Nisin Variants Generated by Protein Engineering and Their Properties

Nisin, a typical lantibiotic, has robust antimicrobial activity combined with limited cytotoxicity, and the development of resistance to it is slow. These properties make nisin a promising antimicrobial agent to control pathogenic microorganisms in dairy foods. However, its low solubility, poor stability and short half-life at neutral pH limit its application within the dairy industry. Protein engineering technology has revealed the potential of modifying nisin to improve its properties, and many valuable variants have emerged. This review summarizes progress in the generation of nisin variants for the dairy industry and for other purposes. These nisin variants with additional modification have improved properties and can even expand the inhibition spectrum range of nisin. Nisin, as the most thoroughly studied lantibiotic, and its variants can also guide the modification of other lantibiotics.


Introduction
Nisin is widely used to treat bovine mastitis, a disease that causes huge economic losses in the dairy industry. Traditional antibiotics have limited utility, and drug-resistant strains lead to milk contamination [1,2]. Nisin is effective against many pathogens causing bovine mastitis, including some with drug resistance [3][4][5]. In addition, nisin has anti-inflammatory actions through activation of the ERK1/2 and p38 mitogen-activated protein kinase (MAPK) signaling pathway and promotes the integrity of the blood-milk barrier [6]. When added to feed, it modulates ruminal microbial ecology, reducing methane production without adversely affecting digestion [7]. However, nisin undergoes proteolytic decomposition in the digestive tract, which is harmless to animal health but reduces its half-life and limits its application as an oral drug. The protein can be used as a food preservative to inhibit pathogens in dairy products and poses no potential health problems for humans [8].
Nisin has also attracted interest as a potential antitumor agent due to its anticancer effect combined with low somatic cytotoxicity. Together with its natural variant, nisin ZP, nisin has been shown to promote apoptosis and inhibit proliferation of head and neck squamous cell carcinoma, colon cancer cells, hepatocellular carcinoma and the chronic myeloid leukemia cancer cell line (K562) [9][10][11] perhaps due to regulation of apoptotic genes [12,13]. The potential immunomodulatory properties of nisin have been discussed in detail in a recent comprehensive review [14].
Nisin is a prototypical type AI lantibiotic encoded by four genes present in clu lanABCT [15]. The structural gene, lanA, encodes precursor peptides or prepeptides, lanB and lanC encode dehydratase B and cyclase C. The LanB dehydratase catalyze dehydration of serine or threonine residues to 2,3-didehydroalanine (Dha) or 2,3-did drobutyrine (Dhb) in the LanA structural region, and the LanC cyclase conjugates cys residues to these dehydro amino acids to form lanthionine or methyl-lanthionine r The fourth gene, lanT, encodes an ABC transporter to translocate the modified prepep across the cytoplasmic membrane for extracellular modification by NisP. NisP is exte to the membrane and removes the leader region to produce mature nisin [16]. E. col been engineered to synthesize lantibiotics by introduction of the biosynthetic path [17] but production levels are limited by the toxicity of the molecules to the host [18]. barrier has been overcome by synthesis of a nontoxic prenisin variant that can be t formed into its active form after production [19,20].
Nisin leads to membrane permeabilization via low-affinity membrane permea plus lipid II-dependent targeted pore formation. The main way for nisin to exert its microbial activity is lipid Ⅱ-dependent mechanism ( Figure 1). Low-affinity memb permeation is achieved by nisin binding to anionic lipids at mmol concentrations in target membrane followed by aggregation at the membrane surface to form pores Lipid II-dependent targeted pore formation is achieved by nisin docking to form a n lipid II complex inserting the C-terminal region by means of a flexible hinge mecha into the membrane to form a pore. The pore complex has a stoichiometry of four lip molecules to eight nisin molecules and inhibits cell wall synthesis, causing leakage o contents [22,23]. Nisin recognizes the pyrophosphate moiety of lipid II, unlike vanco cin, which also targets the pentapeptide of lipid II [24]. Vancomycin resistance is ge ated when the lipid II pentapeptide is modified by the bacterial strain, but pyrophosp modification would be much more difficult to achieve [25], explaining the rarity of resistance [26]. Moreover, nisin has been found to have antimicrobial activity and biofilm activity against multidrug-resistant Streptococcus pneumoniae, methicillin-resi Staphylococcus aureus (MRSA), coagulase-negative Staphylococci and vancomycin-resi Enterococci [27][28][29]. Lipid Ⅱ-dependent antimicrobial mechanism of nisin. When nisin reaches the cell m brane surface (a), rings A and B at the N-terminus of nisin bind to the pyrophosphate group of II (b). The C-terminal region inserts into the membrane to form a pore by means of a flexible h mechanism (c). The pore complex has a stoichiometry of four lipid II molecules to eight nisin m cules (d).
However, Gram-negative bacteria lack lipid II, limiting universal antimicrobia tivity of nisin. Moreover, nisin stability is pH-sensitive, with optimal bioactivity Figure 1. Lipid II-dependent antimicrobial mechanism of nisin. When nisin reaches the cell membrane surface (a), rings A and B at the N-terminus of nisin bind to the pyrophosphate group of lipid II (b). The C-terminal region inserts into the membrane to form a pore by means of a flexible hinge mechanism (c). The pore complex has a stoichiometry of four lipid II molecules to eight nisin molecules (d).
However, Gram-negative bacteria lack lipid II, limiting universal antimicrobial activity of nisin. Moreover, nisin stability is pH-sensitive, with optimal bioactivity and solubility oc-curring at low pH [30], and polymerization at high pH leading to a decline in activity [16,31]. Additional factors, such as instability at neutral pH and fat and emulsifier content limit the antimicrobial effect in dairy products [32,33]. Bioengineering of nisin is aimed at increasing its potency and stability and generating variants with superior properties.
Structural studies of the nisin-lipid II complex [40][41][42] allow predictions to be made for physicochemical properties of variants and their potential interactions with the target molecule by analysis of charge quantity, hydrophobicity and amphiphilicity. Useful tools include Protein Variation Effect Analyzer (PROVEAN) to predict whether amino acid substitutions or indels will influence the biological function of a protein (http://provean.jcvi.org/ index.php, accessed on 9 June 2022). In addition, the Antimicrobial Peptide Database (ADP; https://aps.unmc.edu/, accessed on 9 June 2022) has collated 3324 antimicrobial peptides from various organisms, giving structural statistics and 3D structures. The Database of Antimicrobial Activity and Structure of Peptides (DBAASP, https://dbaasp.org/, accessed on 9 June 2022) gives detailed structural information (chemical, 3D) correlated with antimicrobial activity against specific target molecules. This platform uses amino acid sequences to calculate physical and chemical properties and predict antimicrobial activity, and it also allows strain-specific AMPs to be revealed within microbial genomes. Thus, identification of existing antimicrobial motifs and utilization of antimicrobial peptide databases, combined with advanced machine learning algorithms, allow appropriate nisin variants to be designed.
Nisin is a small peptide of 34 amino acids and less than 5 kDa, making it a suitable candidate for biological engineering ( Figure 2). Modifications have allowed the generation of nisin variants with enhanced antimicrobial activity, solubility, and ability to evade resistance.
Structural studies of the nisin-lipid II complex [40−42] allow predictions for physicochemical properties of variants and their potential interactions wit molecule by analysis of charge quantity, hydrophobicity and amphiphilicity. U include Protein Variation Effect Analyzer (PROVEAN) to predict whether amin stitutions or indels will influence the biological function of (http://provean.jcvi.org/index.php, accessed on 9 June 2022). In addition, the bial Peptide Database (ADP; https://aps.unmc.edu/, accessed on 9 June 2022) h 3324 antimicrobial peptides from various organisms, giving structural statist structures. The Database of Antimicrobial Activity and Structure of Peptides https://dbaasp.org/, accessed on 9 June 2022) gives detailed structural informa ical, 3D) correlated with antimicrobial activity against specific target molecule form uses amino acid sequences to calculate physical and chemical properties antimicrobial activity, and it also allows strain-specific AMPs to be revealed crobial genomes. Thus, identification of existing antimicrobial motifs and ut antimicrobial peptide databases, combined with advanced machine learning allow appropriate nisin variants to be designed.
Nisin is a small peptide of 34 amino acids and less than 5 kDa, making i candidate for biological engineering ( Figure 2). Modifications have allowed tion of nisin variants with enhanced antimicrobial activity, solubility, and abili resistance. We summarized nisin variants with improved anti-pathogen and anti-fo bioactivity and with other enhanced physicochemical properties reported in th research. Given the important role of nisin in the dairy industry, transformi meet the needs of practical applications has become a great trend. Furthermore We summarized nisin variants with improved anti-pathogen and anti-food spoilage bioactivity and with other enhanced physicochemical properties reported in the previous research. Given the important role of nisin in the dairy industry, transforming nisin to meet the needs of practical applications has become a great trend. Furthermore, these nisin variants with enhanced properties will lead to the discovery of promising new drug candidates.

N-Terminal Domain Variants of Nisin
The N-terminal binds lipid II with high affinity at nmol concentrations of nisin via interaction between the lipid II pyrophosphate group and main chain amide bonds of nisin, which form a cage-like structure by intermolecular hydrogen bonds [40]. N-terminal amino acid substitutions have been used to elucidate the structural factors controlling molecular recognition of rings A and B [41]. A truncated peptide (1-12) variant (Thr2, Ser5) failed to form the conformation binding lipid II in solution [42]. The structural integrity of rings A and C is essential for antimicrobial activity [43]. Variant (S3T) has an altered N-terminal conformation that affects lipid II binding, increasing the minimum concentration required for pore formation from 1 nM to 50 nM [22]. Most variants obtained by mutation at position 12 of the nisin core peptide had a decreased production level and incomplete post-translational modification, but variant K12A had enhanced antimicrobial activity against representative target strains [44].
Loss of rings D and E impaired antimicrobial activity [43] due to reduced ability to penetrate the cell membrane [45]. Some ring A variants had enhanced antimicrobial activity, such as nisin I4V with more potent antimicrobial and anti-biofilm activity against Staphylococcus pseudintermedius [46]. Combination of nisin I4V with conventional antibiotics also increased biological activity against Staphylococcus aureus [29]. However, incorporation of large amino acids into ring B had a negative impact on antimicrobial activity. There have also been reports showing no direct correlation between autoinduction and antimicrobial activities of variants [45,47]. Thus, variants with weakened antimicrobial activity but normal inducing activity may be bioengineering targets.
Substitution amino acid physicochemical properties of the first residue of the nisin core peptide [47] have an influence on yield, NisP cleavage efficiency, antimicrobial activity and recognition of nisin immune protein. For example, charged amino acids or those with large hydrophobic side chains at the I1 position reduced the NisP cleavage efficiency [47].

Hinge Region Variants of Nisin
The flexible hinge region connects the N-and C-terminal regions, and structureactivity studies of nisin and other lantibiotics have identified structural features essential for bioactivity in vivo and in vitro [22,40]. Hinge flexibility is crucial for C-terminal membrane translocation, and the three residues in this region affect stability and solubility. Since variation in this region impacts perforation activity and properties of lantibiotics [22,48], it is a promising target for bioengineering to enhance antimicrobial activity.
Saturated hinge mutations result in dramatic changes to nisin structure. Variants generated by this method (SVA, NAK) have been shown to control Listeria monocytogenes in complex media (commercially produced chocolate milk), and variants, AAK, NAI, and SLS displayed enhanced antimicrobial activity against various target strains. It can be inferred that small, chiral amino acids in the hinge region are beneficial in enhancing flexibility and thus antimicrobial activity. Variants AAA and SAA, generated by rational design, displayed enhanced activity against target strains [52].
Decreasing hinge region length led to a sharp decrease in antimicrobial activity. Residual activity was probably due to strong cell membrane deformation caused by nisin crowding at the membrane surface through the lipid II-independent antimicrobial mechanism. However, a nisin variant with reduced hinge length (NK) showed enhanced biological activity against Enterococcus faecalis at higher temperatures, possibly due to thinning of lipid bilayers as the temperature rises [48,53]. However, the predominant impact of reduced hinge region length is to impair pore formation, leading to increased concentrations being required for antimicrobial activity.
Variants with amino acid insertions (+1, +2 variants) have exhibited some enhanced antimicrobial effects against indicator strains with thicker membranes. Others had stronger antimicrobial activity at low temperatures. However, variants with insertion of three amino acids (+3) displayed dramatically reduced activity under all experimental conditions [48]. The variant 20 NIVMK 24 retained some antimicrobial activity and was not recognized by nisin-resistant proteins [54], making it an excellent candidate to deal with nisin resistance.
Most hinge region mutations, including insertions and deletions, did not dramatically affect nisin production in host bacteria, but there were some exceptions [48]. For example, production of variants N20D, K22D, K22E and +1 variants 20 NLMK 24 , 20 NVMK 24 , and 20 NIMK 24 was almost undetectable [49,50,54]. Reduced production may be due to the introduction of negatively charged residues into the hinge region, hindering the formation of ring D [49]. Other workers have shown that the introduction of negatively charged amino acids had no effect on the recognition of NisB and NisC [10].
Decreased hinge region length had a particularly significant impact on posttranslational modification. Dehydration reactions were affected [48], and variant ∆ 21 MK 22 lost its pore-forming activity. Precursor peptide cleavage and formation of lanthionine rings were also affected, indicating impaired recognition by the posttranslational modification enzyme, NisBCP [54].
Hinge region mutations not only affect antimicrobial activity but also alter the antimicrobial spectrum. Variants N20K and M21K displayed antimicrobial activity against Gram-negative Shigella, Pseudomonas and Salmonella [49]. M21F was active against Gramnegative bacteria from the genus Thermus, responsible for higher levels of redness in cheese samples [55]. Nisin variant M21V, alone and in combination with traditional antibiotics, enhanced anti-biofilm activity against Staphylococcus [29]. Subsequently, the combination of M21V and plant essential oils was shown to inhibit the Gram-negative bacteria Cronobacter sakazakii and E. coli O157:H7.
The hinge region plays a key role in biological and pore-forming activity and is a potential target for bioengineering.

C-Terminal Domain Variants of Nisin
The C-terminal region is involved in pore-forming activity and truncation affects antimicrobial activity to a degree proportional to the length of the truncation [43]. This is also the region of the protein associated with nisin resistance.
Some valuable C-terminal variants include N27K and H31K, which exhibit enhanced solubility at neutral pH, and S29A, S29D and S29E, which had enhanced antimicrobial activity against Staphylococcus aureus [16,56]. Nisin resistance protein from Streptococcus agalactiae (saNSR) bound to ring E of nisin and cleaved a site between the 28th and 29th amino acid residues of the core peptide [57]. After cleavage by saNSR, activity of the resulting nisin peptides decreased 400-500 fold [58]. However, some C-terminal domain variants evade the resistance mechanisms, and variants S29P and C28P retained considerable antimicrobial activity to inhibit strains expressing nisin immune protein [59,60].
The C-terminal region is also responsible for the initial interaction with the target membrane [21], and incorporation of negatively charged amino acids in this region has been found to adversely impact antimicrobial activity. Variant V32E had an increased minimal inhibitory concentration for Micrococcus flavus and Streptococcus thermophilus [21]. However, some nisin variants in which negatively charged amino acids were introduced into the C-terminal, such as S29A, S29D and S29E, had enhanced antimicrobial activity against Gram-positive drug-resistant bacteria and some food-borne Gram-negative bacteria.

Nisin Variants Containing Noncanonical Amino Acids
Standard methods for nisin engineering, such as alanine scanning, site-directed or saturation mutagenesis, have usually incorporated canonical amino acids into nisin variants. Noncanonical amino acids (ncAAs) are non-proteinogenic, except for pyrrolysine (pyl) and selenocysteine, and many different side chain structures and backbone configurations have been generated [61]. The possibilities of proteins containing ncAAs have greatly expanded the potential of nisin engineering.
Both residue-specific and site-specific approaches have been taken to incorporate ncAAs into proteins [62]. A residue-specific approach involves the feeding of amino acid analogues to nutritionally deficient strains to replace canonical residues without gene manipulation [63]. Derivatives of tryptophan, phenylalanine, proline and methionine have all been incorporated into target positions of nisin by selective pressure incorporation (SPI) [63][64][65]. NcAAs can be incorporated into specific sites by means of an orthogonal aminoacyl-tRNA synthetase (aaRS)-tRNA pair. Pyl-tRNA synthetase and its cognate tRNA from the methanogenic archaeon Methanosarcina mazei have been used to incorporate Nε-Boc-L-lysine (BocK) and two phenylalanine derivatives into nisin target sites [66,67]. Insertion of methionine derivatives with bio-orthogonal reactivity into nisin enables fluorescence labeling of nisin and coupling of nisin variants [68]. In addition, a variant in which novel macrocyclic topologies replaced the lanthionine ring was derived from reaction between electrophilic ncAAs and cysteine, although no bioactive conformation was formed [65].
Most ncAAs are synthesized by total-chemo or semi-chemo synthesis or extracted from natural substances [69]. The synthesis process of ncAA is complex and expensive, which is not suitable for large-scale production in industry [70]. In vivo synthesis of ncAAs from simple precursors may be a way to overcome the economic barrier and allow the full potential of nisin modified with ncAAs to be realized.

Nisin-Derived Hybrid Peptides
Hybrid peptides, which combine the sequence of one or more different peptides with nisin, have been synthesized [71] to enhance structural diversity and bypass specific bacterial resistance mechanisms [66,72].
A hybrid incorporating the nisin N-terminal sequence (1-22) responsible for lipid II recognition with vancomycin had a 40-fold enhancement of antimicrobial activity against vancomycin-resistant strains compared with native vancomycin [73]. Similarly, a lantibiotic, TL19, with two lipid II binding motifs, one at the N-terminal from nisin and the other at the C-terminal from haloduracin, has been synthesized. This hybrid peptide had 64-fold higher antimicrobial activity against the difficult-to-treat Enterococcus faecium than N-terminal nisin (1-22) and a 2-to 4-fold higher activity against Enterococcus faecium than native nisin. The hybrid peptide had no pore-forming activity, demonstrating that the additional lipid II binding sites completely compensated for loss of pore-forming activity [74]. These results look very promising for the development of new lantibiotics.
Structures that are not naturally occurring have also been introduced into peptides to improve antimicrobial properties. Peptoids (N-substituted glycines) have shown antimicrobial activity against microorganisms, fungi and parasites and have higher resistance to proteases than α peptides [72,75]. Nisin-peptoid hybrids containing the nisin lipid II targeting domain have increased in vivo stability and represent a way of increasing the chemical diversity. Significantly, nisin-peptoid hybrids retained the considerable activity of nisin against MRSA [66].
These promising novel structures have the potential to overcome some shortcomings of nisin and reduce drug resistance.

Production of Nisin Variants
Nisin variants are usually produced by transfection of E. coli with the nisin biosynthesis pathway or the expression of a nisin variant gene in a wild-type nisin producing strain from which the nisin gene has been deleted (Table 1). In addition, nisin can be produced by solid-phase peptide synthesis (SPPS), but this is an expensive method unsuitable for large-scale production. I1X variants influenced antimicrobial activity and the efficiency of the immunity and resistance proteins. [47] L. lactis NZ9800 1 saturation mutagenesis No N/A L. lactis subsp. diacetylactis DRC3 (expressing the nisin resistance protein (NSR))

M21V-M17Aha + M21VM17Hpg
The variant is the most active dimeric nisin construct [68] [55] Traditionally, the identification of lantibiotic variant antimicrobial activity has involved a long process of purification and characterization. However, downscaling and parallelization have been achieved. Peptide modules from 12 natural lantibiotics with different antimicrobial mechanisms have been organized according to the structural distribution of nisin and used to generate a combinatorial lantibiotic library containing almost 6000 variants. A high-throughput screening platform may then be used to obtain combinatorial lantibiotics and screen variants with enhanced antimicrobial activity against target pathogens [82].
Host fermentation conditions and expression of post-translationally modifying enzymes may be optimized to increase yield. For example, increasing tRNA glu expression enhanced NisB expression [65], leading to a higher level of nisin variant production. Resistance to nisin may be engineered by heterologously expressing the nisin resistance protein [83].
In vitro activation of nisin precursors prevents nisin poisoning of hosts, and precursor peptides may then be harvested by a series of steps. First, nisin precursor was purified and digested by the endoprotease NisP [59,68]. Second, indicator strains expressing NisP were used to evaluate the antimicrobial activity of nisin variants [63,78]. Third, nisin precursor was purified and digested by trypsin [48,65] but with some inhibition of antimicrobial activity [84]. An excellent, simplified method for removal of the lantibiotic leader sequence involved the incorporation of hydroxy acid into the precursor by pylRS-pylT pair, resulting in an ester linkage connecting the leader peptide, which may be hydrolyzed [19]. More recently, compartmentalization within the bacterial host has been exploited. Biosynthesis and modification of the precursor peptide took place in the cytosol, but lantibiotic release in the periplasmic space. This method is suitable for high-throughput screening of lantibiotic variants and provides an advanced concept for mass production without the additional step of removing the leader peptide [20].
Residue substitutions can also affect nisin variant production levels. The absence of threonine, serine and cysteine or the introduction of negatively charged residues in the core peptide did not affect NisB and NisC recognition, but antimicrobial activity declined sharply, perhaps due to loss of electrostatic interactions with lipid II [10]. Introduction of glutamic acid into the hinge region decreased production dramatically due to disruption of NisB-and NisC-independent steps [49]. Cleavage by NisP depends on the modification of precursor peptides; hence, factors affecting NisB and NisC will also affect NisP cleavage efficiency [85]. Low cleavage rates of the natural variant, nisin H, by NisP were restored by replacing phenylalanine with isoleucine [86], and integrity of the (methyl-)lanthionine rings is a prerequisite for recognition by the cleavage enzyme [85]. Therefore, the yield of nisin variants requires that the impact on residue changes be considered.
Identification of variant antimicrobial activity requires the following steps ( Figure 3). First, appropriate indicator strains for the specific practical application should be selected. Antimicrobial activity should be evaluated using an agar-based deferred antagonism assay. Factors such as temperature, variant solubility and indicator strain types all have an impact on the extent of the inhibition zone and should be taken into account [87] as expansion of the inhibition zone does not necessarily correlate with specific activity. Growth and kill curve assays should be conducted to determine antimicrobial inhibition within the indicator strains [88]. Following laboratory assays, model food trials would assess the antimicrobial agent activity in more complex media of various food types. In addition, the composition or thickness of the bacterial cell wall and plasma membrane content of negatively charged lipids also determine target cell sensitivity to nisin [21]. R PEER REVIEW 13 of 16

Conclusions
The state of knowledge regarding nisin variants, their physicochemical properties and some trends in production was reviewed. Variants enhance our understanding of the structure and organization of nisin and allow progress to be made in overcoming the inherent shortcomings of nisin. Knowledge of structure-function relationships for the constituent domains of nisin facilitates designing protein variants with specific objectives, and improvements in bioengineering technology have made variant synthesis approachable. However, generating a series of nisin variants with enhanced antimicrobial activity still presents a challenge, especially in complex environments.

Conclusions
The state of knowledge regarding nisin variants, their physicochemical properties and some trends in production was reviewed. Variants enhance our understanding of the structure and organization of nisin and allow progress to be made in overcoming the inherent shortcomings of nisin. Knowledge of structure-function relationships for the constituent domains of nisin facilitates designing protein variants with specific objectives, and improvements in bioengineering technology have made variant synthesis approachable. However, generating a series of nisin variants with enhanced antimicrobial activity still presents a challenge, especially in complex environments.