Trace impurities in sodium phosphate influences the physiological activity of Escherichia coli in M9 minimal medium

In the field of applied microbiology, reproducibility and experimental variability are important factors that influence both basic research as well as process development for industrial applications. Experimental reproducibility and accuracy depend not only on culture conditions such as temperature and aeration but also on raw materials and procedures used for media preparation. The M9 minimal medium is one of the most common synthetic media for culturing Escherichia coli and other bacteria. This synthetic medium can be used to observe and evaluate the physiological activity of microbes under minimal nutritional requirements and determine the limiting factor for the desired phenotype. Although one of the advantages using the M9 medium is that its composition can be modulated, it is difficult to control presence of trace components and impurities from the reagents for preparing this medium. Herein, we showed that trace ingredients present in the reagents used for M9 media preparation affect the bacterial physiological activities (e.g., cell growth, substrate consumption, and byproduct formation). Additionally, we systematically identified the trace ingredient that influenced phenotypic differences. Our results showed that the selection of reagents and accuracy during reagent preparation is important for experimental reproducibility in the field of bio-engineering and systems biology focused on the systematic and continuous development of biomolecular systems (e.g., biorefinery, metabolic engineering, and synthetic biology).

There are several different M9 recipes available in the Cold Spring Harbor Protocols 7 , Helmholtz Zentrum München (https:// www.helmh oltz-muenc hen.de/), SubtiWiki 32 , etc.One of the media components, disodium phosphate (Na 2 HPO 4 ), is available in both anhydrous and hydrated forms, and it is up to the experimenter to decide which reagent to use.Since anhydrous and hydrated disodium phosphate have the same chemical properties once dissolved in water at the same concentration, either recipe theoretically yields the same M9 medium.However, some researchers have shown empirically that microbial growth changes when the reagents and/or water used to make the media are altered 33,34 .
Herein, we compared the physiological activity of E. coli cultured on M9 media prepared using disodium phosphate of different grades and purity.Hydrated and anhydrous disodium phosphate of different grades were purchased from different sources to prepare the M9 media.Bacterial growth was evaluated on these media using standard techniques.Our results suggest that the purity of reagents used for preparing the M9 media might influence the growth of E. coli.
Except for disodium phosphate, all other reagents used in four different types of M9 media (Type A, B, C, and D) were the same.These media were then used to culture E. coli strain BW25113 (Fig. 1).Although the same seed culture broth (overnight LB culture) was passaged with the same passage ratio (0.01% v/v) into each M9 medium, different cell growth profiles were observed in different M9 media (Fig. 1a, b).The time taken by the bacterial cells to enter the logarithmic growth phase (log phase) was not significantly different (approximately 750 min) among all M9 media, although the cell growth rate and final cell density were different (Fig. 1a).The final E. coli density in M9 type C (disodium phosphate heptahydrate) was lowest, and 51% lower than that observed in the M9 type A media.Furthermore, the specific growth rate (μ) of bacterial cells was also different in different M9 media (Fig. 1b).The maximum specific growth rate (μ max ) in M9 type C was significantly lower than that in other M9 media (Fig. 1c).Glucose consumption and metabolic product (acetate) formation were lowest in M9 type C (Fig. 1d, e).Our HPLC analysis showed no significant difference in phosphate concentration of different M9 media used in this investigation (Fig. 1f).
Altogether, our results suggest that differences in cell growth and other physiological activities of E. coli observed in different M9 media are not dependent on the phosphate concentration, but are influenced by differences in disodium phosphate.The reproducibility of this experiment was confirmed through three different experiments (Fig. S1, n = 24 for each experiment).After 24 h of cultivation, approximately 6% of the total media was evaporated (Fig. S2a), and the evaporation volume varied among the 96 wells (RSD = 8.2%) (Fig. S2b-d).Evaporation of the media was especially considerable at the outer periphery of the plate.To investigate the influence of unequal media evaporation, the sample positions were replaced and the cell growth of E. coli on each M9 medium was monitored (Type A-D) (Fig. S3).Consequently, the growth efficiency on each medium was reproduced regardless of the sample position.

Differences in disodium phosphate reagents affect the growth of multiple E. coli substrains under different environmental conditions
Next, we investigated the effects of environmental conditions, such as temperature, aeration, and carbon/nitrogen ratio (C/N ratio) on E. coli growth in different M9 media.Our results showed that altering the culture temperature (30 °C, 37 °C, and 42 °C) affects the bacterial cell growth in different M9 media (Fig. S4).E. coli growth was maximum at the optimum temperature (37 °C) (Fig. S4a-c).Cell growth at 30 °C was drastically suppressed (Fig. S4a), although the final cell density and the μ max were the same in all four media (M9 types A-D) at this temperature (Fig. S4g).Similar results were observed at 42 °C (Fig. S4i).Therefore, the poor growth of E. coli cells in M9 type C was not temperature-or C/N ratio-dependent (Fig. S5), and aeration was regulated by shaking speed (Fig. S6).This was also observed during test tube cultivation (Fig. S7).Increasing the shaking speed and culturing in a test tube improved the oxygen supply in the media and promoted growth in all M9 media types, but poor growth was still observed in medium C. Similar results were observed with other E. coli substrains, including K-12, B, and BL21(DE3) strains (Fig. S8).These results confirmed that differences in cell growth observed in different M9 media were due to variations in the disodium phosphate reagent independently of the E. coli strain and culture conditions.However, the presence of other impurities in the disodium phosphate reagent cannot be ruled out which could have affected the growth of E. coli.
Next, we evaluated E. coli growth on M9 media containing disodium phosphate of different grades.Disodium phosphate reagents were additionally obtained from FUJIFILM WAKO Chemicals, Corp. (Table 1).For Na 2 HPO 4 and Na 2 HPO 4 • 12H 2 O, both WAKO 1st grade (A (1G) and B (1G)) and GR grade (A (GR) and B (GR)) were obtained.Since Na 2 HPO 4 • 2H 2 O was available only with E. Merck KG, we used this reagent (B).For Na 2 HPO 4 • 7H 2 O, we obtained one ACS Grade reagent (C (ACS)) from MP Biomedicals, Inc.As shown in Fig. 2a-d, the growth of E. coli was significantly different in M9 media prepared using different disodium phosphate reagents.Cell growth in M9 type C (ACS) (prepared using Na 2 HPO 4 • 7H 2 O from MP Biomedicals, Inc.) improved to the same extent as other M9 types (A, B, and D) (Fig. 2).In addition, variations in E. coli growth were observed among different M9 media prepared using different grade reagents from the same manufacturer (Fig. 2; A (1G) vs.A (GR)).The amount and purity of the water of hydration in all reagents of different grades were taken into consideration before preparing the media, therefore, the final concentration of phosphoric acid and sodium was constant.This suggests that any differences in these M9 media components must be due to the presence of impurities.

Identification of impurities that affect bacterial cell growth
To investigate the impurities present in Na 2 HPO 4 reagents that affect E. coli cell growth, we performed the elemental analysis of Na 2 HPO 4 reagents using inductively coupled plasma mass spectrometry (ICP-MS).All Na 2 HPO 4 reagents used in this study were analysed, and five trace impurities were detected (Ba, Cr, Fe, Ni, and Zr).As shown in Fig. 2e, the total amount of impurities varied among the reagents, and a positive correlation was found between the total amount of impurities and the maximum specific growth rate (μ max ) (Fig. S9a).A positive correlation was also observed between the amount of iron present and the maximum specific growth rate (μ max ) (Fig. S9d).Iron was detected in all the Na 2 HPO 4 reagents (Table S9).The amounts of other impurities (Ba, Cr, Ni, and Zr) were not significantly correlated with growth rate (Fig. S9b, c, e, f).M9 type C contained the lowest amount of total impurities among the media ([total metals] = 4.1 μg/L), and the cell growth rate was also lowest in this medium (μ max = 0.17 h -1 ) (Fig. 2e).M9 type C contained 2.22 ± 0.09 μg/L of iron, 0.09 ± 0.08 μg/L of barium, and 0.09 ± 0.08 μg/L of zirconium.The iron content was comparable with the media, which supported increased cell growth (M9 type B and D).Of the elements not detected in M9 type C (Cr and Ni), Ni was detected in all the other media.Additionally, Ni was not found in the A5 (Merck KGaA) used in this study.
To investigate the effect of Ni on bacterial cell growth, we then cultured E. coli in M9 Type C with various concentrations of nickel (Fig. 3).The final E. coli cell density after 24 h of shaking culture was 37-43% increased in the presence of nickel (NiSO 4 ) in a specific concentration range (0.06-3.7 μg/L) (Fig. 3a).The maximum specific cell growth (μ max ) was also increased (42-69%) under the same range of NiSO 4 (Fig. 3b).In contrast, excess nickel (> 10 μg/L) significantly suppressed cell growth.These results confirm that nickel in a specific concentration range promotes E. coli growth.

[NiFe] hydrogenases enhance the cell growth on M9 medium
Ni is known to play an important role in the biology of some eubacteria.In E. coli, [NiFe] hydrogenases (Hyd) are oxidative enzymes that catalyse the reversible oxidation of molecular hydrogen to protons and electrons 35,36 .Hydrogenases play an important role in energy metabolism in several microorganisms 37 .Three types of [NiFe] hydrogenases have been identified in E. coli: Hyd-1, Hyd-2, and Hyd-3 (Fig. 4a) 38 .Hyd-1 consists of three subunits (HyaA, HyaB, and HyaC) and is an oxygen-tolerant H 2 oxidiser.Hyd-2 is an O 2 -sensitive [NiFe] hydrogenase involved in anaerobic respiration and consists of subunits HybA, HybB, HybC, and HybO.Hyd-3 is a part of the formate hydrogenlyase complex (FHL) that is produced under fermentative conditions and is required for the disproportionation of formate to CO 2 and H 2 .Several studies have reported that each Hyd is involved in the activation of F O F 1 -ATPase under different non-neutral pH conditions [39][40][41] .Moreover, Hyd enzymes may function as intracellular pH monitors through F O F 1 -ATPase which interacts with the FHL complex composed of FDH and Hyd to regulate the cytoplasmic pH 40 .According to these reports, nickel-dependent hydrogenases can support the growth of E. coli under microaerobic and acidic conditions by facilitating the homeostasis of intracellular energy/redox and pH.To investigate the relationship between these [NiFe] hydrogenases and the enhanced growth of E. coli in the presence of nickel sulfate, we prepared E. coli strains by deleting genes responsible for the activity of [NiFe] hydrogenases (ΔhyaB, hybC, ΔhycE, and ΔhyaB, hybC, hycE).When these strains were cultured in M9 type C with 0.05 µg/L of NiSO 4 , no significant enhancement of cell growth was observed (Figs.4b, c and S10).These results clearly demonstrate that [NiFe] hydrogenases support the growth of E. coli on minimal media containing trace amounts of nickel.It is suggested that these enzymes regulate the physiological activity of E. coli by balancing the intracellular redox and energy metabolism in the presence of appropriate amounts of nickel and iron.
The selection of reagents and accuracy in reagent preparation is important for the reproducibility of systematic experiments in the field of bio-engineering, which requires systematic and continuous development of biosystems (e.g., biorefinery, metabolic engineering, and synthetic biology).Moreover, selection and preparation of culture media is important to systematically evaluate the effects of gene deletion or upregulation using common  www.nature.com/scientificreports/bioresource platforms such as bacterial strain libraries [42][43][44] .This study demonstrated that the physiological activity of E. coli is significantly affected by the trace impurities present in reagents, used in growth media.Trace impurities were supposed to affect enzyme activities of [NiFe] hydrogenase in E. coli.Further validation experiments, such as metabolic 13 C metabolic flux analysis and enzyme activity assay for each [NiFe] hydrogenase, might be helpful to quantify the contribution of these impurity-induced changes of enzymatic activity to bacterial growth and energy metabolism.

Conclusion
Herein, we demonstrated that bacterial physiological activity is affected by impurities present in the reagents used for medium preparation exemplified with Na 2 HPO 4 , even when we intended to prepare the same medium.Furthermore, Ni-dependent hydrogenases assist the growth of E. coli on M9 minimal medium containing glucose as the sole carbon source and a very small amount of nickel as an impurity.Slight differences in medium preparation (purity, grade, manufacturer, and manufacturing lot of reagents) should be considered when performing microbiological experiments to ensure consistency and reproducibility.
The glycerol stocks of each strain were inoculated into 5 mL of LB medium in 30-mL test tubes as the seed culture, which was then incubated at 37 °C with orbital shaking (250 rpm) overnight.The overnight grown seed culture was passaged into fresh M9 medium (0.01% v/v).Afterwards, 200 μL of this solution was dispensed into a 96-well plate, which was covered with a clear plastic lid without sealing to ensure adequate air supply.The covered plate was then incubated with orbital shaking in a multi-mode plate reader Synergy HTX (BioTek Instruments, Winooski, VT).To prevent fogging of the plate lid due to medium evaporation, cultivation was carried out in the condensation control mode of the plate reader.The cultivation temperature and orbital shaking conditions were set to 37°°C and 269 rpm (orbital rotation, φ = 6 mm) unless otherwise specified.The initial inoculum amount for the 96-well main culture was set so that the final cell density (OD 600 ) did not exceed the detection limit of the plate reader (OD 600 ≦ 0.99) after 24 h of cultivation (Fig. S11).Optical density at λ = 600 nm (OD 600 ) was automatically measured every 10 min for 24 h.

Quantification of phosphate, glucose, and organic acids
The extracellular glucose concentration was quantified using high-performance liquid chromatography (HPLC) on LC-20AD, SIL-20ACHT, CTO-20AC, and RID-10A instruments (Shimadzu Corp., Kyoto, Japan) equipped with a ligand-exchange chromatography column (ULTRON AF-HILIC-CD, 4.6 mm × 250 or 150 mm, 5 mm, Shimadzu Corp.).The samples were separated using an 85% acetonitrile aqueous solution.The column temperature was set to 60 °C and the flow rate was adjusted to 0.8 mL/min.Five microliter of each sample was injected.
Extracellular organic acids and phosphate were quantified using HPLC on LC-30AD, SIL-30AC, CTO-20AC, and CDD-10AVP instruments (Shimadzu Corp.) equipped with three tandem ion-exclusion chromatography columns (Shim-pack Fast-OA, 7.8 × 100 mm, 5 mm, Shimadzu Corp.) and a guard column (Shim-pack Fast-OA (G), 4.0 × 10 mm, 5 mm, Shimadzu Corp.).Sample separation was achieved using a 5.0 mM p-toluenesulfonate mobile phase.The column temperature was set at 40 °C and the flow rate was adjusted to 0.8 mL/min.CDD-10AVP was used as the detector for the post-column pH-buffered electrical conductivity.Post-column pH buffering was attained using a pH-buffering solution containing 5.0 mM p-toluenesulfonate, 20 mmol/L Bis-Tris, and 0.1 mmol/L EDTA.Five microliter of each sample was injected.

Statistical analysis and reproducibility of results
All the statistical analyses were performed using Microsoft Excel 2016.Data are shown as individual data points with mean ± standard deviation.Sample size has been mentioned in each figure and/or figure legend.All experiments presented in the manuscript were independently repeated at least two times with consistent results.The statistical significance of each experiment was determined by the Welch's two-tailed t-test and a p-value less than 0.05 was considered statistically significant.

Figure 2 .
Figure 2. Effects of reagent grade of disodium phosphate for M9 medium preparation on bacterial cell growth.Bacterial cell growth on different M9 media are shown in (a)-(d).(e) The maximum specific growth rate (left panel) and amount of impurities (quantified by ICP-MS analysis; right panel) are shown with the Na 2 HPO 4 reagents' information.The amount of impurities are shown in stacked bars; Ba (grey), Cr (yellow), Fe (orange), Ni (blue), and Zr (green).Error bars indicate the standard deviation (n = 24).

*Figure 3 .
Figure 3. Effects of nickel supplementation in M9 medium type C on bacterial cell growth.(a) Maximum cell density after 24 h of shaken culture.(b) The maximum specific cell growth (μ max ).Closed circles represent results at specific nickel concentrations.Green lines indicate results without nickel supplementation ([Ni] = 0 μg/L), and green dashed lines indicate the range of standard deviation.Error bars indicate the standard deviation.*p < 0.05 (Welch's t-test), n = 8.

Figure 4 .
Figure 4. Deletion of genes encoding Ni-dependent enzymes counteracts the enhancement of cell growth by Ni supplementation in M9 type C. (a) Illustration showing E. coli [NiFe] hydrogenases.(b) Maximum cell density and (c) maximum specific cell growth rate (μ max ) of various E. coli strains in 96-well plates cultured using M9 type C with (blue) or without (green) 0.5 µg/L NiSO 4 .Error bars indicate the standard deviation.*p < 0.05 (Welch's t-test), n = 8.Figures are original, created using Microsoft PowerPoint 2019 for Windows version 1808 (https:// www.micro soft.com/).
-S8 summarizes all the disodium phosphate and other reagents (KH 2 PO 4 , NaCl, and NH 4 Cl) used for each 5 × salts solution (Type A ~ D).The Each 5 × salts solution was used for preparation of M9 media (Type A ~ D) supplemented with 4 g/L glucose, 10 ppm thiamine hydrochloride, and 10 ppm trace metal solution A5.

Table 1 .
Disodium phosphate reagents used in this study.