In the early stages of selection, we considered two sump microorganisms commonly used for experiments: E. coli and B. bicolor, and compared them.

Bilobacterial yeast endogenous secreted protein production is limited and easy to purify. And as a eukaryote, it is capable of post-translational modification and glycosylation of bromelain, which is superior to Saccharomyces cerevisiae and E. coli. Despite these advantages, the disadvantages of B. bicolor are more obvious and difficult to address. Firstly bromelain is toxic/inhibitory to the growth of B. aeruginosa. Secondly the studies on localisation signals in B. birchii are not clear. Although it has been shown that STB3 has 88% homology to proteins with lysosomal localisation and that the lysosomal sequence of this protein is conserved, there are no clear studies to demonstrate or find a mechanism for co-localisation, to the detriment of design production. Finally, methanol induction is generally indicated for the expression of recombinant proteins in B. birchii. A minimum concentration of 0.5% methanol is required for the production of recombinant proteins and a maximum concentration of 2-2.5% (wt/vol) is necessary for full protein expression. In addition, high concentrations of methanol (above 5%) are toxic to cells, leading to the accumulation of formaldehyde and hydrogen peroxide, which can lead to cell death. On the other hand, low levels of methanol trigger the proteolytic degradation of heterologous proteins, leading to reduced productivity. On the one hand methanol is difficult to remove completely when extracting recombinant proteins, and on the other hand the adverse effects of products containing methanol added to feed are difficult to assess in feed animals.

E. coli has been studied in rich detail and the modification and utilisation of E. coli is relatively common. Heterologous expression of pineapple protease has been performed in E. coli in the literature. In the literature codon optimisation of pineapple protease has been carried out and the protein has been modified. The yield and activity of the recombinantly expressed pineapple protease were very promising. However, E. coli has a periplasmic space and heterologous expression of recombinant proteins in E. coli tends to form inclusion bodies, which makes subsequent isolation and purification more difficult. Secondly, our ultimate goal is the characterisation of pineapple protease, which cannot be glycosylated by E. coli, which is detrimental to our subsequent work. Moreover, E. coli is a conditional pathogen and recombinant E. coli for industrial production has potential adverse effects on animal and human health when complete removal of the chassis microorganism cannot be guaranteed.

Initially we planned to use E. coli as the sump microorganism, taking into account ease of handling, yield and safety.

After the initial selection of E. coli as the chassis microorganism, we communicated and exchanged ideas with Associate Professor Mengjie Zhang from Tongji University to ask her opinion.

Listening carefully to the pros and cons of both E. coli and B. ricinus, Prof. Zhang recommended that we choose E. coli. On learning that we wanted to apply pineapple protease in feed production, she mentioned a probiotic, Bacillus subtilis, and suggested that we learn a little about Bacillus subtilis before choosing between E. coli and Bacillus subtilis.

After gaining some understanding of Bacillus subtilis, we made a comparison.

Bacillus subtilis is often used as a model strain for bacterial genetics and cellular metabolism studies due to the clear sequencing of its genome and the resolution of essential genes. In addition, Bacillus subtilis has a high safety profile and is recognised as generally recognized as safe (GRAS) by the US Food and Drug Administration (FDA). Because of its clear physiological and biochemical characteristics, simple genetic manipulation, high secretion and expression capacity and ease of culture and fermentation, Bacillus subtilis has also been transformed into a microbial cell factory for the production of industrial enzymes, vitamins, functional sugars, nutraceuticals and drug precursors and other target products, showing strong industrial applications. Compared to Gram-negative bacteria represented by E. coli, the monolayer membrane structure of Bacillus subtilis makes the secretion of heterologous proteins easier, and because B.subtilis' own protein expression and secretion systems are well documented and more comprehensively studied, B.subtilis is widely used in the expression of heterologous protein-based target products. Specifically, the following.

1. Non-pathogenicity. Bacillus subtilis is widely found in the animal gut in nature and is considered to be an intestinal probiotic with a history of preparing fermented foods that do not produce heat-sensitive proteins or various toxins and are safe up to food grade.
2. Bacillus subtilis cells have a strong capacity for extracellular secretion of proteins, with approximately 300 proteins being secreted directly into the extracellular compartment, simplifying and even removing some of the purification steps for the target proteins compared to the commonly used industrial host bacterium, E. coli.
3. Protein products expressed in the Bacillus subtilis expression system do not readily form inclusion bodies. In the E. coli expression system, if the coding sequence of the exogenous protein to be expressed contains a large number of consecutive rare codons, the expression level is usually low or even translation is terminated prematurely; inclusion bodies lead to insoluble proteins, which makes it difficult to isolate and purify the target protein, but the Bacillus subtilis expression system does not have these problems and shows significant advantages in these aspects.
4. Bacillus subtilis is an aerobic, thermophilic bacterium that does not require an anaerobic environment or special media, with a short culture cycle and fast growth rate.
5. Bacillus subtilis is a Gram-positive bacterium and there are many components present in the lysed system, such as peptidoglycan, phosphopeptide and protein. The type and content of phospholipids in Bacillus subtilis are well defined and the ratio between components is little affected by the culture conditions, which is conducive to the design of subsequent isolation and purification experiments.

In addition, unlike the Gram-negative expressing host E. coli, B.subtilis lacks an outer membrane and associated periplasmic space, which gives it a greater capacity for protein secretion. Bacillus subtilis heterologously expressed proteins were more productive compared to other chassis microorganisms, but varied considerably between proteins. Compared to other strains already commercially available there is a gap but it can meet the production requirements. There is no literature available on the heterologous expression of pineapple protease by Bacillus subtilis at this time, there is some literature relating to the expression of other proteases by Bacillus subtilis with more promising yields.

Bacillus subtilis has three possible advantages as a chassis compared to E. coli, as follows.

1. potentially higher expression levels than E. coli
   Since protein products expressed by the Bacillus subtilis expression system do not readily form inclusion bodies and Bacillus subtilis lacks an outer membrane and associated periplasmic space, it has a greater capacity for protein secretion. Therefore, Bacillus subtilis may express recombinant proteases at a higher level than E. coli and is more conducive to enzyme isolation and purification.

2. Biological chassis residues
   As any transgenic chassis must not be used as a feed additive, the final output must be sterile. However, it is not possible for the final output enzyme preparation to be completely free of chassis residues, whereas Bacillus subtilis is a probiotic and E. coli is a pathogenic bacterium, compared to the more harmless residues of Bacillus subtilis. The Chinese feed hygiene standard requires a residual sump dose of bacteria for biological enzyme preparations of <2 x 104 CFU/g. The optimal dose of the common Bacillus subtilis feeding strain PB6 is around 1 x 105 CFU/g or 1 x 104 CFU/g, and its safety as a commonly used Bacillus subtilis preparation has been certified by the European Food Safety Authority. Therefore, it is a win-win situation for Bacillus subtilis as a chassis microorganism as it can still be used for feeding purposes within the standard chassis residue count.

3. Beneficial protein secretion
   While E. coli secretes harmful toxins as it reproduces and grows, Bacillus subtilis produces a number of beneficial proteins. These beneficial proteins include antimicrobial active substances (chymotrypsin, polymyxin, myclobactin, short bacillus peptides), digestive enzymes (alpha-amylase, protease, lipase, cellulase) and B vitamins (B1, B2, B6, niacin).
   After comparing Bacillus subtilis and E. coli, we finally chose Bacillus subtilis as the chassis microorganism due to all these factors.

After choosing Bacillus subtilis as the general direction, we have been informed about the various strains of Bacillus subtilis with the aim of identifying a safe and commonly used strain of Bacillus subtilis. Considering that Bacillus subtilis is currently added and used in feeds, we wanted to find a forage strain of Bacillus subtilis and modify him to produce our target product, bromelain, in a safe and efficient manner. However, in order to keep the experiments on schedule, we split up into two strain types of Bacillus subtilis for feeding and Bacillus subtilis for the common laboratory chassis.

Bacillus subtilis is a Gram-positive model bacterium that has been widely used for metabolic mechanism studies and metabolic engineering. Bacillus subtilis has a well-defined genetic background and a range of genome engineering tools that make Bacillus subtilis an important chassis for biological production. The rapid growth of Bacillus subtilis chassis facilitates shorter fermentation cycles and higher yields of target products. However, engineered Bacillus subtilis typically has a reduced growth rate compared to the initial strain, especially after intensive rewiring of metabolic pathways, which severely hinders its use in bioproduction.

In the section on strains of Bacillus subtilis commonly used in laboratory chassis, we focus on Bacillus subtilis strain 168.

Whereas it has been shown that recombinant expression plasmids constructed in Bacillus subtilis suffer from instability problems, our focus was on an improved strain of Bacillus subtilis strain 168 that performed well in terms of both yield and hardware.

However there is no clear literature proposing an improved strain of Bacillus subtilis 168 that solves, and alleviates, the problem of unstable recombinant expression plasmids in Bacillus subtilis, but we did borrow three genes, oppD, hag and flgD, based on modification methods in the literature for use in subsequent modifications of our strain. We therefore thought of strains of Bacillus subtilis for addition to feed that might be able to be modified for protein production. We have therefore investigated and compared strains for use in feed.

Expression strains have been chosen over feed strains for the following reasons.

1. Neither GM strains nor their recombinant DNA are allowed to be present in fermented products placed on the EU market as food or feed additives. Chinese feed hygiene standards also limit the residual presence of GM chassis. For this reason, feed strains cannot be retained in enzyme preparations as feed preparations.
2. (As most of the forage strains are isolated from animal faeces (i.e. intestinal parasitic strains of livestock), while the expressive strains are mostly derived from indigenous microorganisms in the soil) several common forage strains could not be found in the literature for expressing heterologous proteins, so the final preference was to choose the wild type strain 168 and its modified strains with high expression yield.

However, there are still many similarities between the expressive strains and the feeder strains, with the greatest commonality between them being the production of bacteriocins. The commonly used feeder strain PB6 (ATCC-PTA 6737) has been shown to produce antimicrobial substances and has broad in vitro activity against a variety of strains, including Campylobacter spp. and Clostridium spp. Bacillus subtilis strain 168 can also produce antimicrobial substances.

One of the main factors limiting the application of Bacillus subtilis as an expression host is its production of at least eight extracellular proteases. Bacillus subtilis produces at least eight extracellular proteases (listed below) that have a variety of functions, including degrading proteins in bioorganisms for nutrient supply and proteolytic processing of other proteins. However, extracellular proteases have a high propensity to degrade heterologous secreted proteins, resulting in significant loss of target products.

In order to reduce the effect of endogenous proteases on the heterologous protein pineapple protease, several protease-deficient strains were considered for comparison.

Nevertheless, researchers have also noted that sometimes certain proteases favour the secretion of exogenous proteins. Therefore, in order to maximise the production of exogenous proteins, proteases should be selectively inactivated. Our original plan was to use an evaluation system containing nine Bacillus subtilis 168 mutant strains to identify favourable and unfavourable proteases for target proteins (Zhao L, Ye B, Zhang Q, Cheng D, Zhou C, Cheng S, Yan X. Construction of second generation protease -deficient hosts of Bacillus subtilis for secretion of foreign proteins. Biotechnol Bioeng. 2019 Aug;116(8):2052-2060. doi: 10.1002/bit.26992. (Epub 2019 Apr 24. PMID: 30989640.). The mutant strain PD8 is inactivated for all eight proteases, while each of the other eight mutant strains expresses only one of these eight proteases. The target protein is secreted in these nine mutant strains, and if the yield of the target protein is higher in the mutant strain than in the PD8 strain, the corresponding protease is considered favourable. Thus, an optimal protease-deficient host was constructed by inactivating the unfavourable protease.

Unfortunately, however, due to the Shanghai epidemic, we were unable to return to school early in the summer and the experiment started late, so our original plan could not be implemented due to time constraints and we had to screen various protease-deficient hosts through literature research. Currently the common protease-deficient strains are WB600 (ΔnprE,ΔaprE,Δepr, Δ bpr , mpr::ble , nprB::bsr), WB700 (Δ nprE,Δ aprE,Δ epr,Δ bpr, mpr::ble, nprB::bsr,Δ vpr) and WB800 (Δ nprE,Δ aprE,Δ epr,Δ bpr, mpr::ble, nprB::bsr,Δ vpr, wprA::hyg).

As there is no literature on recombinant expression of bromelain in Bacillus subtilis, we wanted to investigate the literature on recombinant expression of these three strains in order to find an entry point and finally identify the strains. The recombinant enzymes expressed in WB600 had little to no correlation with pineapple protease, while WB700 had little information available and was difficult to obtain, so these two strains were not chosen. WB800 has been applied to produce recombinant enzymes that have a strong correlation with pineapple protease.

In summary, WB800 was chosen as the pineapple protease expressing strain. (Note: All the above protease-deficient hosts are modified strains of the 168 wild strain)

We selected the common shuttle plasmid of Bacillus subtilis: pBE2R, about 7kb.

E. coli-B.subtilis shuttle plasmid is convenient for us to complete the operation in E. coli and then transfer it into Bacillus subtilis for expression. pBE2R has low copy number, simple structure and low price. We chose this plasmid without special requirements.

Promoter is one of the many key factors for the expression efficiency, and the expression level of proteins could be regulated by controlling transcript production at the promoter level. Generally, promoters used for the expression of recombinant proteins in B.subtilis mainly include inducible promoters, stress-specific promoters, auto-inducible (self-inducible) promoters, and constitutive promoters.So far, many native, artificial, and engineered promoters have successfully been used for the expression of recombinant proteins in B.subtilis, they are listed as follows.

Given that there is no need for us to express the protein at a particular time, constitutive promoters may be a decent option in consideration of time and cost. Thus, we chose the P43 promoter at first because it’s easiest to obtain and can meet our fundamental requirements.

The terminators of Bacillus subtilis are rich in GC inverted repeats followed by several A(T)s which are similar to the structures of terminators in E. coli, so we take the T7 terminator, the most frequently used terminator for recombinant protein expression, as our first choice. Besides, we found that artificially designed T31 terminator had the highest stop efficiency among most frequently used known terminators^[2] and had not been used by the iGEM teams before. Thus, we take it into consideration as well. Unfortunately, we were unable to test T31 terminator’s stop efficiency indeed because of the pandemic.

Tat pathway and Sec pathway are two important pathways for Bacillus subtilis to secrete extracellular proteins. Sec pathway is the main exocrine pathway, which secretes proteins that are not well folded; Tat pathway secretes folded extracellular proteins and has higher secretion efficiency for eukaryotic proteins, providing an alternative for proteins that cannot pass through Sec pathway. Proteins secreted through Tat pathway have two characteristics: they fold in the cell and have Tat signal peptide sequence. PhoD is strictly secreted by Tat pathway in Bacillus subtilis, so we found the signal peptide of PhoD through the literature and used it to construct plasmids. After reading the literature, we also found that using different Tat signal peptides will make the exocrine efficiency of the same protein different. If time permits, we also hope to screen the most efficient signal peptide through experiments.