Understanding the NaCl-dependent behavior of hydrogen production of a marine bacterium, Vibrio tritonius

Bacterial strain and pre-cultivation

Pure culture of V. tritonius strain AM2 (JCM 16456^T) was maintained on a marine agar slant at 15 °C and sub-cultured 3 times a month. The cells were then pre-cultured aerobically in marine broth to reach a cell density of 0.8 to 1.0 ± 0.1 OD₆₂₀ measured by a microplate reader (Infinite F200, Tecan, Switzerland), which corresponds to 1.4 to 1.7 ± 0.1 OD₆₂₀ using a UV/Visible spectrophotometer (Ultrospec 2000; Pharmacia Biotech, Piscataway, NJ, USA), with shaking at 130 rpm (MMS-48GR; EYELA, Tokyo, Japan) at 30 °C. The pre-culture medium was a marine broth composed of 0.5% (w/v) polypeptone, 0.1% (w/v) yeast extract, and 75% (v/v) artificial seawater (ASW) composed of 0.7 g/L KCl, 5.3 g/L MgSO₄ ⋅ 7H₂O, 10.8 g/L MgCl₂ ⋅ 6H₂O, 1.3 g/L CaSO₄ ⋅ 2H₂O, and 30 g/L NaCl at pH 7.5.

In a separate experiment, Escherichia coli strain JCM 1649 was used as a reference. Similar to the experiment using V. tritonius AM2, the pure culture E. coli was stored on LB agar slant (1% (w/v) tryptone, 0.5% (w/v) yeast extract, 0.5% (w/v) NaCl, 1.5% (w/v) agar) at 15 °C and sub-cultured three times a month. Pre-culture was prepared using LB broth following a similar method to that used for V. tritonius.

Experimental designs

Batch culture was performed using V. tritonius strain AM2 inoculated in a basal marine media made of 0.5% (w/v) polypeptone, 0.1% (w/v) yeast extract, 100 mM MES (Dojindo, Kumamoto, Japan), and 75% (v/v) basal salt solution composed of 0.7 g/L KCl, 5.3 g/L MgSO₄ ⋅ 7H₂O, 10.8 g/L MgCl₂ ⋅ 6H₂O, and 1.3 g/L CaSO₄ ⋅ 2H₂O. To study the effects of Na⁺ concentration, the amount of NaCl was adjusted accordingly to constitute 0.5%, 1.0%, 2.25%, 3.0% and 5.0% (w/v) of the culture media. One mL of the precultured inoculum was then added to 95 mL marine broth containing 5% (w/v) mannitol in a glass bottle reactor. Batch culture was performed at 37 °C and pH 6.0 with gentle agitation using a magnetic stirrer (speed at 4.5, RO 10 power IKAMAG; IKA, Staufen, Germany). The pH was maintained using a pH controller (DT-1023P; ABLE, Tokyo, Japan) attached to an autoclavable electrode (FermProbe pH electrodes; Broadly-James Corp., Irvine, CA, USA) with the addition of 5 N NaOH. The experiment was repeated in triplicate and samplings of culture and gas were done at 12 h-intervals.

In batch culture experiments for E. coli strain JCM 1649, the bacteria was cultured on three types of media namely LB broth (1% (w/v) tryptone, 0.5% (w/v) yeast extract, 0.5% (w/v) NaCl), LB broth prepared using 60% (w/v) artificial seawater (with total NaCl equal to 2.30% (w/v) and denoted as LB broth with 60% ASW) and marine broth (containing 2.25% (w/v) NaCl). 100 mM MES (Dojindo, Kumamoto, Japan) was supplemented in all media. The strain JCM 1649 was precultured under the same conditions as the strain AM2 culture, which gave a cell density of 1.3 OD₆₂₀ after 24 h incubation. In these batch culture experiments, broth containing 5% (w/v) glucose was used for strain JCM 1649, due to its unsuitability in utilizing mannitol. Other culture conditions were set the same as those for the strain AM2. For both experiments using V. tritonius strain AM2 and E. coli strain JCM 1649, neither external reducing agents nor sparging with inert gas is required because both bacteria are capable of self-installing and maintaining anaerobiosis for hydrogen production using sugars as carbohydrate (Hassan & Morsy, 2015; Matsumura et al., 2015).

Assays

The biogas produced was collected in an aluminum-bag (GL Sciences Inc, Tokyo, Japan) and measured using the water displacement method. Hydrogen content in the biogas was determined using gas chromatography (GC-2014; Shimadzu, Kyoto, Japan) equipped with thermal conductivity detector (TCD) and ShinCarbon ST packed column (Shinwa Chemical Industries Ltd., Kyoto, Japan) with Argon (Ar) as the carrier gas. The operational temperatures of the injection port, oven and detector were 60 °C, 40 °C and 60 °C respectively. Extracellular ethanol was measured using the same gas chromatography equipment together with a flame ionization detector (FID) and HP-Blood Alcohol Capillary column (Agilent, Santa Clara, CA, USA). Helium (He) was used as the carrier gas. The temperatures of the column, oven and detector were 60 °C, 40 °C and 60 °C, respectively.

Organic acids produced, such as formate, acetate, lactate and succinate were measured using high performance liquid chromatography (HPLC) with a conductivity detector and tandem connected to two Shim-pack SCR-102H columns (Shimadzu, Kyoto, Japan). 5 mmol/L p-toluenesulfonic acid was used as the mobile phase with the flow rate at 0.8 mL/min. Mannitol concentration was measured by HPLC with a reflective index detector (RID) and TSKgel Amide-80 column (TOSOH, Tokyo, Japan). 75% acetonitrile was used as the mobile phase with the flow rate at 1.0 mL/min. The optical density (OD) of cell growth was measured using a microplate reader (Infinite F200; Tecan, Männedorf, Switzerland) at a wavelength of 620 nm.

Data analyses

Molar yields of the fermentative products including hydrogen were analyzed using one-way analysis of variance (ANOVA) with Duncan’s multiple range tests. All statistical analyses were performed using the Statistical Analysis System (SAS) University Edition. Probabilities of less than 5% (P < 0.05) were considered statistically significant.

The progress of cumulative hydrogen production, H (mL), in batch tests can also be described using the following modified Gompertz model: (1)${\displaystyle f\left(X\right)=\alpha .Xmax.exp\left(-exp\left[\frac{\alpha .\mu max.e}{\alpha .Xmax}\left(\lambda -t\right)+1\right]\right).}$

Replacing α.Xmax with P as hydrogen production potential (mL), and α.μmax by R m, as maximum hydrogen production rate (mL h⁻¹), Eq. (1) can be expressed by Eq. (2): (2)${\displaystyle H=P.exp\left(-exp\left[\frac{Rm.e}{P}\left(\lambda -t\right)+1\right]\right).}$ Similarly, the substrate consumption could also be expressed using the modified Gompertz model (Mu, Yu & Wang, 2007): (3)${\displaystyle S_0-S=S_{max}.exp\left(-exp\left[\frac{Rms.e}{S_{max}}\left(\lambda -t\right)+1\right]\right).}$ With S ₀ – S as the amount of substrate consumed in g/L; S_max as the maximum substrate consumed and Rms as the maximum rate of substrate consumed (in g/L/h).

λ is the lag time (h) and e is equal to 2.718. All parameters in Eq. (2) were non-linearly estimated by converting the sum square of errors (SSE) between experimental data and estimation from the models to a minimum value. This estimation was carried out using the ‘Solver’ function (Euler integration technique) in Microsoft Excel 2010 (Kemmer & Keller, 2010). The significance of the estimated parameters was tested using analysis of variance (ANOVA) as mentioned above (Chen et al., 2006).

Hydrogen production rate, r (mL h⁻¹) of AM2 grown at various salt levels was calculated following Wang & Wan (2008) in which: (4)${\displaystyle r=\frac{P}{\lambda +\left(\frac{P}{Rm}\right)}.}$

The hydrogen production rate was also denoted as HPR.

Response of the hydrogen production rate to different NaCl concentrations was calculated using a modified Han–Levenspiel model as described in Van Niel, Claassen & Stams (2003) and Zheng & Yu (2005) by taking on the logarithms; (5)${\displaystyle ln\left(r\right)=n\mathit{ln}\left(1-\frac{C}{C\text{crit}}\right)+ln\left(rmax\right)}$ where r is the hydrogen production rate (mL/h); r max is the maximum hydrogen production rate; C is the added Na⁺ concentration (g/L); C crit is the critical Na⁺ concentration at which H₂ production ceases (g/L) and n is the degree of inhibition.

In order to describe the microbial growth in a batch culture, the Gompertz equation was again modified following Zwietering et al. (1990) and Begot et al. (1996): (6)${\displaystyle ln\left(\frac{N}{N_0}\right)=A.exp\left(-exp\left[\frac{\mu .e}{A}\left(\lambda -t\right)+1\right]\right)}$ where A is the logarithmic increase of bacterial population; λ is the lag time; μ is the maximal growth rate; and t is the time.

Cumulative H₂ production, its kinetics and the growth of Vibrio tritonius at various salt levels

The cumulative hydrogen production at different NaCl concentrations over time was fitted with Eq. (2) and is shown in Fig. 1. The experiment confirmed the significant effects of various levels of NaCl on cumulative H₂ production using AM2. Substrate fermentation took place for 72 h and at the end of the process, the culture in 0.5 and 1.0% (w/v) NaCl produced the highest amounts of H₂, i.e., 7,831 ± 295 and 8,561 ± 602 mL/L reactor volume on average, respectively. On the other hand, cells grown in media supplemented with 2.25% (w/v) NaCl produced slightly less hydrogen at 6,693 ± 443 mL/L reactor volume followed by 3.0% and 5.0% (w/v) NaCl at 1,808 ± 79 and 3 ± 2 mL/L reactor volume, respectively. Table 1 shows the P as hydrogen production potential (mL), R m, as maximum hydrogen production rate (mL h⁻¹) and λ (h) of the batch system in response to various salt levels. The results imply the strain AM2 is capable of hydrogen production under a wide range of NaCl concentrations, i.e., 0.5 to 3.0% (w/v). However, the amount of hydrogen produced dropped as NaCl concentrations increased. Maximum H₂ production potential, P, was observed when 0.5 and 1.0% (w/v) NaCl was supplemented in the media followed by the addition of 2.25% (w/v) NaCl and showed further decreases at higher NaCl concentrations in the media.

The efficiency of fermentative biohydrogen is strongly connected to the presence of active cells (Mu, Wang & Yu, 2006). Different NaCl concentrations greatly affected the growth of AM2. Cell growth significantly decreased beyond 2.25% (w/v) NaCl indicating the presence of some degree of inhibition with the addition of 3.0% and 5.0% (w/v) NaCl in the media. Statistical analysis of the growth at the end of fermentation (72 h) for 0.5 to 2.25% (w/v) NaCl supplemented media reveals that they were not significantly different (P < 0.05) to one another. We also calculate the maximum number of organisms, X max, the maximum growth rate, µ_max [ln(N/N₀) h⁻¹] and the lag time, λ (h) of the batch system in response to various salt levels. The maximum specific growth rate, µ_max, was insignificant at 0.5 to 2.25% (w/v) with average µ_max = 0.8 ln(N/N₀)/h but inhibited at 3.0 and 5.0% (w/v) with an average of µ_max = 0.5 ln(N/N₀)/h. V. tritonius showed a classical growth trend under different salt concentrations and the average lag time, λ was at 2.3 h. The cells enter a stationary growth phase after between approximately 14 to 23 h of fermentation and then plateaued until the end of the fermentation period. Table 2 summarizes the kinetics of bacterial growth at different NaCl concentrations as calculated using Eq. (6). The kinetics of bacterial growth and cumulative H₂ produced reveals that hydrogen production started during the exponential phase in which cell growth and H₂ formation may occur simultaneously. Later, hydrogen evolution continues during the stationary phases of bacterial growth and most of the hydrogen was produced during this period.

Specific H₂ production rate and yield

The hydrogen production rate (HPR), r for each NaCl concentration tested was estimated using Eq. (4). The r values of 0.5 and 1.0% NaCl were indistinguishable (Table 1), 2.25 and 3.0% at 140 and 40 mL/h, respectively and 5.0% showed negative value indicating that inhibition had occurred. The logarithm of hydrogen production rate was plotted against the added NaCl concentrations using Eq. (5) and is depicted in Fig. 2. The curve was fitted using a non-competitive inhibition model (Eq. (5)), (r ² = 0.99) at n = 2 to demonstrate the relationship of the hydrogen production rate to different NaCl concentrations using a modified Han–Levenspiel model. The model demonstrates that the maximum hydrogen production rate was 374 mL H₂/h. Figure 2 shows that H₂ production by AM2 is promoted with the addition of 5 to 10.2 g/L (= 0.5 to 1.02% (w/v)) NaCl to the fermentation media. H₂ production is inhibited under higher NaCl concentrations and is critically reduced, C_crit at 46 g/L (=4.6% (w/v)) NaCl. Media containing 22.5 and 30 g/L NaCl resulted in 15 and 78% reduction of the original amount of H₂ produced in 5 g/L NaCl containing media, respectively. The molar yields of hydrogen produced at the end of fermentation (72 h) were compared and presented as mol H₂ per mol mannitol. Statistical analysis by Duncan’s multiple range test revealed that the H₂ yield by cells grown in 5, 10 and 22.5 g/L NaCl supplemented media were not significantly different to each other. The yields were greatly reduced as the NaCl concentration increases.

Substrate consumption and effects of other fermentative products under different salt concentrations

To determine the effects of various salt levels on substrate consumption, substrate conversion efficiency (SCE) were estimated by dividing the amount of mannitol consumed at a particular time by the amount of initial mannitol (expressed as a percentage). Figure 3 shows the effects of NaCl on mannitol consumption based on the kinetic simulation which was calculated following the method described in Mu, Yu & Wang (2007). At 72 h fermentation, almost all mannitol (95.6 and 98.4%) was consumed in the 0.5 and 1.0% (w/v) NaCl media. Consumption was reduced as the NaCl concentration increases with the highest NaCl concentration (5.0% (w/v)) showing a fluctuating pattern in its substrate consumption.

Figure 4 depicts the molar formate concentration throughout the fermentation. For all salt levels tested, formate was accumulated in the first 24 h of fermentation and was later consumed until the end of experiment. Formate accumulation in 5.0% NaCl was significantly delayed and most probably had caused low hydrogen production observed previously. The amount of formate in the first 24 h of fermentation did not show any differences (P > 0.05) among 0.5 to 3.0% NaCl tested with the highest, 2.25% NaCl had achieved 42.3 mM formate. At the end, the average of 10.7 mM formate was left unconsumed/imported in 0.5 to 2.25% NaCl containing media. For both 3.0 and 5.0% NaCl containing media, 20.8 mM formate remained.

Table 3 depicts the molar yields of fermentative products by AM2 in media containing 0.5 to 5.0% NaCl. The ANOVA of raw data shows acetate, lactate and ethanol production was insignificant over the range of NaCl concentrations tested. In relation to cell growth and hydrogen yield, the molar yields of succinate produced were also not significantly different from one another in 0.5, 1.0 and 2.25% NaCl containing media; however, succinate production was significantly reduced up to 3.0% NaCl.

Fermentative hydrogen production behavior of Escherichia coli strain JCM 1649 affected by NaCl and nutrition level of media

The inverted V-shaped pattern in formate production was observed in E. coli strain JCM 1649. Maximum formate production was observed after 6 h-cultivation in normal LB broth which contains 0.5% NaCl and LB broth supplemented with 60% artificial water corresponding to 2.3% NaCl. The formate produced was recorded at 12.2 mM/OD₆₂₀ in 0.5% NaCl LB broth and 11.0 mM/OD₆₂₀ in LB broth with 60% ASW (Figs. 5A–5B). In both cultures, most formate in the supernatant then decreased until 72 h fermentation at which time all formate became undetectable. On the other hand, although the maximum formate concentration in the culture using marine broth reached 26.7 mM/OD₆₂₀ after 24 h-cultivation, the formate were barely consumed by the cells resulting in much lower hydrogen production than those in 0.5% NaCl-LB and LB broth with 60% ASW (Fig. 5C). Note that, hydrogen produced from 0.5% NaCl-LB broth, LB broth with 60% ASW and marine broth were 2,056, 422 and 91 mL/L reactor volume, respectively (Figs. 5A–5C). The rate of sugar consumption after 96 h-cultivation in the 0.5% NaCl-LB broth, LB broth with 60% ASW, and in the marine broth were 93.2%, 93.6% and 46.0%, respectively. Growth was maintained at 1.5, 1.2 and 0.9 OD₆₂₀ after 12, 14 and 48 h-cultivation in the LB broth, LB broth with 60% ASW, and in the marine broth.