FIELD OF THE INVENTION

The invention provides a recombinant probiotic gram-negative bacterium engineered to grow on a rare carbohydrate as a sole carbon source, and, whose ability to colonize the mammalian gut is responsive to the supply of said carbohydrate. The invention further provides a pharmaceutical composition or kit comprising the recombinant probiotic gram-negative bacterium and said carbohydrate.

BACKGROUND OF THE INVENTION

Intestinal bacteria play a crucial role in gastrointestinal health and homeostasis. Many diseases relate to a condition referred to as dysbiosis, a microbial imbalance in the gut associated with a bloom of pathobionts, loss of commensals, and loss of diversity within the gut. Given the implications of the gut microbiome in disease, means for modulating the gut microbiome have been explored with the goal of lowering disease prevalence. Among the approaches tested is the administration of faecal microbial transplants for the treatment of diseases ranging from irritable bowel syndrome to chronic fatigue syndrome. However, such faecal microbial transplants have recently been linked to multiple cases of serious and fatal adverse events due to the transfer of drug-resistant bacteria. In view of concerns as to the safety and efficacy of this treatment, alternative methods for modulating the gut microbiome are needed.

One such method is through the use of prebiotics, which are non-digestible substrates associated with an increased density of health-promoting bacteria. As such, prebiotics serve as a less invasive and short-term mechanism for modulating the gut microbiome for health benefits. Overall, consumption of prebiotics such as inulin, fructo- oligosaccharides, and galacto-oligosaccharides increases the density of beneficial bacteria such as Bifidobacterium and Lactobacilli species.

Probiotics, living organisms that are beneficial to health, have also been shown to help modulate the gut microbiome in order to improve health. As with prebiotics, probiotics are specifically used to alter the gut environment. However, instead of seeking to up- regulate beneficial bacteria by providing prebiotic substrates, such as dietary fibers, probiotics are used to directly introduce beneficial strains. Probiotics function by either directly interacting with the host via chemical and physical signals or by affecting the make-up of the gut microbial community. Previously, probiotics have been useful in treating obesity, diabetes, inflammation, cancer, allergies, and many other ailments. The success of probiotics in benefiting the gut environment, as well as human health as a whole, has led to a new generation of probiotics engineered to augment the innate benefits of probiotics through a wide range of mechanisms such as the production of therapeutics. These next generation probiotics are often referred to as smart probiotics, living therapeutics, or advanced microbial therapeutics. In such systems, microbial production of therapeutics allows for a continuous and inexpensive supply of molecules such as hormones, interleukins, and antibodies. With a potential for secreting a range of molecules, these living therapeutics have a wide scope of possibilities stretching far beyond the already important role of gut microbes. As some therapeutics are unstable or require high doses, utilizing engineered microbials may be a superior alternative to traditional drug delivery as the microbe-produced therapeutic avoids exposure to the harsh acidic conditions of the upper gastrointestinal tract. Additionally, with an ever-expanding toolbox of sensors, kill switches, memory circuits, etc., these bacteria can be fine-tuned to better secrete therapeutics, sense signals within the gut environment, and respond to physiological changes.

Yet, despite these advances, advanced microbial therapeutics still face a variety of challenges in regard to fine-tuning the expression of therapeutics in the gut microbiome. Firstly, most probiotics are poor colonizers, and hence their therapeutic efficacy is transient, due to the failure of such advanced microbial therapeutics to establish long- term stability. Furthermore, precise manipulation of the gut microbiome is currently poorly understood, and, exact dosing of given therapeutics is therefore complicated by a difficulty of controlling microbial density. With the importance of being able to deliver precise doses to patients in order to limit adverse events and cater to specific disease profiles, the ability to control engraftment (colonization and cell density) poses a major challenge to advanced microbial therapeutics.

Accordingly, there exists a need for developing therapeutic probiotic strains and administration methods that allow for controlled engraftment in a subject.

SUMMARY OF THE INVENTION

A first aspect of the invention provides a recombinant probiotic gram-negative bacterium comprising genes encoding:

(a) an outer membrane oligosaccharide transporter, and

(b) one or more periplasmic space glycosidases, wherein said one or more glycosidases are capable of cleaving one or more oligosaccharides imported by said transporter; and wherein said bacterium can utilize said one or more oligosaccharides for growth. A second aspect of the invention provides a composition comprising the recombinant probiotic gram-negative bacterium according to the invention, and said oligosaccharide.

A third aspect of the invention provides a pharmaceutical composition for use as a medicament, comprising a recombinant gram-negative probiotic bacterium according to the invention, wherein said pharmaceutical composition is for oral administration.

A fourth aspect of the invention provides a pharmaceutical kit for use as a medicament, comprising the pharmaceutical composition of the invention, and a prebiotic comprising said one or more oligosaccharides according to the invention, wherein said recombinant probiotic gram-negative bacterium is capable of growth on said prebiotic as sole carbon source, and wherein said prebiotic is for oral administration either separately, sequentially or simultaneously to the administration of said pharmaceutical composition

A fifth aspect of the invention provides a pharmaceutical composition or kit according to the invention for use as a medicament, wherein said recombinant probiotic gram negative bacterium further comprises one or more transgenes encoding: (a) a therapeutic polypeptide, or

(b) one or more enzymes for synthesis of a therapeutic molecule.

A sixth aspect of the invention provides a method for treating and/or preventing a bowel- related disease in a subject, the method comprising administering to the subject diagnosed with a bowel-related disease a recombinant bacteria engineered to express genes encoding:

(a) an outer membrane oligosaccharide transporter, and

(b) one or more periplasmic space glycosidases, wherein said one or more glycosidases are capable of cleaving said one or more oligosaccharides; and wherein said bacterium can utilize said one or more oligosaccharides for growth.

The method of said sixth aspect, may further comprise oral administration of a prebiotic comprising said one or more oligosaccharides, either separately, sequentially or simultaneously to the administration of said recombinant bacteria. DESCRIPTION OF THE INVENTION Brief description of the figures:

Figure 1: Cartoon showing sectional view of a recombinant probiotic gram-negative bacterium comprising an outer membrane transporter protein facilitating diffusion (e.g. passive diffusion) and a secreted glycosidase facilitating the cleavage of a prebiotic oligosaccharide transported into the periplasmic space.

Figure 2: Cartoon showing the genetic components of a pME4.1 plasmid, wherein the sequence encoding the torA signal peptide is fused to a melD gene, which encodes an a-(l 2)- and a-(l 3)-glycosyl hydrolase (EC 3.2.1.10; GH13 glycoside hydrolase family); rafY encodes a glycoporin; pMUTl encodes the plasmid backbone from the native pMUTl plasmid.

Figure 3: Cartoon showing the genetic components of a pFU2.3 plasmid, wherein the sequence encoding the torA signal peptide is fused to a blon_2336 gene, which encodes an a-(l 3)-fucosidase (EC3.2.1.111 GH29 glycoside hydrolase family); ra/Y encodes a glycoporin; pMUTl encodes the plasmid backbone from the native pMUTl plasmid.

Figure 4: Graph showing % plasmid retention in three replicate cultures of E. coli Nissle strains, EcN_MUTl and EcN_ME4.1, comprising, respectively, (A) pMUTl plasmid and (B) pME4.1 plasmid during serial passaging in non-selective LB medium at 24h intervals over a growth period of approximately 100 cell generations. Plasmid retention was evaluated based on CFU counts on non-selective LB medium in comparison to cultivation on selective LB medium supplemented with antibiotics.

Figure 5: Histogram showing the A) growth rate and B) maximum optical density of E. coli Nissle strains, EcN_MUTl and EcN_ME4.1 comprising respectively pMUTl plasmid or pME4.1 plasmid during cultivation on M9 minimal medium supplemented with 1% w/v glucose (Glu) or 1 % w/v melezitose (Mel).

Figure 6: Histogram showing the A) growth rate and B) maximum optical density of E. coli Nissle strains, EcN_MUTl and EcN_FU2.3 comprising respectively pMUTl plasmid or pFU2.3 plasmid during cultivation on M9 minimal medium supplemented with 1 % w/v glucose (Glu) or 1 % w/v 3'-fucosyllactose (3'-FL). Figure 7: Graph showing the growth of E. coli Nissle strains comprising plasmids expressing different variants of the engineered circuit on M9 minimal media + 0.2 % w/v casamino acids + 0.5 % w/v melezitose added to the medium, demonstrating that a combination of both the TorA signal peptide and the RafY outer membrane porin enable the strain to grow efficiently on the rare oligosaccharide. Figure 8: Histogram showing (A) growth rate and (B) maximum optical density (OD) of E. coli Nissle strains, EcN_MUTl_Raf and EcN_ME4.1 comprising, respectively, pMUTl_RafY plasmid or pME4.1 plasmid during cultivation on M9 minimal medium supplemented with 1 % melezitose and a raffinose gradient ranging from 0 to 5 % w/v raffinose (Raf).

Figure 9: Histogram showing (A) growth rate and (B) maximum optical density (OD) of E. coli Nissle strains, EcN_MUTl_Raf and EcN_FU2.3 comprising respectively pMUTl_RafY plasmid or pFU2.3 plasmid during cultivation on M9 minimal medium supplemented with 1 % w/v 3'-fucosyllactose (3'-FL) and a raffinose gradient ranging from 0 to 5 % w/v raffinose (Raf).

Figure 10: Histogram showing (A) growth rate and (B) maximum optical density (OD) of E. coli Nissle strains, EcN_MUTl_Raf and EcN_ME4.1 comprising respectively pMUTl_RafY plasmid or pME4.1 plasmid during cultivation on M9 minimal medium supplemented with a melezitose gradient ranging from 0.05 to 5 % w/v melezitose (Mel).

Figure 11: Histogram showing (A) growth rate and (B) maximum optical density (OD) of E. coli Nissle strains, EcN_MUTl_Raf and EcN_FU2.3 comprising, respectively, pMUTl_RafY plasmid or pFU2.3 plasmid during cultivation on M9 minimal medium supplemented with a 3'-fucosyllactose gradient ranging from 0.05 to 5 % w/v 3'- fucosyl lactose (3'-FL).

Figure 12: Histogram showing (A) aerobic and (B) anaerobic maximum optical density (OD630) of E. coli Nissle strains EcNJ UTl and EcN_ME4.1 comprising respectively pMUTl_RafY plasmid or pME4.1 plasmid during cultivation on 0, 10, or 100 % v/v Gifu Anaerobe Media (GAM) in M9 minimal medium supplemented with 1 % w/v melezitose (Mel).

Figure 13: Histogram showing (A) aerobic and (B) anaerobic maximum optimal density (OD630) of E. coli Nissle strains, EcN_MUTl and EcN_FU2.3 comprising, respectively, pMUTl plasmid or pFU2.3 plasmid during cultivation on 0, 10, or 100 % v/v Gifu Anaerobe Media (GAM) in M9 minimal medium supplemented with 1 % w/v 3'- fucosyl lactose (3'-FL).

Figure 14: Graph showing the ratio of cells of the E. coli Nissle strain, EcN_FU2.3 comprising a pFU2.3 plasmid, relative to cells of the E. coli Nissle strain, EcN_MUTl comprising a pMUTl plasmid, during (A) aerobic or (B) anaerobic cultivation on either M9 minimal medium supplemented with 10 or 100 % v/v Gifu Anaerobe Media (GAM) and 1 % w/v 3'-fucosyllactose, over the course of 4 days with transfers to fresh media each day.

Figure 15: Graph showing growth curves (OD630) of cells of the E. coli Nissle strain, EcN: :FU2.3, comprising an integrated version of the FU2.3 pathway, relative to cells of the E. coli Nissle strain, EcN_FU2.3, comprising a pFU2.3 plasmid, and the E. coli Nissle strain, EcN_MUTl, comprising a pMUTl plasmid, during cultivation on M9 media supplemented with 1 % w/v 3'fucosyl lactose (3'-FL).

Figure 16: A) Cartoon showing the drinking regime for the mouse animal trial 1. Mice, divided into 3 groups (8 animals per group), were subject to the indicated drinking regime, and, at the end of the trial period, were euthanized and their faeces and gastrointestinal samples were collected. B) Histogram showing average water consumption for each group over the course of the control period (week 1) and differentiated drinking period (week 2) for mouse animal trial 1.

Figure 17: Cartoon showing the experimental setup for mouse animal trial 2. Mice, divided into 2 groups (8 animals per group), were fed a low-fiber (low MAC) diet 48 hours prior to exposure to bacteria. From day -1, streptomycin at 5 g/L and the indicated prebiotic concentrations were given in the drinking water. Animals were exposed to a 1: 1 mix of E. coli Nissle strain, EcN_FU2.3 comprising a pFU2.3 plasmid and E. coli Nissle strain, EcNJ UTl comprising a pMUTl plasmid. Figure 18A: Graph of CFU counts per gram of faeces of E. coli Nissle strains, EcN_MUTl and EcN_FU2.3 comprising, respectively, a pMUTl or pFU2.3 plasmid, for mouse animal trial 2 over the course of 5 days after administration of a 1: 1 mix of 10 ⁵ CFU of the two strains (with samples taken at day 1, 3, and 5) with no prebiotic added to the drinking water. Each point represents CFUs from one animal, while the lines indicate the geometric mean for each group and time point.

Figure 18B: Graph of CFU counts per gram of faeces of E. coli Nissle strains EcN_MUTl and EcN_FU2.3 comprising, respectively, a pMUTl or pFU2.3 plasmid, for mouse animal trial 2 over the course of 5 days after administration of a 1: 1 mix of 10 ⁵ CFU of the two strains (with samples taken at day 1, 3, and 5) with 5 % w/v 3'fucossyllactose (3'FL) added to the drinking water. Each point represents CFUs from one animal, while the line indicates the geometric mean for each group.

Figure 18C: Graph directly comparing CFU counts per gram of faeces of E. coli Nissle strains EcN_MUTl and EcN_FU2.3 comprising, respectively, a pMUTl or pFU2.3 plasmid, for mouse animal trial 2 over the course of 5 days after administration of a 1: 1 mix of 10 ⁵ CFU of the two strains (with samples taken at day 1, 3, and 5) with either 0% or 5 % w/v 3'-fucossyllactose (3'-FL) added to the drinking water. Each point represents the group geometric mean (CFUs).

Figure 19 A) Cartoon showing the construct of four genes needed for L-DOPA production, which have been integrated into the E. coli Nissle chromosome. B) Figure showing the measured production levels of L-DOPA compared to a negative control. C) Cartoon showing the experimental setup for mouse trial 3. Mice, divided into 4 groups (8 animals per group), were given differential dosing of prebiotic in the drinking water from day -1 to day 8. On day 0, animals were dosed with the therapeutic EcNcm2_FU2.3_dopa comprising a pFU2.3 plasmid and genes needed for production of the neurotransmitter L-DOPA in competition with EcN_MUTl comprising the pMUTl plasmid. D) Colonization levels (CFU/g faeces) for the therapeutic strain on day 1 to 8 at the different 3'-FL concentrations.

Figure 20: Cartoon showing the experimental setup for a planned mouse trial 4. Mice, divided into 4 groups (8 animals per group), are given differential dosing of prebiotic in the drinking water from day -1. On day 0, animals are dosed with a therapeutic E. coli Nissle strain comprising a pFU2.3 plasmid and BBIJac gene needed for production of an extracellular lactase originating from Bifidobacterium bifidum.

Abbreviations, terms and definitions:

Amino acid sequence identity: The term "sequence identity" as used herein, indicates a quantitative measure of the degree of homology between two amino acid sequences of substantially equal length. The two sequences to be compared must be aligned to give a best possible fit, by means of the insertion of gaps or alternatively, truncation at the ends of the protein sequences. The sequence identity can be calculated as ((Nref- Ndif)100)/(Nref), wherein Ndif is the total number of non-identical residues in the two sequences when aligned and wherein Nref is the number of residues in one of the sequences. Sequence identity between each protein and its homologues is calculated by the BLAST program e.g. the BLASTP program (Pearson W.R and D.J. Lipman (1988)) (www.ncbi.nlm.nih.gov/cgi-bin/BLAST). The identity matrix between any two proteins is determined by the sequence alignment method ClustalW with default parameters as described by Thompson J., et al 1994, available at http://www2.ebi.ac.uk/clustalw/.

Preferably, the numbers of substitutions, insertions, additions or deletions of one or more amino acid residues in the polypeptide as compared to its comparator polypeptide is limited, i.e. no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 substitutions, no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 insertions, no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 additions, and no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 deletions. Preferably the substitutions are conservative amino acid substitutions: limited to exchanges within members of group 1: Glycine, Alanine, Valine, Leucine, Isoleucine; group 2: Serine, Cysteine, Selenocysteine, Threonine, Methionine; group 3: proline; group 4: Phenylalanine, Tyrosine, Tryptophan; Group 5: Aspartate, Glutamate, Asparagine, Glutamine.

E. coli ToplO: E. coli having chromosomal Genotype mcrA, A(mrr-hsdRMS-mcrBC), Phi80lacZ(del)M15, AlacX74, deoR, recAl, araD139, A(ara-leu)7697, galU, galK, rpsL(SmR), endAl, nupG

GI number: (genlnfo identifier) is a unique integer which identifies a particular sequence, independent of the database source, which is assigned by NCBI to all sequences processed into Entrez, including nucleotide sequences from DDBJ/EMBL/GenBank, protein sequences from SWISS-PROT, PIR and many others.

Heterologous gene and heterologous DNA molecule: have a different genetic origin from the recombinant cell in which they are expressed; and this also applies to the transcript thereof. The nucleotide sequence of the heterologous gene or heterologous DNA molecule may be optimized (e.g. codon optimization) with respect to the recombinant cell in which they are expressed. Heterologous gene and heterologous DNA molecule may be located on (and therefore be a part of) the chromosome or a replicon (e.g. plasmid) of the recombinant cell, and may be inserted into this location by recombinant DNA cloning.

Heterologous protein: a protein that is produced in a host cell, where the produced protein differs from any protein that is native to the host cell in question.

Oligosaccharide is a saccharide polymer comprising a small number (typically two to ten or preferably three to ten) of monosaccharides (also called simple sugars).

Origin of replication: (ori) is a particular sequence in a plasmid at which replication is initiated, and which determines the plasmid copy number. Promoter: is a region of DNA that initiates transcription of a particular gene. Promoters are located near the transcription start sites of genes, on the same strand and upstream on the DNA (towards the 5' region of the sense strand). A promoter that is "functionally linked" to a gene is capable of driving expression of said gene. A promoter may drive expression constitutively or only in response to an inducer. Ribosomal Binding Site (RBS): also called the translation initiation region, or translational strength element, refers to a genetic region of the 5' untranslated region upstream of the start codon of a messenger RNA transcript that controls the translation strength of the respective mRNA; and is responsible for the recruitment of a ribosome during the initiation of protein translation. Ribosome Binding Site (RBS) Calculator: provides a method for either predicting or controlling translation initiation rates (TIRs) in bacteria. When used to control translation of a given coding sequence, the RBS calculator generates synthetic DNA sequences that will result in a defined translation initiation rate (and therefore, protein expression strength). For any promoter sequence (controlling transcription strength), many different RBS strengths can be designed to regulate the amount of protein produced from this transcript (Salis at al . , 2009).

RBS strength scale and units: The output values range from 1 to 1000000 on a linear scale with arbitrary units that were experimentally validated using fluorescent protein abundance as a measure of expression strength (Salis at al., 2009). Designed sequences of a particular strength are not unique, i.e. different nucleotide sequences can encode an RBS having the same RBS strength. Furthermore, RBS strength is always context- dependent, therefore a designed sequence of a defined strength is only valid for the coding sequence it was calculated for. The minimum nucleotide sequences defining RBS strength are 35 base pairs upstream and 50 base pairs downstream of the start codon (ATG) of a coding sequence.

Tat signal peptide is one that, when translationally fused to a protein (e.g glycosidase), directs its secretion via the Tat-dependent pathway into the periplasmic space of the bacterium in which it is expressed. A Tat signal peptide comprises a highly conserved twin-arginine leader motif (S/TRRXFLK) located at the boundary between the N-terminal n-region, and the central h-region of Tat-specific signal peptides. The following databases allow the identification of a Tat signal peptide: TatP (http://www.cbs.dtu.dk/services/TatP/),

TatFind (http://sianalfind.org/tatfind.html)· and PRED-Tat (http://www.compaen.ora/tools/PRED-TAT/) (Freudl et al 2018).

Transgene: a gene or genetic material that has been transferred naturally or by any of a number of genetic engineering techniques from one organism to another. The transgene that is transferred to the recipient can be from other individuals of the same species or even from unrelated species.

Detailed description of the invention:

The present invention provides a recombinant probiotic gram-negative bacterium capable of selective enrichment and controlled colonization density in the mammalian gut microbiome. The recombinant probiotic gram-negative bacterium exhibits an advantage in comparison to the competing gut microbes natively inhabiting the gut by virtue of its ability to exploit a rare carbon-source metabolic niche, and thereby specifically colonize the gut by out-competing the otherwise stable gut microbiome only when such an addition to the diet is introduced. Provided that the chosen carbon-source is rare in the human diet, the presence and density of the bacterium may be precisely regulated by the supply of an orally supplemented prebiotic comprising this rare carbon- source. The ability of the probiotic gram-negative bacterium of the invention to successfully compete with the native gut microbiome is also attributable to its energy- efficient genetic design that employs one or several periplasmic space glycosidase(s) to facilitate assimilation of the rare carbon-source.

Given the potential of probiotics to modulate the composition of the gut microbiome, the controlled engraftment of a specific probiotic gram-negative bacterium allows for an improved ability to engineer the gut microbiome for the prevention and treatment of diseases. Furthermore, the recombinant probiotic gram-negative bacterium provides a platform for the development of bacterial derivatives capable of producing therapeutics in a dose-dependent manner. As the bacterium of the invention and its drug-producing derivatives is capable of dose-dependent gut colonization, it is able to confer a controlled therapeutic advantage with a lower need for strain reintroduction. The various embodiments of the invention are described below:

I. Recombinant probiotic gram-negative bacterium

A first aspect of the invention provides a recombinant probiotic gram-negative bacterium comprising genes encoding: (a) an outer membrane oligosaccharide transporter, and

(b) one or more periplasmic space glycosidases, wherein said one or more glycosidases are capable of cleaving said one or more oligosaccharides and wherein said bacterium can utilize said one or more oligosaccharides for growth. As illustrated in Figure 1, the recombinant bacterium has an outer membrane protein capable of transporting oligosaccharides into the periplasmic space, which is encoded by a gene on the genome or replicon of the recombinant bacterium. This outer membrane transporter facilitates the entry of extracellular oligosaccharides into the cell, from the surrounding environment, by means of diffusion (e.g. passive diffusion). The recombinant bacterium further comprises gene(s) encoding at least one glycosidase that is targeted to and localized in the periplasmic space, where each glycosidase is capable of catalysing the cleavage of an oligosaccharide imported into the periplasmic space by means of the transporter. The recombinant bacterium is capable of utilizing the oligosaccharide cleavage products released in the periplasmic space to support its growth. As illustrated in Figure 1, the recombinant bacterium is capable of importing at least some of the small saccharides (in particular mono- and di-saccharides) released by the cleavage activity of the glycosidase across the inner membrane into the cell using endogenous inner membrane transporters, where they are shuttled into hexose fermentation pathways to support growth. Preferably, the recombinant bacterium is capable of utilizing the oligosaccharide(s) as the sole carbon source to support growth. The recombinant probiotic gram-negative bacterium provides a highly versatile vehicle for engineering cells that are capable of assimilating rare carbohydrates by means of the chosen outer membrane oligosaccharide transporter in combination with the chosen periplasmic glycosidase that facilitates cleavage of the assimilated rare carbohydrates into metabolisable sugars, to support growth. Thus, cells of the invention, engineered to grow on a given rare carbohydrate or combination of rare carbohydrates, are conferred with a selective growth advantage sufficient to displace the native microbiome and colonize niche(s) in the mammalian gut when provided with the respective rare carbohydrate(s).

The genes encoding the outer membrane oligosaccharide transporter and at least one periplasmic space glycosidase may individually be located on the chromosome of the recombinant probiotic gram-negative bacterium of the invention or on a replicon (e.g. a plasmid) in the bacterium. Preferably, the genes are either located on the chromosome or on a plasmid of the recombinant bacterium. When the genes are located on a plasmid, the choice of ori conferring either a high or low copy number can advantageously be used to obtain up- or down-regulated expression levels of the respective genes. Furthermore, promoter sequence(s) are operatively linked to the genes to drive their expression in the recombinant bacterium, where the selected promoters may be the same or different. In one embodiment, the promoter sequence(s) drive constitutive expression of the genes, where the promoter sequence(s) are ones capable of driving expression of the encoded proteins in sufficient amount to support cell growth on the oligosaccharide when provided as a sole carbon source. A suitable constitutive promoter includes the Tet promoter [SEQ ID No.: 1], the pTac promoter [SEQ ID NO. 2], and the pMSKL7 promoter [SEQ ID NO. 3]. The expression of an operatively linked gene can be further modulated by selecting an RBS of the desired strength.

In one embodiment, the promoter sequence(s) drive expression of the genes under low oxygen conditions (hypoxia), as is present in the mammalian gut, where the promoter sequence(s) are ones capable of driving expression of the encoded proteins in sufficient amount to support cell growth on the oligosaccharide when provided as a sole carbon source under hypoxic conditions. A suitable hypoxia-driven promoter includes PfnrS [SEQ ID NO: 4]

The recombinant probiotic gram-negative bacterium according to the invention is a live bacterium considered generally safe to consume, and preferably not antibiotic resistant to one or more clinically used antibiotic agents. A non-exhaustive list of suitable bacteria includes members of a phylum selected from the group: Proteobacteria, Bacteriodetes, Firmicutes and Verrucomicrobia,- and preferably are species of a genus selected from the group: Escherichia, Bacteroides, Akkermansia, Alistipes, Prevotella, Parabacteroides, Odoribacter, Enterobacter, Klebsiella, Citrobacter, and Pseudobutyrivibrio .

By way of example, a suitable species of Escherichia includes E. coli,· a suitable species of Bacteroides includes B. fragilis; B. uniformis, B. ovatus, B. stercoris and B. thetaiomicrorr, a suitable species of Akkermansia includes A. muciniphila,· a suitable species of Alistipes includes A. shahii, A. finegoldii and A. indistinctus; a suitable species of Prevotella includes P. copri ; a suitable species of Parabacteroides includes P. goldsteinii,· and a suitable species of Odoribacter includes 0. splanchnicus,· a suitable species of Pseudobutyrivibrio includes P. xylanivorans.

In a preferred embodiment, the recombinant probiotic gram-negative bacterium is E. coli, in particular E. coli Nissle, since members of this species have the added advantage of being easily engineered. In particular, the strain E. coli Nissle is a well-characterized probiotic that is classified as a risk group I organism. Furthermore, E. coli Nissle is known to be a commensal bacterium that is adept at surviving in the gut environment, in particular the upper colon (Grozdanov, 2003). At the same time, although strains of E. coli are typically cleared after a few days, the duration of colonization of recombinant probiotic gram-negative bacterium of the invention (e.g. recombinant E. coli strains of the invention), can be prolonged and controlled by administration and withdrawal of a prebiotic comprising the oligosaccharide as defined herein.

Embodiments of the recombinant probiotic gram-negative bacterium of the invention are detailed below:

Ii. The outer membrane oligosaccharide transporter

In one embodiment, the outer membrane oligosaccharide transporter of the recombinant probiotic gram-negative bacterium of the invention is a porin. Porin is a protein that spans the outer membrane of the gram-negative bacterium to form a passive channel, or pore, through which oligosaccharides and other small molecules can diffuse. Preferably, the porin is a glycoporin capable of providing a channel for oligosaccharide import, wherein said oligosaccharides comprise a small number (typically two to ten, preferably three to ten) of monosaccharides.

A suitable glycoporin includes the RafY single porin protein, which provides a channel for import of oligosaccharides having a range of sizes, including the trisaccharide: raffinose, the tetrasaccharides: maltotetraose and stachyose, the hexasaccharide: maltohexaose and the heptasaccharide: maltoheptaose, as well as several disaccharides. More specifically, growth of the recombinant bacterium of the invention on the trisaccharides, melezitose and 3'-fucosyllactose is herein shown to be facilitated by the expression of a RafY porin (Example 3) that provides an import channel for these trisaccharides.

In one embodiment thereof, the amino acid sequence of the RafY glycoporin encoded by a gene in said recombinant bacterium has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 6.

An alternative suitable glycoporin is the ScrY sucrose porin, which also provides a channel having broad specificity for import of oligosaccharides, including sucrose, raffinose and maltooligo-saccharides. Accordingly, in one embodiment, the primary amino acid sequence of the ScrY sucrose porin encoded by a gene in said bacterium has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 8.

An alternative suitable glycoporin is the LamB maltodextrin porin, which also provides a channel having specificity for import of maltose oligosaccharides up to maltoheptose. Accordingly, in one embodiment, the primary amino acid sequence of the LamB maltodextrin porin encoded by a gene in said bacterium has at least 70, 71, 73, 74, 75,

76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97,

98, 99 or 100 % sequence identity to SEQ ID NO: 10. An alternative suitable glycoporin is the OprB carbohydrate porin from Pseudomonas putida, which also provides a channel having specificity for import of simple sugars as well as maltotriose, sucrose and lactose. Accordingly, in one embodiment, the primary amino acid sequence of the OprB carbohydrate porin encoded by a gene in said bacterium has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ

ID NO: 12.

The N-terminal amino acid sequence of the oligosaccharide transporter comprises a signal peptide that facilitates secretion of the transporter across the inner membrane, leading to removal of the signal peptide, after which the transporter inserts into the outer membrane of the recombinant probiotic gram-negative bacterium to form a pore.

Iii. The oligosaccharides imported by the oligosaccharide transporter The one or more oligosaccharides, according to the present invention, are those whose import into the recombinant probiotic gram-negative bacterium is facilitated by the oligosaccharide transporter (e.g. glycoporin); and that are both cleavable by the one or more periplasmic space glycosidases, and capable of providing a carbon source to support growth of the bacterium.

The one or more oligosaccharides, according to the present invention, are those that confer the bacterium with a selective growth advantage in the mammalian gut sufficient for the bacterium to colonize niche(s), while not being harmful to the commensal microbiome and is preferably also one that: (1) is rare in the human diet (2) remains largely un-degraded after passage through the upper gastrointestinal tract; and (3) is safe for mammalian consumption in therapeutically relevant amounts. When the oligosaccharide is provided for oral administration to a mammal (e.g. human), such as in the form of a prebiotic, the oligosaccharide needs to be well-tolerated by humans, so as to limit toxicity and adverse events.

Preferably the one or more oligosaccharides comprises a small number (typically two to ten; preferably three to ten) of monosaccharides, capable of import by the glycoporin in the outer cell membrane. By way of example, a suitable oligosaccharide is one present in human milk, or a derivative of said oligosaccharide, that can support the growth of a gram-negative probiotic bacterium of the invention engineered to express a porin and said one or more glycosidases that together facilitate import and cleavage of the one or more oligosaccharides into simple sugars capable of metabolism by the bacterium.

Accordingly, a suitable oligosaccharide includes one selected from fructo-, galacto, gluco, xylo-, manno-, and arabino-oligosaccharides or seaweed-derived oligosaccharides such as oligo-alginate, -fucoidan, -lamarin, -xylan, -agar, - carreageenan, and porphyrin- or fucosylated- or sialylated-oligosaccharides.

By way of example, a suitable oligosaccharide includes the trisaccharide, 3'- fucosyl lactose, a-L-Fuc-(l 3)-[ -D-Gal-(l 4)]-D-Glc [CAS Number 41312-47-4], whose cleavage yields galactose, glucose, and fucose. 3'-fucosyllactose is found in human milk, and thus safe for consumption. Importantly, 3'-fucosyllactose is a rare carbohydrate in adult diets, and correspondingly any members of the adult mammalian microbiome capable of degrading this trisaccharide would not be enriched prior to its administration. Alternatively, a suitable oligosaccharide includes the trisaccharide 2'-fucosyllactose, a- L-Fuc-(l 2)-p-D-Gal-(l 4)-D-Glc [CAS Number 41263 94 91 whose cleavage yields galactose, glucose, and fucose. Since 2'-fucosyllactose is also found in human milk, it is safe for consumption. Further since 2'-fucosyl lactose is a rare carbohydrate in adult diets, any members of the adult mammalian microbiome capable of degrading this trisaccharide would also not be enriched prior to its administration. Alternatively, a suitable oligosaccharide includes the trisaccharide, Lacto-neo-N-tetraose (LnNT), -D-Gal-(l 4)- -D-GlcNAc-(l 3)- -D-Gal-(l 4)-D-Glc [CAS Number 13007 32 41 whose cleavage yields galactose, glucose, and N-Acetylglucosamine. LnNT is found in human milk, and thus safe for consumption. Further since, LnNT is a rare carbohydrate in adult diets, any members of the adult mammalian microbiome capable of degrading this trisaccharide would not be enriched prior to its administration.

Alternatively, a suitable oligosaccharide includes the trisaccharide, Melezitose, a-D-GIcp- (l®3)-p-D-Fruf-(2®l)- a-D-GIcp, [CAS Number: 10030-67-8], whose cleavage yields two glucose molecules and one fructose molecule, and is safe for consumption. In nature, it is found in honeydew or manna produced by many plant sap-eating insects. Melezitose is uncommon in the adult diet, and correspondingly any members of the adult mammalian microbiome capable of degrading this trisaccharide would not be enriched prior to its administration.

Alternatively, a suitable oligosaccharide includes the trisaccharide, erlose, a-D-GIc- (l 4)-a-D-Glc-(l 2)-8-D-Fru [CAS number 13101-54-71. which is a glucosylsucrose found in honey dew consisting of sucrose and having an alpha-D-glucopyranosyl residue attached at the 4-position of the glucose ring.

Alternatively, a suitable oligosaccharide includes the trisaccharide 1-kestose, b-D- fructofuranosyl-(2 l)-p-D-fructofuranosyl a-D-glucopyranoside [CAS Number 470-69- 9], whose cleavage yields fructose and glucose. 1-kestose is found in low amounts in many fruits and vegetables, and is thereby rare in the adult diet in large amounts.

Alternatively, suitable oligosaccharides include those derived from seaweed polysaccharides. Sodium alginate [CAS Number 9005-38-3] can be treated either enzymatically (with alginate lyase) or via acid hydrolysis to yield Alginate Oligosaccharides (AOS) in accordance with the protocol described by Wang 2006. The degradation product of alginate has been shown to stimulate growth of Bifidobacterium bifidum ATCC 29521 and Bifidobacterium longum SMU 27001. Therefore, suitable enzymatic activities can be isolated from these species as has been shown for 3'fucosyl lactose and melezitose. Iiii. The periplasmic space giycosidase

The one or more glycosidases, also known as glycoside hydrolases (GH) [EC 3.2.1.], are enzymes that catalyse the hydrolysis of glycosidic bonds between sugar moieties or between sugar and non-sugar moieties in complex sugars including oligosaccharides, in the periplasmic space of the recombinant probiotic gram-negative bacterium, and thereby releases sugars that can be transported across the inner membrane by endogenous transporters and shuttled into cellular hexose fermentation pathways to support growth. Hydrolysis of the oligosaccharides, imported via the outer membrane oligosaccharide transporter, in the periplasmic space by the one or more glycosidases avoids the necessity and energy burden of expressing an additional tailored oligosaccharide transporter for their transport across the inner membrane. Additionally, hydrolysis in the periplasm provides an advantage of increased substrate exclusivity which would not be the case were the glycosidase(s) to be secreted to the extracellular space.

Said glycoside hydrolases (EC: 3.2.1] comprise a wide spread group of enzymes whose members are classified into 160 members (http://www.cazv.org/Glvcoside- Hydrolases. f https://www.amul.ac.Uk/sbcs/iubmb/enzvme/EC3/2/l/ ^' ).

In one embodiment the one or more glycosidases is an a-glucosidase. Suitable a- glucosidases include those capable of cleaving a-(l®3) and a-(l®2) glucosidic bonds in melezitose, as well as in isomaltulose, turanose, maltose, sophorose and xylobiose to yield monosaccaride moeities. Additional suitable a-glucosidases include those capable of cleaving a-(l®6) glucosidic bonds in panose, isomaltotriose, as well as isomaltose, isomaltulose (palatinose), trehalulose, turanose, maltulose. Additional suitable a- glucosidases include those capable of cleaving a-(l®4) glucosidic bonds in sucrose xylobiose, and erlose.

In a preferred embodiment, the glycosidase encoded by a gene in said bacterium exhibits a-(l®3) and a-(l®2) glucosidase activity and cleaves melezitose, wherein the amino acid sequence of said mature glycosidase polypeptide has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 14.

In one embodiment the glycosidase is an a-fucosidase. Suitable a-fucosidases include those capable of cleaving a-(l®3) fucosidic bonds as found in 3'-fucosyllactose to yield fucose and lactose. Additional suitable a-fucosidase include those capable of cleaving a-(l®2) fucosidic bonds as found in 2'fucosyllactose to yield fucose and lactose.

In a preferred embodiment, the glycosidase encoded by a gene in said bacterium is an a-fucosidase that exhibits a-(l®3) fucosidase activity and cleaves 3'-fucosyl lactose, wherein the amino acid sequence of said mature a-fucosidase polypeptide has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 16.

In a preferred embodiment, the glycosidase encoded by a gene in said bacterium is an a-fucosidase that exhibits a-(l®2) fucosidase activity and cleaves 2'-fucosyl lactose, wherein the amino acid sequence of said mature a-fucosidase polypeptide has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 18.

In one embodiment the glycosidase is a b-fructofuranosidase. Suitable b- fructofuranosidases include those capable of cleaving inulin-type b-(2 1) bonds as found in 1-kestose to yield fucose and glucose. In a preferred embodiment the b- fructofuranosidase is derived from from Lactobacillus piantarum, wherein the amino acid sequence of said mature b-fructofuranosidase polypeptide has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 20. The gene(s) in the recombinant probiotic gram-negative bacterium, for said one or more glycosidases, each encode an N-terminal heterologous signal peptide sequence that is translationally fused to the respective glycosidase, whose expression yields a primary translation product. The signal peptide facilitates secretion of the glycosidase across the inner membrane, leading to removal of the signal peptide, such that the mature glycosidase (following signal peptide removal) is localized in the periplasmic space.

In a preferred embodiment, the N-terminal heterologous signal peptide comprises a highly conserved twin-arginine leader motif (S/TRRXFLK), such that secretion of the glycosidase in said bacterium is mediated by a Tat-dependent pathway. This pathway facilitates cytoplasmic folding of the glycosidase prior to secretion. Since the catalytic activity of the cytoplasmic-folded glycosidase is significantly higher than when the glycosidase is secreted in an unfolded state via the general secretory (Sec) pathway, the Tat-dependent route provides a more energy efficient expression system (Example 3B). In a recombinant probiotic gram-negative bacterium of the invention, the Tat- dependent pathway is mediated by a family of native Tat translocases, each composed of three essential membrane proteins; TatA, TatB, and TatC.

In one embodiment the signal peptide is the TorA signal peptide, wherein the amino acid sequence of the peptide comprises the motif S/TRRXFLK and has at least 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO: 21. Iiv Plasmid based expression of genes in a recombinant probiotic gram-negative bacterium

In one embodiment, the genes in the recombinant probiotic gram-negative bacterium of the invention, encoding the outer membrane oligosaccharide transporter and one or more periplasmic space glycosidase, can be located on a plasmid, where they are operatively linked to a promoter. A suitable plasmid is one comprising an ori that is compatible with replication in the chosen gram-negative bacterium, and optionally a selectable reporter gene operably linked to a promoter, to monitor plasmid maintenance in the recombinant bacterium of the invention.

In a preferred embodiment, the plasmid is a native plasmid of the probiotic gram negative bacterium of the invention. By way of example, a suitable native plasmid is the pMUTl or pMUT2 plasmids of E. coli Nissle, which are known to be stable (see examples).

Since expression of the genes in the recombinant probiotic gram-negative bacterium of the invention places a burden on the bacterium, there is a risk that non-expressed gene variants will arise through mutation. When the genes are located on plasmids in the bacterium, there is a risk of plasmid loss. Since long-term plasmid stability is advantageous for the various therapeutic applications of the recombinant probiotic gram-negative bacterium, means for enhancing plasmid stability, known to the skilled person, are envisaged. By way of example, plasmid stability can be enhanced by introducing genes encoding a toxin-antitoxin system into the plasmid. A toxin-antitoxin gene cassette may be cloned into the plasmid, to prevent survival of cells lacking the plasmid. Alternatively, a gene, essential for the survival of the bacterium may be inserted into the plasmid and operably linked to a promoter. Provided that the sole copy of this essential gene is located on the plasmid, then any cell losing the plasmid will be unable to proliferate. Suitable essential genes are those whose products directly or indirectly are essential for bacterial cell growth, irrespective of the growth medium or growth conditions under which the cell is cultivated - such as proteins involved in cell- wall synthesis or structural protein components of the cell wall (Goodall et al ., 2018; Chen 2015; Mori et al. 2015)

Iv. Deletions of native carbohydrate metabolism in the recombinant probiotic gram negative bacterium In order to increase dependency of the recombinant probiotic gram-negative bacterium of the invention on the delivered prebiotics in the diet, its native carbon metabolism can be limited.

In one embodiment the recombinant probiotic gram-negative bacterium is impaired in synthesis of the glucose storage molecule glycogen. Glycogen is synthesized during growth in excess carbon, and provides storage of energy for periods of starvation. Impairment of glycogen biosynthesis leads to increased sensitivity to carbon starvation, thereby increasing the cell's dependency on a continuous flow of carbohydrates from the environment for continued growth. By way of example, the gene encoding glycogen synthase glgA [SEQ ID NO: 22] can be deleted from the E. coli Nissle 1917 chromosome (Example 15). In one embodiment the recombinant probiotic gram-negative bacterium is impaired in the utilization of mucin-derived carbohydrates. E. coli Nissle 1917 cannot release usable carbohydrates from mucin, but relies on anaerobic members of the microbiota to perform the initial release. E. coli Nissle 1917 can then utilize these mucin-derived carbohydrates when colonizing the mucosal layer in the mammalian gut (Chang 2004). Impaired mucin-derived carbohydrate metabolism increases the cell's dependency of diet-derived carbon sources, such as the rare oligosaccharide, for efficient growth in the mammalian gut. By way of example, the gene encoding N-acetylneuraminate lyase nanA [SEQ ID NO: 23] has been deleted from the E. coli Nissle chromosome. NanA catalyses the formation of pyruvate and N-acetylmannosamine (ManNAc) from N-acetylneuraminic acid (Neu5Ac) (Example 16).

II. A composition or kit comprising the recombinant probiotic gram-negative bacterium of the invention A second aspect of the invention provides a composition or kit comprising the live recombinant probiotic gram-negative bacterium according to the invention (as described in section I) and at least one of said oligosaccharides.

In one embodiment, the composition comprises said live recombinant probiotic gram negative bacterium in combination with the one or more oligosaccharides, e.g., 3'- fucosyl lactose, (3'-FL), 2'-fucosyllactose (2'-FL), Lacto-neo-N-tetraose (LnNT), melezitose (Mel), 1-kestose, or erlose, where the selected oligosaccharide may comprise 0.01% to 10% by weight of the composition, e.g., 0.01 % to 1.00 %, 0.05 % to 2.00 %, 0.10 % to 5.00 %, 0.50 % to 10 %, 1.00 % to 5.00 %, or 5.00 % to 10.00 % by weight of the composition. In one embodiment, the kit provides a first container comprising said live recombinant probiotic gram-negative bacterium and a second container comprising the one or more oligosaccharides, where the oligosaccharide(s) may comprise between 0.01 g and 10 g oligosaccharide per 10 grams of composition, where the oligosaccharide(s) may, for example, be one or more of 3'-fucosyllactose, (3'-FL) 2'-fucosy I lactose (2'-FL), Lacto- neo-N-tetraose (LnNT), melezitose (Mel), 1-kestose, and erlose. For example, the prebiotic composition may comprise between 0.1 g and 5 g, between 0.5 g and 5 g, between 1 g and 5 g, or between 1.5 g and 3 g of any one of 3'-fucosyllactose, (3'-FL) 2'-fucosyl lactose (2'-FL), Lacto-neo-N-tetraose (LnNT), melezitose (Mel), 1-kestose, and erlose per 10 grams of composition. Alternatively, the oligosaccharide(s) (selected from 3'-fucosyl lactose, (3'-FL) 2'-fucosyllactose (2'-FL), Lacto-neo-N-tetraose (LnNT), melezitose (Mel), 1-kestose, and erlose, or a combination thereof) may comprise 90% to 100% by weight of the composition, e.g., 92% to 100%, 95% to 100%, 96% to 100%, 97% to 100%, 98% to 100%, or 99% to 100% by weight of the composition.

III. A pharmaceutical composition or kit for use as a medicament, comprising a live recombinant probiotic gram-negative bacterium of the invention

A third aspect of the invention provides a pharmaceutical composition for use as a medicament, comprising a live recombinant probiotic gram negative bacterium as described in section I, wherein said pharmaceutical composition is for oral administration to an animal in need thereof. The animal can be any mammal, e.g., a human, a primate, a mouse, a rat, a dog, a cat, a horse, as well as livestock or animals grown for food consumption, e.g., cattle, sheep, pigs, chickens, and goats. In a preferred embodiment, the mammal is a human.

A fourth aspect of the invention provides a pharmaceutical kit for use as a medicament, comprising the pharmaceutical composition and a prebiotic comprising the oligosaccharide according to the invention as described in sections I and II, wherein said recombinant probiotic gram-negative bacterium is capable of growth on said prebiotic as sole carbon source, and wherein said prebiotic is for oral administration either separately, sequentially or simultaneously to the administration of said pharmaceutical composition to an animal ( as defined in the third aspect) in need thereof. Furthermore, prolonged administration and controlled withdrawal of the prebiotic comprising the oligosaccharide (as defined herein) to a subject to whom the recombinant probiotic gram-negative bacterium is administered, provides a means for maintaining and controlling colonisation and/or therapeutic activity of said recombinant gram-negative probiotic bacterium in the subject.

Administration of a pharmaceutical composition comprising the live recombinant probiotic gram-negative bacterium and the separate, sequential or simultaneous administration of a prebiotic comprising the oligosaccharide of the invention facilitates a robust colonization of a subject's gut to whom the composition has been administered, which in turns facilitates the therapeutic treatment provided by the live recombinant gram-negative probiotic bacterium. As demonstrated herein, the gram-negative probiotic bacterium of the invention is capable of robust growth on a rare carbohydrate, both as sole carbon source (Example 3) and as a component of a mixture of carbohydrates (Example 6). Furthermore, the gram-negative probiotic bacterium of the invention is capable of competitive and prebiotic dose-dependent growth, such properties being essential for regulating its niche colonization in the gut of a subject (Example 7). In preferred embodiments, the pharmaceutical composition comprising the live recombinant gram-negative probiotic bacterium, for separate, sequential or simultaneous administration with a prebiotic comprising the oligosaccharide of the invention, is for use in the prevention or treatment of a bowel-related disease such as enteric pathogen infections of the gut, diarrhoea, constipation, Inflammatory Bowel Disease (IBD), Irritable Bowel Syndrome (IBS), colitis, Ulcerative Colitis (UC), Crohn's Disease, diverticulosis, diverticulitis and toxic catabolites in the gut. In one example, the subject is an animal (e.g. mammal) in need of treatment, e.g., a mammal that has been diagnosed with infectious diarrhoea, acute infectious diarrhoea, antibiotic-associated diarrhoea (AAD), traveler's diarrhoea (TD), necrotizing enterocolitis (NEC), inflammatory bowel disease, Helicobacter pylori infection.

As demonstrated herein, the gram-negative probiotic bacterium of the invention is capable of robust and competitive growth in the gut microenvironment, as shown in a synthetic gut medium (Example 8 and 9). Furthermore, the gram-negative probiotic bacterium of the invention is shown to colonize the mammalian gut when, and so long as, it is co-administered with prebiotic comprising the respective oligosaccharide (Examples 11 to 13). As a result, it is able to establish a stable population in the gut as well as displacing a pathogenic microbiome from a subject's gut; or reduce inflammatory bowel diseases by displacing the subject's endogenous microbiome and thereby reducing abnormal immune responses in the gut. The therapeutic efficacy of the probiotic bacterium of the invention, for prevention or treatment of the aforementioned diseases is consistent with reported benefits of administering probiotics, found in nature, in the prevention or treatment of diarrhoea, acute diarrhoea in children, IBS, IBD and UC (Islam, S., 2016). In particular, the therapeutic efficacy of administering probiotic bacterium of the invention in combination with its respective prebiotic oligosaccharide is consistent with reported benefits of administering probiotics and symbiotic prebiotics in the treatment of UC (Asto et a I . , 2019). The properties of probiotics that underlie their therapeutic uses in treatment of these diseases has been attributed to their ability to positively modulate the intestinal epithelial barrier formed by intestinal epithelial cells (IECs) and intercellular junctions which are severely compromised in these diseases (Ukena et a I . , 2007)

IV. A pharmaceutical composition or kit for use as a medicament, comprising a live recombinant gram-negative probiotic bacterium capable of producing a therapeutic polypeptide or therapeutic molecule

A fourth aspect of the invention provides a pharmaceutical composition or kit for use as a medicament, according to the invention, wherein the live recombinant probiotic gram-negative bacterium of the pharmaceutical composition further comprises one or more transgenes encoding:

(a) a therapeutic polypeptide, or

(b) one or more enzymes for synthesis of a therapeutic molecule. The strategy disclosed herein for facilitating robust colonization by a probiotic bacterium of the invention can facilitate a range of therapeutic applications, where the ability to establish and control levels of colonisation in the gut by the administered engineered therapeutic recombinant probiotic gram-negative bacterium in a subject is important. Furthermore, prolonged administration and controlled withdrawal of the prebiotic comprising the oligosaccharide (as defined herein) to a subject to whom the recombinant probiotic gram-negative bacterium is administered, provides a means for maintaining and controlling colonisation and/or therapeutic activity of said recombinant gram negative probiotic bacterium in the subject; wherein said therapeutic activity is at least partly or fully due to the therapeutic polypeptide or therapeutic molecule produced by said bacterium.

As an example, this may be important when attempting to, for instance, change the short chain fatty acid profile or reduce the accumulation of harmful chemicals that are produced in the gut. Achieving controlled levels of colonization can also be important for therapeutics that must be carefully dosed, such as bacteria with pro- or anti- inflammatory activities, which could be used in the treatment of Inflammatory Bowel Diseases, or therapeutics where there is a maximum threshold for tolerance to the therapeutic that must not be reached, as is the case for treatment of e.g. Parkinson's disease and diabetes.

Diseases and disorders that can be treated in a subject with the engineered therapeutic recombinant probiotic gram-negative bacterium include, but are not limited to, diseases or disorders that are impacted by the gut microbiota, which include obesity, diabetes, heart disease, central nervous system diseases or disorders, autoimmune-related disorders (e.g. rheumatoid arthritis), metabolic disorders, CNS related disorders and cancer. For example, in some cases, the individual has gut inflammation, and in some such cases the individual has an inflammatory disease (e.g., Crohn's disease, ulcerative colitis, and the like), and in some cases gut inflammation can indirectly impact the disease, such as colorectal cancer or obesity.

In one embodiment, the expressed therapeutic peptide is a digestive enzyme leading to the degradation of non-tolerated carbohydrates in the diet. By way of example, the therapeutic polypeptide is beta-galactosidase, which catalyses the degradation of lactose into glucose and galactose, and thereby relieves symptoms of lactose intolerance (whose production by recombinant probiotic bacteria of the invention can be mediated by expression of a beta-galactosidase gene described in example 14.

In one embodiment, the expressed therapeutic peptide is the human hormone TFF3 affecting appetite, host metabolism, host gut physiology, or other host mechanisms, which promotes mucosal healing by promoting gut epithelial integrity thereby combatting the effects of inflammatory bowel disease, fistulae and ulcers (Praveschotinunt et al . , 2019); or the therapeutic polypeptide is insulin for the treatment of diabetes.

In one embodiment, the synthesized therapeutic molecule is a human neurotransmitter capable of crossing the Blood-Brain-Barrier (BBB), and modulating the central nervous system of the host, such as Dopamine (L-DOPA); Serotonin; Norepinephrine; Epinephrine; GABA and vitamin B9. Secretion of these molecules by an administered recombinant gram-negative probiotic bacterium of the invention can be used to prevent and/or treat disorders or diseases related to treatment-resistant depression, Parkinson's disease, Alzheimers Pterin deficiency, and Cerebral folate deficiency. or other neurological disorders. In some cases, the therapeutic molecule is 3,4-Dihydroxy-L- phenylalanine (L-DOPA) (whose production by recombinant probiotic bacteria of the invention can be mediated by expression of an L-DOPA synthesis pathway genes for use in treatment of Parkinson's disease or dopamine-responsive dystonia. The expression of the following four enzymes: a tyrosine hydroxylase (EC: 1.14.16.2), which catalyses the conversion of L-tyrosine into L-DOPA in the presence of tetrahydrobiopterin; a GTP cyclohydrolase I (EC:3.5.4.16) and a Dihydromonapterin reductase EC: 1.5.1.50 which catalyse the synthesis of tetrahydromobipterin; and a pterin recycling enzyme EC:4.2.1.96, in a recombinant gram-negative probiotic bacterium of the invention confers the bacterium with the ability to produce L-DOPA (Example 13).

Suitable enzymes for L-DOPA synthesis pathway include:

(1) a tyrosine hydroxylase (EC: 1.14.16.2), preferably a non-rate-limited enzyme such as one lacking a functional "control domain" of catalytic activity located in the N-terminal region e.g. one lacking the first about 150 amino acids of the native enzyme, as is the case where the amino acid sequence of said enzyme has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID NO:44; that is derivable from the native rat tyrosine hydroxylase (Uniprot: P04177) by substitution of the first 157 amino acid residues with MK. (2) a GTP cyclohydrolase I (EC:3.5.4.16), where the amino acid sequence of said enzyme has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87,

88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID

NO: 48, having a (T198I) substitution; (3) a Dihydromonapterin reductase EC: 1.5.1.50, where the amino acid sequence of said enzyme has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87,

88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID

NO: 50; and

(4) a pterin recycling enzyme EC:4.2.1.96, where the amino acid sequence of said enzyme has at least 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87,

88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99 or 100 % sequence identity to SEQ ID

NO: 46.

Genes encoding the L DOPA synthesis pathway, are operably linked to promoters and RBS sequences known to the skilled person in order to provide a suitable level of expression. In some cases, the therapeutic molecule is b-carotene, a precursor of the essential vitamin A (Miller 2013), used in prevention and or treatment of blindness, xerophthalmia, susceptibility to infections, impaired immune function, cancer, and birth defects. EXAMPLES

Example 1: Engineering a rare carbohydrate metabolic pathway in a probiotic gram-negative bacterium

1.1 Plasmid constructs: The plasmids pME4.1 (Figure 2) and pFU2.3 (Figure 3) comprise genes conferring the ability to assimilate and metabolize two trisaccharides, melezitose and 3'-fucosyl lactose that are rarely found in mammalian diet. Both plasmids comprise a backbone derived from the native E. coli Nissle 1917 plasmid, pMUTl, by removal of its mobility elements; having a native pMUTl origin of replication (Ori) and additionally comprising a kanamycin resistance gene. Both plasmids further comprise a rafY gene encoding a porin, namely a raffinose transporter native to certain E. coli strains, cloned from an E. coli clinical isolate.

Specifically, pME4.1 further comprises the coding sequence of the TorA signal peptide fused to the coding sequence of the melD gene derived from the melABCDE operon of Bifidobacterium breve UCC2003 (O'Connell, et a I . , 3013), where the fused gene expresses a secreted alpha-glucosidase capable of cleaving alpha-l->3 and alpha-l->2 glucosidic linkages, and thereby degrading melezitose into glucose and fructose.

Specifically, pFU2.3 further comprises the coding sequence of the TorA signal peptide fused to the coding sequence of the blon_2336 gene derived from Bifidobacterium iongum subspecies infantis ATCC 1597, where the fused gene expresses a secreted alpha-fucosidase capable of cleaving alpha l->3 linkages and thereby degrading 3'- fucosyl lactose into fucose and lactose.

The plasmids pME4.1 and pFU2.3 comprise the TetA promoter driving expression of both carbohydrate assimilation and metabolism genes.

The plasmids pMEO.1-5 are less- or non-functional variants of pME4.4, which were constructed to demonstrate the need for a TorA secretion tag allowing for cytoplasmic folding as well as the RafY outer membrane porin. All genes are under control of the pTet promoter. pMEO.l contains the melD gene without a signal peptide coding sequence to guide secretion, and thus only expresses cytoplasmic MelD. pME0.2 contains a melD gene fused to the MalE secretion tag coding sequence, where the secretion tag transports unfolded peptides for periplasmic folding. pME0.3 contains the melD gene fused to the TorA secretion tag coding sequence, where the tag allows for cytoplasmic folding prior to translocation to the periplasm. pME0.4 is identical to pMEO.l except that it also comprises the RafY outer membrane porin gene. pME0.5 is identical to pME0.2, except that it also comprises the RafY outer membrane porin gene.

1.2 Cloning and transformation: Plasmids (see Table 1) were assembled using USER cloning in accordance with Nour-Eldin et al. (2010), inserting their respective genes, each gene being amplified using Phusion U Hot Start DNA Polymerase from Thermo Fisher Scientific using sequence specific Uracil primers.

The plasmids were first cloned and transformed into E. coli ToplO cloning strain and subsequently into E. coli Nissle 1917 (https://www.mutaflor.com/index.html) or EcN- GFP (a modified E. coli Nissle 1917 with a chromosomal point mutation conferring streptomycin resistance. The point mutation was introduced using single stranded lambda recombineering in accordance with Sharan (2009). In addition, this strain also carries a gene encoding Green Fluorescent Protein (GFP) engineered into the chromosome using a T7 system in accordance with McKenzie (2006). Clones were selected by growth on media supplemented with antibiotics corresponding to the respective plasmid antibiotic resistance genes. Correct cloning was verified by colony PCR, function, and/or sequencing. The E. coli Nissle strain EcN: :FU2.3 has the genes needed for uptake and metabolism of 3'fucosyl lactose integrated into the chromosome by means of the pOSIP integration system in accordance with St-Pierre et al. (2013). More specifically, the operon comprising the tet promoter, torA-Blon2336 fusion gene, and rafY gene where amplified with specific primers containing USER-enabled overhangs compatible to primers used to amplify the backbone of the pOSIP-KO plasmid, which comprise the elements needed for integration into a single site in E. coli Nissle 1917, and which co-integrates a kanamycin resistance marker. This marker can be removed by the use of the plasmid pFLP, which allows recombination between two FRT sites flanking the resistance marker, thereby removing it from the chromosome (St-Pierre, 2013).

For in vivo competition experiments, engineered and control strains of EcN were prepared with different selection markers integrated into the chromosome to be able to monitor both strains in the same animal, and allow for stable selection without interfering with plasmid architecture and ensuring minimal influence on host cell fitness. Two selection markers conferring resistance to either spectinomycin or chloramphenicol (SpecR and CamR) were integrated via the Tn7 integration system in accordance with McKenzie (2006). Correct integration was verified by PCR amplification of the integration cassette with one primer binding within the cloned gene and one primer binding the chromosome. Additionally, strains were functionally verified to be resistant to their respective antibiotics at concentration that allowed for clean differentiation of the strains. The streptomycin resistance point mutation did not confer resistance to spectinomycin. Alternatively, one strain carries the CamR marker as described above, and the control strain comprises EcN_pMUTl, which carried GFP on the chromosome.

L-DOPA producing variants of the oligosaccharide-degrading strains were generated by chromosomal integration of 4 genes in the phage 186 site of E. coli Nissle 1917 via the pOSIP integration system (St-Pierre et al. 2013). The four genes introduced comprise the tyrosine hydroxylase gene TyrH, which catalyses the conversion of L-tyrosine into L-DOPA, two genes from the tetrahydromobipterin biosynthetic pathway from E. coli encoding a GTP cyclohydrolase I (EC:3.5.4.16) and a Dihydromona pterin reductase EC: 1.5.1.50 (FolE(T198I) and FolM respectively) and the pterin recycling enzyme, EC:4.2.1.96 (PhhB) from Chromobacterium violaceum.

The lactase production strains are constructed by integrating a beta-galactosidase originating from Bifidobacterium bifidum, BBIJac, into the chromosome of E. coli Nissle 1917 under the control of the strong constitutive promoter pMSKL7 into the phage 186 site of E. coli Nissle 1917 via the pOSIP integration system (St-Pierre et al. 2013). The native signal peptide is removed from the gene, and replaced by one of two secretion peptides: torA or ompA, and both variants are screened for most efficient secretion. In both cases, the BUJac gene is further modified to also comprise a Hi BiT detection tag to allow quantification and distinction from native beta-galactosidases in a faecal matrix via the NanoLuc system (https://dk.promega.com/resources/technologies/nanoluc- luciferase-enzyme/). The strains are then further engineered by introduction of the 5 pFU2.3 plasmid comprising the torA_blon2336 fusion gene and rafY outer membrane porin.

Table 1: Plasmids and E. coli strains

1.3 Media composition: For culturing of cells, 2xYT media (16g/L tryptone; 10 g/L yeast extract; 5 g/L NaCI) was used. For growth assays assessing the ability of an engineered cell to grow on one of the listed prebiotics, M9 minimal media was used (Cold 5 Spring Harbour protocols 2010). M9 minimal media herein contains lx M9 salts, ImM thiamine hydrochloride, 1 w/v % glucose, 0.2% w/v casamino acids, 2mM MgS04, 0.1 mM CaCl2. LB medium (Cold Spring Harbour protocols 2009) was used for plating cultures. Example 2: Plasmids comprising genes for metabolism of the prebiotics melezitose and 3'-fucosyllactose are stably maintained in engineered gramnegative probiotics

The stable inheritance of genes in a host probiotic cell that confer the capacity to metabolize a given carbohydrate is crucial for niche colonization in the gut of a subject.

Methods: Stability of the plasmid pME4.1 compared to the empty backbone plasmid pMUTl was tested by cultivation on non-selective media (i.e. antibiotic free). Colonies of each E. coli Nissle transformed strain, comprising the respective plasmid, were first streaked out on LB plates with kanamycin to select for plasmid presence at the beginning of the experiment. Individual colonies were inoculated in triplicate in 210 mI of 2xYT media in a 96-well plate; and 10 mI from each said replicate was used to perform dilutions up to 10 ^-16 . These dilutions were subsequently spotted on LB plates as well as LB plates supplemented with antibiotics corresponding to the respective plasmid's antibiotic resistance gene. The remaining 200 mI of each inoculate was placed on a 37° C tabletop shaker at 800 rotations per minute (rpm) for approximately 24 h and then used to plate dilutions on LB plates, and LB plates supplemented with kanamycin. Additionally, 4 mI of the 24 h culture was inoculated in 196 mI of 2xYT media and placed into the tabletop shaker. Each day, colonies appearing on the respective LB plates (with or without kanamycin) were counted; the protocol being repeated until each strain reached approximately 100 generations. The percent-loss of the plasmid for each strain was mapped in R studio (www.rstudio.com) in relation to the number of generations for each strain, based on the colony counts from the initial inoculation and assuming a 2% transfer for each new inoculation. For replicates in which 0 colonies were detected, the 0 value was replaced with half of the detection limit (Croghan et al. 2003). Results: E. coli Nissle strain comprising plasmid pME4.1 maintained nearly 100% plasmid stability over 100 generations, comparable to a strain comprising a control plasmid pMUT, which lacks the torA-melD fusion gene and rafY gene (Figure 4A and B). Similar plasmid stability was observed for the E. coli Nissle strain comprising plasmid pFU2.3 (not shown). Example 3: Gram-negative probiotics comprising a RafY porin and periplasmic alpha-glucosidase or alpha-fucosidase grow on rare carbohydrates

The expression of porin and periplasmic carbohydrate hydrolase in a gram-negative probiotic cell is shown to facilitate assimilation and metabolism of the respective carbohydrate, which is crucial for niche colonization in the gut of a subject. This growth property of a probiotic bacterium of the invention is illustrated with respect to the ability to grow on the rare carbohydrates, melezitose and 3'-fucosyllactose. Methods: Growth experiments were conducted utilizing colonies comprising plasmids streaked onto LB agar plates supplemented with the corresponding antibiotic for plasmid selection. Colonies from these plates were picked individually after overnight growth and inoculated in a solution of either M9 Minimal Media with 1% w/v melezitose and 1% w/v glucose or M9 Minimal Media with 1% w/v 3'-fucosyllactose and 1% w/v glucose depending on carbon source utilization. For experiments using a strain with an empty plasmid as a control, the strains were inoculated in the same media as the carbon utilizing strains it was tested alongside.

Colonies of each strain were inoculated in triplicate at an optical density (OD) of 0.01 from the overnight culture on multi-well plates using a BioTek ELx808 Microplate Reader. The plates were shaken continuously at 37 ⁰ C, and the OD (630 nm) of each strain was read every 10 minutes. Data was analysed using R studio. For growth analysis, the minimum OD for each well was subtracted from all reads to account for initial differences in OD measurements, using a growth/no growth threshold of OD 0.1. This threshold accounts for preliminary growth due to co-transferred nutrients from the inoculation as well as ability to grow on nutrients such as casamino acids in the media. For growth curves, error bars represent the standard error from amongst the 3 replicates for each condition. Growth rates were calculated utilizing local growth rates over the course of 5 measurement points. For each well, the greatest growth rate with an r ² value greater than 0.90 was used. The maximum OD was calculated by finding the highest OD630 value recorded for each growth curve and subtracting the corresponding minimum OD. The error bars again represent the standard error from amongst the 3 replicates.

Results: E. coli Nissle strain comprising plasmid pME4.1 is capable of growth on M9 Minimal Media supplemented with either the carbon source 1% w/v glucose or w/v 1% melezitose (Figure 5A, B). Furthermore, expression of the carbohydrate assimilation and metabolism genes in pME4.1 only conferred a minor fitness cost, since the growth of this strain on glucose was only slightly less than a control strain comprising the empty plasmid pMUTl. The inability of the control strain to grow on 1% w/v melezitose confirms that the expression of the torA-melD fusion gene and rafY gene on the pME4.1 plasmid can support niche colonization by a probiotic bacterium in vivo.

E. coli Nissle strain comprising plasmid pFU2.3 is capable of growth on M9 Minimal Media supplemented with either the carbon source 1% w/v glucose or 1% w/v 3'-fucosyllactose (Figure 6A, B). Furthermore, expression of the carbohydrate assimilation and metabolism genes in pFU2.3 also only conferred a minor fitness cost, since the growth of this strain on glucose was only slightly less than a control strain comprising the empty plasmid pMUTl, while being essential for growth on 3'-fucosyllactose. Example 4: Optimisation of the engineered gram-negative probiotic for growth on melezitose

Cytoplasmic folding of the glucosidase MelD and expression of the RafY porin are shown to be necessary for optimal growth of the recombinant gram-negative probiotic bacterium of the invention on melezitose.

Methods: Single colonies were used to inoculate cultures in 2xYT media with kanamycin added at 50ug/ml. Precultures were grown overnight at 37C with shaking at 225 rpm. After 24 hours, the cultures were spun down and resuspended in lx Phosphate Buffered Saline (PBS). Cultures were normalized to OD=0.1 and 2ul of each culture was used to inoculate triplicate wells in a 96 well microtitter plate. Each well contained 200 ul M9 medium + 0.2 % w/v casamino acids and 0.5 % w/v melezitose. The growth of the cultures was tracked in an ELx808 plate reader for 24 hours with fast shaking and OD630 readings every 10 minutes.

Results: In order for the engineered gram-negative probiotic to grow maximally on melezitose, a torA secretion tag is preferred over the malE secretion tag. Having no secretion tag is also preferred over malE, but is less efficient than torA. This result demonstrates the need for cytoplasmic folding of the glucosidase for activity. Additionally, expression of the rafY outer membrane porin greatly enhances growth on the prebiotic (Figure 7). Example 5: An engineered gram-negative probiotic with PfnrS-controlled torA_blon2336 and rafY can only grow on 3'fucosy I lactose under hypoxic conditions

When the periplasmic glycosidase and the outer membrane porin are put under control of PfnrS, the genes necessary for growth on the corresponding prebiotic will be restricted to hypoxic conditions, such as those present in the mammalian gut.

Methods: Overnight cultures of EcN_MUTl and EcN_FU2.3_fnrs comprising the pMUTl and the pFU2.3fnrs plasmids respectively, are inoculated from single colonies and grown aerobically in 2xYT media with kanamycin added at 50 ug/ml. The overnight cultures are then we spun down and resuspended in sterile 1 x Phosphate Buffered Saline (PBS). Both cultures are normalized to OD630=0.1, and 2ul and used to inoculate triplicate wells in a 96 well microtitter plate containing M9 minimal media either with or without 0.5 % w/v 3'fucosyllactose added. This is done both aerobically and anaerobically to demonstrate that the genes for 3'FL metabolism are only expressed anaerobically. Example 6: Robust growth of engineered gram-negative probiotics on rare carbohydrates in the face of competition from porin transported substrates

Assimilation of rare carbohydrates by the engineered probiotic gram-negative bacterium of the invention is mediated by a carbohydrate-specific porin transport protein. The RafY porin, however, is a promiscuous trisaccharide transporter native to certain E. coli strains, principally known for its capacity to transport raffinose. Since rare carbohydrate assimilation is a crucial component of the engineered probiotics ability to establish niche colonization in the gut of a subject, the following example illustrates that the engineered probiotics retain this ability when exposed to trisaccharide transport competitors. Methods: E. coli Nissle strains comprising the indicated plasmids were streaked and inoculated in M9 minimal media with 1% w/v glucose. After overnight growth, they were inoculated in a 96-well plate at an OD630 of 0.01. Strains comprising pME4.1 and pMUTl_rafY plasmids were inoculated in M9 minimal media with 1% w/v melezitose and a gradient of raffinose. Similarly strains comprising pFU2.3 and pMUTl_rafY were inoculated in M9 minimal media with 1% w/v 3'-fucosyllactose and a gradient of raffinose. The growth experiment and data analysis was conducted as described in Example 3.

Results: The trisaccharides, raffinose, 3'-fucosyl lactose and melezitose, are all rare carbohydrates in the gut of adult humans. Hence, the E. coli Nissle strain comprising the plasmid (EcN_RafY) expressing the rafY porin, but lacking genes for expression of either Blon_2336 alpha-fucosidase or MelD alpha-glucosidase show very limited growth on raffinose or the substrates 3'-fucosyllactose or melezitose, respectively since native E. coli Nissle lacks the enzymes for their metabolism. The growth rate and maximum optical density of E. coli Nissle strains (EcN_ME4.1) expressing both the rafY porin, and MelD alpha-glucosidase was essentially unaffected by the presence of an increasing abundance of raffinose (Figure 8A, B). The growth rate of E. coli Nissle strains (EcN_FU2.3) expressing both the rafY porin, and Blon_2336 alpha-fucosidase, on 1 % w/v 3'-fucosyllactose was maintained, in the presence of an increasing abundance of raffinose (Figure 9A), although the maximum optical density was lowered (Figure 9B). This confirms that promiscuous porin transporters (e.g. rafY) can be used to facilitate the assimilation of a range of rare trisaccharide carbohydrates in a probiotic bacterium, and that the fitness cost of assimilating non-metabolisable carbohydrates is either insignificant, or, at most partial. Furthermore, since the growth rate of E. coli Nissle strains comprising plasmids pMUTl_rafY and pMUTl on M9 minimal media + 1 % w/v glucose was similar, the expression of the porin transporter, rafY, perse, has a minimal or insignificant fitness cost (date not shown). Example 7: Growth of engineered gram-negative probiotics on rare carbohydrates is dose-dependent and competitive

Successful niche colonization by engineered probiotic gram-negative bacterium of the invention relies on the bacterium's ability to assimilate and grow on a given rare carbohydrate and thereby out-competes the resident gut microbiome. Correspondingly, as illustrated in the following example, the ability to regulate the abundance of the bacterium by dosing the supply of the rare carbohydrate will facilitate their regulation in a given niche in a subject's gut.

Methods: E. coli Nissle strains comprising the indicated plasmids (pME4.1, pFU2.3 and pMUT_rafY) were grown as described in Examples 3 and 4; and tested for growth on M9 minimal media supplemented with a gradient of melezitose or 3'-fucosyllactose.

Results: The growth rate and maximum OD attained by E. coli Nissle strains comprising the plasmids pME4.1 (Figure 10 A, B) or pFU2.3 (Figure 11 A, B) increased as the concentration of melezitose or 3'-fucosyllactose in the growth medium was raised. Furthermore, their growth and maximal OD was significantly higher than the control strains unable to express either MelD alpha-glucosidase or Blon_2336 alpha-fucosidase. Growth of the E. coli Nissle strains comprising pME4.1 or pFU2.3 on these carbon sources was saturated around 1% melezitose, and 0.1 % 3'-fucosyllactose respectively; while even higher concentrations showed only limited growth reduction for 5 % 3'- fucosyl lactose, which could be caused by the high osmolarity of the resulting medium.

This confirms that a probiotic gram-negative bacterium expressing a RafY porin transporter and a periplasmic MelD alpha-glucosidase or alpha-fucosidase can successfully compete for growth when nutrient levels are low, and when the concentrations of their respective substrates, melezitose and 3'-fucosyllactose, are increased. Furthermore, the growth and growth rate of these probiotic gram-negative bacteria is dose responsive with respect to their respective carbohydrate substrates, which is an important property for regulating their niche colonization in the gut of a subject. Finally, high levels of their respective substrates, melezitose and 3'- fucosyl lactose that may arise in vivo through cyclical feeding patterns do not have a detrimental effect on these bacterial stains.

Example 8: Robust growth of engineered gram-negative probiotics in a synthetic gut environment

Growth conditions that best mimic that of the gut microenvironment are anaerobic on the growth medium, Gifu Anaerobe Media (GAM), which is found to produce the least bias in cultured faecal bacteria, and is therefore deemed to be an acceptable method for in vitro reproduction of the gut environment. The cecum region of the gut, by contrast, is populated by a largely aerobic microbiome. Accordingly GAM was used to demonstrate the growth properties of the engineered probiotic gram-negative bacterium of the invention under both aerobic and anaerobic conditions that mimic the gut environment.

Methods: Growth conditions mimicking the gut were simulated using Gifu Anaerobe Media (GAM) from Nissui Chemical Company, Tokyo, Japan. The growth experiment was conducted as described in Examples 3 and 4, but the respective colonies were plated, selected and cultured under either aerobic or anaerobic conditions. Specifically, the colonies were inoculated at an OD630 of 0.01 in a 96-well plate on GAM supplemented with melezitose for E. coli Nissle strains comprising plasmids pMUT or pME4.1; or supplemented with 3'-fucosyl lactose for E. coli Nissle strains comprising plasmids pMUT or pFU2.3.

Results: When cultivated in the nutrient-enriched GAM, all tested strains grew well under both aerobic and anaerobic conditions (Figures 12 A, B and 13 A, B), despite the lower energy efficiency of anaerobic growth. Potential fitness costs, due to replication and expression of the plasmid encoded genes, were eliminated in the GAM, indicating that when exposed to a more complex nutrient environment, a higher proportion of the nutrients can be used to support growth to a higher cell density.

On the basis of this data, the engineered E. coli Nissle strains expressing a RafY porin transporter and a periplasmic MelD alpha-glucosidase or alpha-fucosidase demonstrate their ability to exploit a rich nutrient supply for cell growth equivalent to non-engineered strains. This property, combined with the ability of these strains to exploit a rare carbon source when other nutrients are limited or exhausted (Figures 10 and 11), strongly supports the expectation that they will outcompete a resident gut microbiome if provided with a prebiotic comprising this rare carbon source. Example 9: Competitive growth of engineered gram-negative probiotics in a synthetic gut environment

Gifu Anaerobe Media (GAM) was further used to demonstrate the ability of engineered probiotic gram-negative bacterium of the invention to outcompete a control bacterial strain under both aerobic and anaerobic growth conditions that mimic the gut environment.

Methods: Colonies of E. coli Nissle strains comprising plasmids pFU2.3ng or pMUTng, were inoculated together at an OD630 of 0.01 in a 96-well plate on M9 growth medium supplemented with 10% v/v GAM and 1% w/v 3'-fucosyl lactose or 100% v/v GAM supplemented with 1% w/v 3'-fucosyllactose. The respective colonies were plated, selected and cultured under either aerobic or anaerobic conditions for a period of 24 hours at which point 2% of each culture was transferred to fresh media and cultured under the same conditions for a further 24 hour period, these steps being repeated a total of 3 times. At the end of each 24 hour period a further culture aliquot was plated out on LB agar plates containing either no antibiotics to allow for growth of both strain, or LB agar + 50ug/mL streptomycin to select for the strain comprising pFU2.3. The difference in CFUs counted on the two plates was used to determine the cell population ratio between E. coli Nissle strains comprising the plasmid, pFU2.3, and E. coli Nissle strains comprising the plasmid, pMUT.

Results: Under aerobic conditions, E. coli Nissle strains comprising plasmid pFU2.3 is able to significantly outcompete the control strain comprising plasmid pMUT on both 100% v/v GAM medium and on M9 medium comprising only 10% v/v GAM, which has a comparatively lower nutrient content (Figure 14A). The competitive growth of the E. coli Nissle pFU2.3 strain may be attributed to its ability to assimilate and metabolize 3'- fucosyllactose.

When cultured under anaerobic conditions, the ability of the E. coli Nissle pFU2.3 strain to outcompete the control pMUT strain was greatest under the nutrient limiting conditions provided by M9 medium comprising only 10% GAM (Figure 14B), which is again attributable to the ability of the E. coli Nissle pFU2.3 strain to assimilate and metabolize 3'-fucosyllactose in the medium, which is most advantageous when nutrient supply is limiting. Thus, in a synthetic gut environment, engineered probiotic gram-negative bacterium of the invention show the ability to out-compete control bacteria, when provided with the rare carbohydrate they are engineered to assimilate and metabolize, this being an important property for colonization of the gut. Furthermore, the supply of the rare carbohydrate as a prebiotic is shown to provide a tool for regulating growth and colonization by the engineered probiotic gram-negative bacterium in a synthetic gut environment.

Example 10: Integration of 3'fucosyllactose genes into the chromosome of the engineered gram-negative probiotics allows for utilization of 3'FL as a sole carbon source. When genes needed for metabolism of 3'fucosyllactose (3'FL) are integrated into chromosome of cells of the gram-negative probiotic and then expressed, the cells are capable of growth on 3'FL as a sole carbon source.

Methods: The genetic element comprising the TetR promoter, the TorA signal peptide, Blon_2336, and the RafY outer membrane transporter, where amplified from pFU2.3 and inserted into the pOSIP-KO vector as described in Example 1. Confirmed integrants of EcN: :FU2.3 were compared with EcN_FU2.3 comprising the plasmid-borne version of the genes and EcN_MUT comprising the empty control plasmid. Characterization of growth was done using the same method as for Example 3, except that the overnight cultures of EcN: :FU2.3, EcN_FU2.3, and EcN_MUTl were grown in 2xYT media. On the following day, 200ul volumes in a 96 well plate containing M9 minimal media supplemented with 1 % w/v 3'FL were inoculated to OD630 of 0.01, and optical density (OD630) was tracked for 1400 minutes (24 hours) on an BioTek ELx808 plate reader with fast orbital shaking and readings every 10 minutes.

Results: As can be seen from Figure 15, the integrated version of the FU2.3 pathway is capable of above-background growth on the prebiotic 3'-FL, demonstrating that chromosomal integration is a feasible strategy for carrying genes needed for assimilation of rare carbohydrates, in cases where plasmid stability or the availability of suitable plasmids is an issue.

Example 11: Animal feeding trial demonstrates tolerance and safety of melezitose and 3-fucosyllactose as a prebiotics The clinical use of the rare carbohydrate substrates, melezitose and 3'-fucosyllactose, to confer a competitive growth advantage to probiotic gram-negative bacteria expressing a RafY porin transporter and a periplasmic MelD alpha-glucosidase or alpha- fucosidase in the gut environment, requires that these carbohydrates are both safe and tolerated by a subject. This example illustrates the tolerability and safety of administering melezitose and 3'-fucosyllactose as prebiotics to a mammal.

Methods: Conventional NMRI female mice, approximately 6 weeks old, were utilized for testing the effects of adding the prebiotics to mice drinking water. Mice were randomized by weight and age and split into their relevant groups. Within each group, mice were caged in pairs with shared drinking water and access to food. Over the course of the trial, groups were given either plain drinking water or drinking water with 5% w/v melezitose or 5% w/v 3'-fucosyllyactose in accordance with the experimental plans outlined in Figure 15. Drinking water consumption was measured throughout the trial, and the collected data was grouped together for the control period compared to the experimental period. Error bars were calculated based upon the standard error of the average water consumption per day amongst the 4 cages of 8 mice total in each group.

Results: The preliminary mouse animal trial study confirmed that the drinking habits, in terms of water consumption, of mice provided with 5% w/v melezitose water, or 5% w/v 3-fucosyllactose was the same as mice provided with plain water over the course of these 7 days, (Figure 16A, B). Likewise, the mice showed no change in weight nor demonstrated any adverse events over the course of this time period. The absence of adverse advents and the taste tolerance for 5% w/v melezitose water, and 5% w/v 3-fucosyllactose shown in this animal trial fulfils an important prerequisite for the successful clinical application of prebiotics comprising these rare carbohydrates in mammals. Example 12: Competitive colonization of gram-negative probiotics engineered to utilize 3'-fucosyllactose against a control strain increases of the animal gut by co-administration of 3'-fucosyllactose as a prebiotic

The ability of an engineered probiotic gram-negative bacterium to compete with a control strain in the streptomycin-treated mouse gut is shown to increase by the administration of the rare oligosaccharide, 3'-fucosyllactose.

Methods: The animal trial was conducted on 2 groups of conventional NMRI female mice (see example 9). All groups received sterile filtrated drinking water with 5 g/L streptomycin for 24 hours prior to onset of prebiotic supplementation. At day -1, group 1 continued receiving plain drinking water with 5g/L streptomycin, while group 2 received 5 % w/v 3'-fucosyllactose supplemented with 5g/L streptomycin. Additionally, the animals were fed a modified diet (SF09-028), in which dietary fibers and starch were replaced by dextrose to achieve increased carbon-starvation in the mouse gut. 24 hours after start of prebiotic supplementation (day 0), all animals were dosed by oral gavage with lOOul of a 1:1 mix of control and engineered bacteria in a PBS suspension containing 10 ⁵ CFU/ml of each strain, as indicated in the protocol in Figure 17. The engineered and control strains had been tagged with different chromosomal antibiotic selection markers (chloramphenicol and spectinomycin, respectively) which allowed them to be distinguished when faecal samples were plated on LB agar plates containing either of the two antibiotics. Faecal samples were collected in 1 ml Phosphate-Buffered Saline (PBS) at indicated time points, then each was weighed, vortexed, spun down for 1 min at lOOxg, diluted, and plated on LB agar plates with streptomycin, kanamycin, and either spectinomycin (to detect the control strain, EcN_MUTl) or chloramphenicol (to detect the engineered strain, EcN_FU2.3). Detected colonies were counted across all countable dilutions, and averaged amongst replicates and dilutions.

The detection limit was calculated in accordance with Croghan et al. (2003) by calculating the CFU/mL/g of 1 colony on an undiluted plate. Since plating of faecal samples was done in triplicate, only instances where at least 2 out of 3 replicates showed CFU counts were scored as a positive result, while instances of 1 out of 3 replicates with colonies were scored as negative results, and therefore not counted. The faecal or gut weight used for calculating the detection limit was the average faecal or gut weight measured for that experiment.

Results:

This competitive animal study confirmed that the recombinant probiotic bacterium has a competitive advantage when the corresponding prebiotic is provided in the drinking water. Figure 18A shows that the engineered strain is not competitive without addition of 3'FL to the drinking water. At day 5, EcN_FU2.3 is outcompeted, while the control strain colonises the gut. Figure 18B shows the competitive colonization of EcN_FU2.3 when 5% w/v 3'FL is added to the drinking water. 7 out of 8 animals displayed CFU at day 5, albeit approximately two-fold lower than the control strain.. Figure 18C gives an overview of the data comprised in Figure 18A and B, by showing lines defined as the geometric mean and grouping all CFU counts. This experiment supports the hypothesis that engineering rare oligosaccharide metabolism in E. coli can provide a competitive advantage in vivo when said oligosaccharide is supplied in the drinking water. Example 13: Use of a dietary prebiotic for dose-controlled administration of a therapeutic molecule by recombinant probiotic gram-negative bacteria engineered to produce L-DOPA

When the recombinant probiotic gram-negative bacterium is capable of production of the neurotransmitter L-DOPA via the further introduction of two transgenes (tyrH_rat_OH and cvPhhB) and overexpression of two native co-factor genes (folE(T198I) and folM). Administration of the prebiotic allows for control of L-DOPA production levels in the mammalian gut.

Method:

Introduction of L-DOPA pathway : L-DOPA production was achieved by the introduction of four genes into the chromosome of EcNcm_FU2.3, thus generating the strain EcNcm_FU2.3_dopa as stated in Table 1. The four genes introduced comprise the tyrosine hydroxylase gene TyrH, which catalyses the conversion of L-tyrosine into L- DOPA, two genes from the tetrahydromobipterin biosynthetic pathway from E. coli (FolE(T198I) and FolM) and the pterin recycling enzyme (PhhB) from Chromobacterium violaceum (Figure 19A).

L-DOPA production cultures : L-DOPA production cultures were carried out in 96 deep well plates and 350 pi media. Biological triplicates of each strain were used to inoculate precultures in M9 media with 0.4% glucose + vitamin solution + trace elements. Precultures were grown for 24 hours at 37°C in a shaking incubator at 250 RPM. Production cultures were carried out by inoculating the preculture with 1: 100 ratio of the total volume and incubated at 37°C in a shaking incubator at 250 RPM for 24 hours. After 24 hours the cultures were centrifuged at 4700 RPM and the supernatant was collected and frozen until further analysis.

Measurement of in vitro production via HPLC\ Quantitative analysis of L-DOPA in cell- free supernatant was performed by High-Performance Liquid Chromatography (HPLC) on an UltiMate 3000 UHPLC system (ThermoScientific). The system consisted of an LPG- 3400RS quaternary pump and a WPS-3000RS autosampler with a TCC-3000 column oven and a DAD-3000 diode array detector. Samples were run at a pressure of 600 bar through a CORTECS column (1.6 pm, 2.1x150 mm) at 30D with an injection volume of 1 pi and a flowrate of 0.350 ml/min in 10 mM ammonium formate as mobile phase.

Animal study: Four groups of animals (n=8) were dosed with the recombinant therapeutic strain EcNcm_FU2.3_dopa, with the ability to produce L-DOPA and further comprising torA_blon2336 and rafYe nabling utilization of 3'FL for growth, in combination with EcN MUTl comprising pMUTl and chromosomal GFP as a control strain. Both strains were dosed at 10 ⁵ CFU, and the gavage mix had a concentration of 0.5 % w/v 3'FL to ensure prebiotic access at the onset of colonization. Both strains were also treated with the TDC inhibitor Carbidopa via intraperitoneal injection (10 mg/kg body weight) every 24h. Fresh fecal samples were collected daily for 16 days to quantify colonization and metabolite levels. Plasma samples were taken on day 2 (submandibular sampling) and day 7 (vena cava) after gavage, and urine samples were taken on day 3, 10 and 15.

Each group of animals received different amounts of 3'FL in the drinking water (0, 2.5, 5, 10 % w/v) from day -1 until day 8. On day 8, all animals were switched to drinking water with 5g/L streptomycin without any prebiotic to observe the wash-out rate of the engrafted strain. (Figure 19C).

Results:

In vitro production of L-DOPA: Production levels of L-DOPA are shown in Figure 19B. Comparison between L-DOPA levels observed for EcNcm_FU2.3_dopa and EcN_pMUTl_dopa shows that the plasmid for oligosaccharide degradation does not interfere with production levels (Figure 19B).

Animal study: Data is shown from the first 8 days post gavage, where prebiotic was added to the drinking water (Figure 19D). With increasing levels of prebiotic in the drinking water, a corresponding increase in CFUs of the engineered strain was observed. It is expected upon analysis of faecal material and blood samples from the animals, that levels of L-DOPA produced by the engineered strain will differ significantly between groups 1 and 2/3/4 as a direct effect of the prebiotic induced difference in cell density found in the faces. Any observed differences between groups 2, 3, and 4, might stem from carbon saturation regarding colonization, leading to increased carbon flow towards L-DOPA. Example 14: Use of a dietary prebiotic for dose-controlled administration of a therapeutic peptide by recombinant probiotic gram-negative bacteria engineered to produce secreted beta-galactosidase

When the recombinant probiotic gram-negative bacterium is capable of production of the enzyme beta-galactosidase via the further introduction of an overexpressed copy of an extracellular beta-galactosidase originating from Bifidobacterium Bifidum, administration of the prebiotic allows for control of secreted lactase production levels in the mammalian gut.

Method: Four groups of animals (n = 8) are given the recombinant therapeutic bacteria EcNcm_FU2.3Jacl or Iac2 in addition to different dosages of the prebiotic in the drinking water (figure 20). Fresh faecal samples are collected daily throughout the experiment to track CFU numbers, extracellular beta-galactosidase and stool consistency (to observe symptoms of lactose intolerance). This experiment is carried out with adult NMRI mice, which are lactose intolerant post-weening.

With increasing levels of prebiotic in the drinking water, it is expected that a corresponding increase in both CFUs and extracellular lactase is observed. This should result in relief of lactose intolerance symptoms in adult mice when given high amounts of lactose in the diet, and this will act in a dose-dependent manner, as higher levels of secreted lactase will result in greater clearance of lactose from the murine gastrointestinal tract. Example 15 Deletion of the glgA gene in the engineered probiotic bacteria increases dependency on prebiotic in the media

E. coli colonizing the mammalian gut experiences fluctuations in nutrient levels, which can necessitate the storage of energy during times of excess, which can then be utilized during times of starvation. Removal of the ability to synthesize the carbon storage molecule glycogen by deletion of glgA renders the strain more sensitive to the presence of prebiotic supplement in order to maintain colonization.

Methods: In vitro competition experiments are carried out by co-inoculating equal amounts of EcN_GFP with either EcN_glgA_FU2.3 or EcN_FU2.3 under conditions of carbon starvation by growth in LB medium to stationary phase, which has low levels of carbohydrates, with and without the addition of 3'FL. It is expected that the control strain will outcompete the engineered probiotic with no prebiotic present, and that this will happen at a faster rate for EcN_glgA_FU2.3 than for EcN_FU2.3. Competitive index of the pairwise strains will be determined by CFU counts. .

Example 16 Deletion of the nanA gene in the engineered probiotic bacteria increases dependency on administered prebiotic in the media

E. coli depends on microbial neighbours in mixed biofilm for mucin-derived carbohydrates, that E. coli can utilize, but not release. By removing the ability of E. coli to grow on N-acetylneuraminate, the engineered gram-negative bacterium is expected to be more depend on diet-derived carbohydrates when colonizing the mammalian gut. This adds a layer of safety, as the engineered strain will be a poor colonizer in the absence of the prebiotic.

In vitro competition experiments are carried out by co-inoculating equal amounts of EcN GFP with either EcN_nanA_FU2.3 or EcN_FU2.3 under conditions of carbon starvation by growth in LB medium, which has low levels of carbohydrates, with and without the addition of 3'FL, and with the addition of N-acetylneuraminate to mimic the availability of this carbon source in the mammalian gut. It is expected that the control strain will outcompete the engineered probiotic with no prebiotic present, and that this will happen at a faster rate for EcN_nanA_FU2.3 compared to EcN_FU2.3 since it won't be able to utilize the mucin-derived carbon source. Competitive index of the pairwise strains will be determined by CFU counts.

References

Asto E., Mendez I., Audivert S., Farran-Codina A., and Espadaler J., 2019: The Efficacy of Probiotics, Prebiotic Inulin-Type Fructans, and Synbiotics in Human Ulcerative Colitis: A Systematic Review and Meta-Analysis. Nutrients , 11, 293; doi: 10.3390/null020293 Chen, Lei et al ., 2015 Identifying Essential Streptococcus sanguinis Genes Using Genome-Wide Deletion Mutation in "Gene Essentiality Methods and Protocols" in Methods of Molecular Biology vol. 1279 Editors: Lu, Long Jason (Ed.) ISBN 978-1-4939- 2398-4

Cold Spring Harbour Protocols (2010) Recipe M9 minimal medium (standard) doi: 10.1101/pdb.recl2295

Cold Spring Harbour Protocols (2009) Recipe LB agar doi: 10.1101/pdb.recll683

Croghan, C. W. and P. P. Egeghy. (2003) Methods of Dealing with Values Below the Limit of Detection using SAS, Presented at Southeastern SAS User Group, St. Petersburg, FL, September 22-24, 2003. Freudl, R. (2018) Signal peptides for recombinant protein secretion in bacterial expression systems Microb Cell Fact. 17:52. doi: 10.1186/S12934-018-0901-3

Emily C. A. Goodall, Ashley Robinson, Iain G. Johnston, Sara Jabbari, Keith A. Turner, Adam F. Cunningham, Peter A. Lund, Jeffrey A. Cole, Ian R. Henderson, The Essential Genome of Escherichia coli K-12, mBio Feb 2018, 9 (1) e02096- 17; DOI: 10.1128/mBio.02096-17

Grozdanov, L., Raasch, C., Schulze, J., Sonnenborn, U., Gottschalk, G., Hacker, J., & Dobrindt, U. (2004). Analysis of the genome structure of the nonpathogenic probiotic Escherichia coli strain Nissle 1917. Journal of bacteriology, 186(16), 5432-5441. https://doi.org/10.1128/JB.186.16.5432-5441.2004 Hayes, F. D. A. S. (2003). Toxins-Antitoxins: Plasmid Maintenance, Programmed Cell Death, and Cell Cycle Arrest. Science, 301(5639): 1496-1499

Islam, S.U., (2016) Clinical Uses of Probiotics. Medicine (Baltimore). 95(5): doi: 10.1097/MD.0000000000002658

McKenzie, G. J., 8i Craig, N. L. (2006). Fast, easy and efficient: site-specific insertion of transgenes into enterobacterial chromosomes using Tn7 without need for selection of the insertion event. BMC microbiology, 6, 39. https://doi.org/10.1186/1471- 2180-6-39

Mori, Hirotada et al., 2015 Identification of Essential Genes and Synthetic Lethal Gene Combinations in Escherichia coli K-12 in in "Gene Essentiality Methods and Protocols" in Methods of Molecular Biology vol. 1279 Editors: Lu, Long Jason (Ed.) ISBN 978-1-4939-2398-4

Nour-Eldin H.H., Geu-Flores F., Halkier B.A. (2010) USER Cloning and USER Fusion: The Ideal Cloning Techniques for Small and Big Laboratories. In: Fett-Neto A. (eds) Plant Secondary Metabolism Engineering. Methods in Molecular Biology (Methods and Protocols), vol 643. Humana Press, Totowa, NJ O'Connell, KJ., O'Connell Motherway M., O'Callaghan J., Fitzgerald GF., Ross RP., Ventura M., Stanton C., and van Sinderen D. (2013) Metabolism of Four a-Glycosidic Linkage- Containing Oligosaccharides by Bifidobacterium breve UCC2003; Appl Environ Microbiol. 79(20): 6280-6292. Praveschotinunt P., Duraj-Thatte AM., Gelfat I., Bahl F., Chou DB., and Joshi NS., (2019) Engineered E. coli Nissle 1917 for the delivery of matrix-tethered therapeutic domains to the gut. Nature Communications 10, 5580,

St-Pierre, F., Cui, L., Priest, D.G., Endy, D., Dodd, I.B. (2013) One-Step Cloning and Chromosomal Integration of DNA, ACS Synthetic Biology 2013 2 (9), 537-541, DOI: 10.1021/sb400021j

Ukena Ukena SN, Singh A, Dringenberg U, Engelhardt R, Seidler U, et al (2007) Probiotic Escherichia coli Nissle 1917 Inhibits Leaky Gut by Enhancing Mucosal Integrity. PLoS ONE 2(12): el308. doi: 10.1371/journal. pone.0001308

Sharan SK, Thomason LC, Kuznetsov SG, Court DL. Recombineering: A Homologous Recombination-Based Method of Genetic Engineering. Nature protocols. 2009;4(2):206-

223. doi: 10.1038/nprot.2008.227. PubMed PMID: 19180090. PubMed Central PMCID: PMC2790811.

Dong-Eun Chang, Darren J. Smalley, Don L. Tucker, Mary P. Leatham, Wendy E. Norris, Sarah J. Stevenson, April B. Anderson, Joe E. Grissom, David C. Laux, Paul S. Cohen, Tyrrell Conway (2004) Carbon nutrition of Escherichia coli in the mouse intestine Proceedings of the National Academy of Sciences, 101 (19) 7427- 7432; DOI: 10.1073/pnas.0307888101

Miller JK, Harrison MT, D'Andrea A, et al. b-Carotene Biosynthesis in Probiotic Bacteria. Probiotics Antimicrob Proteins. 2013;5(2):69-80. doi: 10.1007/sl2602-013- 9133-3