering each well with 250 µl of protein solution at concentrations varying between 20 µg/ml to 10 mg/ml. Similarly, 250 µl of FBS and HS were also used at concentrations varying from 0.1% to 10%. The plate was incubated at 4°C overnight. Following incubation, the protein solution was removed and each well was washed twice with 250 µl of DMEM. We verified that the proteins used were adsorbed to the plate by staining the wells with 250 µl of Coomassie blue diluted 1∶5 in double-distilled water (Dye reagent concentrate; Bio-Rad Laboratories, Hercules, CA, USA). After 10 min of incubation, the shift to blue color was monitored directly by visualization of the plate and compared to controls without proteins which did not produce the blue color. Following adsorption of the proteins, 1 ml of DMEM was deposited in each well and the plate was incubated in cell culture conditions for several months. Alternatively, the protein solutions were deposited in each well and left to dry overnight under a laminar flow hood. Each well was washed twice with DMEM. 1 ml of DMEM was pipetted into each well and the plate was incubated in cell culture conditions. The coating agents poly-lysine (Sigma) and octadecyltrichlorosilane (OTS; Sigma) were also used to promote adherence of the proteins to the bottom of each well. For coating with poly-lysine, 250 µl of a 0.5% (w/v) solution of poly-lysine was incubated into each well. The plate was incubated for 5 min and the solution was removed. The plate was then incubated with the protein solution at 4°C overnight or left to dry overnight, followed by the same procedure described above. For coating with OTS, 250 µl of a solution of 0.5% (w/v) OTS was pipetted into each well and dried for 5 min. The protein solutions were then deposited into each well and the plate was processed as described above. Precipitation was monitored regularly by A650 turbidity readings, by visual inspection, and by using a cell culture inverted microscope (Diaphot; Nikon, Tokyo, Japan) at a magnification of 400X.

Estimation of the Number of Apatite Crystals Bound to Each Fetuin-A or Albumin Molecule under Conditions that Produce Optimal Turbidity of Seeded Nanoparticles
The number of apatite crystals bound to each molecule of BSF or HSA was calculated for the experiment described in Fig. 6, under conditions that produced maximal precipitation of seeded nanoparticles following one month of incubation. For instance, when 0.7 mM of precipitating reagents was added to the three different protein solutions, followed by incubation for one month, maximal turbidity was observed at 21 µg/ml for BSF and at 0.4 mg/ml for HSA. An example of the calculation performed is described here for 0.7 mM of precipitating reagents added to DMEM in the presence of BSF at a concentration of 21 µg/ml. We estimated the number of apatite crystals bound to each protein molecule for these conditions by dividing the number of phosphate ions by the number of protein molecules present in the well. Since the molar concentration of phosphate ions present in the well was lower than that of calcium ions, we considered that phosphate would be the limiting factor for the formation of apatite crystals under these conditions. To calculate the number of phosphate ions present in the well, we added the concentration of phosphate ions used (0.7 mM) to the concentration of phosphate already present in DMEM (0.9 mM). This number was then converted to the number of moles of phosphate present in the well which contained 1 ml of solution (1.6×10−6 moles/well). This value was multiplied by Avogadro's number (6.023×1023 atoms/mole) in order to obtain the number of phosphate ions present in the well (9.64×1017). We assumed that all the phosphate ions present in the well would form apatite crystals after prolonged incubation. Thus, the number of phosphate ions present in the well (9.64×1017) was divided by the number of phosphate ions present in a single crystal unit of apatite (this latter value was averaged to 63 phosphate ions based on earlier estimations, see ref. 94). The estimated number of apatite crystals (1.53×1016) was then divided by the number of BSF molecules present in the well at a concentration of 21 µg/ml (2.64×1014 molecules of BSF) in order to obtain the number of apatite crystals bound to each molecule of BSF (58 apatite units/BSF molecule).

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE)
NB-like particles containing fetuin-A and/or albumin like the ones shown in Fig. 8 were prepared by diluting BSF (Sigma) at a final concentration of 20–160 µg/ml in DMEM or HSA (Talecris) at a final concentration of 0.2–1.6 mg/ml in DMEM. The precipitating reagents CaCl2 and NaH2PO4 were then added each at a final concentration of 3 mM. A final volume of 1 ml of DMEM was used. Incubation was done in cell culture conditions for 1 month. The particles were then pelleted by centrifugation at 16,000× g for 15 min at room temperature. The pellet was washed twice with HEPES buffer using the same centrifugation steps. The particles were resuspended in 50 µl of 50 mM EDTA. Each sample was mixed with the 5X “loading buffer” (0.313 M Tris-HCl pH 6.8, 10% SDS, 0.05% bromophenol blue, 50% glycerol, 12.5% β-mercaptoethanol) to obtain a final concentration of “loading buffer” of 1X in a volume of 20 µl. 50 mM EDTA was used to dilute the samples. The protein solutions were heated at 95°C for 5 min and were subsequently loaded on a 10% SDS-polyacrylamide gel. For Fig. 8A, BSF-NLP were prepared by using DMEM (final volume of 1 ml) containing BSF (Sigma) at 20 µg/ml (lane 1), 40 µg/ml (lane 2), 80 µg/ml (lane 3), and 160 µg/ml (lane 4), followed by addition of CaCl2 and NaH2PO4 each to 3 mM and incubation in cell culture conditions for 1 month. Following incubation, the particles were pelleted as described above and washed twice with DMEM. The pellet was resuspended in 50 µl of 50 mM EDTA and 4 µl of each sample was loaded in the lanes described above. For Fig. 8B, HSA-NLP were prepared in a similar manner by using DMEM containing HSA (Sigma) at 0.2 mg/ml (lane 1), 0.4 mg/ml (lane 2), 0.8 mg/ml (lane 3), and 1.6 mg/ml (lane 4). For Fig. 8C, BSF-HSA-NLP were prepared similarly by using DMEM containing both proteins at the concentrations mentioned above. Gel electrophoresis was performed using a mini-gel system (Hoefer, Holliston, MA, USA). The gels were stained with Coomassie blue as described earlier [2].

Protein Quantification
To quantify proteins, a standard curve was prepared by diluting a stock solution of BSA (Sigma) in double-distilled water at various concentrations varying from 10 to 50 mg/ml. Four volumes of each protein solution were mixed with one volume of dye reagent concentrate (Bio-Rad Laboratories) using a final volume of 1 ml of solution. The solutions were mixed and the optical density was monitored with a spectrophotometer (Molecular Devices) at a wavelength of 595 nm. A “blank” solution containing only the dye reagent diluted in double-distilled water as described above was used to subtract the optical density seen without proteins. A graph of optical density at 595 nm was plotted against the concentration of proteins from the standards. The samples with unknown concentration of proteins were processed the same way as the standard solutions. The protein concentration of the unknown was evaluated based on the optical density of the protein solutions and interpolation from the graph prepared. The solutions of unknown concentration were diluted until the optical density value obtained was within the linear part of the graph. The values of proteins mentioned represented average of experiments performed in triplicates.

Preparation of Calcium Granules from Serum
Calcium granules were prepared as described earlier for serum pellets [3]. Briefly, the granules shown in this study (labeled as “Calcium Granules”) were prepared by adding sterile solutions of either 0.25 M CaCl2 and/or 0.25 M Na2HPO4 (both at pH 7.4) to FBS (2.5 ml) at final concentrations of either 48 mM CaCl2 (used for Figs. 9G, 10D, 12D, 13D, 14D, and 15G), 24 mM Na2HPO4 (Figs. 9H, 10E, 12E, 13E, 14E, and 15H), or 2 mM of both CaCl2 and Na2HPO4 (Figs. 9I, 10F, 12F, and 14F). Calcium granules were also prepared in HS (2.5 ml) by diluting 2 mM of both CaCl2 and Na2HPO4 (Figs. 13F and 15I). The ion solutions were filtered through both 0.2-µm and 0.1-µm membranes prior to use and were added in a drop-wise manner with vigorous shaking in order to avoid spontaneous precipitation. Treated sera were incubated at room temperature overnight, followed by centrifugation at 16,000× g for 1 hour, and washing steps using HEPES buffer and the same centrifugation steps. The granules were resuspended in 100 µl of HEPES buffer and were used for the various microscopy and spectroscopy analyses.

Electron Microscopy
For scanning electron microscopy (SEM), washed particles, calcium granules, and NB specimens were resuspended in double-distilled water. A small aliquot of the sample was deposited on formvar carbon-coated grids (Electron Microscopy Sciences, Fort Washington, PA, USA). The excess liquid was removed with an absorbent paper and the grids were dried overnight under a laminar flow hood. Prior to observation, the specimens were coated with gold for 90 sec. SEM observations were conducted using a SEM S-5000 field-emission scanning electron microscope (Hitachi Science Systems, Tokyo, Japan).
For transmission electron microscopy (TEM), washed particles and NB samples were deposited on formvar carbon-coated grids and were dried overnight as described above. All TEM observations were performed without staining. Commercially available calcium carbonate (CaCO3, A.C.S. grade reagent, purity 99.6%, Mallinckrodt Baker, Inc., Phillipsburg, NJ, USA), calcium phosphate tribasic (Ca3(PO4)2, Kanto Chemical Co., Tokyo, Japan) and HAP (buffered aqueous suspension, 25% solid, Sigma) were diluted into DMEM or double-distilled water and processed like the other samples as controls.
For thin-sections, washed particles and NB samples were dehydrated with two washes of 100% ethanol. The samples were incubated with Epon 812 resin (Electron Microscopy Sciences) with gentle end-to-end agitation overnight at room temperature. The samples were then centrifuged at 16,000× g for 15 min and incubated at 72°C for 2 days to allow resin polymerization. Thin-sections were prepared using a Leica Ultracut UCT microtome (Leica Microsystems GmbH, Wetzlar, Germany). Thin-sections were deposited on formvar carbon-coated grids. TEM observations and electron diffraction patterns were performed with a JEOL JEM-1230 transmission electron microscope (JEOL, Tokyo, Japan) operated at 120 keV.

Energy-Dispersive X-Ray Spectroscopy
Washed particles, calcium granules, and NB specimens were resuspended in double-distilled water and deposited on formvar carbon-coated grids. The grids were dried overnight in a laminar flow hood. The samples were observed with a SEM S-3000N scanning electron microscope (Hitachi Science Systems) and the energy-dispersive X-ray spectroscopy (EDX) analysis was performed using an EMAX Energy EX-400 EDX device (Horiba, Tokyo, Japan). Each sample was irradiated for 30 sec and data acquisition was performed with the EMAX software (Horiba) in point mode analysis. Three different areas of each sample were analyzed to ensure homogenous readings.

Nanoparticle Sizing by Dynamic Laser Scattering
Sizing of mineral nanoparticles was performed using DLS according to established protocols [222], [223]. Briefly, the particles were prepared by adding calcium and phopshate ions into DMEM (final volume of 1 ml) each at a concentration of 1 mM. Alternatively, particles were also prepared by adding 1 mM of calcium and phopshate ions each into DMEM containing 2 mg/ml of either BSF (Sigma) or HSA (Sigma). The samples were prepared directly in disposable plastic cuvettes and were shaken vigorously prior to reading. Some samples were incubated for various periods of time varying from 5 min to several days before measurement. Measurements consisted of an average of 10 individual measurements taken at an interval of 10 sec. Each measurement was performed in triplicates using the Malvern Zetasizer Nano-Series (Malvern Instruments Ltd., Malvern, Worcestershire, UK). A laser consisting of a helium-neon lamp with a wavelength of 633 nm was used and the measurements were performed at an angle of 90° from the sample.

Fourier-Transformed Infrared Spectroscopy
Washed and dried samples were mixed with potassium bromide (Sigma) to obtain a ratio of 1∶100 (w/w). The samples were compressed with a hand press to form a thin transparent pellicle. FTIR spectra were acquired with a Nicolet 5700 FTIR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a deuterated triglycine sulfate (DTGS) detector. The spectra were obtained at a resolution of 4 cm−1 and at wavelengths ranging between 4,000 cm−1 to 400 cm−1. Each spectrum represented an average of 32 consecutive scans. For comparison, the commercially available controls of CaCO3, Ca3(PO4)2, and HAP described earlier were diluted and washed in either double-distilled water (shown in Fig. 13J, K, L, respectively) or DMEM (data not shown), followed by processing for FTIR analysis. The use of DMEM produced dampening of signals and significant depressions/troughs that could not be explained at this time.

Micro-Raman Spectroscopy
Aliquots of washed particles and NB samples were processed for micro-Raman spectroscopy as described before [2]. Briefly, the samples were resuspended in double-distilled water, deposited on glass slides, and dried overnight. Micro-Raman spectra were obtained using the inVia Raman confocal microscope (Renishaw, Stonehouse, UK) equipped with a charge-coupled device (CCD) detector. A laser beam of 633 nm operated at 17 mW was used as the excitation source. Controls of CaCO3, Ca3(PO4)2, and HAP were diluted and washed in either double-distilled water (shown in Fig. 14J, K, L, respectively) or DMEM (data not shown), followed by processing for micro-Raman analysis. In this case, the use of DMEM produced increased carbonate signals for the Ca3(PO4)2 sample which may have originated from prolonged contact with CO2 from air.

Powder X-Ray Diffraction Spectroscopy
X-ray diffraction spectroscopy was performed as described earlier [2]. Briefly, washed particles and NB specimens were deposited on glass slides and dried overnight. X-ray diffraction was performed using a D5005 X-ray diffractometer (Bruker AXS, Madison, WI, USA) with a source of X-ray consisting of a copper tube operating at 40 kV. X-ray diffraction spectra were searched against the database of the Joint Committee on Powder Diffraction and Standards (JPCDS) in order to identify the chemical formula of the crystalline compound under study.