The increasing need for low-carbon society makes hydrogen gas (H2) an optimal energy carrier featuring high energy content (122 kJ/g) and clean combustion product (Akhlaghi and Najafpour-Darzi, 2020). Hydrogen can be produced by a variety of methods (e.g., steam reformation of CH4, coal gasification, electrolysis of water) (Akhlaghi and Najafpour-Darzi, 2020; Valle et al., 2019), but biotechnological approaches provide important additional environmental sustainability of H2-based energy policy since they do not rely upon usage of fossil fuels (Wang and Yin, 2021). Currently available technologies for biological production of H2 suffer from low yield and economic viability (Lepage et al., 2021; Nirmala et al., 2023). The production cost of bio-H2 has recently been estimated at 1.2–4.3 €/Kg H2 while that of H2 obtained by traditional approaches such as natural gas/coal reforming and partial oxidation is in the range 0.8–1.6 €/Kg H2 (Wang and Yin, 2021). However, H2 production from fossil fuels generates massive greenhouse gas emissions (e.g., 9 Kg CO2/Kg of H2 for steam CH4 reforming) (Lepage et al., 2021) and, by 2030, because of carbon emission taxes, is predicted to be more expensive than green processes with lower CO2 emissions such as water electrolysis or biotechnological approaches (Hassan et al., 2024). In this perspective, biological H2 production is generally considered to have a higher global warming potential (i.e., higher CO2 emission) than water electrolysis (Busch et al., 2023; Wilkinson et al., 2023) but lower cost (2.35–10.36 €/Kg H2 for water electrolysis) (Hassan et al., 2024; Lepage et al., 2021; Nazir et al., 2020).

Generation of H2 provides a means for microorganisms to dispose of excess reducing equivalents (generated by fermentative metabolism or anoxic photosynthetic activity) or is obtained as a by-product of N2 fixation by nitrogenase. Bio-H2 production can be obtained through light-driven processes (i.e., direct and indirect biophotolysis, photofermentation) by means of photosynthetic microorganisms (e.g., green microalgae, cyanobacteria, purple non‑sulfur bacteria) or by the so called dark processes, that is anaerobic fermentation of organic compounds by a number of heterotrophic microbes (mainly bacteria) (Arizzi et al., 2021; Cao et al., 2022; Das and Veziroglu, 2008; Mudhoo et al., 2011). Each of these biotechnological alternatives is limited by specific issues. Light-driven processes take advantage of cheap and abundant resources such as sunlight, water and biomass but their productivity is typically low due to limited light conversion efficiency (Suresh et al., 2023). Dark fermentation features much higher H2 productivity, can be used to dispose of organic waste biomass, and is supported by existing bioreactor technology although suffers from low yield (Cao et al., 2022; Mudhoo et al., 2011). Light-driven processes have been extensively reviewed elsewhere (Akhlaghi and Najafpour-Darzi, 2020; Schumann et al., 2023; Suresh et al., 2023). The present review will focus on dark fermentation with special emphasis on metabolic improvement of microbes involved.

Among the microorganisms producing H2 through dark fermentation, facultative anaerobes such as Escherichia coli and Enterobacter sp., have a maximum theoretical yield of 2 mol H2/mol glucose (Wang et al., 2011). In these strains H2 is generated through formate oxidation by formate hydrogen lyase. Obligate anaerobes such as Clostridium spp. have 2-fold higher maximum theoretical H2 yield from glucose fermentation. In these bacteria, electrons from both NADH (derived from glycolysis) and reduced ferredoxin (Fdred, produced by pyruvate ferredoxin oxidoreductase, Pfor) can be used by hydrogenases (H2ases) to synthesize H2. Maximum H2 yield can be obtained when sugars are fermented to acetate (and CO2), while it is lower when more reduced products (e.g., propionate, butyrate, lactate, ethanol) are accumulated (Islam et al., 2015; Ortigueira et al., 2015; Wang et al., 2011, Wang et al., 2021). Advantageously, some strains such as C. perfringens ATCC 13124 (Wang et al., 2014) and C. pasteurianum DSM525 (Zhang et al., 2019) have been reported to produce H2 at yields close to the theoretical maximum. However, observed H2 yields in most Clostridia are far lower (≤ 25%) than the theoretical maximum and widely vary from strain to strain and depending on culture conditions (Wang and Yin, 2021).

In fact, H2 production is significantly affected by a number of culture parameters such as the amount of carbon sources, agitation, H2 partial pressure (PH2), pH, bioreactor operation mode (batch, fed-batch, continuous) and type of bioreactor (e.g., stirred-tank, upflow anaerobic sludge blanket, anaerobic fluidized bed or membrane bioreactor) (Cai et al., 2013; Julkipli et al., 2023; Łukajtis et al., 2018; Son et al., 2021; Wang et al., 2014). For instance, PH2 can significantly affect the thermodynamics of biological H2 production (Schink, 1997; Van Lingen et al., 2016). The ΔrG'm (i.e., the change in Gibbs free energy associated with a metabolic reaction/pathway when all the reactants have a concentration of 1 mM) for the production of H2 by oxidation of glucose to acetate (glucose +4 ADP + 4 Pi ⇆ 2 acetate +2 CO2 + 4 ATP + 2 H2O + 4 H2) can vary from −205.1 to 0.2 kJ/mol for dissolved H2 concentration ranging from 10−9 M to 1 M (Flamholz et al., 2012). Different technologies are available for lowering PH2 in a bioreactor (e.g., stirring the growth medium, sparging the growth medium with an inert gas, removing gas by a vacuum pump, selectively removing H2 by active membranes) but further research is necessary to develop more effective and inexpensive methods (Jayachandran et al., 2022; Kisielewska et al., 2015; Łukajtis et al., 2018; Nemestóthy et al., 2020).

Independently from H2 yield, dark fermentation processes, including those catalyzed by clostridia, result in large accumulation of carbon by-products which need to be disposed. Advantageous approaches to remove dark fermentation by-products include two-stage processes that couple dark-fermentation with, for instance, photo-fermentation, leading to additional H2 production, or anaerobic digestion, thus generating methane (Ramprakash et al., 2022). In addition syntrophic microbial co-cultures have been established which resulted in improved H2 production by overcoming thermodynamic barrier and/or eliminating the by-product inhibition effect (Laxman Pachapur et al., 2015; Li et al., 2021; Lu and Lee, 2015). Advantageously, experimental evidence has been brought that several clostridia such as C. butyricum, C. pasteurianum, and C. bifermentans are electroactive bacteria, that is are able of extracellular electron transfer (Choi et al., 2014; Park et al., 2001). Improved fermentative H2 production has been obtained by co-culturing Clostridium strains with acetate-oxidizing electron-donating Geobacter bacteria through either indirect (i.e., through artificial extracellular electron shuttles) (Zhang et al., 2013a, Zhang et al., 2013b) or direct (through pili-like structures) (Zhang et al., 2023) interspecies electron transfer. However, it is worth remembering that maintaining stable artificial microbial consortia is frequently challenging (especially at the industrial scale) for instance because of the different conditions (e.g. temperature, pH) that may be required for the growth or metabolic activity of the microbial partners (Jiang et al., 2019; Johns et al., 2016; Zhang et al., 2023).

Alternatively, the same goal can be pursued by improving clostridial strains through metabolic engineering as summarized by a number of recent reviews (Goyal et al., 2013; Husaini et al., 2023; Krishnan et al., 2023; Latifi et al., 2019; Ramprakash et al., 2022). Universal systems have been developed which are able to overcome the traditional recalcitrance of Clostridia to gene manipulation (Minton et al., 2016; Wen et al., 2020; Yang et al., 2016). These technologies have enabled systems metabolic engineering of Clostridia leading to improved production of several industrially relevant compounds (Mazzoli et al., 2020) that include a number of biofuels such as ethanol (Hon et al., 2017; Tian et al., 2016), butanol (Re and Mazzoli, 2023), isobutanol (Higashide et al., 2011; Lin et al., 2015), medium chain esters (Lee and Trinh, 2020; Seo et al., 2023). More recently, additional tools for editing clostridial genome based on Clustered Regularly-Interspaced Short Palindromic Repeat (CRISPR)/cas (CRISPR associated) technology (Husaini et al., 2023; Walker et al., 2020; Wilding-Steele et al., 2021) or for fine tuning gene expression by riboswitches (Marcano-Velazquez et al., 2019) or CRISPR interference (Ganguly et al., 2020) have become available.

Metabolic engineering also promises to be a key tool for increasing H2 production in Clostridia. Global rearrangement of the metabolism is likely to expand the panel of substrates and increase maximum theoretical yields. The superior native levels of H2 generation and current availability of efficient and reliable gene tools position Clostridia among the best candidates for industrial production of biohydrogen. The present review will focus on strategies for enhancing H2 production by clostridia through metabolic engineering. This will be preceded by an analysis of the central and redox metabolism and the enzymes directly involved in catalyzing H2 biosynthesis in these bacteria, as described in the next sections.