Enhancing Hydrogen Productivity of Photosynthetic Bacteria from the Formulated Carbon Components of Lignocellulose

To develop an efficient photofermentative process capable of higher rate biohydrogen production using carbon components of lignocellulosic hydrolysate, a desired carbon substrate by mixing xylose with glucose was formulated. Effects of crucial process parameters affecting cellular biochemical reaction of hydrogen by photosynthetic bacteria (PSB), i.e variation in initial concentration of total carbon, glucose content in initial carbon substrate, as well as light intensity were experimental investigated using response surface methodology (RSM) with a Box-Benhnken design (BBD). Hydrogen production rate (HPR) in the maximum value of 30.6 mL h -1 L -1 was attained under conditions of 39 mM initial concentration of total carbon, 59% (mol/mol) glucose content in initial carbon substrate and 12.6 W m -2 light intensity at light wavelength of 590 nm. Synergic effects of metabolizing such a well formulated carbon substrate for sustaining the active microbial synthesis to sufficiently accumulate biomass in bioreactor, as well as stimulating enzyme activity of nitrogenase for the higher rate biohydrogen production were attributed to this carbon substrate can enable PSB to maintain the relatively consistent microenvironment in suitable culture pH condition during the optimized photofermentative process.


Introduction
Hydrogen has been suggested as a promising alternative to fossil fuels for a source of clean energy due to its nature of high energy content (122 kJ g -1 ), recyclability and eco-friendly [1][2] . Methods of producing hydrogen through microbial processes have attracted increased attentions of bioenergy engineers in recent years due to its advantage of less energy consuming as compared to traditional ways such as the thermo-chemical or electro-chemical processes, and potential benefits of utilizing natural resources including biomass, water as well as various organic wastes from agriculture, municipalities or industries [3][4] . It was considered as a striking technology to produce renewable bioenergy for a sustainable industrial development. The common bioprocesses used for biohydrogen production are biophotolysis of water, dark-or photo-fermentation of organic compounds. Biophotolysis of water into hydrogen using microalgae under sunlight is considered as the cleanest approach.
However, oxygen inhibition on hydrogenase activity during the water-splitting reaction reduces its productivity. Dark fermentative conversion of carbonhydrate-rich substances into hydrogen by light-independent process with anaerobic bacteria is found to be a faster process, and can occur spontaneously with no external energy requirement from a thermodynamic point of view due to its changes in negative Gibb ' s free energy [5] . While, organic acids formed by dark fermentative hydrogen production could not be further decomposed to hydrogen any more, which resulted in an incomplete degradation of organic wastes. Photosynthetic bacteria (PSB) could use the captured light energy through the action of bacterial photosynthesis in a type of metabolism called photofermentation to overcome the thermodynamic barrier for further conversion of organic compounds into biohydrogen, thus nearly complete biodegratation could be attained. Moreover, PSB have displayed their versatile cellular metabolic pathways in the ability of utilizing a wide variety of carbon substrates, which making photofermentation an attractive technique for simultaneous biohydrogen generation and organic waste treatment [6][7][8] . However, a slower rate of hydrogen productivity accompanied with microbial synthesize of PSB during photofermentation have been recognized by previous studies as compared to those of anaerobic bacteria in dark fermentative process [9] . Therefore, more investigations were necessary to explore the potential capability in term of a high rate phototrophic hydrogen production.
Investigations on microbial metabolisms of PSB during bioprocess of photofermentative hydrogen production have revealed that there were a series of biochemical reactions in which the organic matter underwent biodegradation, and biohydrogen generation was a result of the protons reduction by the electrons that mainly catalyzed by nitrogenase enzyme [10] . Organic carbon can be decomposed during an anaerobic light-depended tri-carboxylic carbon cycle (TCA cycle) to release electrons. Working in parallel was the their photosynthetic membrane apparatus, which transforms the absorbed light energy into the biochemical energy in form of adenosine triphosphate (ATP). Protons were supplied in part by TCA cycle, and the remaining by the action of ATP-synthase within photosynthetic apparatus as aforementioned. Finally, nitrogenase combined the derived electrons with protons to produce hydrogen [11][12] . Such an enzymatic biochemical reaction of hydrogen generation involved in photofermentation can be expressed as: Thus, specific regulations for plenty supply of reducing power (e -) and ATP to promote nitrogenase in their activity of hydrogen production to achieve an efficient photofermentation were necessary.
Lignocellulosic biomass (LCB) was the most abundantly carbohydrates-rich raw material on earth stored plenty of cellulose and hemicellulose, which containing approximately 70-80% polysaccharide sugars (mainly pentose and hexose) [13] .
Glucose and xylose were the two most common monomeric sugars that could be liberated after their hydrolysis to be considered as abundance carbon substrates for biohydrogen production [14] . Considering PSB exhibited different characteristics in their microbial growth and biohydrogen production from different carbon substrates, investigations on formulating a desired carbon substrate contianing the main carbon components of lignocellulosic hydrolysate, i.e., glucose and xylose to attain an efficient photofermentation for a higher rate biohydrogen production were promising.
From the engineering point of view, it can offer the prospect of the reduction of the overall cost for the further scale-up of the photofermenative hydrogen production [15].
Accordingly, strategies on encourage PSB for efficient conversion of the formulated carbon substrate into biohydrogen via the optimized photofermentation process were needed.
Conventional methods of changing a single factor while setting all of other factors at fixed values to investigate multifactor bioprocess were not only time-consuming but also failed to obtain true optimal operational conditions because possible interactions between these factors were ignored. Thus, statistical experimental design methods were necessary to design experiments for identifying the synergistic effects of process parameters and determining the optimum conditions of photofermentation. As a statistical technique, response surface methodology (RSM) has been successfully applied in the field of biotechnology for designing experiments, building empirical models and evaluating the simultaneous effects of influencing factors [16][17] . Therefore, in this study, three factors three levels of a Box-Behnken design (BBD) using response surface methodology (RSM) was applied to investigate impacts of key operational parameters of photofermentative process affecting nitrogenase-catalyzed biochemical reaction of hydrogen generation, which included initial concentration of total carbon, glucose content in initial carbon substrate and light intensity, along with their interactions for maximizing the process performance of hydrogen production rate (HPR).

Photosynthetic bacteria and experimental procedure
The seeding PSB cells used in this study were isolated from local municipal sewage sludge, and identified as Rhodopseudomonas palustris with steady and high ability for photoheterotrophic hydrogen production [18] . The modified RCVBN medium using glucose (6.3 g L -1 ) and urea (1.26 g L -1 ) as the carbon and nitrogen source was used for maintenance of the strain [19] . For biohydrogen production, the seed was anaerobic cultivated for 48 h at pH 7.0, 30±1 ℃ under 50 W m -2 illumination intensity by tungsten filament lamps to arrive their mid-exponential growth period. Experiments of photofermentative biohydrogen production were performed in batch reactor using conical flasks (Suxiu Biotech Ltd, China) with 1000 mL working volume. Pre-grown cells were inoculated into reactor with a initial cell concentration of 30 mg L -1 and sparged with oxygen-free argon to achieve an anaerobic environment [20][21] .
The schematic diagram of experimental setup was depicted in Fig proved to be the suitable wavelength according to our previous studies [18,22] . The incident intensities of light are considered as those on the outer surface of bioreactor, which was adjusted by changing the distance between bioreactor and LED lamps.

Analytical methods
The volume of biogas generated was measured by water-replace method after nm was used for determing the concentrations of glucose and xylose within the sample supernatant, respectively [23] .

Data analysis
Performance and microbial behaviours of PSB on cell growth and biohydrogen production during photofermentation process were evaluated by the logistic model and modified Gompertz equation as described in equation (2) and (3), respectively [21,22,24] : Thus, hydrogen production rate (HPR) was defined as Where, x is the cell concentration (g L -1 ), x0 is the initial cell concentration (g L -1 ), xmax is the maximum cell concentratin (g L -1 ), kc is the apparent specific growth rate (h -1 ), t is the photofermentation time (h), e denoted constant (2.818282), H(t) denoted cumulative H2 productin (mL Lmedium -1 ), P is H2 production potential (mL Lmedium -1 ), Rm is the maximum H2 production rate (mL Lmedium -1 h -1 ), λ is the duration of lag phase (h), and V is the working volume of the culture (L).

Experimental design and regression analysis
A statistical 3 k Box-Behnken design (BBD) using RSM was applied to optimize photofermentation process for higher rate biohydrogen production, and investigate interactive impacts of three critical process variables, i.e., initial concentration of total carbon (x1, mM), glucose content in initial carbon substrate (x2, %mol/mol) and light intensity (x3, W/m 2 ) due to they can significantly influence carbon source and energy source of photofermentation, thus affect HPR. Experiments were employed by varying initial concentration of total carbon (x1) from 20 to 50 mM, glucose content in initial carbon substrate (x2) from 0 to 100 %mol/mol, and light intensity (x3) from 5.8 to 17.4 W/m 2 , their designed range were determined based on our preliminary batch experiments (shown in Table 1).
The actual values of independent factors (xi) were coded as Xi (i.e. -1, 0, 1) by applying the following equation: Where xi is the actual value of independent factor, x0 is the actual value at the center, Δx is the step change value.
HPR was selected as the response variable. The predicted response was calculated according to a second-order polynomial equation [18,25] : Where HPR is the predicted response variable, β0 is the constant coefficient, i and j are the index number of patterns, k is the number of factor, and βj, βjj and βij are the estimated coefficients from regression of linear, quadratic and interaction effects, respectively. The obtained regression equations were then analyzed using the analysis of variance methodology (ANOVA) to measure the significance and fit of experimental data to the model. Thereafter, the three-dimensional (3D) response surface graphs and two-dimensional (2D) contour plots can be obtained.

Photo-H2 production performance
The experimental design consisted of 15 runs including three replicates at center point was conducted. Each run of the experimental design was performed in triplicates with exception of the three central points, then various HPRs (as tabulated in Table 2) were described as the average value from triplicate experiments with standard deviations. In order to obtain the optimal operational conditions leading to the maximum HPR and reveal the interaction effects among three variables, RSM was employed to get the empirical relationship between the response variable of HPR and the influencing process parameters, and two regression quadratic models for the values of variables in the actual and coded experimental units were obtained as: HPRactual = -6.96+0.981x1+0.22x2+1.88x3-0.00113x1x2+0.0138x1x3+0.0019x2x3 -0.0143x1 2 -0.00172x2 2 -0.1x3 2 Statistical analysis of variance (ANOVA) was used to estimated significance of the fitted BBD model (7) in terms of linear, quadratic and interactions (demonstrated in Table 3). the F-value of 53.7 for this model showed it was significant, proven by the low p-value of 0.0002 which was less than 0.05, indicating a high statistical significance of the model terms. The linear (x1, x2 and x3), quadratic (x1 2 , x2 2 and x3 2 ) and the interactive effects of x1x2 and x1x3 were significant (p <0.05). While, x2x3 was considered not significant by the p-value greater than 0.05, implying the interactive effect between glucose content in initial substrate (x2) and light intensity (x3) has little impact on HPR, whereas it still be retained in Eq. (7) due to it is a hierarchical model. 53% showed that this model can be considered reasonably reproducible because the CV is less than 10%. Consequently, the model can be employed to navigate the design space.

Effects of initial concentration of total carbon and light intensity on HPR
3D surface graph and 2D contour plot indicating HPR with respect to initial concentration of total carbon (x1) and light intensity (x3) at constant glucose content in initial carbon substrate (x2) of 59% mol/mol were illustrated in Fig. 2. As shown from the 3D surface graph, for the studied range of x1 from 20 to 39 mM, and x3 from 5.8 to 12.6 W m -2 , HPR was observed to increase as the two variables increased. While, after reaching a peak value of 30.0 mLh -1 L -1 , HPR decreases even if the two variables continue to increase. The 3D surface graph also revealed that HPR was more sensitive to the change of light intensity from 5.8 to 12.6 W m -2 than that of initial substrate concentration from 20 to 39 mM. Moreover, the 2D contour plot was elliptical with elongated diagonals indicating there is a significant interactive effect between initial concentration of total carbon and light intensity. Experimental results confirmed that there was a close relationship among the cellular metabolisms of organic carbon degradation for electron release, microbial photosynthesis for the conversion of light energy into ATP, and the nitrogenase-catalyzed reaction of biohydrogen generation.
From the perspective of process operation, a higher HPR is likely the result of process optimization for the sufficient providing of both electrons and ATPs by an appropriate combination of initial concentration of total carbon and light intensity to promote the activity of nitrogenase in catalyzing the reaction of biohydrogen generation.

Effects of initial concentration of total carbon and glucose content in initial carbon
substrate on HPR Effects of initial concentration of total carbon (x1) and glucose content in initial carbon substrate (x2) on HPR were depicted in Fig. 3. Significant individual effects of the two variables and their interactions that influence biohydrogen production were demonstrated, and the highest value of HPR was achieved within the range of process operational process paramters investigated. It is also revealed that HPR was more sensitive to change in glucose content in initial carbon substrate, as it exhibited a steeper slope on the axis of x2 compared to the axis of x1. So, type of carbon source is considered to be a significant factor affecting intracellular metabolisms of both cell synthesis and biohydrogen production. The different components of the mixed carbon likely partition to synthesis of new cells for microbial growth and production of hydrogen to different extents due to their different cellular metabolic pathways to be utilized by PSB during the photofermentative process. A mixed carbon source that was well formulated by RSM with a particular combination of initial concentration of total carbon (x1) and glucose content in initial carbon substrate (x2) can likely balance the eletrons allocation between cell growth and biohydrogen production during the optimized operational conditions. Accordingly, both efficient growth of PSB cells to increase level of nitrogenase within bioreactor, i.e. to produce sufficient bomass, and stimulation of the enzyme activity for nitrogenase to catalyze biohydrogen production can be attained, which resulted in the highest HPR.

Optimization analysis and verification test
The optimum operational parameters of photofermentation for maximizing HPR, determined by setting the partial derivatives of Eq(7) to zero with respect to the corresponding variables. Initial concentration of total carbon (x1) at 39 mM, glucose content in initial carbon substrate (x2) at 59% mol/mol and light intensity (x3) of 12.  Experimental investigations demonstrated that the strain of PSB showed their metabolic versatility and capable of consuming both glucose and xylose for microbial growth, as well as biohydrogen production. The RSM with BBD can be applied as an efficient and feasible tool to formulated a carbon source in suitable content and initial total carbon concentation for their sufficient degradation during the optimized photofermentation process under a relative neutral pH culture condition, thus maintaining higher activity of their various intracellular enzymes to facilitate biomass accumulation to attain an appropriate cell concentration of PSB within bioreactor (i.e. to obtain a suitable amount of nitrogenase within biosystem), as well as maintain the specific catalytic activity of nitrogenase by saturating their demand of reducing power (electrons) and ATP energy in driving the cellular hydrogen producing reaction, which resulted in the enhanced performance of biohydrogen production.

Ethical Approval
Not applicable.

Consent to Participate
Informed consent was obtained from all individual participants included in the study.

Consent to Publish
All the authors accept that the copyright of the manuscript will be transferred to the journal upon acceptance for publication.

Authors Contributions
Conceptualization, methodology, data curation, supervision and writing-review and  Experimental setup for photofermentative H2 production using mixed carbon Effects of initial concentration of total carbon and light intensity on HPR at constant glucose content in initial carbon substrate of 59% mol/mol (a) 3D surface graph, and (b) 2D contour plot Effects of initial concentration of total carbon and glucose content in initial carbon substrate on HPR at constant light intensity of 12.6 W/m2 (a) 3D surface graph, and (b) 2D contour plot   Substrate consumption rate at optimum conditions from (a) pure xylose, (b) pure glucose and (c) the formulated carbon substrate