Introduction

To transition the global economy towards sustainable biobased production, a range of organisms with different properties will be required to fulfil a range of bioproduction niches. Clostridium spp. were used widely during the early 20th century for the industrial production of acetone, ethanol, and butanol, but bioproduction was later replaced in favour of cheaper production from fossil carbon [1]. Recently, Clostridium spp. have been undergoing a renaissance, building on historic examples that were proven feasible at scale with the application of modern bioprocess engineering and strain engineering.

As well as their natural production of useful products, the ability of Clostridium spp. to utilize a range of carbon sources is of interest, including industrial waste products such as lignocellulose, CO, and CO2. These feedstocks do not compete with food production and are often available at low or potentially even negative cost (in the case of wastes with a disposal cost) making them ideal for the production of low-value, high-volume chemicals such as solvents [2]. LanzaTech and others are developing Clostridium-based fermentation technologies that can be retrofitted to high carbon-emitting industries, such as steel works, and also developing the use of gasification of biomass as a feedstock for fermentation [3]. The prospects and challenges of using Clostridium spp. at industrial scale are well reviewed by Vees et al. [2].

As well as a range of organisms, the bioeconomy also requires the ability to build, test, and optimize metabolic pathways in these organisms. Clostridium spp. were historically limited by a lack of tools for strain engineering, but are now relatively well served by tools including modular shuttle vectors [4], genetic parts such as promoters and selection markers (reviewed by Joseph et al. [5]), directed genomic modification methods such as the bacterial Group II intron-based ClosTron system [6, 7], integration using allele-coupled exchange (ACE) asymmetric homologous recombination [8] and examples of CRISPR in species including Clostridium beijerinckii [9], Clostridium cellulolyticum [10], and Clostridium ljungdahlii [11]. These foundational genetic tools for Clostridium spp. enable work towards implementation of more advanced approaches, such as highly parallel combinatorial assembly.

Combinatorial DNA design and construction is an efficient and pragmatic way to achieve high-performing metabolic pathway designs quickly, leading to recombinant strains which produce desired products effectively, without prior knowledge of the optimal expression levels of individual coding sequences (CDSs). This is achieved by building libraries of metabolic pathway-encoding construct variants in which enzyme CDSs are placed under the control of a varied set of expression control parts [in bacteria mainly promoters, ribosome binding sites (RBSs) and terminators] to sample different designs in a ‘design space’. Golden Gate type multi-part DNA assembly systems using a standard framework, such as Start-Stop Assembly [12, 13], are ideal for design and assembly of such libraries, which may use monocistronic, operon or hybrid designs. In Start-Stop Assembly (and other systems), any chosen mixture of parts in the appropriate format, at any chosen stoichiometry amongst the parts, can be used instead of an individual part at any position. The choice represents a trade-off between library diversity and screening practicality. We usually begin with standard mixtures of six promoters and six RBSs [12] spanning a range of strengths, which have proven effective in many cases. These libraries can be screened for variants that exhibit the desired traits, typically high yield and productivity, caused by combinations of enzyme expression levels that provide high flux towards the product, cause minimal accumulation of potentially toxic by-products or intermediates, avoid excessive protein expression, and show little or no growth impairment [14]. These potential negative effects of heterologous expression have been formalized and studied as ‘metabolic burden’ [15, 16]. Combinatorial methods have been used for functional optimization of gene clusters [17] and improving production of a variety of products, including carotenoids, terpenoids, violacein, riboflavin, 2,3-butanediol, and fatty acids (reviewed in Refs. [14, 18]).

In this work we aimed to establish an approach allowing the construction of combinatorial metabolic pathway libraries in Clostridium spp. We identified Clostridium promoters which are excessively strong in Escherichia coli as a barrier to combinatorial assembly, and we developed a new general approach to overcome this barrier. As an example to develop and test this approach, we constructed a hexanol production library in Clostridium acetobutylicum, building upon the native CoA-dependent branched fermentation pathway in this organism. This work provides methodological foundations to apply combinatorial assembly for improved bioproduction in Clostridium and other organisms.