**3.1. Biosynthesis of carotenes**

As a colorless carotene, phytoene is the common precursor to all the C40 carotenoids (**Figure 2**). It exhibits excellent anti-UV activity [33] and is clinically proved to have activities of skin whitening and wrinkle reduction [34]. Hence, there are increasing cosmetic products developed based on phytoene. Phytoene is an intermediate carotenoid in plants and exists only as a minor product; hence, it is expensive to extract phytoene from plant materials. Consequently, it is promising to engineer microbes to produce higher concentrations of phytoene and more importantly, to produce it at high purity without other carotenoids. By deleting the *crtI* gene, encoding phytoene desaturase (see below), from an engineered lycopene-producing strain of *Escherichia coli* previously developed in our laboratory [19], it was relatively simple to generate strains of producing more than 50 mg/L of high-purity phytoene in simple low-cell density shake flasks [35]. Although this carotene with a high-potential market in cosmetics could be relatively simple to transfer to the industry, this is only just the beginning to attract interest, as witnessed by a French company Deinove (www.deinove.com) [36]. Despite certainly being more efficient than the use of the tomato strain developed to this end, there could still be considerable progress made by optimizing the engineered strains such as that used in our study and coupling this to high cell density fermentation processes to achieve a more cost-effective process.

Lycopene, a red color pigment most commonly associated with tomatoes, belongs to one of the top six commercial carotenoids. It is produced from the dehydrogenation of phytoene catalyzed by different types of phytoene desaturases (*crtI*, *PDS* or *ZDS,* **Figure 2**). Lycopene has been used as animal feed, food coloring and nutritional products. Some clinical studies have suggested that lycopene functions in reducing the risk of prostate cancers [37, 38]. In recent years, multiple research groups reported relatively high concentrations of lycopene produced in *E. coli* and yeasts. In *E. coli*, Kim et al. have used a mixture of carbon sources containing glucose, glycerol and arabinose to produce lycopene at 1.35 g/L [39]. Our laboratory initially optimized the MEP pathway which enabled the *E. coli* strain to produce at 20 mg/g dry cell weight (DCW) of lycopene [40] and more recently reconstituted the MVA pathway in *E. coli* to produce lycopene in a glucose-defined medium, reaching 1.5 g/L in a simple nonoptimized fed-batch process [19]. Xie et al. evolved the bifunctional enzyme crtYB to acquire only the phytoene synthase function. By applying this mutated enzyme and optimizing the copy number of crt genes, the engineered *Saccharomyces cerevisiae* strain produced 1.61 g/L of lycopene [41]. In addition, *Yarrowia lipolytica*, the oleaginous yeast, has also been engineered to produce lycopene but at slightly lower yields [42, 43].

Moving further along the carotenoid biosynthetic pathway, lycopene is usually cyclized into β-carotene or α-carotene by a lycopene cyclase (**Figure 2**). Both α- and β-carotenes are yellow pigments; β-carotene is more commonly marketed, being one of the most important commercial carotenoids. As mentioned earlier, β-carotene is a direct precursor of vitamin A (**Figure 2**). It has been widely used as a colorant, nutritional supplement, animal feed and in pharmaceutical and personal care industries. Chemically synthesized β-carotene is less popular among consumers than that extracted from natural sources or so-called 'bio-based' sources. At the same time, naturally derived β-carotene has gradually taken over the market. Currently, β-carotene is produced mainly in the microalga *Dunaliella* [21] and the fungus *Blakeslea trispora* [2]. As summarized in **Table 1**, many groups have engineered fast-growing microorganisms and achieved high titers of β-carotene. Yang et al. have applied a hybrid MVA pathway in *E. coli* to overproduce β-carotene at 3.2 g/L [44]. Zhao et al. have engineered the central metabolic pathway to increase cofactor supply in an *E. coli* strain, which enabled the strain to produce at 2.1 g/L of β-carotene [45]. *Y. lipolytica* has shown potential as a better host for producing β-carotene; 4 g/L of β-carotene was achieved in an *Y. lipolytica* strain by integrating multiple copies of key enzymes (hmgr in **Figure 1** and the bi-functional enzymes phytoene synthase/ lycopene cyclase carRP) [46]. Recently, based on an engineered lipid overproducing strain of *Y. lipolytica*, Larroude et al. have rewired it to produce at 6.5 g/L and 90 mg/g DCW of β-carotene [47]. These results are relatively better than those previously achieved [48–50]. It would not be surprising if some of these examples would lead to the successful commercialization notably of novel β-carotene sources in the near future.
