Genetically modified microorganisms

Abstract

The present invention relates to genetically modified microorganisms comprising one or more heterologous nucleic acid molecules together encoding at least three different proteins, each protein comprising an enzymatic domain and a bacterial microcompartment-targeting signal polypeptide, wherein said enzymatic domains each catalyse a different substrate to product conversion in the same metabolic pathway, and wherein said microorganisms are essentially free of bacterial microcompartments (BMCs); and to cell free systems comprising aggregates comprising at least three different proteins, each protein comprising an enzymatic domain and a bacterial microcompartment-targeting signal polypeptide, wherein said enzymatic domains each catalyse a different substrate to product conversion in the same metabolic pathway, and wherein said system does not comprise bacterial microcompartments; and to methods for the production of said microorganisms and cell free systems and their use in methods of producing a product of interest.

Claims

1. A genetically modified microorganism comprising one or more heterologous nucleic acid molecules together encoding at least three different proteins, each protein comprising an enzymatic domain and a bacterial microcompartment-targeting signal polypeptide, wherein said enzymatic domains each catalyse a different substrate to product conversion in the same metabolic pathway, wherein each of said proteins is a fusion protein comprising an enzymatic domain and a bacterial microcompartment-targeting signal polypeptide that are not present in the same protein in nature, wherein said microorganism is lacking bacterial microcompartments (BMCs), and wherein either (i) said microorganism does not naturally comprise the genes necessary for the expression of BMCs and has not been genetically engineered to provide BMC production capability, or (ii) said microorganism naturally comprises the genes necessary for the expression of BMCs but wherein said microorganism is a BMC null mutant.

2. The genetically modified microorganism of claim 1, wherein said microorganism comprises one or more heterologous nucleic acid molecules together encoding at least four different proteins, each protein comprising an enzymatic domain and a bacterial microcompartment-targeting signal polypeptide, wherein said enzymatic domains each catalyse a different substrate to product conversion in the same metabolic pathway.

3. The genetically modified microorganism of claim 1, wherein each of said heterologous nucleic acid molecules encodes only one of said proteins.

4. The genetically modified microorganism of claim 1, wherein each protein is over-expressed relative to the level of expression of said protein in a microorganism of the same strain which lacks said one or more heterologous nucleic acid molecules.

5. The genetically modified microorganism of claim 1, wherein said microorganism comprises aggregates comprising said at least three proteins.

6. The genetically modified microorganism of claim 1, wherein each protein comprises an enzymatic domain and a BMC-targeting signal polypeptide linked by an amino acid linker, preferably wherein said amino acid linker lacks stable secondary structure.

7. The genetically modified microorganism of claim 1, wherein each BMC-targeting signal polypeptide comprises an amphipathic alpha helix.

8. The genetically modified microorganism of claim 1, wherein each BMC-targeting signal polypeptide comprises the following amino acid sequence: X.sub.1X.sub.2X.sub.3X.sub.4X.sub.5X.sub.6X.sub.7X.sub.8X.sub.9 (SEQ ID NO: 76), wherein: X.sub.1, X.sub.4, X.sub.5, X.sub.8, and X.sub.9 are each independently hydrophobic amino acids selected from the group consisting of I, L, V, M, F, Y, A and W; X.sub.2, X.sub.3 and X.sub.6 are each independently polar or charged amino acids selected from the group consisting of Q, N, T, S, C, D, E, R, K and H; and X.sub.7 is any amino acid.

9. The genetically modified microorganism of claim 1, wherein each BMC-targeting signal polypeptide comprises a sequence selected from the group consisting of LEQIIRDVL (SEQ ID NO:1), LETLIRTIL (SEQ ID NO:2), LETLIRNIL (SEQ ID NO:3), LRQIIEDVL (SEQ ID NO:4), IEEIVRSVM (SEQ ID NO:5), IEQVVKAVL (SEQ ID NO:6), VEKLVRQAI (SEQ ID NO:7), IQEIVRTLI (SEQ ID NO:8), VEEIVKRIM (SEQ ID NO:9), IESMVRDVL (SEQ ID NO:10), VQDIIKNVV (SEQ ID NO:11), IRQVVQEVL (SEQ ID NO:12), VRSVVEEVV (SEQ ID NO:13) and ARDLLKQIL (SEQ ID NO:14) or a variant sequence thereof.

10. The genetically modified microorganism of claim 1, wherein said microorganism is a bacterium.

11. The genetically modified microorganism of claim 1, wherein said microorganism does not naturally comprise the genes necessary for the expression of BMCs and has not been genetically engineered to provide BMC production capability.

12. The genetically modified microorganism of claim 1, wherein said microorganism naturally comprises the genes necessary for the expression of BMCs but wherein said microorganism is a BMC null mutant.

13. The genetically modified microorganism of claim 12, wherein said microorganism comprises a mutation in the regulatory region of the BMC operon(s), preferably in the operon's promoter.

Description

(1) The invention will now be further described in the following non-limiting Examples and the Figures in which:

(2) FIG. 1 illustrates a pathway for the synthesis of 1,2-propanediol from glycerol. Glycerol dehydrogenase and dihydroxyacetone kinase catalyse the conversion of glycerol to dihydroxyacetone phosphate, via the intermedia dihydroxyacetone. Methylglyoxal synthase catalyses the conversion of dihydroxyacetone phosphate to methylglyoxal. Glycerol dehydrogenase and 1,2-propanediol oxidoreductase catalyse the conversion of methylglyoxal to 1,2-propanediol, via the intermediate lactaldehyde.

(3) FIG. 2 shows the specific activity of the enzymes involved in the microbial synthesis of 1,2-propanediol: (a) glycerol dehydrogenase (b) dihydroxyacetone kinase (c) methylglyoxal synthase (d) 1,2-propanediol oxidoreductase, when untagged (i.e. not linked to a BMC-targeting signal sequence), when tagged with the BMC-targeting signal sequence D18 and when tagged with the BMC-targeting signal sequence P18.

(4) FIG. 3 provides a statistical analysis showing the percentage of cells that contan inclusion bodies when said cells express either untagged GldA, DhaK, MgsA or FucO, or P18- or D18-tagged versions of said proteins.

(5) FIG. 4 shows the results of a protease protection assay of GFP fused to a C-terminal proteolysis tag (SsrA) and an N-terminal tag being either a BMC targeting signal peptide (P18 or D18) or a non-targeting His-tag, in the presence and absence of BMCs (AU). E. coli competent cells were transformed with plasmids encoding the protein fusions with and without shell proteins and the resulting strains were cultured for 24 hours, samples were taken and run on a 15% denaturing polyacrylamide gel. Gels were subsequently submitted to Western blotting using an anti-GFP primary antibody. Total lysates were analysed by SDS-PAGE and subsequently western blotted with an anti-GFP primary antibody. Cell densities were normalised to an OD.sub.600=2.5 for loading of samples.

(6) FIG. 5 shows in vivo 1,2-propanediol production. The graph shows the 1,2-propanediol content (normalised to OD.sub.600=1) over 96 h in the growth medium of strains that lack shell proteins and 1,2-propanediol producing enzymes (.circle-solid.), Shell proteins only (), un-tagged 1,2-propanediol producing enzymes (.Math.), 1,2-propanediol producing enzymes tagged with targeting sequences (), un-tagged 1,2-propanediol producing enzymes and shell proteins (.square-solid.), 1,2-propanediol producing enzymes tagged with targeting sequences and shell proteins (). Data points represent an average of three independent experiments; standard deviations are represented by error bars.

(7) FIG. 6 shows in vivo 1,2-propanediol production. The graph shows the 1,2-propanediol content (not normalised to OD.sub.600=1) over 96 h in the growth medium in of strains that lack shell proteins and 1,2-propanediol producing enzymes (.circle-solid.), Shell proteins only (), un-tagged 1,2-propanediol producing enzymes (.Math.), 1,2-propanediol producing enzymes tagged with targeting sequences (), un-tagged 1,2-propanediol producing enzymes and shell proteins (.square-solid.), 1,2-propanediol producing enzymes tagged with targeting sequences and shell proteins (). Data points represent an average of three independent experiments; standard deviations are represented by error bars.

(8) FIG. 7 shows thin sections of E. coli strains labelled with an anti-his antibody and then with a secondary antibody conjugated to 10 nm gold particles viewed under TEM (a) Wild type (b) Shell proteins only (c) 1,2-propanediol producing enzymes tagged with targeting sequences (d) 1,2-propanediol producing enzymes tagged with targeting sequences and shell proteins (e) un-tagged 1,2-propanediol producing enzymes (f) un-tagged 1,2-propanediol producing enzymes and shell proteins.

(9) FIG. 8 shows TEM analysis of strains expressing (A) His-tagged GldA (B) P18-tagged GldA (C) His-tagged DhaK (D) P18-tagged DhaK (E) His-tagged MgsA (F) D18-tagged MgsA (G) His-tagged FucO (H) D18-tagged FucO. Scale bar shows 0.2 M.

EXAMPLES

(10) In this study we are creating fusion proteins between Pdu targeting peptides and the four 1,2-propanediol producing enzymes to target the enzymes to recombinant Pdu microcompartment shells. We explore how the targeting peptides affect the activity of the different enzymes and their properties, particularly solubility. Strains are engineered for the targeting of all enzymes to microcompartments and compared for their 1,2-propanediol production to strains containing the native enzymes and also to a strain containing enzymes with targeting peptides but no shell proteins. The protein solubility of these strains is investigated by TEM analysis and protein aggregation is found to play an unexpected but important role in the efficiency of our pathway. Finally, we propose an alternative pathway engineering approach alongside compartmentalisation in protein shells.

(11) Materials and Methods

(12) Strains

(13) The strains used in this study are shown in Table A below:

(14) TABLE-US-00004 TABLE A Strains used in this study Strain Genotype Source BL21*(DE3) F- ompT hsdSB (rB- mB-) gal dcm (DE3) Novagen BL21*(DE3) F- ompT hsdSB(rB- mB-) gal dcm (DE3) Novagen pLysS pLysS (CamR)

(15) The BL21*(DE3) strain comprises genes encoding Eut BMCs. To ensure that Eut BMCs were not produced by the microorganisms during this study, ethanolamine was not included in any fermentation media. The absence of BMCs was confirmed by TEM.

(16) Plasmid Construction

(17) Plasmids were constructed to include each of the genes of interest with an N-terminal tag comprising a BMC-targeting signal polypeptide (P18 or D18) and/or a hexa-histidine tag.

(18) The genes of interest were the four enzymes of the metabolic pathway for the production of 1,2-propanediol from glycerol, as outlined in FIG. 1, i.e. glycerol dehydrogenase (GldA), dihydroxyacetone kinase (DhaK), methylglyoxal synthase (MgsA) and 1,2-propanediol oxidoreductase (FucO).

(19) The BMC-targeting signal polypeptide-containing tags used in the study were as follows:

(20) TABLE-US-00005 D18: (SEQIDNO:64) MEINEKLLRQIIEDVLSEPMGSSHHHHHHSSGLVPRGSH

(21) (N-terminal 18 amino acids of PduD from Citrobacter freundii (the BMC-targeting signal polypeptide) followed by flexible linker PMGSS, 6-his linker, flexible linker SSGL, thrombin cleavage site LVPRGS and amino acid linker H)

(22) TABLE-US-00006 P18: (SEQIDNO:65) MNTSELETLIRNILSEQLAMGSSHHHHHHSSGLVPRGSH

(23) (N-terminal 18 amino acids of PduP from Citrobacter freundii (the BMC-targeting signal polypeptide) followed by flexible linker AMGSS, 6-his linker, flexible linker SSGL, thrombin cleavage site LVPRGS and amino acid linker H)

(24) All primers used in this study are listed in Table B below.

(25) TABLE-US-00007 TABLEB Oligonucleotideprimers usedinthisstudy,restriction sitesareunderlined Name Sequence5-3 GldA_NdeI_FW CATCATATGGACCGCATTATTC AATCACC (SEQIDNO:66) GldA_SpeI_RV CATACTAGTTTATTCCCACTCT TGCAGG (SEQIDNO:67) dhaK_NdeI_FW CGCCATATGTCTCAATTCTTTT TTAACCAACGCACC (SEQIDNO:68) dhaK_SpeI_RV CATACTAGTTTAGCCCAGCTCA CTCTCCGC (SEQIDNO:69) mgsA_NdeI_FW CATCATATGGAACTGACGACTC GCACTTTACC (SEQIDNO:70) mgsA_SpeI_RV CATACTAGTTTACTTCAGACGG TCCGCGAG (SEQIDNO:71) fucO_NdeI_FW CCGCATATGGCTAACAGAATGA TTCTG (SEQIDNO:72) fucO_SpeI_RV CCTACTAGTTTACCAGGCGGTA TGG (SEQIDNO:73) GFP_NdeI_FW GTACATATGAGCAAAGGAGAAG AACTTTTC (SEQIDNO:74) GFP-SsrA_SpeI_RV GACACTAGTTTAAGCTGCTAAA GCGTAGTTTTCGTCGTTTGCTG CTTTGTACAGCTCATCCATGCC (SEQIDNO:75)

(26) All genes were amplified with flanking Ndel and Spel restriction sites and each was ligated into pET14b, pET14b-D18 and pET14b-P18 vectors using Ndel and Spel restriction sites.

(27) Plasmids pML-1 to pML-6 as outlined in Table C, were constructed by a Link and Lock approach utilizing the compatible sticky ends formed by digestion with Xbal and Spel (McGoldrick et al., (2005) J Biol Chem 14:1086-1094).

(28) TABLE-US-00008 TABLE C Plasmids used in this study Plasmid name Genotype Description Source pET14b pET14b Overexpression vector Novagen containing N-terminal polyhistidine-tag pET14b-D18 pET14b-D18 Overexpression vector This study containing an N-terminal D18 targeting tag and an N-terminal polyhistidine-tag pET14b-P18 pET14b-P18 Overexpression vector This study containing an N-terminal P18 targeting tag and an N-terminal polyhistidine-tag pLysS PlysS Overexpression vector Novagen pLysS-PduABBJKNU pLysS-PduABBJKNU Construct for expression Parsons et al., of empty Pdu BMC 2010 pET14b-gldA pET14b-gldA PCR product of gldA This study ligated into Ndel/Spel sites of pET14b pET14b-dhaK pET14b-dhaK PCR product of dhaK This study ligated into Ndel/Spel sites of pET14b pET14b-mgsA pET14b-mgsA PCR product of mgsA This study ligated into Ndel/Spel sites of pET14b pET14b-fucO pET14b-fucO PCR product of fucO This study ligated into Ndel/Spel sites of pET14b pET14b-GFP-SsrA pET14b-GFP-SsrA PCR product of gfp-ssrA This study ligated into Ndel/Spel sites of pET14b pET14b-D18-gldA pET14b-D18-His-gldA Ndel/Spel fragment of This study pET14b-gldA ligated into Ndel/Spel sites of pET14b-D18 pET14b-D18-dhaK pET14b-D18-His-dhaK Ndel/Spel fragment of This study pET14b-dhaK ligated into Ndel/Spel sites of pET14b-D18 pET14b-D18-mgsA pET14b-D18-His-mgsA Ndel/Spel fragment of This study pET14b-mgsA ligated into Ndel/Spel sites of pET14b-D18 pET14b-D18-fucO pET14b-D18-His-fucO Ndel/Spel fragment of This study pET14b-fucO ligated into Ndel/Spel sites of pET14b-D18 pET14b-D18-GFP-SsrA pET14b-D18-His- Ndel/Spel fragment of This study GFP-SsrA pET14b-GFP-SsrA ligated into Ndel/Spel sites of pET14b-D18 pET14b-P18-gldA pET14b-P18-His-gldA Ndel/Spel fragment of This study pET14b-gldA ligated into Ndel/Spel sites of pET14b-P18 pET14b-P18-dhaK pET14b-P18-His-dhaK Ndel/Spel fragment of This study pET14b-dhaK ligated into Ndel/Spel sites of pET14b-P18 pET14b-P18-mgsA pET14b-P18-His-mgsA Ndel/Spel fragment of This study pET14b-mgsA ligated into Ndel/Spel sites of pET14b-P18 pET14b-P18-fucO pET14b-P18-His-fucO Ndel/Spel fragment of This study pET14b-fucO ligated into Ndel/Spel sites of pET14b-P18 pET14b-P18-GFP-SsrA pET14b-P18-His- Ndel/Spel fragment of This study GFP-SsrA pET14b-GFP-SsrA ligated into Ndel/Spel sites of pET14b-P18 pML-1 pET14b-His-gldA-His- Xbal/EcoRI fragment This study fucO from pET14b-His-fucO ligated into Xbal/EcoRI sites of pET14b-His- gldA pML-2 pET14b-P18-gldA- Xbal/EcoRI fragment This study D18-fucO from pET14b-D18-His- fucO ligated into Spel/EcoRI sites of pET14b-P18-His-gldA pML-3 pET14b-His-dhaK- Xbal/HindIII fragment This study His-mgsA from pET14b-His-mgsA ligated into Spel/HindIII sites of pET14b-His- dhaK pML-4 pET14b-P18-dhaK- Xbal/HindIII fragment This study D18-mgsA from pET14b-D18-His- mgsA ligated into Spel/HindIII sites of pET14b-P18-His-dhaK pML-5 pET14b-His-dhaK- Xbal/Clal fragment from This study His-mgsA-His-gldA- pML-3 ligated into His-fucO Spel/Clal sites of pML-1 pML-6 pET14b-P18-dhaK- Xbal/Clal fragment from This study D18-mgsA-P18-gldA- pML-4 ligated into D18-fucO Spel/Clal sites of pML-2

(29) Overexpression and Purification of Recombinant Protein

(30) BL21*(DE3) pLysS competent cells were transformed with a plasmid containing the gene(s) of interest. 1 L of LB supplemented with ampicillin (100 mg/L) in baffled flasks was inoculated from an overnight starter culture. The cultures were grown at 37 C. with shaking for 7 hours; protein production was induced by the addition of IPTG to a final concentration of 400 M. The cultures were then incubated overnight at 19 C. with shaking. Cells were harvested by centrifugation at 3320g for 15 minutes at 4 C., pellets were resuspended in 20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 5 mM Imidazole. Cells were lysed by sonication and cell debris removed by centrifugation. Recombinant protein was then purified from the soluble fraction by immobilized metal ion affinity chromatography.

(31) Activity Assays

(32) Glycerol Dehydrogenase

(33) The activity of GldA for the oxidation of glycerol to dihydroxyacetone was measured by following the initial rate at 340 nm for the reduction of NAD+ to NADH. Activity assays were carried out in 1 ml reactions containing 0.1 M potassium phosphate buffer pH 8.0, 500 M NAD+, 2 mM MgCl.sub.2 and 200 nM GldA. The activity of GldA for the reduction methylglyoxal to lactaldehyde, was measured by following the initial rate of the oxidation of NADH to NAD+ at 340 nm. Activity assays were carried out in 1 ml reactions containing 0.1 M potassium phosphate buffer pH 8.0, 0.1 mM NADH, 2 mM MgCl2, 200 nM GldA.

(34) Dihydroxyacetone Kinase

(35) The activity of DhaK for the conversion of dihydroxyacetone to dihydroxyacetone phosphate was measured in a coupled reaction with Glyceraldehyde 3-phosphate dehydrogenase (G3PDH) by following the oxidation of NADH to NAD+ at 340 nm. Activity assays were carried out in 1 ml reactions containing 50 mM Tris-HCl, 100 mM NaCl, 1 mM ATP, 0.1 mM NADH, 2.5 mM MgCl.sub.2, 7.2 U G3PDH, 125 nM DhaK.

(36) Methylglyoxal Synthase

(37) The activity of MgsA was monitored in a colorimetric assay over a time course. 25 l 0.5 mM MgsA was incubated in a reaction mixture containing 400 l 50 mM imidazole pH 7.0, 25 l 15 mM dihydroxyacetone phosphate, 50 l dH2O, the reaction mixture was incubated at 30 C. with shaking. At time intervals 50 l of the reaction mixture was removed and added to a detection mixture containing 450 l dH2O, 165 l 0.1% 2,4-Dinitrophenylhydrazine hydrochloric acid solution. The detection mixture was incubated at 30 C. with shaking for 15 minutes. 835 l of 10% (w/v) NaOH was added to the detection mixture which was incubated at room temperature for 15 minutes. Absorbances were then measured at 550 nm.

(38) 1,2-Proanediol Oxidoreductase

(39) The activity of FucO was determined for the NADH dependant reduction of glycolaldehyde to elthylene glycol was measured by following the initial rate of the oxidation of NADH to NAD+ at 340 nm. Activity assays were carried out in 1 ml reactions containing 100 mM Hepes, 10 M NADH, 100 M MnCl2, 200 nM FucO.

(40) Embedding of Strains for TEM Analysis

(41) 50 ml of LB was inoculated with one colony and grown at 37 C. with shaking to an OD600 of 0.4, cells were harvested by centrifugation at 3000g for 10 minutes. The cell pellet was resuspended in 2 ml 2.5% Glutaraldehyde in 100 mM cacodylate pH 7.2 and incubated for 2 hours with gentle spinning. Cells were pelleted by centrifugation at 6000g for 2 minutes and were washed twice with 100 mM cacodylate pH 7.2. Cells were stained with 1% osmium tetroxide in 100 mM cacodylate pH 7.2 for 2 hours and subsequently washed twice with dH2O. Cells were dehydrated by incubation in an ethanol gradient, 50% EtOH for 10 minutes, 70% EtOH overnight followed by two 10 minute washes in 100% EtOH. Cells were then washed twice with propylene oxide for 15 minutes. Cell pellets were embedded by resuspention in 1 ml of a 1:1 mix of propylene oxide and Agar LV Resin and incubated for 30 minutes with spinning. Cell pellets were washed twice in 100% Agar LV resin. The cell pellet was resuspended in fresh resin and transferred to a 0.5 ml mould, centrifuged for 5 minutes at 3000g to concentrate the cells to the tip of the mould and incubated for 16 hours at 60 C. to polymerise.

(42) Sectioning and Visualisation of Samples

(43) Samples were thin sectioned on a RMC MT-XL ultramicrotome with a diamond knife (diatome 45) sections were placed on 300 mesh copper grids. Grids were stained by incubation in 4.5% uranyl acetate in 1% acetic acid solution for 1 hour followed by 2 washes in dH.sub.2O. Grids were then stained with 0.1% lead citrate for 8 minutes followed by a wash in ddH.sub.2O

(44) Electron microscopy was performed using a JEOL-1230 transmission electron microscope.

(45) Culture Medium and Conditions for 1,2-Propanediol Production

(46) The culture medium designed by Neidhardt et al., 1974 was supplemented with 30 g/L glycerol, 10 g/L tryptone, 5 g/L yeast extract and appropriate antibiotics. Strains were cultured in sealed serum bottles with a working volume of 100 ml at 28 C. with shaking. Cultures were inoculated from starter cultures to starting OD.sub.600 of 0.05. During growth 1 ml samples were removed at 0, 6, 12, 24, 48, 72 and 96 hours.

(47) Western Blot Analysis

(48) Nitrocellulose membranes following transfer and blocking were incubated in primary antibody (mouse anti-GFP) followed by incubation in a secondary coupled antibody (Anti-mouse IgG AP). Bands were visualised by incubation in substrate 5-Bromo-4-chloro-3-indolyl phosphate/Nitro blue tetrazolium (BCIP/NBT).

(49) Analysis of 1,2-Propanediol Production

(50) In-vivo 1,2-propanediol production was determined by GC/MS analysis of the growth medium at time intervals (0, 6, 12, 24, 48, 72 and 96 hours). The supernatant after centrifugation, was boiled for 10 minutes at 100 C. followed by centrifugation at 19,750g. The sample was then acidified with trifluoroacetic acid to a final concentration of 0.01% followed by a second centrifugation at 19,750g. The supernatant following centrifugation was diluted 1:4 in acetonitrile for GC/MS analysis.

(51) Visualisation of Engineered Strains

(52) Embedding of Strains for Immunolabeling

(53) Strains were grown as described previously (Culture medium and conditions for 1,2-propanediol production, above) overnight, cells were harvested by centrifugation for 10 minutes at 3000g. The cell pellet was resuspended in 2% formaldehyde, 0.5% gluteraldehyde in 100 mM sodium cacodolate buffer pH 7.2 and incubated for 2 hours with gentle spinning. Cells were pelleted by centrifugation at 6000g for 2 minutes and were washed twice with 100 mM sodium cacodylate pH 7.2. Cells were dehydrated by incubation in an ethanol gradient, 50% EtOH for 10 minutes, 70% EtOH for 10 minutes, 90% EtOH for 10 minutes, followed by three 15 minute washes in 100% EtOH. Cell pellets were then resuspended in 2 ml LR white resin and incubated overnight with spinning at room temperature after which the resin was changed and incubated for a further 6 hours. Cell pellets were resuspended in fresh resin and transferred to 1 ml embedding tubes and centrifuged at 4000g to pellet the cells at the tip and incubated for 24 hours at 60 C. to polymerize.

(54) Samples were thin sectioned on a RMC MT-XL ultramicrotome with a diamond knife (diatome 45) sections were placed on 300 mesh gold grids.

(55) Immunolabeling of Sections

(56) Grids were equilibrated in one drop of TBST (20 mM Tris-HCl pH 7.2, 500 mM NaCl, 0.05% Tween (RTM) 20, 0.1% BSA) before being transferred into a drop of 2% BSA in TBST and incubated at room temperature for 30 minutes. Grids were then immediately transferred into primary antibody (Anti-His) and incubated for 1 hour. Grids were washed in a fresh drop of TBST followed by washing in a stream of TBST. Grids were equilibrated in a drop of secondary antibody (Goat anti-mouse IgG 10 nm) then incubated for 30 minutes in a fresh drop. Excess antibody was removed by washing in two drops of TBST before washing in a stream of ddH.sub.2O and dried.

(57) Staining

(58) Grids were stained for 15 minutes in 4.5% uranyl acetate in 1% acetic acid solution followed by 2 washes in dH.sub.2O. Grids were then stained with 0.1% lead citrate for 3 minutes followed by a wash in ddH.sub.2O.

(59) Electron microscopy was performed using a JEOL-1230 transmission electron microscope.

(60) Results

(61) Effect of Fusing Enzymes to BMC-Targeting Peptides on Enzyme Specific Activities

(62) The effect of fusing either of the two targeting peptides P18 and D18 to heterologous enzymes on the functionality of those enzyme had not previously been investigated in detail. In this study the enzymes involved in the microbial synthesis of 1,2-propanediol from glycerol, namely glycerol dehydrogenase (GldA), dihydroxy acetone kinase (DhaK), methylglyoxal synthase (MgsA) and 1,2-propanediol oxidoreductase (FucO) were cloned with both N-terminal targeting peptides (P18 or D18) followed by a hexa-histidine tag. Proteins of interest were purified by IMAC and the kinetic parameters of each of the protein fusions were subsequently determined and compared to enzymes containing only the N-terminal hexa-histidine tag. In this Example, the term tagged proteins refers to P18-his and D18-his containing proteins, while the his-only containing proteins as referred to as untagged proteins.

(63) It was found that the targeting peptides effect the specific activities of some of the proteins studied (FIG. 2). For instance, tagging GldA (the enzyme that catalyses the oxidation of glycerol) with the D18 targeting peptide resulted in a reduction of its specific activity by 90% compared to un-tagged GldA. Tagging GldA with the P18 targeting peptide reduced the activity of the protein by approximately half (55% reduction) compared to un-tagged GldA (FIG. 2A). GldA's ability to reduce methylglyoxal to lactaldehyde was similarly affected, with D18 having the greatest negative effect (83% reduction compared to un-tagged GldA) and P18 causing a loss of 53% of specific activity compared to untagged GldA. The activity of DhaK for the ATP dependant phosphorylation of dihydroxyacetone phosphate was determined by a coupled reaction involving a second enzyme, glyceraldehyde 3-phosphate dehydrogenase, in excess. In contrast to GldA, kinetic analysis of DhaK fused with either a P18 or D18 targeting peptide had no significant effect on the enzyme's activity (FIG. 2B).

(64) Tagging MgsA with either a P18 or D18 targeting peptide had a negative effect on enzyme activity, reducing the activity by 18% and 15% respectively in comparison to untagged MgsA, as shown in FIG. 2C. When fused to D18, FucO's specific activity decreased by 58% in comparison to untagged FucO and when fused to P18, FucO's specific activity decreased by 76% in comparison to untagged FucO (FIG. 2D).

(65) It is concluded that the fusion of targeting peptides to the N-termini of proteins is likely to have an effect on the specific activities of a significant proportion of said proteins. Without wishing to be bound by theory, the inventors consider this is most likely due to changes in structural and chemical properties and potential changes in protein folding as a result of the fusion.

(66) Targeting Peptides Cause Protein Aggregation that can be Visualised by TEM

(67) The production levels and solubility of GldA, DhaK, MgsA and FucO with and without targeting peptides fused thereto were investigated by subjecting samples of the purification process, including the soluble and insoluble fractions after clarification of the crude cell lysate as well as the final purified protein samples, to denaturing polyacrylamide gel analysis (data not shown). DhaK and MgsA were well produced and soluble irrespective of the presence of a P18 or D18 tag. The solubility of FucO was not affected by targeting peptides, but the yield of un-tagged FucO appeared slightly lower compared to FucO containing the targeting peptides. In contrast, although, both P18 and D18-tagged GldA appeared to be produced, the protein bands were predominantly detected in the insoluble fractions of the SDS gels.

(68) This suggests that the fusion proteins D18-GldA and P18-GldA were aggregating compared to GldA. P18-GldA was also found to be eluted from the IMAC column with an additional band of smaller molecular weight, indicative of protein degradation.

(69) In order to investigate the aggregation behaviour of the tagged proteins further, the most active protein fusions (P18-GldA, P18-DhaK, D18-MgsA and D18-FucO), the candidates for the construction of the 1,2-propanediol production pathway targeted to microcompartments, were chosen to be visualised by TEM.

(70) Strains encoding each of the tagged proteins and strains encoding the un-tagged proteins were cultured overnight without induction. Subsequently, the cells were harvested, embedded in low viscosity resin, thin sectioned and visualized using TEM. For each strain 100 cells were examined for protein aggregation, statistical analysis of each of the strains is shown in FIG. 3 and representative TEM micrographs were compiled (FIG. 8).

(71) Control strains producing un-tagged proteins (GldA, DhaK, MgsA, FucO) displayed a normal phenotype, with only 1% of observed cells containing electron dense areas indicative of aggregated proteins. In contrast, half of all observed cells (52%) producing P18-GldA showed protein aggregates located at the pole of the cells (FIG. 8). The addition of the P18 targeting peptide to the N-terminus of DhaK resulted in aggregate formation in 8% of the observed cells. Fusion of the D18 targeting peptide to MgsA and FucO resulted in the presence of protein aggregates in 12% and 4% of cells respectively.

(72) These results confirm that the fusions between the enzymes of the 1,2-propanediol production pathway and targeting peptides cause protein aggregation.

(73) Enzymes Fused to BMC-Targeting Peptides are Recruited to BMCs

(74) It was investigated whether fusion proteins comprising an enzyme of interest fused to either the P18 or D18 targeting peptide were targeted to bacterial microcompartments.

(75) Strains co-expressing the individual genes of the 1,2-propanediol pathway (P18-gldA, P18-mgsA, D18-dhaK, D18-fucO) and the construct for empty shell formation (pLysS-PduABBJKNU) were cultured and the recombinant microcompartments were purified as described previously (Lawrence et al., (2014) ACS Synth. Biol. 3: 454-465). Samples were taken throughout the purification and analysed on 15% denaturing polyacrylamide gels for the protein profile. Analysis of the resulting SDS-PAGE gels reveals that tagging each of the proteins with a targeting peptide facilitates their co-purification with the microcompartment proteins. This was further confirmed by kinetic assays of the final purified BMC fraction.

(76) Further evidence of protein targeting to microcompartments was provided by a protease protection assay that was previously reported by Sargent et al., (2013) Microbiology 159: 2427-2436.

(77) Plasmids were constructed containing GFP fused to an N-terminal P18 or D18 tag and a C-terminal SsrA proteolysis tag (AANDENYALAA*). The C-terminal SsrA tag targets proteins for degradation by the E. coli proteases ClpAP and ClpXP. E. coli competent cells were transformed with plasmids encoding the protein fusions with and without shell proteins and the resulting strains were cultured for 24 hours, samples were taken and run on a 15% denaturing polyacrylamide gel, adjusted to cell number as determined by OD.sub.600 measurements. Gels were subsequently submitted to Western blotting using an anti-GFP primary antibody.

(78) The results show that the co-expression of GFP-SsrA fused to targeting peptides and produced with shell proteins have the highest amount of GFP (FIG. 4 lane 7+8). In the absence of a targeting peptide GFP is effectively degraded as represented by only a faint band present in lane 6 of FIG. 4. In the absence of shell proteins, all GFP fusion proteins are present to a much lesser extent than fusion proteins in the presence of microcompartments. The faint bands seen could be a result of protein aggregation, which would protect the GFP fusions from proteolytic cleavage. The band seen in lane 1 of FIG. 4 (shell only) most likely represents unspecific binding of the antibody. The difference seen in band intensity between lanes 3 and 6 of FIG. 4 is likely due to differences in expression levels as a result of the co-expression of shell proteins.

(79) These results are consistent with microcompartments providing protection from cytosolic proteases for proteins internalised therein.

(80) Construction of 1,2-Propanediol Producing Strains and Comparative Analysis of Bacterial Growth

(81) For the in vivo production of 1,2-propanediol, single plasmids were engineered by using link and lock cloning combining firstly the genes coding for the most active protein fusions (pML-6 containing P18-his-gldA, P18-his-dhaK, D18-his-mgsA, D18-his-fucO) and secondly the same genes but without targeting sequences (pML-5 containing his-gidA, his-dhaK, his-mgsA, his-fucO). Both plasmids were used to transform the E. coli strain BL21*(DE3).

(82) With the aim of targeting the 1,2-propanediol producing enzymes to recombinant microcompartments, strains were engineered to co-express the 1,2-propanediol production plasmids with the genes coding for the protein shell (pLysS-PduABBJKNU). The shell protein construct allows for the formation of a microcompartment shell to which the fusion enzymes are recruited by virtue of their BMC-targeting peptide. Additionally, the following control strains were set up: firstly BL21*(DE3) transformed with pET14b and pLysS; and secondly, a shell only strain transformed with pET14b and pLysS-PduABBJKNU. All strains were compared for the production of 1,2-propanediol.

(83) The culture medium designed by Neidhardt et al., 1974 (Neidhardt F C, Bloch P L, Smith D F. Culture Medium for Enterobacteria. Journal of Bacteriology. 1974; 119(3):736-747) was supplemented with 30 g/L glycerol, 10 g/L tryptone and 5 g/L yeast extract and appropriate antibiotics. Strains were cultured in sealed serum bottles with a working volume of 100 ml at 28 C. with shaking The cultures were started with an initial OD.sub.600 of 0.05 by inoculation from 5 ml starter cultures. During growth, 1 ml samples were collected at 0, 6, 12, 24, 48, 72 and 96 hours and optical densities at 600 nm were measured. The resulting growth curves (not shown) indicate that strains encoding proteins with targeting peptides (either with shell proteins or without) grow slower and reach a lower final optical density in comparison to strains expressing un-tagged proteins and control strains. Furthermore, cell densities declined from 24 hours in strains producing un-tagged enzymes whereas the cell densities of the strains with tagged enzymes remained constant, which indicates that cells with tagged enzymes cells are being protected from a toxic intermediate.

(84) In Vivo 1,2-Propanediol Production is Elevated in Strains Producing Enzymes with Targeting Sequences

(85) The 1,2-propanediol content in the growth media of the various strains was quantified by gas chromatography-mass spectrometry (GC-MS). Whole cell samples were collected at 0, 6, 12, 24, 48, 72 and 96 hours and the supernatant following centrifugation was prepared for GC-MS analysis as described in materials and methods. The measured 1,2-propanediol content as shown in FIG. 5 is expressed for a cell density of OD.sub.600=1. FIG. 6 shows the measured 1,2-propanediol content not adjusted for a cell density of OD.sub.600=1.

(86) Strains encoding un-tagged 1,2-propanediol producing enzymes with and without shell proteins showed low 1,2-propanediol production despite growing well and reaching the highest cell densities at 96 hours. Both strains reached the maximum product concentration (at 96 hours) of 3.59 mM/OD.sub.600=1 in the absence of shell proteins and 1.95 mM/OD.sub.600=1 in the presence of shell proteins.

(87) The highest product concentrations were detected in the growth media of strains producing proteins tagged with targeting peptides. Although both of these strains (with and without shell proteins) grew to lower density than the strains harbouring un-tagged proteins, and despite the negative effect the targeting peptide has on the specific activities of the individual enzymes, they produced significantly more 1,2-propanediol (FIG. 5) than the strains lacking the targeting peptides. The strain containing tagged 1,2-propanediol enzymes and shell proteins reached a final yield of 7.10 mM/OD.sub.600=1. However, the highest final yield of 11.56 mM/OD.sub.600=1 was observed when the shell proteins were not present. 1,2-propanediol was not detected in control stains (wild type E. coli and a strain producing shell proteins only).

(88) A comparison of FIG. 5 and FIG. 6 shows that the increase in 1,2-propanediol production levels is not simply an effect of differences in cell densities i.e. even though the strains with tagged enzymes don't grow as well as the strains without tagged enzymes, they still produce more 1,2-propanediol.

(89) The higher product yield exhibited by the strain producing tagged enzymes in the absence of shell proteins was unexpected.

(90) To investigate if aggregation of the proteins were causing this effect, electron microscopy and immunolabeling were used to visualise the subcellular organisation and location of the recombinant proteins in the various strains. Sections of the strains were labelled with anti-histidine primary antibody designed to bind to the hexa-histidine tag on the N-terminus of proteins in our pathway thereby revealing the intracellular location. A secondary antibody conjugated to 10 nm gold particles was used to bind to the primary antibody thereby revealing the intracellular location of 1,2-propanediol producing enzymes.

(91) Control strains (wild type and shell only) showed a small amount of antibody binding around the membrane of the cells (FIGS. 7A and 7B); this is likely due to unspecific binding. Aggregates are visible in approximately 100% of observed cells expressing P18/D18-tagged proteins, regardless of the presence or absence of shell proteins, and it is in these areas that the vast majority of antibody binding occurs (FIGS. 7C and D). Such structures cannot be seen in cells expressing un-tagged enzymes (FIGS. 7E and F), suggesting that it is the presence of the targeting peptides that facilitates the aggregation of proteins.

(92) It is concluded that aggregation occurs due to tagging proteins of interest with targeting peptides and it is this aggregation that results in a significant increase in product yield despite the reduction in specific activities of the individual tagged pathway enzymes.

(93) It has been determined that the fusion of BMC-targeting peptides to the individual enzymes in the pathway for the production of 1,2-propanediol lowers the specific activities of the enzymes in some (most) cases. The solubility of each enzyme was also affected to varying degrees, with GldA forming large inclusion bodies in the majority of cells observed by TEM when fused with a targeting peptide compared to un-tagged GldA. It has also been demonstrated that the addition of a BMC-targeting tag recruited the enzymes to BMCs and that purified samples thereof remained metabolically active.

(94) Despite the significant decrease of enzyme activity seen with the addition of targeting peptides to both GldA and FucO, expression of the complete tagged pathway enzymes led to an increase in product formation as compared to strains in which untagged enzymes were expressed. Rather unexpectedly, the presence of the microcompartment shell was not required for the increased product formation and, furthermore, the strain generating the most 1,2-propanediol produced tagged 1,2-propanediol pathway enzymes, but no shell proteins. This strain showed an increase in product formation of 245% OD-adjusted in comparison to the strain producing un-tagged enzymes; despite the lower in vitro activity of the individual tagged proteins compared to the un-tagged proteins.

(95) TEM analysis showed that co-production of all four tagged enzymes resulted in protein aggregation and deposition at the poles of nearly all cells observed and it is this aggregation that appears to provide a significant benefit to the efficiency of the pathway. Aggregation of our proteins of interest is likely due to the amphipathic helical nature of the BMC-targeting sequences and/or by their coiled coil structure. Without wishing to be bound by theory, it can be considered that the aggregation creates a scaffolding effect that result in increased channeling of substrates and products between enzymes, similarly to the environment inside a microcompartment.

(96) This study is the first to demonstrate that the presence of short targeting peptides can not only convert individual fusion proteins but also whole pathways into active aggregates that allow for increased product yield in vivo. These aggregations of multiple enzymes allow for increased localised concentrations of enzymes and intermediates and possibly channeling between them thereby resulting in a higher product yield. This is the first study to demonstrate that increased product yields can result from tagging enzymes in a metabolic pathway for the production of said product with a BMC targeting sequence in a cell lacking BMCs themselves.