Optimizing Fermentation Strategies for Enhanced Tryptophan Production in Escherichia coli: Integrating Genetic and Environmental Controls for Industrial Applications

Abstract

Tryptophan is an essential aromatic amino acid widely used in the pharmaceutical, agricultural, and feed industries. Microbial fermentation, mainly using Escherichia coli, has become the preferred method for its production due to sustainability and lower costs. Optimizing tryptophan production requires careful control of various fermentation parameters, including nutrients, pH, temperature, and dissolved oxygen (DO) levels. Glucose, as the primary carbon source, must be fed at controlled rates to avoid metabolic overflow, which leads to by-product accumulation and reduced production efficiency. Nitrogen sources, both organic (such as yeast extract) and inorganic (like ammonium), influence biomass growth and tryptophan yield, with ammonium levels requiring careful regulation to avoid toxic accumulation. Phosphate enhances growth but can lead to by-product formation if used excessively. pH is another critical factor, with an optimal range between 6.5 and 7.2, where enzyme activity is maximized. Temperature control promotes growth and production, particularly between 30 °C and 37 °C. High DO levels increase tryptophan titers by boosting the pentose phosphate pathway and reducing by-products like acetate. Furthermore, surfactants and supplements such as betaine monohydrate and citrate help alleviate osmotic stress and enhance precursor availability, improving production efficiency. Careful manipulation of these parameters allows for high-density cell cultures and significant tryptophan accumulation, making microbial fermentation competitive for large-scale production.

1. Introduction

L-tryptophan is an aromatic amino acid discovered in casein extracts by Hopkins and Cole in 1901 [1,2]. This amino acid acts as a precursor in the synthesis of niacin and neurotransmitters and hormones such as 5-hydroxytryptophan, serotonin, and melatonin. Its use is widespread in the pharmaceutical industry [3,4,5,6,7]. 5-hydroxytryptophan regulates memory, appetite, digestion, sexual behavior, mood, and sleep. Depressed people have low levels of 5-hydroxytryptophan, while patients with anxiety have high levels promoting tension and fear [8]. Pellagra is a rare disease caused by niacin deficiency; the symptoms are dermatitis, photosensitivity, gastrointestinal symptoms, and neuropsychiatric disorders [9]. Tryptophan consumption can restore the levels of these compounds and improve the symptoms produced by alterations in their metabolism [8,10]. Additionally, it has sedative effects when administered at night, improving sleep quality [11] and enhancing the effectiveness of lithium in patients with bipolar disorder [12].

Tryptophan is used in the food and feed industry for its essential nutritional properties [3,13,14,15]. Humans and animals are unable to synthesize it internally, as its production entails a substantial metabolic cost; therefore, it is more advantageous to acquire it from feed [3,16,17]. Tryptophan improves livestock productivity; adding tryptophan in feed reduces the demand for traditional protein sources, such as corn-soybean meal, allowing the formulation of suitable feeds for each growth stage [18,19,20]. Ingested tryptophan increases appetite by regulating the production of 5-hydroxytryptophan, serotonin, and insulin. Feed intake is also favored by the production in the intestine of cytokines and indoles with anti-inflammatory properties, which reduce the population of pathogenic bacteria, improve peristaltic movements, and facilitate the absorption of nutrients [8,18,19]. Increased production of 5-hydroxytryptophan minimizes the production of cortisol, epinephrine, and norepinephrine, which are stress hormones; relaxed animals increase their resting time and reduce fighting events, which are favored by the stress of high stocking density, weaning, and transport [18,19]. Tryptophan supplementation increases the litter survival rate and weight of each piglet [18] while it strengthens the reproductive system of poultry by increasing egg frequency and weight; 5-hydroxytryptophan increases calcium absorption to improve eggshell quality [19]. This amino acid promotes the development of mammalian glands and milk production [19], and the concentration of growth-promoting hormones such as prolactin and insulin in milk is increased [18]. It is also necessary for T-cell proliferation and maturation and is responsible for increasing blood immunoglobulins, favoring the immune response [18,19].

Tryptophan is a precursor of auxins, essential plant growth regulators [21]. It can be applied by foliar spraying, seed priming, or soil application; tryptophan can be converted to auxins by microbes in the soil and absorbed by the roots [22]. In carrot cultivation, tryptophan applied to leaves and soil increases productivity in amounts as low as 7.5 g/ha and is enhanced when combined with glutamic acid [23].

Tryptophan can be obtained by chemical synthesis, protein hydrolysis, enzymatic conversion, and fermentations [1,20,24]. Glutamic acid [25] or γ-glutaraldehyde acid [26] can be chemically converted to tryptophan after reacting with other chemicals with the help of catalysts. Chemical synthesis is unsuitable for complex compounds [20,27], as it produces mixtures of L- and D- isomers that reduce their quality and specificity [7,28]. Obtaining tryptophan from protein hydrolysates using acids or enzymes depends on the availability of protein-rich sources [20]. Like most amino acids, tryptophan can be produced by microorganisms that consume low-cost raw materials such as sugars and ammonium [7,29,30,31,32,33], promising an affordable and sustainable alternative [15,34]. Tryptophan is one of eight amino acids produced in food grade from fermentations [20]. Commercial tryptophan is classified into high-purity powders (>98%) and low-purity granules (20–80%). Both can be obtained from fermentations, but powders require further processing [20]. High-quality and purity forms are preferred for formulating products; however, their use increases production costs [19]. Supplementary Table S1 lists companies producing and distributing tryptophan, highlighting the role of China, Japan, Germany, and South Korea in the commercialization. Tryptophan produced in these countries is exported worldwide in various products such as raw materials, feed additives, and capsules. These countries are also the most productive in research publications on tryptophan production through fermentation, as shown in Supplementary Figure S1.

High levels of production are desired at industrial scale [35]. Annual tryptophan production is estimated at 4.1 × 10⁴ metric tons yearly [36] and continues to increase [30,36,37]. Although other microorganisms have been used for tryptophan production [28,30,37,38,39,40,41] Corynebacterium glutamicum and Escherichia coli are the main study models and have been well-characterized genetically [1,14,34,37,42,43,44].

The tryptophan metabolic pathway in C. glutamicum is similar to that in E. coli. However, there are some genetic differences, such as that there is not a tryptophan operon repressor (TrpR) in C. glutamicum; instead, tryptophan and leucine production is regulated by a transcriptional repressor LtbR [45], and there is an internal constitutive promoter in the tryptophan operon (trp operon) that drives the expression of the tryptophan synthase genes trpBA [46]. C. glutamicum has some advantages over E. coli; for example, it can naturally utilize sucrose to produce biomass and metabolites [47] and is generally recognized as safe (GRAS) [20]. Tryptophan production in C. glutamicum can reach 58 g/L [48]. However, the overproducing strains described in scientific articles are auxotrophs for tyrosine and phenylalanine. These strains were obtained by random mutagenesis and selected using aromatic amino acid analogs [1,43,45,48,49,50,51,52,53,54,55,56,57,58,59]. This approach makes it difficult to recreate the general scheme on tryptophan production in C. glutamicum for rational strain design; although some strains have been designed with defined genetic modifications, these use randomly mutagenized strains, and some changes are adapted from studies on E. coli [43,48,49,53,54,56,59].

E. coli offers advantages such as rapid proliferation, ease of cultivation, extensive research background, metabolic plasticity, and a broad array of genetic tools [36,42]. The knowledge of tryptophan metabolism in E. coli is greater than that of C. glutamicum. Supplementary Figure S2 represents the most significant publications related to tryptophan in E. coli compared with C. glutamicum.

The production efficiency of a strain can be represented by the conversion rate of the carbon source to the desired product (Y_p/s) [60]; this rate should be maximized by modifying its metabolism to reduce production costs and be able to use them industrially [3,32,60,61]. There are two theoretical maximum conversion rates of glucose to tryptophan [6]. The first is 0.227 g/g of glucose and is used for strains that import glucose through the phosphoenolpyruvate:glucose phosphotransferase system (PTS) [38,61,62] since this consumes 50% of the phosphoenolpyruvate (PEP) produced by the cell and cannot be directly channeled to tryptophan production. The second value of 0.46 g/g of glucose is for strains that incorporate glucose by a system independent of PEP consumption [32]. Tryptophan production is still low compared to lysine and glutamate production [28]. The maximum conversion rates achieved are around 0.238 g/g of glucose, representing 50% of the theoretical maximum conversion rate, making industrial production processes more expensive and encouraging research to improve tryptophan production [3,15,27].

The production titer is the most commonly used parameter representing a strain’s efficiency. Paradoxically, the strains with the highest conversion rate are not always those with the highest titers [6]. To define which parameter is most important in tryptophan production, the cost of the carbon source must be considered versus the cost of purification in the subsequent processes.

Although much effort has been invested in developing new engineered strains and mathematical models to improve tryptophan production [28,35,63,64], their metabolic behavior is affected by specific environmental parameters and substrate availability [65]. The most notable example is the strain FB-04 harboring the plasmid pSV03, which produced 13.3 g/L of tryptophan [66], while a later publication managed to increase production to 38.1 g/L by modifying only the culture conditions [67]. The optimization of tryptophan production should be considered based on the culture medium formulation, carbon source supply, fermentation parameters, precursor supply, and simultaneous regulation of biosynthetic pathways and gene expression [35,68,69,70]. Genetic modification strategies used for tryptophan production have been previously reviewed in depth [71,72]. However, fermentation strategies have not been analyzed in the same detail [73]. This review will focus primarily on describing how fermentation conditions affect tryptophan production and, in a second part, on how these culture conditions can complement genetic modifications.

2. Obtention of Overproducer Strains

Overproducing strains have been developed through two main approaches: selection of randomly generated mutants and rational design. The latter approach is increasingly favored due to advances in systems biology, synthetic biology, and computational modeling, which allow more precise manipulation of metabolic networks.

2.1. Generation of Overproducing Strains from Random Mutagenesis

The creation of random mutations in the genome by physical, chemical, or biological agents generates genetic variability in the bacterial population, resulting in individuals with different levels of tryptophan production. To isolate the strains with the highest tryptophan production, it is necessary to place them in a selective environment [5,32]. This principle is the same as in Adaptive Laboratory Evolution (ALE), in which a microorganism is cultivated for long periods in an environment that exerts a selection pressure that favors the desired phenotypes, making their characteristics more pronounced [74].

Strains grown in the presence of tryptophan analogs are generally overproducers of tryptophan [37,75]. Tryptophan analogs have a high similarity with tryptophan, which allows them to substitute it in its functions, for example, by being incorporated interchangeably in most proteins, but with lethal consequences by inactivating some essential enzymes [76,77]. E. coli prefers to use tryptophan over its analogs, and its overproduction helps reduce toxicity [76]. Spontaneous mutations can confer resistance to tryptophan analogs through different oligogenic mechanisms [77]. Tryptophan analogs interact through their α-carboxyl group with the TrpR aporepressor, inhibiting the transcription of the trp operon in the same way that tryptophan inhibits its synthesis. Compounds with this property include 4-methyl-, 5-methyl-, 6-methyl-, 5-fluoro- and 6-fluoro-tryptophan. 4-fluoro-tryptophan-resistant mutants have a nonsense mutation in the trpR gene that derepress transcription of the trp operon [78]. 5-methyl-tryptophan is not incorporated into proteins but can reduce the activity of anthranilate synthase (trpE), which requires a complex with anthranilate phosphoribosyl transferase (trpD) to be fully active [79]. These enzymes catalyze the first two reactions of the tryptophan synthesis pathway. Anthranilate synthase has a conserved region around the Ser40 residue where tryptophan binds and changes the enzyme’s conformational structure, affecting the affinity of the binding sites for glutamine and chorismate, preventing the binding of these substrates. This change propagates to anthranilate phosphoribosyl transferase, preventing the formation of phosphoribosyl anthranilate [80,81,82]. Mutations in Ser40 of trpE reduce its affinity for tryptophan and its analogs, making it resistant to feedback inhibition [83].

Mutations in aromatic transporter proteins reduce toxicity by reducing the uptake of tryptophan analogs. The mutation in the aromatic amino acid permease (aroP) confers resistance to 4-fluoro-tryptophan [76]. Similarly, the tryptophan-specific transporter mtr is implicated in the 5-methyl-tryptophan resistance phenotype, from which it derives its name [53]. Mutation of the glucose transporter (ptsG) increases carbon flux toward serine production, conferring resistance to 5-fluoro-indole [77].

Genetic circuits have been designed to detect tryptophan production using fluorescent proteins to indicate its production level (see Figure 1). They are based on the tnaC regulatory element responsible for the transcription of the tnaAB operon, which promotes the intake and degradation of tryptophan when present [5] or by using riboswitches such as riboswitch-based high-throughput screening (HTS) [84]. With these biosensors, bacteria obtained after exposure to mutagenic agents can be sorted based on tryptophan production, which is indirectly measured by fluorescence using a Fluorescence-Activated Cell Sorter (FACS) [5,84].

Another strategy is CRISPR/Cas9-facilitated engineering with Growth-coupled and Sensor-guided in vivo Screening (CGSS), in which the CRISPR/Cas9 technique is leveraged to create a library of mutations targeting the aroG gene, which encodes the first enzyme of the shikimate pathway. The strains’ growth depended on their insensitivity to phenylalanine inhibition and a fluorescent protein estimated production. This strategy does not require equipment such as FACS to select cells showing enhanced production since strains not resistant to feedback inhibition do not grow when placed in a phenylalanine-rich medium [85].

The mutations obtained can occur anywhere in the genome and affect genes that have not previously been linked to tryptophan production, so it is necessary to take advantage of new sequencing technologies to identify them and be able to use them in the rational design of strains that improve tryptophan production [5].

2.2. Rational Design of Strains

The tryptophan biosynthetic pathway is extensive and complex, so several authors have arbitrarily divided it into modules that include the tricarboxylic acid (TCA) cycle, glycolysis (EMP pathway), the pentose phosphate pathway (PPP), the shikimate pathway, the chorismate pathway, as well as pathways involving glutamine, serine, and the terminal branch of tryptophan synthesis [6,15,42,60,86,87,88]. These pathways are tightly regulated because tryptophan formation consumes many resources and energy [89]. The regulatory mechanisms that have been described are attenuation, feedback inhibition, feedback repression, and feed-forward inhibition [15,27,42,84,86,87]. To enhance tryptophan production, researchers have rationally eliminated attenuation and repression of critical enzymes, overexpressed enzymes involved in rate-limiting steps, blocked competitive and degradative reactions, improved precursor supply, modified the transporter system, and enhanced the tryptophan biosynthesis pathway. Figure 2 resumes the essential genes, reactions, metabolites, and pathways involved in the production of tryptophan for several overproducer strains [14,36,89,90,91].

More than 40 individual genetic modifications have been identified in E. coli to enhance tryptophan production [92]. However, a small group of these modifications have been consistently used to achieve titers of more than 30 g/L, as shown in Figure 3. These essential modifications are closely related to the final pathway of aromatic compound production, particularly tryptophan, so their use is stable in diverse genetic backgrounds. TRTH is one of the most representative strains, which has been tested under different fermentation conditions and has served for the development of new strains used in other studies, achieving productions between 32.24 and 60.2 g/L [65,91,93,94], which is the highest titer documented in the academy.

Although tryptophan production in genome-reduced organisms has not been explored, it could be advantageous because a rational design has eliminated genes that are not relevant under laboratory culture conditions, reducing the cost of cell maintenance. An example is the strain MDS-205, an efficient producer of threonine [95].

3. Fermentation Process for Tryptophan Production

To simplify the tryptophan production process description, we will go through a typical fermentation process describing the main parameters that must be controlled at each stage. To avoid redundancy, the parameters are grouped into a specific stage; however, some may be important throughout the entire fermentation process; for example, acetate production must be controlled throughout the fermentation.

3.1. Fermentation Preparations

Fermentation to produce tryptophan is performed from an isolated colony activated in a rich medium, commonly Lysogen Broth (LB), and then transferred to a seed culture medium similar to the production medium but differing mainly in having a higher amount of yeast extract and nitrogen. This medium is intended to produce biomass and adapt the strain to the production medium. Various inoculum (seed) and production media used with different strains are described and listed in Supplementary Table S2. Depending on the fermentation scale, it can be performed in a flask controlling only the temperature and agitation, or it can be performed in a fermenter controlling all the relevant parameters by scaling up the fermentation to achieve the appropriate amount of inoculum. Fermentation in seed medium is used to inoculate the production medium, which is more focused on tryptophan production, so parameters such as dissolved oxygen (DO), growth rate, feed rate, pH, agitation, aeration, and temperature are controlled, as described in Figure 4.

In strain T13, biomass and tryptophan production increased as the temperature increased from 33 °C to 36 °C; 36 °C was the highest temperature tested, with the highest production obtained [15]. This elevated temperature could affect serine production, as a temperature of 30 °C has been reported to be better for serine production [96].

When the initial glucose in the fermentation medium is depleted, a concentrated glucose solution, sometimes mixed with other substrates, is pumped to feed the fermenter. The capacities of the fermentation equipment need to be considered, which can be a limitation for producing high-density cell cultures, as they can be physically limited in the distribution of nutrients in the medium [70]. For example, due to the large scale of industrial reactors, zones with significantly high glucose concentrations are formed near the feed inlet, triggering the formation of acetate [97]. Increasing the stirring speed cannot correct this phenomenon because high shear forces cause mechanical damage to the bacteria [44]. In the following subsections, we will describe the main conditions that should be observed that affect tryptophan production and, therefore, shall be considered in media preparations and fermentation parameters.

Fermentation is stopped when tryptophan accumulation stops increasing.

Table 1. Parameters and genes that are involved in affecting tryptophan production.

Parameter	Pathways	Genes *	Effect	Ref.
Low DO	Acetate, glyoxylate shunt, TCA cycle	pta, ackA, poxB, tdcD, gltA, icd	Acetyl-CoA, pyruvate, lactate, and acetate accumulation.	[14,65,68,93,98,99,100]
High DO	Acetate degradation, glycolysis, gluconeogenesis, PTS, PPP, glutamate and tryptophan	fba, rpoS, acs, pckA, ppsA, ppc, maeA, pfk, zwf, pgl, tktA, araD, aroG, aroK, trpEDCBA, gdhA, pntAB	Acetate consumption, production of PEP, E4P, NADPH, tryptophan, and glutamate, reduction in glucose consumption, and oxidative stress.	[14,65,68,98,101,102,103,104]
Low pH	Glycolysis, acetate, and TCA cycle	pfk, sucB, sucC	Low growth rate, stable plasmids, low acetate production, and tryptophan crystallization.	[14,37,44,68,105]
High pH	Acetate, shikimate, and tryptophan	pta, aroK, tnaA, tnaB	Accumulation of acetate, NH4+, and K+, and tryptophan production.	[7,68,105]
Low growth rate	Glycolysis, acetate		Plasmid stability, low acetate production, increased tryptophan production.	[13,28,91,105]
High growth rate	Glycolysis, acetate		Biomass and acetate accumulation.	[13,97,105,106,107]
Low feed rate	Glycolysis, acetate, and tryptophan		Low production of acetate, biomass, and tryptophan.	[14,38,65,68,93,98,106,107,108]
High feed rate	Acetate and glutamate	gdhA	Acetate and glutamate accumulation.	[14,38,86,97,100,105,107,109]
Calcium	TCA cycle	gltA, icd, sucC	Increase ATP and biomass production	[67,69]
Betaine	PPP, stress response, and K+ uptake	zwf	The activation of PPP, osmotic stress relief, reduced glutamate accumulation, and reduced K+ uptake.	[110,111]
Citrate	Ca2+ uptake, glycolysis, PPP, and TCA cycle	pykAF, zwf, gltA	Reduce production of acetate and glutamate while increasing accumulation of PEP, NADH, NADPH, and ATP.	[87,112]
High organic nitrogen			Increase acetate and biomass production.	[94,112]
Inorganic nitrogen	Glutamate, tryptophan, glutamine	glnA	Increase energy waste, change pH, affect ionic strength, provide NH4+ for tryptophan production, and reduce plasmid stability.	[13]
Phosphate	Glycolysis		Increase growth rate, increase acetate accumulation.	[7,70,113]
Surfactants	PPP, tryptophan transport, glutamate	zwf, yddG, trpBA	Enhance crystallization and secretion of tryptophan, reduce inhibition, and make it a sink for indole.	[4,7,113]
Cation’s accumulation			Changing pH and ionic strength makes the purification of tryptophan difficult and increases glutamate accumulation.	[68,105,111,114,115]
Precursors	Tryptophan, isoleucine, threonine	trpC, serA, thrA	Improve tryptophan production and feedback inhibition of precursor pathway.	[4,68,106]
Phenylalanine			Relieves serine inhibition and blocks common pathways of aromatic amino acids.	[15,116]
Leucine			Relieves serine inhibition.	[116]
Isoleucine	Isoleucine		Relieves serine inhibition.	[116]
Methionine	Methionine		Increase production of proteins and reduce accumulation of acetate.	[3,31,84,107,112]
The high volume of inoculum			Increase rate growth and reduce lag-phase.	[70,117]

* Genes symbols are the same as in Figure 2.

summarizes the main parameters analyzed in this review for tryptophan production.

3.2. Acetate Metabolism

Despite efforts that have been made to improve tryptophan production, the generation of by-products is a common problem, with acetate being the main inhibitory metabolite [13,14,65,67,118]. Acetate profoundly affects bacterial metabolism, reducing glucose consumption, inhibiting cell growth, and causing autolysis that affects the production of desired metabolites [93,98]. Growth is inhibited by the depletion of the bacteria’s energy, which is represented by its adenosine triphosphate (ATP) reserves. When acetate is protonated, it can diffuse through the membrane into the cytoplasm, lowering the pH by dissociating from its H⁺, and to compensate for acidity, bacteria need to pump protons into the extracellular milieu, expending ATP, becoming a useless cycle that causes an energy imbalance [42,93]. Acetate accumulation stimulates the formation of other by-products and is toxic in the mid-log phase, making it challenging to achieve high cell density [65,108], concentrations higher than 40 mM [93,100] affect the expression of recombinant proteins, reducing gene overexpression, especially in modified strains [65,107]. Longer growth phases are achieved in fed-batch than in batch fermentations, so acetate accumulation significantly affects biomass production [107].

Acetate reduces the activity of isocitrate dehydrogenase by promoting its phosphorylation by the enzyme phosphatase/kinase (aceK), which triggers the activation of the glyoxylate shunt [101,102]. Increased flux in the glyoxylate shunt reduces oxidative respiration, which affects growth, a result also seen when the genes that form the glyoxylate shunt (aceBA) are overexpressed. Furthermore, the aldehyde group of glyoxylate makes it highly reactive and toxic, so it is rapidly metabolized by malate synthase (aceB) [102].

3.2.1. Overloaded Metabolism Driving Acetate Accumulation

Acetate is produced via two pathways composed of genes pta-ackA and poxB. In the main path, acetyl phosphotransferase (pta) converts acetyl-CoA to acetyl phosphate, which is then dephosphorylated to acetate by the acetate kinase (ackA) with concomitant production of ATP. The second pathway involves a pyruvate oxidase reaction that converts pyruvate to acetate (poxB) (see Figure 2). Acetate production responds to at least three environmental conditions: metabolic overflow when excess glucose is available, low DO, and high pH values [105,106]. Some mechanisms to reduce acetate production are described in Figure 5.

E. coli tends to overload its metabolism when glucose is in excess, which is inconvenient because the high flux through glycolysis generates pyruvate and acetyl-CoA that cannot be entirely processed by the TCA cycle, favoring the conversion of excess to acetate in a process known as acetogenesis [13,14,97,100,105,106,107].

Acetate accumulation is related to glucose consumption and growth rate. A low growth rate reduces acetate formation and favors the availability of DO [93,107]. Many fed-batch fermentation strategies seek to reduce acetate production to achieve high-density cultures. One of the most used strategies is to reduce the feed rate to maintain glucose concentration below 2 g/L. This limitation reduces growth and prevents metabolic overflow [14,65,93,98,106,107,108]. Acetate accumulation above 0.5 g/L begins to prevent cultures from reaching high cell density because it starts to inhibit the growth of E. coli [14,67,86,107,119]. In the case of strain TRJH, adding 2 g/L acetate reduces biomass and tryptophan production by 11.0% and 17.9%, respectively [107].

3.2.2. Acetate Production Is Regulated by pH

Acetate production increases with a high pH. Acetate concentration increased from 6 to 12 g/L when the pH changed from 6.0 to 7.5 in the Dpta/mtr-Y strain [68]. This increase in production could serve as a mechanism to neutralize high pH by activating the transcription of the pta gene [105]. Conversely, reduced pta activity may redirect carbon flux through the PPP and the common aromatic amino acid pathway [67].

A low pH induces the activation of the sucB and sucC genes, which encode subunits of the 2-oxoglutarate dehydrogenase and succinyl-CoA synthetase complexes, respectively. This activation enhances the TCA cycle’s efficiency, enabling the exploitation of the high proton potential. Thus, the TCA cycle’s increased consumption of acetyl-CoA may reduce the metabolic overflow that produces acetate [14,68,105].

3.2.3. Dissolved-Oxygen Affects the Acetate Metabolism

DO regulate genes involved in acetate production at the transcriptional level. The main acetate synthesis pathway is expressed at low and high DO levels [68,86,98,118], so it has been considered a constitutive pathway [100]. However, this pathway is promoted at low DO levels with the transcriptional activation of the tdcD gene, whose product has acetate kinase activity like ackA [68]. This reversible pathway can produce trace amounts of acetyl-CoA when acetate is too high [86]. The secondary pathway contributes less to acetate production [93]. However, it is activated in low oxygen levels [68,98] or in the presence of acetate [109].

The citrate synthase enzyme encoded by the gltA gene incorporates acetyl-CoA and oxaloacetate into the TCA cycle, oxidizing acetyl-CoA to CO₂ [100]. This enzyme can carry out a high metabolic flow; under aerobic conditions, about 70% of the total glucose is consumed via glycolysis and the TCA cycle [42]. In the JB102/pBE7 strain, up to 47.78% of the carbon atoms can be consumed to produce CO₂ [38]. Low DO levels decrease the TCA cycle activity by low-regulating the citrate synthase, causing acetyl-CoA to accumulate and be redirected toward acetate production [93,100]. Acetyl-CoA accumulation, in turn, inhibits the pyruvate dehydrogenase complex, increasing the accumulation of its substrate pyruvate [93]; in strains that have a reduced capacity to produce acetate, pyruvate is redirected to lactate and glutamate synthesis [14,65,93,98].

Under aerobic conditions, the enzyme isocitrate lyase (ICL) activity is lower than the isocitrate dehydrogenase (ICD), channeling the carbon flow to the TCA cycle. In contrast, when DO decreases, the isocitrate dehydrogenase activity is reduced, redirecting the carbon flow to the glyoxylate shunt, which can minimize acetate accumulation [102].

High oxygen levels reduce the transcription of the fba gene, which limits the activity of fructose-1,6-bisphosphate aldolase in glycolysis and prevents the accumulation of pyruvate and acetyl-CoA, both precursors of acetate. In addition, it promotes the consumption of these precursors by increasing the transcription of genes, such as pckA, ppsA, ppc, and maeA, which encode enzymes that interconvert pyruvate, phosphoenolpyruvate (PEP), oxaloacetate, and malate, playing key roles in gluconeogenesis, glycolysis, and Nicotinamide Adenine Dinucleotide Phosphate Reduced (NADPH) production [68]. Reducing the accumulation of these metabolites increases the efficiency of the conversion of glucose to tryptophan [68,98].

Although acetate is considered a waste product, it can be consumed with glucose and plays a vital role in the overall regulation of the E. coli metabolic network. However, glucose consumption decreases when acetate accumulates [108] because acetate represses the transcription of genes involved in the PTS for glucose uptake and those for the TCA cycle [109]. Increasing DO concentration drives acetate consumption, decreasing its accumulation [14]. The rpoS gene is required to transcribe several stress response genes that prepare bacteria to enter the stationary phase; one of these genes is acs, which is repressed at low DO levels and activated in the late fermentation phase when there is more DO [98,100]. In the stationary phase, acetate accumulation inhibits its recycling through the pta-ackA-acs pathway [98,109].

Biomass and tryptophan production does not follow the trend of acetate about DO levels. The highest biomass and tryptophan yields are 30% and 20%, respectively, requiring an intermediate DO concentration to achieve maximum production [98].

3.3. Lag Phase of Fermentation

After being inoculated in a new medium, bacteria undergo a temporary period of adaptation, without replication, and with low nutrient consumption, which prepares them to grow optimally, taking advantage of the new nutrients [120]. Higher amounts of inoculum reduce the lag phase and allow it to enter the exponential phase quickly because it contains many extracellular enzymes that favor using fresh medium and allow it to adapt more rapidly. Increasing the inoculum also increases the final production of biomass and tryptophan [70,117]. However, too much inoculum causes excessively accelerated growth, favoring the autolysis that occurs in the last stages of fermentation. Hence, a 20% inoculum (v/v) equivalent to 5–6 g/L of dry cells was the optimal value for the TRTH strain [70].

Most fermentations to produce tryptophan have been efficiently carried out using glucose, which is cheap but has the disadvantage of being obtained through agro-industrial processes, which compete with the food and feed industries with socio-economic implications [27]. Alternatively, other carbon sources have been used, such as glycerol, which is obtained as a by-product of biodiesel production, has high reducing power, is inexpensive, and enters the cell through protein-assisted facilitated diffusion without consuming PEP, which is an essential precursor of aromatic amino acids [27,60,89].

Strategies implemented to use other carbon sources may vary due to changes in the distribution of metabolic fluxes; for example, glycerol is incorporated into cellular metabolism through gluconeogenesis [89]. When glycerol is used as a carbon source, and 3-deoxy-D-arabino-heptulosonate 7-phosphate (DAHP) accumulates, or there is a phosphate deficiency, the highly toxic methylglyoxal, an electrophile that reacts with proteins and DNA, is generated [27].

In strains FB-04 and FB-04 (ΔfruR), using mixtures of sugars such as xylose, fructose, and glucose produces lower tryptophan titers than when glucose is used alone [88]. Feeding xylose is promising because it is directly metabolized by the PPP to produce the precursor D-erythrose 4-phosphate (E4P) [71].

3.4. Early Exponential Phase

Agitation, oxygenation, and feeding rates must be controlled in the early exponential phase to meet the needs of the growing bacterial population. The fermentation medium is formulated with an initial carbon source, which is depleted in this phase, prioritizing the accumulation of biomass that consumes rich nutrients from the medium (trophophase) [32,71]. Increasing initial glucose enhances biomass production; for the TRTHΔpta strain 10 g/L, initial glucose accumulates the highest biomass [93]. Higher glucose concentrations generate osmotic stress, reduce TCA cycle activity through the Crabtree effect, and induce acetate accumulation, affecting biomass production [93].

Initial glucose depletion means an active metabolism and reduces DO levels. DO can be maintained by increasing the agitation rate and oxygen feed; however, after the optical density (OD) reaches a certain level, agitation, and oxygenation are insufficient to control DO, and the growth rate must be limited [14]. The growth rate is a critical parameter in fed-batch fermentations. It can be controlled by manipulating the feed rate of a limiting substrate, commonly glucose, to produce the metabolite of interest and maintain the biomass in a stage known as idiophase [32,93,105,119].

By reducing the growth rate, bacteria require fewer amino acids for protein production, allowing resources to be redirected toward tryptophan synthesis and secretion [28,91,121]. If feed rate and growth rate are not coordinated, low substrate concentrations result in low productivity, with prolonged culture times; conversely, overfeeding induces metabolic overflow that favors the generation of by-products, such as small organic acids, which can result in carbon flux is not being adequately incorporated into the tryptophan production pathway, causing bottlenecks [14,38,86,97,105,109]. Feeding strategies for tryptophan production are discussed in Section 3.5.

3.4.1. Growth Rate Affects Plasmid Stability and By-Product Formation

The specific growth rate should be controlled below 0.25 1/h to avoid the formation of by-products and favor the accumulation of the desired metabolites [13,105,107]. A threshold growth rate between 0.14 and 0.17 1/h has also been recommended, although this value may depend on the strain and the culture medium [13,107,113]. For the TRTHΔpta strain, a rate of 0.25 1/h produces the highest biomass and tryptophan. A low growth rate, especially during the early fermentation phase, reduces acetate accumulation, extending the fermentation for a longer time and maintaining growth in the late phase, which favors biomass accumulation [65]. Lactic acid and alanine are other by-products that can accumulate when the growth rate is too high. This also leads to the accumulation of NH₄⁺ due to the utilization of NH₃∙H₂O to neutralize them, which affects biomass production [13].

Plasmids are commonly used to overexpress genes that enhance metabolite production. Bacteria that maintain a plasmid show impaired growth rates compared to those that do not, and bacteria with unstable plasmids tend to lose them or accumulate adaptive mutations that affect antibiotic resistance or gene expression [122]. Plasmid stability is reduced when the growth rate increases, affecting metabolite production [13,105]; this instability may be due to the accumulation of short-chain organic acids such as acetate that interfere with nucleotide, lipid, and protein synthesis and to the accumulation of NH₄⁺ [13]. However, highly stable plasmids could be used in industrial-scale fermentations even without adding antibiotics [123].

3.4.2. High Phosphate Boosts Glucose Uptake

Optimizing the activity of glycolytic enzymes improves tryptophan production [35]. Phosphate is essential for cellular metabolism; it is part of ribonucleic acids, membrane phospholipids, and intermediates glycolysis and PPP [70]. Adding KH₂PO₄ to the medium activates glycolysis and enhances biomass production by increasing the growth rate in the initial fermentation stage; its participation in glycolysis makes its consumption proportional to glucose. However, excessive KH₂PO₄ values enhance glucose consumption so much that it causes an overflow of the metabolism that ends in the accumulation of organic acids. Tryptophan production increased as KH₂PO₄ increased to certain levels and then decreased; concentrations higher than 10 g/L facilitated the accumulation of acetic and lactic acid, exceeding 5 g/L and affecting growth, possibly due to a metabolic overload of the TCA cycle and the tryptophan pathway due to the high activity of glycolysis. Optimizing only the amount of inoculum and KH₂PO₄ in the TRTH strain allowed it to reach one of the highest titers of tryptophan in E. coli (59.55 g/L) [70]. In the W3110-ZDrr strain, 20 g/L of phosphate instead of 7.5 g/L increased tryptophan production from 15.8 to 26.8 g/L, but 25 g/L of phosphate resulted in growth inhibition [113].

Due to its ubiquity, phosphate can affect plasmid expression in E. coli, affecting the expression of recombinant proteins [70]. Tryptophan synthase activity improves as the concentration of phosphate buffer increases, and phosphate buffer turns out to be more suitable than borate or Tris [7].

3.4.3. Sodium Citrate Boosts Cofactor Production, Enhancing Biomass Growth

Adding sodium citrate enhances tryptophan production in E. coli, with an optimal concentration of 2 g/L improving tryptophan titers from 33.8 to 35.7 g/L in strain TRTH. The production of other metabolites in bacteria suggests that sodium citrate increases Ca²⁺ uptake, inhibiting phosphofructokinase, and pyruvate kinase enzymes, resulting in less glucose processing in glycolysis. Inhibition of pyruvate kinase increases the availability of the critical precursor PEP by reducing the production of pyruvate and acetyl-CoA; it also reduces the formation of acetate and lactate. Glycolysis is also weakened by increased activity of the competing PPP due to activating the enzyme glucose-6-phosphate dehydrogenase. The oxidative branch of the PPP produces NADPH, which is necessary to produce tryptophan by balancing the availability of precursors [112].

Another pathway affected by citrate is the TCA cycle. Citrate inhibits the citrate synthase, reducing the incorporation of acetyl-CoA into the TCA cycle [112]. Decreased TCA cycle activity prevents glutamate accumulation, which has been observed in some strains [117].

The addition of citrate promotes the production of ATP, NADH, and NADPH, which are vital for energy production, metabolic regulation, and the maintenance of cellular functions. In addition, it reduces the accumulation of by-products, promoting the formation of biomass [112]. Adding 6 g/L of citric acid to the medium provides intermediates to reactivate the TCA cycle in strain TRP04, which lacks the enzymes pyruvate kinase and phosphoenolpyruvate carboxylase, responsible for converting PEP into precursors that feed the TCA cycle [87].

3.5. Feeding Strategies During Fermentations

The supply of glucose and other important substrates for fermentation must be balanced so as not to deprive the bacteria of nutrients without generating a metabolic overload. In addition, the growth rate must be controlled so that the availability of oxygen does not stimulate the formation of by-products. For this reason, there are several strategies that seek to adapt to the requirements of each strain.

3.5.1. Exponential Feeding

The exponential feeding strategy aims to tune the specific growth rate (µ) to a predetermined value by controlling the feed rate; this is calculated using an equation that considers how the glucose demand grows exponentially during the mid-logarithmic phase after biomass production [13,113,124]. One of the goals is to limit the specific growth rate below a threshold value (µ_critical) to avoid acetate accumulation. For example, for strain TRJH, a growth rate of 0.2 1/h has been observed to give the maximum tryptophan and biomass production [107], while Liu and collaborators selected a value of 0.15 1/h for several strains [60]. Although this approach seems simple in theory, slight variations in growth rates trigger glucose starvation and overfeeding intervals, which reduce biomass and tryptophan production, contributing to by-product formation [97,113]. Attempts have been made to improve this strategy by eventually correcting the feeding function with real-time measurements of glucose concentration by varying the growth rate at different stages of fermentation in a process known as pseudo-exponential feeding rate [113] or by combining the pseudo-exponential feeding with the glucose-stat strategy at the end of fermentation [107].

The exponential feeding strategy is very effective during the exponential growth phase, where E. coli can take advantage of high feeding rates without acetate accumulation. However, when the growth rate is reduced at late stages, this strategy causes an accumulation of up to 80 g/L of glutamate due to metabolic overflow, as demonstrated in strain JB102/pBE7. For this reason, decreasing the feeding rate when E. coli reduces its growth rate while still producing tryptophan improves the glucose conversion rate [38].

3.5.2. Glucose-Stat Feeding

Lack of glucose in fermentation impairs bacterial metabolism, affecting the production of metabolites [107]. Glucose concentrations close to 1 g/L produce high concentrations of biomass and tryptophan [113], although concentrations lower than 0.2 g/L have also been used to reduce acetate production [38]. In the TRJH strain, it was possible to produce 51.4 g/L of biomass and 36.3 g/L of tryptophan by maintaining the glucose concentration at 0.15 g/L with a maximum specific growth rate of 0.275 1/h [107]. Adjusting glucose concentration is an effective control to regulate microbial metabolism, mainly if performed by computers that predict glucose consumption behavior. Performing measurements and controls manually is a tedious process if online measuring instruments are not available, so other strategies have been adapted to regulate glucose concentration through indirect measurements of its concentration, taking advantage of fermenters’ most used sensors.

3.5.3. DO Feedback Control

DO levels affect tryptophan biosynthesis and glucose consumption [68]. Metabolic reactions deplete available oxygen; a high oxygen demand indicates an accelerated growth rate with high glucose consumption. As glucose is depleted, metabolism is reduced, increasing DO levels; therefore, glucose availability and growth rate can be indirectly controlled by maintaining DO levels [98]. In this strategy, when an upper limit of DO is reached, the system is fed with glucose until DO reaches a lower limit, effectively avoiding glucose overfeeding and DO limitation. Limits are typically between 20 and 30% because E. coli can suffer from starvation glucose when DO exceeds 40% saturation [106]. This control strategy can reduce acetate excretion and increase tryptophan and biomass production. However, significant variation in glucose and DO throughout fermentation can lead to glucose starvation and metabolic overflow with subsequent acetate production [106]. Despite this, one of the highest titers in the TRTH strain, 59.55 g/L, was achieved following this strategy [70].

3.5.4. DO-Stat Control

DO-stat feeding can prevent substrate overfeeding and oxygen limitation [93,98,107]. When glucose is the limiting reagent, a DO level of 5% corresponds to 0.35 g/L residual glucose, while 50% corresponds to less than 0.10 g/L [98]. The DO-stat control strategy aims to maintain the DO level at a constant value by regulating glucose feed so that the glucose concentration and growth rate remain fixed; this strategy has been applied for phenylalanine production with almost zero acetate accumulation [98,107,110]. A commonly used value for DO in this type of fermentation is 20% [93,106].

Maintaining oxygen consumption at an optimal fixed value reduces the risks of glucose starvation and overfeeding. This strategy leads to low glucose concentrations by reducing acetate levels to less than 0.61 g/L. This strategy can achieve better results for biomass and tryptophan production than DO feedback control; however, it has a lower conversion rate from glucose to tryptophan than E. coli mutants for the acetate production pathway [106].

3.5.5. DO Stage Control Strategy

The DO stage control strategy can take advantage of the metabolic state of E. coli by using different DO levels for each fermentation stage [98]. Fermentation can be divided into two phases: biomass and tryptophan production. For the E. coli TRTH strain, the highest tryptophan production is achieved at a DO level of 20% (0–20 h) and 30% (20–40 h); these conditions lead to reduced accumulation of by-products such as lactate, acetate and glutamate at the beginning of fermentation, allowing to produce more biomass and stimulate the transcription of genes required for tryptophan biosynthesis at the end of fermentation when biomass production is less active [68,98].

3.5.6. pH-Stat Control

Glucose consumption and organic acid production in E. coli are linearly correlated. Also, the amount of alkali required to neutralize these organic acids is proportional to the amount of glucose consumed. The pH-stat feed control strategy uses a feed solution in which glucose and alkali are mixed to regulate the pH of the fermentation and maintain the glucose concentration at a steady state [105].

3.5.7. pH Stage Control Strategy

As with the DO control strategy, fermentation can be divided into phases to shift the range of optimal pH values, allowing for better tryptophan and biomass production [105]. A low pH in the initial fermentation phase reduces glycolysis activity, acetate accumulation, glucose consumption, and growth rate. A high pH increases the activity of enzymes involved in tryptophan production, but more ammonia is required to maintain the high pH. Using more alkaline solution to raise the pH inhibits growth due to NH₄⁺ accumulation, reducing energy efficiency [68].

The pH with the highest tryptophan and biomass production in the Dpta/pta-Y strain is pH 7.0 (0–20 h) and 6.5 (20–40 h), leading to a high glucose conversion rate, producing 52.15 g/L biomass and 52.57 g/L tryptophan. The pH shift between fermentation stages reduced the activity of the critical gene fbp, decreasing the flux through glycolysis and glucose consumption. In contrast, genes involved in PPP (zwf, pgl, and araD) and tryptophan biosynthesis (pckA, ppsA, tktA, aroG, aroK, and trpEDCBA) were more actively transcribed. The transcription of genes involved in acetate production (poxB, ackA, and tdcD) decreased, and the acetate-consuming gene acs was transcribed more actively [68].

In the BCTRPG strain, fermentation was also divided into stages to adjust the pH, starting with low values and increasing them over time (6.5 from 0 to 12 h, 6.8 from 12 to 24 h, and 7.2 from 24 to 39 h). This strategy achieved higher tryptophan production than when a constant pH of 7 was used during fermentation [14].

3.6. Mid-Exponential Phase

In the mid-exponential phase, bacteria actively divide at a constant and maximal rate, increasing biomass. The biomass in fermentation contains all the enzymes, precursors, and energy required for tryptophan synthesis [27,32]. Increased tryptophan production depletes cellular resources and can produce metabolic overload that limits growth [32,42,102]. During the mid-exponential phase of fed-batch fermentations, tryptophan production is proportional to the available biomass [118]. Theoretically, if the production rate is maintained in high-density cell cultures, titers of the desired metabolite and fermentation efficiency will be improved while production costs are reduced [70,87].

Exponential growth increases oxygen consumption in this phase. DO availability profoundly affects gene transcription and metabolic flux distribution. Among the main pathways affected by DO levels are glycolysis, PPP, TCA cycle, and cytochrome activity [68,98]. High DO level enhances the PEP, E4P, and NADPH supply, positively affecting tryptophan production [98].

PEP availability can be increased by reducing glucose consumption through PTS [34,98,125]. High DO levels reduce glucose uptake because the TCA cycle is activated, which increases ATP production that inhibits the activity of phosphofructokinase, encoded by the pfk, which catalyzes one of the first steps of glycolysis [98]. Glycolysis and PPP branch at glucose-6-phosphate, determining the flux between these two pathways [35]. Changes in DO from 5% to 20% in the THTR strain can increase carbon flux to the PPP by up to 2.25-fold, providing more E4P for aromatic amino acid production. Thus, by favoring the oxidative branching of the PPP, the cofactor NADPH is produced, which drives biomass and tryptophan synthesis [98].

Elevated DO levels stimulate gluconeogenesis by increasing transcription of the genes pckA and ppsA, which recover intermediates from the TCA cycle and channel them toward glucose synthesis, thereby increasing PEP availability. The PPP is also activated by increasing transcription levels of its genes zwf, pgl, tktA, and araD, which provide E4P for tryptophan synthesis [68,98]. Finally, elevated DO levels stimulate transcription of the genes aroG, aroK, and trpEDCBA, which are involved in the aromatic amino acids and tryptophan synthesis [68].

3.6.1. Escherichia coli Metabolism Lowers pH, Impacting Glucose Uptake and Tryptophan Production

Fermentation pH can influence DO levels, growth rate, and metabolite production [14,68]. E. coli can adapt its metabolism to grow at a pH between 4.4 and 9.2, although it usually acidifies the medium by producing organic acids because of glucose metabolism [105]. Low pH reduces its growth rate due to decreased respiratory activity that increases DO, improves plasmid stability, and reduces acetate production [37,44,105]. Low pH reduces the activity of enzymes such as phosphofructokinase (pfk), becoming a rate-limiting step in glycolysis, reducing glucose consumption, growth rate, and tryptophan production. However, attenuation of glycolysis also reduces acetate accumulation because it reduces pyruvate and acetyl-CoA production [68].

The tryptophan accumulation is also affected by pH. Enzymes of shikimate and tryptophan degradation pathways encoded by the genes aroK, tnaA, and tnaB change their activity depending on the pH. For example, the enzyme shikimate kinase I (aroK) is induced at high pH [105], while the highest tryptophan productions from indole and serine have been achieved at pH 8 [7].

Multiple factors determine the optimal pH for tryptophan production: lowering the pH to prevent acetate accumulation affects stable and optimally active enzymes at near-neutral pH [68]. In Aureobacterium flavescens, maintaining pH at near-neutral levels extends the active growth period and enhances tryptophan production [37]. Using high pH to activate enzymes involved in tryptophan production requires using alkaline solutions, which causes the accumulation of cations such as NH₄⁺, Na⁺, or K⁺ that make the medium toxic to bacteria. On the contrary, acidic pH reduces the solubility of tryptophan, favoring its crystallization and preventing feedback inhibition [68]. Although studies have determined that the optimal pH for tryptophan production is between 6.5 and 7.2 and maintaining the pH at 6.8 can produce 32.15 g/L of tryptophan in the TRJH/ΔPB strain, fermentations can be divided into several stages to first improve biomass production at low pH by avoiding acetate production and then increase the pH for tryptophan production [105].

CaCO₃ can be used as a buffer in fermentations for tryptophan production [67,126]. Ca²⁺ is a catalytic cofactor that activates enzymes of specific metabolic pathways, affecting the distribution of metabolic fluxes, cell growth, morphology, and accumulation of desired metabolites. CaCO₃ has been observed to strengthen the TCA cycle and increase ATP production, which may benefit biomass production. The core genes activated by CaCO₃ are gltA (citrate synthase), icd (isocitrate dehydrogenase), and sucC (succinyl-CoA synthase), which belong to the TCA cycle [69]. Isocitrate dehydrogenase can complement the activity of PPP in NADPH production by increasing its supply for amino acid production and biomass formation [35,69]. Activation of gltA has been reported to reduce acetate production in non-glucose-limited media [97].

3.6.2. Balanced Precursor Supplementation Is Essential for Tryptophan Production

The supply of precursors is critical for amino acid productivity [123]. However, it is essential to control their feeding and production rate to avoid toxic effects due to the accumulation of metabolites. The supply of precursors must be balanced to achieve efficient tryptophan production; the biosynthesis of 1 mol of tryptophan requires 1 mol of E4P and PEP as initial precursors and consumes the equivalent of 1 mol of PEP in each of glutamine, phosphoribosyl-5-pyrophosphate, and serine pathways, so favoring any of these pathways can unbalance the production of the other precursors [29]. For this reason, from 1 mol of glucose, 0.86 mol of DAHP can be produced, but only 0.45 mol of tryptophan [35].

The addition of precursors can reveal new bottlenecks. Anthranilate is derived from chorismate and is a direct precursor of tryptophan; anthranilate supplementation improves glucose conversion rate to tryptophan by preventing acetate production at high glucose concentrations of up to 2.0 g/L. However, anthranilate can inhibit biomass production at concentrations higher than 0.2 g/L [106]. Anthranilate inhibits the growth of E. coli, possibly by blocking the tryptophan production by inhibiting the indole-3-glycerol-phosphate synthase (IGPS) enzyme, since, in wild-type bacteria, the addition of tryptophan restores growth [62].

Anthranilate supplementation has been reported to cause indole accumulation when tryptophan accumulation is higher than 30 g/L, as it inhibits tryptophan synthase activity. Adding serine to the medium did not improve tryptophan production, as it was not a limiting reagent for tryptophan synthase [68]. Increased tryptophan production is also associated with decreased indole accumulation [127].

Finally, adding precursors such as serine can make fermentation processes more expensive, forcing the search for cheap and easily accessible sources such as cane molasses [4]. For this reason, most current strains seek production only from glucose and ammonium [29,30,31].

3.7. Late Exponential Phase

Depleting limiting nutrients can shorten the exponential phase. Supplementation of E. coli 10303 strain cultures with potentially limiting metabolites revealed several substrates that enhance biomass production by delaying the onset of the death phase, including vitamins B1 and B2, the amino acids methionine, cysteine, leucine, and threonine, and the nucleotides guanosine monophosphate (GMP) and guanine. Among these, leucine had the most significant effect, boosting tryptophan production to 0.983 g/L, an improvement of 34.8% [31].

Serine is essential as a precursor of tryptophan and in biomass production. If serine is insufficient, tryptophan production is impaired, while if there is an excess in its accumulation, homoserine dehydrogenase (thrA) is inhibited, which participates in the isoleucine synthesis pathway [15]. Adding isoleucine or threonine can correct the growth inhibition caused by an excess of serine [15,116]. The addition of corn liquor enhances growth in strain T13. After analyzing its composition, it was identified that its high isoleucine content was responsible for this effect. For this strain, adding 0.5 g/L isoleucine increased glucose consumption by 41.2% [15]. The addition of casamino acids also promotes biomass production when serine is accumulated. Phenylalanine in casamino acids is known to be responsible for restoring growth. However, the mechanism of action is not yet fully understood [116].

Yeast extract supplementation can improve biomass production [117], but results may vary depending on the commercial supplier [31]. Yeast extract is a common ingredient of culture media, obtained from autolysis, hydrolysis, or plasmolysis, serving as a source of organic nitrogen. Due to its complex composition, it provides amino acids, peptides, carbohydrates, nucleic acids, vitamins, and minerals for the growth of microorganisms [31]. Adding yeast extract to the culture medium can reduce acetate formation and enhance protein production [112]. However, cultures in rich media based on acid-hydrolyzed casein have been reported to stimulate tryptophanase activity [16].

Due to the complex composition of yeast extracts, levels of critical nutrients can vary widely between suppliers and even batches, affecting tryptophan production [31,94]. Impurities in yeast extract, such as pigments, proteins, and endotoxins, also can reduce tryptophan production [94].

Residual E. coli bacteria from fermentation can be chemically or enzymatically hydrolyzed to serve as an alternative organic nitrogen source to yeast extract. It provides amino acids, nucleotides, and other components at levels appropriate for E. coli growth while enhancing tryptophan production. However, rapid nitrogen uptake from the hydrolysate leads to acetate accumulation [94].

3.7.1. Accumulation of Toxic Compounds

Conform fermentation advance, inhibitory metabolites such as acetate and NH₄⁺ accumulate. These metabolites can be removed using a disk stack centrifugal separator that recycles the cells to continue fermentation [118]. An alternative process is to use a vacuum scraper concentrator evaporation, which yields the highest reported titer of tryptophan in E. coli (60.2 g/L) [119]. Cell recycling reduces operating costs by lowering fermentation times. They are most efficient when used exclusively in the final fermentation stage, as inhibitory metabolites accumulate and tryptophan begins to crystallize [118].

The amino acid methionine is the precursor of S-adenosyl methionine (SAM), a methyl group donor involved in the biosynthesis of nucleic acids, phospholipids, proteins, and other molecules [69]. In tryptophan-producing strains, it can drive biomass synthesis [31]. Some publications report 1 g/L of methionine in the fermentation medium [3,84]. For tryptophan production, the effect of methionine has not yet been studied in detail. However, adding methionine or glycine to the culture medium reduces acetate formation and improves protein production [112]. Adding methionine can alleviate the effects of acetate accumulation because its accumulation inhibits methionine synthesis, leading to homocysteine, which is an inhibitor of methionine synthesis [107].

An imbalance in the carbon-nitrogen ratio can reduce tryptophan production [108]. NH₄⁺ assimilation in E. coli occurs through two assimilation pathways: the glutamine synthetase/glutamate synthase (GS/GOGAT) cycle and the glutamate dehydrogenase (GDH) pathway, which involve three key metabolites: glutamine, glutamate, and α-ketoglutarate [128]. For fermentation, one source of NH₄⁺ is aqueous ammonia, the preferred alkaline solution for pH regulation. NH₄⁺ has a high similarity to K⁺ in size and charge, allowing E. coli to exchange them in some cellular functions, reducing the importance of K⁺. This similarity allows NH₄⁺ in high concentrations to be transported by the high-affinity potassium uptake system (Kpd); when NH₄⁺ is incorporated into the metabolism, it dissociates from its proton, lowering the pH of the cytoplasm and the cell has to expend ATP to pump protons out in a wasteful cycle that increases the rate of oxygen consumption, decreases the energetic efficiency of the bacteria and inhibits their growth [13,93,129]. For optimal biomass and tryptophan production, the NH₄⁺ concentration should be less than 120 mmol/L [13,68,105,119].

In modified strains using plasmids to overexpress genes, NH₄⁺ accumulation decreases tryptophan production because it decreases the stability of the plasmids. The population of bacteria that have lost plasmids does not contribute to tryptophan production [119].

Alkaline solutions of NaOH and KOH that are commonly used to regulate the pH of fermentations release Na⁺ and K⁺ cations that affect nutrient uptake and growth when present above 40 mmol/L and inhibit biomass and tryptophan production when present above 60 and 78 mmol/L, respectively [105]. Purification and production costs of tryptophan can be increased because the presence of K⁺ in the medium reduces the adsorption capacity of organolites for tryptophan [68]. To avoid high K⁺ concentrations, a KOH/NH₃∙H₂O mixture in a 1:2 v/v ratio was used to control the pH, achieving high tryptophan titers [105].

K⁺ is a cofactor for many vital enzymes [13] and is the most critical cation for the growth of non-halophilic microbes; growth at low K⁺ concentrations requires high energy expenditure to maintain the ionic gradient generated across the membrane when the optimal K⁺ concentration is reached inside the bacteria [129]. High concentrations of salt ions interact non-covalently with proteins and DNA, modifying the pH and ionic strength and negatively affecting enzymatic activity [111,114]. Under conditions of osmotic stress, such as those present when tryptophan is produced in high concentrations, E. coli modifies its enzymatic activity and transport systems to accumulate ionic solutes such as K⁺, glutamate, and other anions that allow it to maintain the turgor of its membrane [115]. In minimal media, K⁺ plays the leading role in regulating osmolarity [111]; however, ionic solutes have adverse effects on growth, so they only accumulate at low osmolarities, to be replaced at high osmolarities by compatible solutes that interfere little with most enzymes [115]. As osmolarity increases, K⁺ is neutralized by glutamate to form potassium glutamate, which serves as an activation signal for other osmotic stress response mechanisms [114]. Glutamate accumulation is probably due to the inhibition of glutamate-consuming pathways rather than an increase in glutamate synthesis [115]. Glutamate accumulation impairs tryptophan production.

High concentrations of tryptophan can be toxic to bacteria, accelerating their death phase. Intracellular concentrations of tryptophan in overproducing strains can exceed the concentration at which it crystallizes (11.4 g/L in water at 24 °C) [43,117,126]. Adding betaine monohydrate to the culture medium could also protect E. coli from the effects of tryptophan [117]. Betaine monohydrate is used as an osmoprotectant of bacterial cells, reverses the inhibitory effect of salts and urea on enzymes, protects proteins from chemical and thermal denaturation, is a donor of methyl groups, prevents caramelization of glucose, among other applications in fermentations [110,111]. An unexpected effect of betaine on E. coli fermentation is the activation of the transcription of the zwf gene, improving the activity of the glucose-6-phosphate dehydrogenase enzyme to produce NADPH, which is required as a cofactor to produce several metabolites, for example, serine production in the THRD strain increases when NADPH availability is high. This transcriptional activation is specific to the zwf gene and is not observed in the other central metabolic pathway genes gnd, icd, pfkA, or pykF [110]. Osmotically active solutes (osmolytes) such as betaine, proline, K⁺, glutamate, trehalose, or 3-(N-morpholino) propane sulfonate (MOPS) allow E. coli to incorporate free water to recover its cytoplasmic volume as an adaptation mechanism to osmotic stress conditions, avoiding dehydration. When betaine is present, E. coli uses it preferentially, replacing the other osmolytes, preventing K⁺ and MOPS from being taken from the medium and decreasing the amount of glutamate and trehalose synthesized. Betaine at 1 mM increases the growth rate of E. coli more efficiently than other osmoprotectants, such as proline [111]. Adding betaine could reduce the harmful effects of K⁺ accumulation under osmotic stress conditions [111], which can occur when KOH is used as an alkaline solution to regulate pH in fermentations.

3.7.2. Surfactants Reduce Intracellular Accumulation of Tryptophan

Surfactants influence tryptophan production in several ways. They reduce the solubility of tryptophan to promote its crystallization and facilitate its accumulation in the medium, preventing intracellular accumulation that would otherwise inhibit its biosynthesis. A schematic description is shown in Figure 6.

Microorganisms use feedback inhibition as a mechanism to avoid producing more metabolites than required; intracellular concentrations of more than 4.5 g/L of tryptophan can affect its accumulation because it inhibits tryptophan synthase by up to 50% [126]; this enzyme can also be inhibited by indole [4]. Anthranilate synthase is another critical enzyme that is regulated by tryptophan negative feedback [66]; its inhibition when tryptophan reaches 30 g/L in strain AGX1757 favors indole accumulation [126], which inhibits the growth of E. coli at concentrations as low as 0.04% [7]. Blocking tryptophan and anthranilate synthases can redirect carbon flux toward phenylalanine and tyrosine synthesis [66].

The solubility of tryptophan in the culture medium is 22 g/L but increases to 32.1 g/L in the culture supernatant, possibly due to interaction with macromolecules secreted by E. coli [126]. Rationally designed E. coli strains can accumulate more than 40 g/L of tryptophan, which is above the saturation concentration under typical culture conditions [62]. High intracellular concentrations of metabolites such as tryptophan can inhibit cell growth [29,36,42], which could be prevented by enhancing tryptophan export to reduce feedback inhibition [36,90].

Aromatic amino acids can diffuse through the cell membrane due to their hydrophobic properties [15,91]. Surfactants can affect membrane fluidity and permeability by facilitating metabolite exchange [4,113]. The surfactant triton X-114 has been shown to enhance tryptophan production but is a poor candidate for use in fermentation because it generates large amounts of foam [113]. Triton X-100 served as an indole reservoir to reduce tryptophan synthase inhibition, boosting the non-overproducing strain E. coli ATCC 11303 when cane molasses was used as a substrate [4]. Using Triton X-100 allowed production levels of 141.4 g/L to be achieved in the E. coli B 10 strain from indole and serine. Another indole reservoir used to improve tryptophan production is Amberlite XAD-2 because it does not encapsulate tryptophan [7].

The surfactant Pluronic L-61 also improves tryptophan production; its mechanism of action is through the precipitation of tryptophan in the medium when it reaches concentrations of 25 g/L, breaking its interaction with macromolecules that improve its solubility; this prevents the total inhibition of tryptophan synthase at 30 g/L, facilitating its secretion and recovery by centrifugation [113,126]. Tryptophan production in fed-batch fermentation improves as the surfactant Pluronic L-61 increases up to 0.5% (v/v); higher concentrations inhibit growth. In batch fermentations, the addition of this surfactant has no effect [113]. It is recommended to add Pluronic L-61 after tryptophan reaches its solubility threshold to allow biomass accumulation [126].

Tween 60 also improves tryptophan production in fed-batch and batch fermentations. The mechanism of action appears to be related to the activation of transmembrane exporters that, by secreting tryptophan, reduce negative feedback. Tween 60 also promotes glutamate production. The improvement in tryptophan production with surfactants does not generalize to other amino acids; for example, it impairs the production of lysine, proline, leucine, valine, and the aromatic amino acids tyrosine and phenylalanine [113].

3.8. Stationary Phase

Tryptophan production by overproducing strains continues during the stationary phase. High DO fermentations allow high-density growth in E. coli, favoring tryptophan production [68,98]. An adverse effect of high DO levels is that they cause oxidative stress by generating Reactive Oxygen Species (ROS), such as hydrogen peroxide, which promotes the formation of free radicals through Fenton reactions, damaging proteins, membranes, and DNA, resulting in cell death [65,68,101,104]. Oxidative stress causes a redistribution of metabolic fluxes in E. coli; among the pathways that are attenuated are the TCA cycle and glycolysis, while the PPP, the acetate pathway, and the glyoxylate shunt are activated, resulting in a high NADPH/NADH ratio and the production of α-ketoglutarate [102,104].

The glyoxylate shunt plays an essential role in the response to oxidative stress [101,102,104], participating in the metabolism of acetate and other substrates that can be converted to acetyl-CoA for redirection toward gluconeogenesis and cell growth. The overall effect of this cycle is to form one mole of malate from two moles of acetyl-CoA by avoiding the oxidative decarboxylation steps of the TCA cycle and preserving the carbon skeleton for biomass generation [101,102]. The glyoxylate shunt competes with the TCA cycle by reducing the production of α-ketoglutarate and glutamate [98,102]. Under conditions of oxidative stress, α-ketoglutarate and glutamate accumulate even though the TCA cycle is attenuated and the glyoxylate shunt is active because the enzyme α-ketoglutarate dehydrogenase is almost inactive, forming a bottleneck [98,104].

Glutamate can be found at extremely high levels in overproducing strains such as S028 and JB102/pBE7 [38,117]. Glutamate accumulation diverts resources that could be used to produce tryptophan, reducing the efficiency with which glucose is converted [15,38,98,118]. Glutamate production from α-ketoglutarate by glutamate dehydrogenase (gdhA) consumes NADPH as a cofactor [130], which causes an imbalance in the reducing power of the cell when produced in large amounts [68,98]. Why E. coli accumulates glutamate in tryptophan-overproducing strains is not very clear. However, it could be due to an imbalance in the redox state that prevents the regeneration of NADP from NADPH. E. coli has a high demand for NADPH during its growth, which PPP and isocitrate dehydrogenase supply; however, when its growth is reduced, the excess NADPH is consumed, producing glutamate [38]. Although glutamine synthase (glnA) requires NH₄⁺ to produce glutamine, its activity is regulated by several mechanisms, such as feedback inhibition by the end product. At high ammonium concentrations, such as those in late fermentation phases, adenylation of glutamine synthase causes it to lose activity, making it a limiting reagent for producing tryptophan [103].

Glutamate accumulation in the late fermentation phase can be decreased by reducing the glucose feed rate and enhancing tryptophan production [38,68,98]. It is not efficiently converted to glutamine, even though the accumulation of its precursor, α-ketoglutarate, has been reported to increase glutamine synthesis [88]. Glutamine acts as an amino group donor in the conversion of chorismate to anthranilate, making it an essential precursor for chorismate synthesis and potentially a limiting factor in tryptophan production [103].

Cells become more oxidized when they transition from the exponential to the stationary phase [131]. ROS stress conditions trigger the activation of NAD(P) transhydrogenases sthA and pntAB, which are essential for regulating the NADH/NADPH ratio [104]. These enzymes have also been previously overexpressed to enhance the supply of cofactors required for tryptophan production [103].

4. Industrial Purification of Tryptophan

Feed supplements can be obtained by directly drying the fermentation broth and solidifying it, however, by this method the remaining impurities produce hygroscopy, decreasing the quality of the final product [132].

To eliminate solid impurities such as bacteria, the fermentation broth can be centrifuged or filtered [20]. To increase the yields in purification, it is convenient to increase the solubility of tryptophan by adding a calcium source [132].

Because tryptophan is stable at high temperatures, the culture broth can be evaporated under vacuum for two purposes: to eliminate the water that is the main component of the broth and to take advantage of the fact that tryptophan is at the solubility threshold, favoring crystallization [20].

Then, the crystals are separated from the mother liquid by filtration or centrifugation. The exhausted liquid is recycled through evaporation to reconcentrate it and extract the non-crystallized tryptophan. To obtain a purer preparation of tryptophan, chromatography can be applied at higher costs. The final step in the purification of tryptophan is the drying of the crystals by means of temperature and vacuum [20]. In this step, granules with a lower tryptophan content can be produced by adding condensation seeds [132].

5. Fermentation Parameters and Genetics Modifications as Complementary Approaches to Achieve High Tryptophan Production

Tryptophan production can be improved by reinforcing its biosynthetic pathway by adjusting fermentation parameters or genetic modifications. Observing fermentation behavior can guide the adjustment of fermentation parameters to accommodate bacterial metabolism, albeit with limitations, which can then be overcome by rational strain design, as is the case with the observation that acetate is commonly accumulated and can be controlled by adjusting fermentation conditions, although silencing the genes involved in acetate synthesis is a better alternative. We revise the main parameters and processes improved by genetic and media modifications in the following sub-sections.

5.1. Activation of the Final Branch of Tryptophan Production

The enzymes that produce tryptophan from chorismate are expressed from the trp operon [28]. This operon is regulated by the trpL attenuator in their mRNA 5′ UTR, whose activity depends on tryptophan. Their transcription is regulated by the repressor TrpR, which is activated when bound to tryptophan. The overexpression of this operon enhances tryptophan production and can be achieved by removing the trpL attenuator, swapping the trp promoter, and cloning the operon in stable and multicopy vectors [6,13,14,36,38,39,61,62,65,68,70,84,86,90,91,93,94,98,105,106,107,108,112,113,117,118,119,123,125,126].

Plasmids are commonly used to overexpress the trp operon, so increasing plasmid stability by reducing the growth rate [13,105], lowering the pH [37,44,105], and preventing the accumulation of NH₄⁺ [13] or phosphate [70] improves the expression of these genes. The availability of Mg²⁺ [81] and pH values close to 8 [7] increase the activity of tryptophan synthase.

The anthranilate synthase enzyme encoded by the trpE gene is, in fact, the main bottleneck in the operon. This enzyme catalyzes the first reaction of the pathway and is inhibited by tryptophan. The trpE^S40F mutant is insensitive to inhibition and has been extensively used to create overproducing strains. The trpE^S40 and trpD genes have been overexpressed in the FB-04/pSV03 strain and their derivatives without modifying the native trp operon, achieving yields of up to 45.6 g/L [60,66,67]. Another mutant allele insensitive to feedback inhibition is trpE^{M293T/N168D/A478T/C237R} [89]. The regulation of the final tryptophan pathway is very susceptible to the accumulation of intermediates, so tryptophan must be produced in a balanced manner with their precursors. The heterologous expression of the trpC^R378F gene from Aspergillus niger improves tryptophan production, mainly when anthranilate accumulates because it inhibits the IGPS of E. coli [6,62].

Adding surfactants to the medium can reduce the inhibition caused by tryptophan and precursors on its synthesis [4,113]. The increase in available Mg²⁺ decreases the sensitivity to feedback inhibition of anthranilate synthase to tryptophan [80].

5.2. Reduction in Tryptophan Degradation

The tryptophan degradation pathway comprises the tryptophanase enzyme encoded by the tnaA gene. This degradative pathway converts tryptophan to pyruvate, ammonium, and indole. In low concentrations of tryptophan, this enzyme has low activity, so it does not interfere with the metabolism of E. coli. However, high concentrations of tryptophan, such as those in overproducing strains, activate the transcription of tryptophanase, preventing reaching high production levels [133]. Since deleting the tnaA gene is one of the most widely used modifications to enhance tryptophan production [3,67,133], changing fermentation parameters might be irrelevant. However, reducing yeast extract reduces the activation of tryptophan degradation [16]. Surfactants also reduce the available tryptophan within the bacteria, thereby reducing tryptophan degradation [4,113].

5.3. Modifications to the Central Metabolic Pathways

Regarding tryptophan synthesis, the central metabolism can be divided into four main pathways: glycolysis, TCA cycle, PPP, and glyoxylate shunt [42]. These pathways provide the metabolites and energy the cell needs to function. For the biosynthesis of aromatic amino acids, glycolysis is responsible for providing PEP, while PPP provides E4P, phosphoribosyl pyrophosphate (PRPP) and the reducing cofactor NADPH. Glutamine, which donates an amino group to tryptophan production, is derived from the TCA cycle [42].

PEP is a crucial precursor of aromatic compounds and the TCA cycle and is central to glucose uptake [42]. To transport a mol of glucose by PTS a mol of PEP is consumed, generating pyruvate that can be channeled to the production of acetate [14,34,42,60,61,98,125]. In this condition, 50% of the cellular PEP is used to import sugar, while around 1.5–3% is redirected toward producing aromatic amino acids [3,34]. The PTS can be inactivated by deleting the ptsI [6], ptsG [3,14,90,125], or ptsH genes [15,108]. Still, it reduces glucose consumption so significantly that it affects growth, impairing tryptophan production, although it favors the conversion rate of glucose to tryptophan [3,6,14,34,90,125]. The PTS has been attenuated with the ptsH^N12S mutant instead of deleting it [60]. In strains with the PTS inactive, glucose transport has been complemented by overexpressing galactose permease (galP) [15,108,117] or expressing the glucose facilitator (glf) of Zymomonas mobilis [3]. When transporting glucose through both systems, overexpression of glucose kinase (glk) is required to incorporate glucose into glycolysis efficiently [3,117]. Parameters that can be adjusted to reduce glucose intake and increase PEP production are increasing DO levels [68], adding sodium citrate to the culture medium [112], and lowering the pH [68].

PEP is consumed by pyruvate kinase (pykF or pykA) to perform its primary function: to supply the TCA cycle [60,86,134]. Deleting these genes allows the accumulation of PEP; however, when made simultaneously, it attenuates the TCA cycle, affecting biomass production [3,42,61,87]. Deletion of pykF reduces the flux on glycolysis and the acetate production pathway, redirecting it toward the PPP [60]. Eliminating pyruvate kinase activity forces the redistribution of metabolic fluxes to generate pyruvate through other enzymes such as anthranilate synthase; by this means, tryptophan production can be growth-coupled, as it compels the organism to produce anthranilate to support its growth [134]. In the fermentation process, adding sodium citrate can reduce the activity of pyruvate kinase [112].

PEP can be regenerated from pyruvate or oxaloacetate by overexpression of PEP synthase (ppsA), which catalyzes the reverse reaction to pyruvate kinase, or by overexpressing PEP carboxykinase (pckA), which consumes oxaloacetate [36,68,84,123]. The ppc gene encodes phosphoenolpyruvate carboxylase, which consumes PEP to produce oxaloacetate conversely to pckA; deleting it reduces PEP consumption [86]. Pyruvate has also been channeled to oxaloacetate production by expression of C. glutamicum pyruvate carboxylase (pyc^P458S) to create a PEP-pyruvate-OAA node in which oxaloacetate is converted to PEP by pckA [3]. Similarly, PEP has been regenerated from acetyl-CoA by overexpressing aceB, mdh, and pckA genes. The aceB gene encodes malate synthase in the glyoxylate shunt that consumes acetyl-CoA, while mdh encodes malate dehydrogenase, producing oxaloacetate (see Figure 2 and Figure 5) [86]. Increasing oxygen levels makes it possible to activate the gluconeogenesis pathway to regenerate PEP [68] and the glyoxylate shunt to recycle available acetyl-CoA [101,102].

When strategies to increase PEP production deplete the supply of the TCA cycle, this can be reactivated by adding citrate to the medium. To efficiently metabolize citrate in the TRP07 strain, it was necessary to overexpress the citrate transporter citT and the TCA cycle genes acnAB and icd [87].

A decrease in tryptophan production has been observed when CO₂ production is increased [48]. The CO₂ produced through the TCA cycle removes carbon atoms from the metabolism that could be used to form the carbon skeleton of tryptophan. In the TCA cycle, the reactions that produce CO₂ are isocitrate dehydrogenase and α-ketoglutarate dehydrogenase; the activity of these reactions has been previously reduced by making a bypass that redirects the flow of acetyl-CoA to malate to regenerate PEP (see Figure 2) [86]. Adding sodium citrate helps reduce the TCA cycle activity [112], and the increase in DO can activate the glyoxylate shunt [101,102].

In E. coli, aromatic metabolites are generated from the condensation reaction between PEP and E4P, which gives rise to DAHP [60,84,87,89,123]. The shikimate pathway requires equal proportions of E4P and PEP, but E4P cellular levels are lower than PEP and can be a limiting substrate [3]. E4P can be increased by overexpression of the tktA gene, which encodes a transketolase, increasing tryptophan production [15,36,61,68,89,90,123,125]. An alternative that is more efficient in yeast to increase the supply of E4P is the introduction of the xfpk gene from Bifidobacterium adolescentis, which converts a molecule of fructose-6-phosphate and phosphate into E4P, with the production of acetyl phosphate. This strategy has also increased tryptophan production in E. coli and can bypass the carbon-losing enzyme 6-phosphogluconate dehydrogenase that produces CO_2. However, it is necessary to reinforce the incorporation of acetyl phosphate into the metabolism to avoid acetate accumulation [3]. The PPP can be enhanced by adding betaine monohydrate [110] and sodium citrate [112] and increasing the level of DO [98]. This increases the activity of the PPP genes zwf, pgl, tktA, and araD [68,98].

Another function of the PPP is to provide the NADPH required for tryptophan synthesis. The pntAB genes have been overexpressed, which encode a membrane-bound complex with transhydrogenase function that transfers a proton from NADH to NADPH. This allows for a more suitable NADPH/NADH ratio for tryptophan production [3,103]. The sthA gene encodes a transhydrogenase that catalyzes the reverse reaction to that of pntAB and helps control NADPH accumulation [103]. The increase in PPP activity caused by the rise in DO also favors the production of NADPH [98].

5.4. Reduction in Acetate Accumulation

Acetate accumulation is detrimental to biomass production and constitutes a waste of carbon that prevents achieving high conversion rates for tryptophan production. Deleting the pta gene effectively decreases acetate production and favors tryptophan production [14,36,68,84,93,94,105,118]. However, pta removal has also been reported to affect biomass and tryptophan production in some strains. As alternatives, the transcription of the pta gene has been attenuated using a small RNA, or the wild-type gene has been replaced by a less active allele called pta1 so that biomass production is not affected [60,67,108]. Another strategy to reduce acetate production is eliminating the acetate kinase activity catalyzed by two enzymes encoded by the genes ackA and tdcD [65]. Silencing of poxB has little impact on acetate accumulation [86]; in strain TRP07, it was used as an insertional site, and its elimination alone was not evaluated [87]. However, two or more genes involved in acetate production can be silenced to increase its effect, as in strains TRTHBPA, TRP12, or TRP30 [36,84,94,118]. Expression of genes involved in acetate production can be reduced by increasing DO levels [68,86,98,118] and reducing the pH [68]. Reducing glucose feeding lowers the growth rate and prevents the metabolic overflow that leads to acetate accumulation [14,65,93,98,106,107,108]. Avoiding high phosphate concentrations prevents acetate formation [70].

Gecse and colleagues propose that eliminating poxB and pta is more effective in glucose-limited culture media. In contrast, overexpression of the gltA gene for a non-limited medium is better for preventing the metabolism overflow that generates acetate [15,97]. However, the harmful effect of pta deletion is also influenced by DO levels; at low DO levels, the impairment of growth is more pronounced [99,135]. Adding CaCO₃ can increase citrate synthase activity to reduce metabolic overload [69].

In E. coli, acetate can be inefficiently consumed by its conversion to acetyl-CoA through two irreversible reactions catalyzed by the enzyme acetyl-CoA synthetase. The first step consumes one molecule of ATP to form acetyl-adenosine monophosphate. The second step converts acetyl-AMP to acetyl-CoA. Overexpression of the acs gene allowed strain T16 to consume the acetate produced [86]. Transcription of the acs gene can also be increased by elevating DO levels [98,100].

By decreasing acetate production, the accumulation of pyruvate and acetyl-CoA may increase, which can be funneled into the synthesis of other by-products [14,65,93,98]. After acetate, the most produced compounds are lactate, ethanol, and formate; to avoid their accumulation in the TRP12 strain, the genes ldhA, adhE, and pflB that code for the enzymes that branch toward the synthesis of these compounds have been deleted [36].

Finally, the acetate produced can be removed from the medium using vacuum evaporation or recycling the cells to continue fermentation in a fresh medium [118,119].

5.5. Activation of the Common Aromatic Amino Acid Pathway

For the shikimate pathway, there are three isoforms of the enzyme 3-deoxy-7-phosphoheptulonate synthase, encoded by the genes aroG, aroH, and aroF, and are mainly regulated by phenylalanine, tryptophan, and tyrosine, respectively. The enzyme encoded by the gene aroG has the most significant activity in E. coli, followed by aroF and aroH [72]. High oxygen levels activate the transcription of the aroG gene [68]. To increase the activity of this enzyme, it has been overexpressed, and N-terminal coding sequence libraries have been used to improve their translation levels [84]. Overexpression of the feedback inhibition-resistant mutant aroG^S180F is commonly used to enhance tryptophan production [3,14,15,36,38,61,62,68,84,87,90,108,117,123]; other mutants used for tryptophan production are aroF394 [33], aroF^I11- [66,67] and the aroG^D6G/D7A_, which seems to be a better alternative to the aroG^S180F [6,85].

After overexpressing the common aromatic amino acid pathway genes aroB, aroD, aroE, aroK, aroA, and aroC, the enzymes shikimate dehydrogenase (aroE) and shikimate kinase I (aroK) have been identified as rate-limiting steps in shikimate production and have therefore been successfully overexpressed [15]. The insufficient activity of shikimate kinase has also been corrected by overexpressing the isoform II encoded by the gene aroL. This gene was identified after introducing perturbations in the supply of the carbon source and allowed to identify the gene aroB encoding a 3-dehydroquinate synthase as a control step in tryptophan production [89]. The expression of the aroK gene is increased at high pH [105] and DO values [68].

5.6. Supply of Tryptophan Precursors

5.6.1. Glutamine

The glutamine pathway comprises the enzyme glutamine synthase, which combines glutamate and NH₄⁺. The gene glnA encodes this enzyme and may be a rate-limiting step in tryptophan production [15]. E. coli can supply glutamine to the enzyme anthranilate synthase during the first hours of fermentation but then become scarce as glutamate accumulates [42]. Wild-type glutamine synthase from E. coli can be inhibited by high concentrations of NH₄⁺. This does not happen in Gram-positive bacteria such as Bacillus subtilis and Bacillus megaterium, although they are sensitive to glutamine [103]. Mutations have been introduced into the glnA gene of B. subtilis (glnA_Bs) to eliminate the inhibition by glutamine and increase its affinity for ammonia and glutamate [136]. Overexpression of the glnA^L159I/E304A mutant of B. subtilis has improved tryptophan production [15]. Another modification used to improve tryptophan production is glnA^Y405F from C. glutamicum [5]. NH₄⁺ buildup can be prevented by lowering the pH using a less alkaline solution [68], or by using mixtures of alkaline solutions [105]. In addition, accumulated NH₄⁺ can be removed from the medium as described previously.

Glutamine can be deaminated by glutamate synthase encoded by the gltB gene. This gene has been deleted in the TRTHBPA strain to prevent this reaction from consuming glutamine, improving tryptophan production [94,118].

5.6.2. Serine

Serine is another precursor that can be limiting in tryptophan production [15,38]. The serine synthesis pathway involves dehydrogenation, transamination, and finally, dephosphorylation of 3-phosphoglycerate from glycolysis [88,127]. Its production has been improved through the overexpression of the genes constituting its synthesis pathway: serA, serB, and serC [15,36]. The serA gene encodes a phosphoglycerate dehydrogenase, which is inhibited by serine accumulation, so a serA^H344A/N364A mutant is commonly used [3,5,15]. Another mutant used is serA^T372N, accompanied by deleting the sdaB gene, which codes for the serine degradation reaction to boost serine supply for tryptophan production [36].

Serine is an amino acid that can be obtained from other strains capable of overproducing it and added to the medium. Regulating temperature can boost serine production [96]. Low oxygen levels reduce TCA cycle activity, allowing the accumulation of acetyl-CoA and pyruvate [93], and pyruvate accumulation favors serine production [88].

5.7. Silencing of Transcriptional Regulators

The TrpR repressor regulates the transcription of genes involved in the production, degradation, transport, and regulation of tryptophan; among these are aroL, aroH, trpEDCBA, mtr, and trpR. Eliminating the trpR gene is a minimum requirement for a marked production of tryptophan by releasing the expression of multiple genes responsible for its production [3,39,67]. The addition of surfactants reduces the availability of tryptophan within the cell by reducing its effect on the regulatory mechanisms for tryptophan [4,7,113].

Similarly, the tyrosine repressor of transcription (tyrR) controls the production of common precursors for tryptophan and tyrosine, so silencing it has favorable effects on tryptophan production [3,87]. TyrR represses the aroF gene, a limiting factor in DAHP production [33].

The fructose repressor, FruR, is a global regulator of the central metabolic pathway. Its deletion activates genes involved in glycolysis, alternative sugar catabolism, the Entner–Doudoroff pathway, and PPP. Furthermore, it represses gluconeogenesis, the TCA cycle, and the electron transport chain. Although the fruR mutant produces more pyruvate and acetate by activating glycolysis, its deletion improves the production of tryptophan in a stable manner using sugars in different mixtures. The accumulation of pyruvate can stimulate the conversion of 3-phosphoglycerate to serine-phosphate, which can be converted to serine, improving tryptophan production and reducing the accumulation of precursors of the common pathway of aromatic amino acids such as quinone [88]. Because the FruR regulator acts on multiple genes involved in metabolism, its effect is complex. However, some responses may be consistent with the impact of changing some fermentation parameters. Silencing of FruR activates glycolysis, as does the addition of phosphate to the medium [70]. The activity of the TCA cycle, gluconeogenesis, and the PPP pathway may coincide with how increased oxygen activates these pathways [68].

Tang and collaborators identified transcriptional regulators whose expression changes in overproducing strains. To test their effect, they modified several genes but only found an improvement in tryptophan production when deleting the putative regulator yihL, whose function needs to be better characterized [84].

5.8. Modulating Pathways Competing for Precursors

Chorismate is a common precursor for the synthesis of aromatic amino acids [84]; when it is not efficiently directed toward tryptophan synthesis, it accumulates, causing feedback inhibition, and phenylalanine and tyrosine can be produced as by-products [35,38]. Blocking phenylalanine and tyrosine production by silencing the pheA and tyrA genes can improve tryptophan production [33,67]. However, auxotrophic strains are generated, which increases the manufacturing cost, and supplementation of the medium with these amino acids inhibits the common pathway for synthesizing aromatic amino acids, reducing the supply of precursors for synthesizing tryptophan [15]. To reduce competition for precursors and avoid generating auxotrophic strains, the native promoters of the tyrA and pheA genes have been replaced by the self-regulated promoter P_fliC, increasing tryptophan production without affecting growth [84]. Achieving optimal tryptophan production requires combining the parameters described above to achieve a balanced metabolic flow in which metabolic overloads that lead to the production of by-products are avoided. The fine regulation of the tryptophan production pathway can cause bottlenecks that favor the production of other aromatic amino acids, so using surfactants can be an excellent strategy to prevent the accumulation of precursors from being redirected toward the synthesis of other undesirable metabolites.

6. Models for Tryptophan Production

Constructing mathematical models from experimental observations and measurements facilitates studying biochemical processes. Simulating variations in these models can guide the selection of regulatory elements that allow the generation of stable tryptophan-producing strains [121,137]. The main mechanisms that regulate the genetic network for tryptophan production were discovered in the 1960s, and since then, they have become one of the most studied [17].

These models have been built using different approaches. Models based on Ordinary Differential Equations (ODE) are inherently quantitative and, therefore, require precise equations to describe biochemical reactions; the constants of these functions are difficult to obtain through experiments and can be inferred from mathematical techniques that adjust the equations to a defined behavior [17]. Models based on ODE for tryptophan production have been gradually developed by different research groups that incorporate various components and regulatory mechanisms, making them increasingly complex as the understanding of the functioning of the trp operon advances [64]. Due to this complexity, the models must be solved numerically [138]. The representation of the tryptophan production models has focused mainly on the analysis of its synthesis through the trp operon, transport, and degradation [17]. The effect of central metabolism [35], the demand for tryptophan by the cell [28,137] and the impact of the availability of extracellular tryptophan [139] have also been added. Santillan and Mackey included the delay time and highlighted the importance of anthranilate synthase in the production of tryptophan [64]. The delay time between transcription and translation creates an oscillation in the transcription of the genes of the trp operon, which is not linear but exists in both on and off states; this oscillation is characteristic of repressed systems [140]. For the expression of the tna operon, it was also found that it has a bi-stable behavior that depends on the concentration of glucose and tryptophan; however, when the trp operon is included in the simulation, this behavior is no longer observed, possibly because the system enters in a state of homeostasis [17,141]. Xiu and colleagues included attenuation in their model and the effect of the growth rate [140]. Sun and collaborators considered tryptophan secretion an essential factor in tryptophan accumulation and included the interaction between gene structure, regulatory elements, and regulators [142]. Simulations in these models have found that feedback inhibition increases trp operon stability. However, several models omit it, while attenuation reduces the response time when the bacteria undergo a nutritional shift [138,142]. The ODE models have also been used to describe the behavior of strains that overproduce tryptophan [63].

Gene regulatory networks can be represented as Boolean networks, where nodes represent genes associated with Boolean values, 0 represents inactive genes, and 1 represents active genes [17,141]. Changes in the values of Boolean networks represent their dynamics and are defined by Boolean functions. The order in which these values are updated is called the update scheme: they can be updated simultaneously, in sequential order, or asynchronously. It has been observed that each update scheme can give information about different properties of the network [141]. Boolean networks have the advantage of being wholly qualitative and do not require precise values to carry out their simulations [17].

A variant of Boolean networks is the threshold network, in which a weight is assigned to each of the Boolean functions of each node belonging to a directed graph [141].

Unlike Boolean representations, complex metabolic networks can be presented as qualitative models to take advantage of the large amount of scientific information. These models are known as Petri Nets (PNs), which facilitate the identification of attractors and the recognition of critical regulatory elements thanks to bioinformatics tools and techniques specific to PNs [142].

Machine Learning has gained momentum in recent years and permeated different branches of science. It can complement other tools, such as metabolic flux balance models, and incorporate experimental data, as is the case of the optimization of tryptophan production in yeast carried out by Zhang and collaborators [143]. Another Machine Learning model is the Qualitative Perturbation Analysis Machine Learning (QPAML) classification model, fed by metabolic fluxes generated by introducing perturbations in a metabolic network [92].

Very shortly, fermentations will be monitored in real-time using Machine Learning models, which will allow the determination of the state of the fermentation and the accumulation of metabolites to adjust their parameters [144].

7. Conclusions

Optimizing tryptophan production in E. coli through fermentation requires a multifaceted approach that integrates genetic, metabolic, and environmental factors. This review highlights that precise control over nutrients, pH, temperature, and DO is crucial for maximizing production efficiency. The balance between growth rate and substrate consumption is essential, as excessive feed rates or poor oxygen control can lead to by-product accumulation, reducing overall yields. Key strategies include adjusting glucose feed rates to avoid metabolic overflow, managing nitrogen and phosphate levels to support biomass and tryptophan production, and maintaining optimal pH and temperature conditions. Additionally, supplements like betaine and surfactants enhance production by reducing stress and increasing precursor availability. Combining these fermentation strategies with genetic modifications that enhance the biosynthetic pathway makes it possible to achieve high-density cultures capable of producing significant tryptophan titers. Continued research in fine-tuning these parameters will further improve microbial tryptophan production’s economic viability and scalability for industrial applications.

As perspectives, transporting glucose through systems that do not consume phosphoenolpyruvate increases the theoretical maximum of tryptophan production, encouraging further research to improve production yields and conversion rates. New technologies such as CRISPR/Cas systems should be considered to achieve these improvements. CRISPR/Cas has already been employed in tryptophan production to create mutant gene libraries, efficiently integrating beneficial alleles into the genome. Technologies derived from CRISPR/Cas also allow for the precise adjustment of gene expression, allowing metabolism to be balanced. Reduced genome organisms offer great potential as cell factories due to the absence of non-essential genes. This will enable them to allocate more resources to metabolite synthesis, as demonstrated by the E. coli strain MDS-205, which exhibits higher threonine production than its full-genome counterparts with the same genetic modifications. Furthermore, the application of artificial intelligence is expanding in several scientific fields, providing predictive models that can accurately identify target genes for modification. This technology can accelerate the use of alternative substrates and improve real-time fermentation monitoring. Adaptive evolution in the laboratory (ALE) remains a widely used strategy to optimize tryptophan production, mainly through exposure to tryptophan analogs. However, advances in high-throughput sequencing and bioinformatics are accelerating the discovery of new genetic modifications that further improve tryptophan production.