2.1 Cloning

For all experiments, we used the previously described shuttle plasmids pIM and pM9 (Gaida, Liedtke, Jentges, Engels, & Jennewein, 2016). The Populus alba ispS gene (GenBank accession number: Q50L36) was modified to delete the first 51 codons and to introduce the mutation P494L. The sequence was then codon-optimized by ATG:biosynthetics GmbH (Germany) and cloned downstream of the C. acetobutylicum phosphotransbutyrylase (ptb) promoter (Tummala, Welker, & Papoutsakis, 1999) using the Gibson Assembly process (Gibson et al., 2009) with the following overlaps: forward 5′-GTACCGGTGGTGGCTCCGGTGATGACGACGACAAG-3′ and reverse 5′-GGCTTTGTTTAGCAGCCTAGGTATTAATCAATTAG-3′. The E. coli BL21 (DE3) idi gene was cloned downstream of the thiolase promotor in the modified vector pSOS95 (GenBank accession number: AY18768) by restriction digest and ligation. The constructs were named pIM-repL::P_ptb-ispS and pSOS95::P_thl-Idi_Ec, respectively (Table 1). We used E. coli NEB10B cells (New England Biolabs GmbH, Germany) for cloning.

2.2 Cultivation and electroporation of Clostridium cellulolyticum and genomic integration

The C. cellulolyticum strain H10 (DSM 5812; DSMZ, Braunschweig, Germany) was cultivated essentially as previously described (Gaida et al., 2016), but only 6 g/L of crystalline cellulose was provided as a carbon source. Following electroporation (Gaida et al., 2016), genomic integration of the P_ptb-ispS expression cassette was achieved by adding 0.5% xylose to induce the expression of Himar1 transposase under the control of a xylose-dependent promoter as previously described (Philipps, Vries, & Jennewein, 2019). The resulting strain was named Cce(P_ptb-ispS) (Table 1).

2.3 GC-MS/MS and IspS activity assay

GC-MS/MS experiments were carried out using the TQ8030 system (Shimadzu) fitted with a headspace syringe (Gerstel) and ZB624 column (Phenomenex). We used toluene (0.001%, dissolved in acetone) as an internal standard, and isoprene concentrations were determined against external standards. The 68 > 67 Da transition was used for quantification, assuming that 100% of the isoprene was found in the gas phase. Multiple reaction monitoring (MRM) was used to quantify substances with transitions of 68 > 67 and 67 > 41 Da for isoprene and a retention time of 2.2 min. The GC-MS/MS system was held at 40°C for 1 min, followed by a gradient of 40°C/min to reach a plateau at 200°C for 8 min. For the activity assay, cells were harvested after cultivation for 72 hr (early stationary phase) and were disrupted at 4°C with a sonicator (Hielscher). The quantity of total soluble protein was determined using a Bradford assay (Bio-Rad). Enzyme activity was determined by mixing 50 µg of total protein and 5 µg DMAPP with sufficient buffer (50 mM Tris-HCl, 10 mM MgCl₂, pH 8) to reach a total volume of 500 µl in 10-ml headspace tubes. The headspace was sampled at 8 min intervals following the addition of DMAPP and 12 samples were taken in total, each in triplicate.

2.4 Targeted proteomics

Targeted proteomics was carried out as previously described (Stergachis, MacLean, Lee, Stamatoyannopoulos, & MacCoss, 2011) using either a 1200 Series LC (Agilent Technologies, USA) combined with a QTRAP 3200 MS/MS system (Sciex, USA), or a Shimadzu HPLC combined with a QTRAP 6500 MS/MS system (Sciex). Bovine serum albumin (BSA) was used as an internal control (10 ng/µl; target peptide: YLYEIAR). This was added to the crude extract before 200 µg of total protein was digested per sample. The targeted proteomics method was developed with Skyline (MacLean et al., 2010), and the peptide VDITQIK was used to detect chloramphenicol acetyltransferase (CatP, backbone marker), whereby the peptide VLSVITEPILPFER was used to detect IspS. The peptides were separated in a linear gradient by changing the ratio between buffer A (2% acetonitrile, 98% water, 0.1% formic acid) and buffer B (98% acetonitrile, 2% water, and 0.1% formic acid) from 95:5 to 5:95 with a flow rate of 0.4 ml/min.

3.1 Plasmid cloning, transformation, and genomic integration of Isoprene synthase

The ispS gene was inserted downstream of the ptb promoter using a Gibson Assembly process. The promoter is part of a plasmid segment flanked by internal recognition sites for the Himar transposase. The plasmid DNA was transferred by electroporation into competent C. cellulolyticum cells and stabilized by adding thiamphenicol and clarithromycin. This episomal strain was named Cce(pIM-repL::P_ptb-ispS). The plasmid DNA was reisolated and sequenced, confirming that the ispS gene was intact in all clones. Genomic integration was then induced by adding xylose (0.5%) to activate the transposase gene. This step eliminates the thiamphenicol resistance gene, and we confirmed that the clones were no longer able to grow in the presence of this antibiotic. This strain was named Cce(P_ptb-ispS), and the integrity of one clone was confirmed by sequencing.

3.2 Protein expression, isoprene synthesis, and in vitro IspS activity

The quantity of IspS protein was determined by targeted proteomics, with two MRM transitions and the correct retention time for each unique peptide taken as the minimum evidence for the presence of the target. The IspS peptide transitions were not detected in our control strain Cce(pM9) (Figure 1). However, the transitions were observed in strains Cce(P_ptb-ispS) and Cce(pIM-repL::P_ptb-ispS), confirming that IspS was present in both (Figure 1). The genetic difference between these strains was confirmed by testing for the plasmid backbone marker CatP, which was detected only in strain Cce(pIM-repL::P_ptb-ispS). BSA was added to all crude extracts as an internal standard, and there were no significant variations between different strains.

To verify that IspS was expressed in its active form, we measured isoprene emissions after incubating batch cultures grown in headspace vials for 7 days on cellobiose (Figure 2a) or 14 days on crystalline cellulose due to the slower growth (Figure 2b). We detected higher isoprene levels for all recombinant strains compared to controls, and the cells growing on cellobiose produced up to 10-fold more isoprene than cells growing on crystalline cellulose. Furthermore, strain Cce(pIM-repL::P_ptb-ispS) produced twice as much isoprene as the genomic integrant strain Cce(P_ptb-ispS), probably reflecting the higher gene dosage of the episomal plasmid compared with the single-copy integrated construct. Even so, this is a relatively small difference, probably due to a low plasmid copy number or simply the consequence of a substrate limitation. The latter assumption was supported by the results of introducing the E. coli gene encoding Idi, which catalyses the isomerization of IPP to DMAPP. The transformation of strain Cce(P_ptb-ispS) with pSOS95::P_thl-idi_Ec resulted in an eightfold increase in the quantity of isoprene, probably reflecting the much faster regeneration of the DMAPP pool (Lv, Xu, & Yu, 2013).

The combined proteomics, enzyme activity, and GC-MS/MS data provided strong evidence for the expression of an active IspS protein leading to the accumulation of isoprene, but the background levels of isoprene (≤10 µg/L) made it impossible to exclude the autocatalytic formation of isoprene from DMAPP in larger amounts in our recombinant strains. We therefore carried out in vitro enzyme tests to prove that IspS was responsible for the conversion of DMAPP to isoprene. We confirmed the DMAPP-dependent emission of isoprene from crude extracts of the recombinant strains expressing IspS, with the extract of strain Cce(pIM-repL::P_ptb-ispS) revealing activity of 0.23 µmol/min. In contrast, the recombinant control strain carrying vector pM9 (Gaida et al., 2016) revealed no background activity within the chosen time frame (Figure 3). Moreover, the addition of 5 µg IPP, which is an isomer of DMAPP but not a substrate for IspS, did not lead to any detectable activity during the experiment (Figure 3).

Overall, our combined results therefore provide strong experimental evidence for the stable expression of an active IspS in C. cellulolyticum. The plant enzyme is the central biological catalyst for isoprene production. Hence, its introduction into C. cellulolyticum is the initial and crucial step for the production of isoprene in industrial quantities. However, further steps will be necessary. Usually, the heterologously expressed MVA pathway is used for the supply of DMAPP for isoprene production in bacteria (Yang et al., 2012). Modern approaches try to integrate both pathways, the endogenous MEP and MVA pathway, to increase productivity (Yang et al., 2016). Next to well-designed genetic constructs and proteomics, the analysis of pools of pathway intermediates will be equally important to identify the bottlenecks of isoprene production and to further evaluate C. cellulolyticum as a chassis for it. Therefore, a holistic approach is required for future studies to determine the actual limits of the biological system C. cellulolyticum for industrial-scale production. Clostridia are often considered to be genetically difficult to access and to manipulate. Our experiments strongly indicate the stable expression of an active IspS. Thus, C. cellulolyticum should be considered as a well-suited chassis for the expression of the central enzyme, the IspS from poplar, for isoprene production.