Biosynthesis, acquisition, regulation, and upcycling of heme: recent advances

Abstract

Heme, an iron-containing tetrapyrrole in hemoproteins, including: hemoglobin, myoglobin, catalase, cytochrome c, and cytochrome P450, plays critical physiological roles in different organisms. Heme-derived chemicals, such as biliverdin, bilirubin, and phycocyanobilin, are known for their antioxidant and anti-inflammatory properties and have shown great potential in fighting viruses and diseases. Therefore, more and more attention has been paid to the biosynthesis of hemoproteins and heme derivatives, which depends on the adequate heme supply in various microbial cell factories. The enhancement of endogenous biosynthesis and exogenous uptake can improve the intracellular heme supply, but the excess free heme is toxic to the cells. Therefore, based on the heme-responsive regulators, several sensitive biosensors were developed to fine-tune the intracellular levels of heme. In this review, recent advances in the: biosynthesis, acquisition, regulation, and upcycling of heme were summarized to provide a solid foundation for the efficient production and application of high-value-added hemoproteins and heme derivatives.

GRAPHICAL ABSTRACT

Keywords:

• Heme
• biosynthesis
• acquisition
• regulation
• upcycling
• hemoproteins
• heme-derived chemicals

Introduction

Heme, the cofactor of hemoproteins, plays essential roles in prokaryotic and eukaryotic cells, including: oxygen transport [1], electron transfer [2], gas sensing [3], signal transduction [4], biological clock [5], and microRNA processing [6]. Currently, hemoproteins have been widely applied in the fields of food and medicine. For example, bovine myoglobin (bovine-Mb) and soy hemoglobin (soy-Hb) can be used as food-grade coloring and flavoring agents [7,8]; lactoperoxidase can be used as a bioavailable bacteriostatic agent [9]; human hemoglobin (human-Hb) can be used as an acellular oxygen carrier [10]; and P450 enzymes can be used as efficient biocatalysts [11]. In addition, heme-derived compounds, such as biliverdin (BV, usually refers to BVIXα), bilirubin (BR), and phycocyanobilin (PCB), have received considerable attention for their potential in fighting various viruses and diseases, especially COVID-19 [12–14]. Due to their antioxidant [15,16], anti-inflammatory [17,18], and neuroprotective effects [19], these compounds can enhance antioxidant enzymic activity, upregulate anti-inflammatory cytokines, and inhibit the formation of α-synuclein and amyloid-β fibril.

All these biotechnological applications rely on the sufficient supply of heme in microbial cell factories. Heme is incorporated into hemoproteins during the co-translational folding process and assists in polypeptide-folding [20]. The insufficient intracellular heme can cause the degradation of subunits or apoproteins [21]. In addition, the biosynthesis of heme-derived chemicals starts from the catabolism of heme by heme oxygenase (HO), which leads to the generation of BV. Subsequently, biliverdin reductase (BvdR) and phycocyanobilin: ferredoxin oxidoreductase (PcyA) utilize BV as a substrate to generate BR and PCB, respectively (Figure 1(b)). Therefore, the intracellular availability of heme is crucial for producing highly active hemoproteins [22,23] and high-value-added heme derivatives [24–26].

Figure 1. Schematic overview of the biosynthetic pathway of heme (a) and its derivatives (b). G-6-P: glucose-6-phosphate; 3-PG: 3-phosphoglycerate; Pyr: pyruvate; CoA: coenzyme A; A-CoA: acetyl-CoA; CIT: citrate; ICT: isocitrate; α-KG: α-ketoglutarate; SUC: succinate; MAL: malate; TCA: tricarboxylic acid cycle; GLO: glyoxylate; LDH: lactate dehydrogenase; PTA: phosphate acetyltransferase; KDH: α-ketoglutarate dehydrogenase; SCS: succinyl‑CoA synthase; ICL: isocitrate lyase; Fd_red: reduced Fd (ferredoxin); Fd_ox: oxidized Fd; Fnr: Fd-NADP⁺ reductase.

There are two ways to improve intracellular heme supply. The first approach is the enhancement of endogenous heme biosynthesis. Although heme biosynthesis can be enhanced by the exogenous addition of 5-aminolevulinic acid (ALA, the key precursor of heme), the high cost of ALA makes it unsuitable for large-scale industrial production [23]. Therefore, many researchers have tried to increase the intracellular accumulation of ALA and activate the downstream enzymes responsible for heme synthesis. The second method is the supplementation of heme during the synthesis of hemoproteins or heme-derived chemicals. However, the ability of heme uptake for common microbial hosts (Saccharomyces cerevisiae, Pichia pastoris, Escherichia coli, and Corynebacterium glutamicum) is extremely weak. Therefore, it is necessary to introduce heme importers in these hosts to overcome this drawback [27–30].

With the increase of intracellular heme levels, the production of heme derivatives and hemoproteins can be improved. However, the excess free heme that exceeds the demand for product synthesis and cell growth is cytotoxic [23], which has an adverse effect on the sustainable production of heme derivatives and hemoproteins. Hence, the heme-responsive biosensor was designed by the incorporation of a heme-sensing regulator with Clustered Regularly Interspaced Short Palindromic Repeats interference (CRISPRi) or small regulatory RNA (sRNA) systems to fine-tune intracellular heme levels [23, 31].

With the rapid development of synthetic biology and system biology in recent years, many new advances have been reported in the: biosynthesis, acquisition, regulation, and upcycling of heme. Many new heme importers and heme-sensing regulators were discovered. Based on these discoveries, a series of strategies were developed to enhance the intracellular heme supply for synthesizing highly active hemoproteins and high-value-added heme derivatives. However, these advances have not been comprehensively reviewed. Thus, this review provides an overall summary in the: biosynthesis, acquisition, regulation, and upcycling of heme, laying a solid foundation for the development of valuable hemoproteins and heme derivatives.

Enhanced heme biosynthesis and its application in hemoprotein expression

Heme biosynthetic pathway

Heme biosynthesis begins with ALA, which has two main synthetic pathways depending on the species, C4 or C5 pathways (Figure 1(a)). In the C4 pathway that exists in: humans, animals, fungi, and α-proteobacteria, ALA synthase (ALAS) catalyzes the condensation of glycine (Gly) and succinyl-CoA (Suc-CoA) to form ALA. In the C5 pathway found in archaea, plants, and many bacteria, ALA is synthesized from l-glutamate (l-Glu) via a series of enzymes containing: glutamyl-tRNA synthetase (GluRS), glutamyl-tRNA reductase (GluTR), and glutamate-1-semialdehyde 2,1-aminomutase (GSAM). Then, ALA is condensed into porphobilinogen (PBG) by porphobilinogen synthase (PBGS; also named as ALA dehydratase, ALAD). Next, PBG undergoes a polymerization reaction triggered by porphobilinogen deaminase (PBGD) to form an unstable linear tetrapyrrole 1-hydroxymethylbilane (HMB), which is utilized as a substrate by uroporphyrinogen III synthase (UROS) and cyclized to produce uroporphyrinogen III (UPG III). In the classical protoporphyrin-dependent pathway (PPD) found in eukaryotes and α-proteobacteria, UPG III is converted into heme through the consecutive generation of intermediates, including: coproporphyrinogen III (CPG III), protoporphyrinogen IX (PPG IX), and protoporphyrin IX (PP IX), which are synthesized by: the uroporphyrinogen III decarboxylase (UROD), coproporphyrinogen III oxidase (CPO), protoporphyrinogen oxidase (PPO), and ferrochelatase (FECH) catalyze, respectively. In the noncanonical coproporphyrin-dependent pathway (CPD) found in Actinobacteria and Firmicutes, the initial step for the conversion of UPG III into heme is similar with the classical pathway. Subsequently, CPG III is transformed into coproporphyrin III (CP III) by CPO, which is gradually converted into coproheme and heme through two sequential reactions catalyzed by coproporphyrin ferrochelatase (CPFC) and hydrogen peroxide-dependent heme synthase (HS), respectively [32,33].

The strategies of metabolic engineering to synthesize ALA

The key precursor ALA is essential for the biosynthesis of heme [23]. Therefore, the recent metabolic engineering strategies for ALA production were summarized (Table 1). Initially, the researchers focused on engineering the shorter and simpler C4 pathway to improve ALA production, but it encounters a critical requirement of supplementing main precursors (Gly and Suc-CoA), which limits the productivity and yield of ALA [36]. For example, the titer of ALA reached 2.46 g/L in E. coli by co-expressing ALAS_RS (derived from Rhodobacter sphaeroides) and ALA exporter RhtA with the addition of 3 g/L Gly and 1 g/L Suc [34]. To overcome this problem, the biosynthesis of CoA and precursors (Gly and Suc-CoA) was enhanced and the expression of ALAD was downregulated, resulting in the titer of ALA increased to 2.81 g/L [35]. Based on these strategies and further utilization of 30 g/L glycerol as carbon source, 6.93 g/L of ALA was synthesized in E. coli, reaching 50.9% of the theoretical maximum yield [38]. In addition, the soluble expression and activity of ALAS were improved by the fusion expression with TrxA tag or co-expression of GroEL/ES chaperones, leading to a significant enhancement in ALA production from 1.21 to 3.67 g/L [37]. Meanwhile, the supplementation of 30 μM PLP (the cofactor of ALAS) was applied to increase the ALA titer from 1.4 to 2.41 g/L [39]. Furthermore, ALA can cause severe damage to E. coli cells and alters their morphology by generating reactive oxygen species [40], thus it is necessary to strengthen the antioxidant defense system of hosts by co-expressing catalase KatE and superoxide dismutase SodB, resulting in an improved ALA titer of 11.50 g/L. Another research showed that the removal of NCgl0580 that encodes a permease of the drug/metabolite transporter in C. glutamicum can redistribute the central carbon metabolism fluxes and increase the supply of Suc-CoA. As a result, there was a 2.49-fold increase in the titer of ALA, which opens up a new avenue for achieving high-level ALA production [54]. Finally, in the latest research, the highest titer of ALA (30.70 g/L) in E. coli was reported by developing a C75A/R365K variant of Rhodopseudomonas palustris ALAS with the property of anti-feedback inhibition and enhanced catalytic activity and discovering a new ALA exporter (EamA) [41].

Table 1. Metabolic engineering strategies to synthesize ALA and heme in recent years.

Strain	Engineering strategy	Titer	Ref.
	ALA biosynthesis using C4 pathway	g/L
E. coli	Co-expression of ALASRS and ALA exporter RhtA, supplementation of 3 g/L Gly and 1 g/L Suc, and 2.5 L baffle flask fermentation	2.46	[34]
	Overexpression of ALASRC, optimization of CoA and precursors (Gly and Suc-CoA) biosynthesis, downregulation of ALAD expression, and 3 L batch-fermentation	2.81	[35]
C. glutamicum	Optimization of Gly biosynthesis pathway and addition of 7.5 g/L Gly	3.40	[36]
E. coli	Co-expression of ALASRC, TrxA fusion tag, and GroEL/ES chaperones and supplementation of 6 g/L Gly, 2 g/L Suc, and 30 μM PLP	5.66	[37]
	Repression of ALAD expression, optimization of Suc-CoA biosynthesis, and 1 L fed-batch cultivation	6.93	[38]
	Co-expression of ALASRS, RhtA, and GroEL/ES chaperones and supplementation of 3 g/L Gly, 1 g/L Suc, and 30 μM PLP	7.47	[39]
	Co-expression of catalase KatE and superoxide dismutase SodB, supplementation of 4 g/L Gly, and 5 Lfed-batch fermentation	11.50	[40]
	Downregulation of SCS, ICL, and ALAD expression, enhancement of CoA supply, overexpression of the ALAS mutant C75A/R365K and the new ALA exporter EamA, supplementing Gly with 1–2 g/L/h, and 5 L fed-batch cultivation	30.70	[41]
	ALA biosynthesis using C5 pathway	g/L
E. coli	Co-expression of GluTRSty variant and GSAMEco	1.22	[42]
	Fine-tuning of ALAD expression using CRISPRi and 5 L fed-batch fermentation	2.00	[43]
C. glutamicum	Flux redistribution of TCA cycle toward l-Glu and introduction of RhtA	2.90	[44]
	Co-expression of GluTRSar variant and GSAMEco, weakness of KDH expression, cofactors (ATP, NADPH, and PLP) regeneration, and dynamically regulation of RhtA expression	3.16	[45]
E. coli	Co-expression of GluTRSty variant and GSAMEco, deletion of KDH, and activation of the glyoxylate cycle	3.40	[46]
	Balance of GluTRSar and GSAMEco expression, downregulation of ALAD expression, enhancement of PLP biosynthesis, and pH two-stage fermentation in 3 L bioreactor	5.25	[47]
	Assembly of GluRSBA, GluTREco, and GSAMBA, fine-tuning of KDH and ALAD expression, NADPH regeneration, overexpression of RhtA, and 5 L fed-batch cultivation	11.40	[48]
	Heme biosynthesis	mg/L
E. coli	Reinforcement of C5 pathway, overexpression of PBGS, PBGD, UROS, UROD, CPO, PPO, and FECH, knockout of byproduct pathway genes pta, ldhA, and yfeX, overexpression of heme exporter CcmABC, supplementation of 20 g/L glucose and 20 g/L l-Glu, and 6.6 L fed-batch fermentation	239.20	[49]
	Use of the engineered strain HAEM7[49], optimization of fermentation media, induction time, pH control, feeding solution and feeding strategies, and 6.6 L fed-batch fermentation	1034.30	[50]
	Overexpression of GluTR and GSAM, knockout of competitive pathway genes cysG, hemX, and cyoE, overexpression of heme exporter CcmA and DppA, and 3 L fed-batch fermentation	28.20	[51]
C. glutamicum	Combinational use of C4 and C5 pathways, reinforcement of CPD pathway using global transcriptional regulator DtXR, elimination of UPG III accumulation, enhancement of heme efflux by modifying cell membrane and overexpressing heme exporter HrtBA, and 2 L fed-batch fermentation	309.18	[52]
S. cerevisiae	Obtaining a heme-overproducing strain IMX581-FECH-PPO-PBGD-Δshm1-PBGS-Δhmx1-FET4-Δgcv2-ALAS-Δgcv1-CPO using genome-scale metabolic models Yeast8 and ecYeast8	53.50	[53]

ALAS_RS/RC: ALAS, derived from Rhodobacter sphaeroides or Rhodobacter capsulatus; Suc: succinate; GluRS_BA: GluRS, derived from Bacillus amyloliquefaciens; GluTR_Sty/Sar/Eco: GluTR, derived from Salmonella typhimurium or Salmonella arizona or Escherichia coli; GSAM_Eco/BA: GSAM, derived from E. coli or B. amyloliquefaciens; ATP: adenosine triphosphate; NADPH: nicotinamide adenine dinucleotide phosphate; PLP: pyridoxal phosphate; pta: encoding phosphate acetyltransferase; ldhA: encoding lactate dehydrogenase A; yfeX: encoding a putative heme-degrading enzyme; cysG: encoding siroheme synthase; hemX: encoding uroporphyrinogen-III C-methyltransferase; cyoE: encoding heme O synthase; shm1: encoding glycine hydroxymethyltransferase; hmx1: encoding HO; FET4: encoding putative iron/copper ion transporter; gcv2: encoding glycine decarboxylase subunit P; gcv1: encoding glycine decarboxylase subunit T.

Alternatively, the C5 pathway has attracted more attention to produce ALA from glucose. It has been shown that N-terminal modification of the GluTR increases its stability and reduces the effect of feedback inhibition from heme. Indeed, the feedback-resistant version of GluTR resulted in a 1.76-fold increase in ALA production compared to its wild-type [42]. Based on this research, ALA production was further improved by weakening the expression of ALAD or pushing the flux of the TCA cycle toward l-Glu [43,44]. However, the titer of ALA remained low, ranging between 2.0 and 3.0 g/L. Therefore, the following studies combined the previous strategies and focused on regenerating the cofactors of GluRS, GluTR, and GSAM (ATP, NADPH, and PLP) and activating the glyoxylate cycle, which increased the titer of ALA to 3.0–5.0 g/L [45–47]. Additionally, due to more steps involved in the C5 pathway, spatial assembly of GluRS, GluTR, and GSAM using the protein scaffold was applied to eliminate the obstacles in the transfer of intermediates, resulting in a significant increase in ALA titer to 11.40 g/L [48]. In the future, the useful strategies utilized in various species could be integrated together to further boost ALA production.

The strategies of metabolic engineering to synthesize heme

Based on the accumulation of ALA, the enhancement of the downstream heme biosynthetic pathway has been achieved in various microorganisms (Table 1). At first, an engineered E. coli HAEM7 strain that can efficiently synthesize and secrete heme was constructed by enhancing the C5 pathway and the other subsequent seven genes, inhibiting the formation of byproducts and the degradation of heme, and overexpressing the heme exporter CcmABC, producing 87.8 mg/L intracellular and 151.4 mg/L extracellular heme in the 6.6 L fed-batch fermentation [49]. In the following, the same group further optimized the fermentation conditions and developed a novel feeding strategy, leading to the highest reported titer of heme (1.03 g/L) in E. coli [50]. In addition, another recent study focused on the elimination of competitive pathways (the formation of siroheme and heme O) and the reinforcement of heme efflux pathway (Dpp and Ccm), increasing the titer of heme by 2.08-fold to 28.20 mg/L in E. coli [51].

Furthermore, the metabolic engineering of C. glutamicum has shown promising results in heme synthesis. Through combining the C4 and C5 pathways, optimizing the noncanonical CPD pathway, reducing the accumulation of intermediate (UPG III), knocking out the heme-binding proteins (HBPs) (HrrS, HtaA, and HmuT) in the cell membrane, and overexpressing the heme exporter HrtBA, the engineered strain produced 66.23 mg/L intracellular and 242.95 mg/L extracellular heme in the 2 L fed-batch fermentation [52]. Moreover, S. cerevisiae was considered as a suitable food-grade host for the production of Hb and Mb to develop artificial meat. Applying the genome-scale metabolic models (Yeast8 and ecYeast8) to identify the genes that significantly affect heme biosynthesis, the engineered S. cerevisiae strain showed a 70-fold increase in the intracellular titer of heme (53.5 mg/L), which can be used as a powerful heme-supplying chassis [53].

Efficient expression of hemoproteins through enhancing heme biosynthesis

Recently, enhancing heme biosynthetic pathway has emerged as a promising approach for synthesizing highly active hemoproteins (Table 2). Soy-Hb and bovine-Mb, synthesized in an engineered P. pastoris strain overexpressing eight genes involved in heme biosynthesis, have been approved by the USA Food and Drug Association (FDA) as flavoring and coloring agents (GRN nos. 737 and 1001), and applied by Impossible Foods Inc. (Redwood City, CA) and Motif Food Works Inc. (Boston, MA) to develop the popular artificial meat products of “Impossible Burgers” and “HEMAMI”, respectively. The following research used the similar strategy to increase intracellular availability of heme in P. pastoris, increasing the titer of soy-Hb to 3.5 g/L [55]. However, a recent study revealed that there was a noticeable accumulation of heme intermediates (PBG, HMB, CPG III, PPG IX, and PP IX) rather than heme in the engineered P. pastoris strain overexpressing all heme biosynthetic enzymes [61]. Therefore, in our latest result, the intracellular supply of heme was moderately enhanced in S. cerevisiae by improving ALA synthesis (multi-copy integrated expression of ALAS), removing the spatial barrier during heme synthesis (the assembly of CPO with mitochondrial localization signal truncated PPO and FECH), and knocking out the HMX1(HO). Applying this engineered S. cerevisiae strain, several highly active Hb and Mb with a superior heme-binding ratio were successfully synthesized, including: soy-Hb, clover-Hb, bovine-(Hb/Mb), and porcine-(Hb/Mb) [22]. Based on these engineering strategies, the rate-limiting steps in heme biosynthesis mediated by PBGS, PBGD, and UROS were further eliminated in our subsequent study, resulting in the enhanced synthesis of high-activity porcine-Mb, soy-Hb, and Vitreoscilla-Hb in P. pastoris [56]. In addition, a metabolically engineered S. cerevisiae strain produced human-Hb at the level of 18% total intracellular proteins by overexpressing PBGD and deleting HMX1 and the heme-dependent repressor ROX1 [57]. Furthermore, in C. glutamicum, the latest study showed that the heme-binding ratio of synthesized soy-Hb could reach 28% by adding 1.0 g/L ALA [58].

Table 2. Enhanced hemoprotein expression through improved intracellular heme biosynthesis.

Strain	Engineering strategy	Hemoproteins	Titer/activity	Ref.
P. pastoris	Overexpression of all heme biosynthetic enzymes	Soy-Hb	ND	GRN no. 737
	Overexpression of all heme biosynthetic enzymes	Bovine-Mb	ND	GRN no. 1001
	Overexpression of all heme biosynthetic enzymes and 10 L fed-batch fermentation	Soy-Hb	3.5 g/L	[55]
	Enhancement of ALA synthesis, assembly of PBGS, PBGD, and UROS, cytoplasmic localization of PPO and FECH, and inhibition of heme degradation	Porcine-Mb, soy-Hb, Vitreoscilla-Hb, and mutated P450-BM3	Titers reached 247.0, 286.5, 66.3, and 57.3 mg/L, respectively, while functionalized activities improved by 38.1%, 29.3%, 85.2%, and 7.5-fold, respectively	[56]
S. cerevisiae	Multi-copy integrated expression of ALAS, assembly of CPO with truncated PPO and FECH, and inhibition of heme degradation	Soy-Hb, clover-Hb, bovine-(Hb/Mb), and porcine-(Hb/Mb)	108.2, 13.7, 19.3, 68.9, 11.7, and 85.9 mg/L, respectively	[22]
	Overexpression of PBGD and deletion of HMX1 and the heme-dependent repressor ROX1	Human-Hb	18% of intracellular proteins	[57]
C. glutamicum	Supplementation of 1.0 g/L ALA	Soy-Hb	20% of intracellular proteins	[58]
E. coli	Supplementation of 0.5 mM ALA	CYP153A33	Whole-cell catalytic activity increased by 1.2-fold	[59]
	Overexpression of GluTR^fbr and GSAM	trLUT5/trLUT1 and trCYP93B16	Whole-cell catalytic activities improved by 71.5% and 99.0%, respectively	[60]
	Assembly of PBGS, PBGD, and UROS and integrated expression of GluTR, GSAM, and FECH	Mutated P450-BM3	Whole-cell catalytic activity enhanced by 2.0-fold	[23]
	Overexpression of GluRS, GluTR, GSAM, UROD, CPO, PPO, and FECH	Three P450s derived from M. marinum, Polaromonas sp. JS666, and M. aquaeolei	Whole-cell catalytic efficiencies increased by 1.8- to 6.5-fold	[61]

trLUT5/trLUT1: truncated β-carotene 3-hydroxylase (LUT5) and carotene ε-monooxygenase (LUT1); trCYP93B16: truncated CYP93B16; ND: not determined.

In the field of whole-cell P450s catalysis, 1,12-dodecanediol synthesized by CYP153A33 increased from 0.26 to 0.58 mM with the supplementation of 0.5 mM ALA [59], the co-expression of ALAS improved the heme-binding ratio in CYP119 from 66 to 99% [62], as well as the synthesis of lutein and apigenin by P450s increased by 71.5% and 99.0%, respectively, when GluTR^fbr (a feedback-resistant version) and GSAM were overexpressed [60]. Moreover, the combination of assembling three rate-limiting enzymes (PBGS, PBGD, and UROS) and integrally expressing another three biosynthetic enzymes (GluTR, GSAM, and FECH) enhanced the enzymic activity of mutated P450-BM3 by approximately 2.0-fold [23], overexpression of: GluRS, GluTR, GSAM, UROD, CPO, PPO, and FECH, could increase the whole-cell catalytic efficiencies of three P450 enzymes (from Mycobacterium marinum, Polaromonas sp. JS666, and Marinobacter aquaeolei) by 1.8- to 6.5-fold [61], and: overexpression of ALAS, assembly of PBGS, PBGD, and UROS using protein scaffolds, cytoplasmic localization of PPO and FECH, and inhibition of heme degradation, improved the whole-cell catalytic activity of mutated P450-BM3 by 7.5-fold [56].

Improving intracellular heme supply by introducing heme importers

Improving the ability of heme uptake is beneficial for increasing intracellular heme supply, producing high-activity hemoproteins and high-value-added heme derivatives [23, 30]. In this section, the new discovered heme importers and their latest applications in biotechnology were introduced.

The mechanism of heme acquisition in pathogenic bacteria and fungal

Most of heme importers were identified in pathogenic bacteria, fungal, parasites, and mammals (Table 3). In Gram-negative bacteria, the classic heme uptake system (HUS) consists of TonB-dependent heme receptors (TBDRs) in the outer membrane, the TonB–ExbB–ExbD complex, periplasmic HBP and ATP-binding cassette (ABC) transporters [89] (Figure 2(a)). The TBDRs can either bind free heme or accept heme from Hb and hemophore (HMP), and then internalize it into the periplasm through the TonB–ExbB–ExbD complex [66]. Subsequently, heme is transferred by the periplasmic HBP to ABC transporters in the inner membrane and is further carried to the cytoplasm for the synthesis of hemoprotein or the utilization of iron [89].

Figure 2. The mechanism of heme uptake and its diverse applications in biotechnology. (a) The HUS in Gram-negative and Gram-positive bacteria. There are both Isd and non-Isd systems present in Gram-positive bacteria, such as S. aureus and C. diphtheriae. OM: outer membrane; IM: inner membrane; CW: cell wall. (b) The applications of heme importers in the production of hemoprotein and heme-derived compounds, the assembly of ArMs, and the development of vaccines and antibodies.

Table 3. The latest identified heme importers and their potential targets for fighting pathogens.

Strain	Heme importers			Ref.
Gram-negative bacteria	Hemophores	TBDRs	ABC transporters
T. forsythia	Tfo	Tanf_RS09470	ND	[63]
P. intermedia	PinO and PinA	ND	ND	[64]
B. vulgatus	Bvu	BVU_RS11040	ND	[65]
R. anatipestifer	ND	RhuR	ND	[66]
C. crescentus	ND	HutA	ND	[67]
Roseobacter	ND	HmuR	HmuTUV	[68]
V. vulnificus M2799	ND	HupA and HvtA	HupBCD	[69]
Acinetobacter baumannii	HphA	HemTR and HphR	ACICU_1637/1638/1639	[70,71]
Chromobacterium violaceum	ND	ChuR	ChuTUV	[72]
Gram-positive bacteria
Streptococcus pneumoniae	Cytosolic heme-binding protein SPD_0310 and ABC transporters SPD_0090, SPD_1609, and SPD_1621			[73]
S. lugdunensis	ECF transporters LhaSTA and its homolog SLUG_19150/19160/19170			[74,75]
Group A Streptococcus	ECF transporters SiaFGH			[76]
L. sakei	ECF transporters Lsa_1836-1840			[77]
Eukaryotes
S. pombe	Cell-surface membrane protein Str3			[78]
I. ricinus	Solute carrier protein IrHRG			[79]
Leishmania	Solute carrier protein LmFLVCRb			[80]
Mammals cell	Plasma membrane heme importer SLCO2B1 and solute carrier protein SLC46A1			[81,82]
	Targets
S. aureus	Out membrane heme receptor IsdA			[83]
Bacillus anthracis	Out membrane heme receptor NEAT proteins (IsdX1, IsdX2, and Bslk)			[84]
A. baumannii	Out membrane heme receptor HemTR			[70]
Mycobacterium tuberculosis	Out membrane heme receptor PPE36 and PPE62			[85,86]
Group A Streptococcus	Out membrane heme receptor Shr and ABC transporter HtsA			[87,88]

ND: not determined.

In Gram-positive bacteria, classic HUSs are mainly divided into iron-regulated surface determinants (Isd) and non-Isd systems (Figure 2(a)). In the Isd system of Staphylococcus aureus, the cell-wall-anchored surface receptors (IsdH, IsdB, and IsdA) contain the special domain of heme-binding near transporter (NEAT) [90]. These receptors are the main binding proteins for haptoglobin (HP)–Hb complexes, Hb, and heme, respectively. IsdA can bind free heme or receive it from IsdH and IsdB and transfer it to the NEAT-containing IsdC buried in the cell wall [91]. IsdC then delivers heme to the membrane-embedded transporter IsdDEF that transports heme across the membrane and into the cytoplasm [91]. As for non-Isd systems, taking Corynebacterium diphtheriae as an example, it primarily includes the cell-wall-anchored surface receptors HtaA and HtaB, as well as the ABC transporter HmuTUV [92]. HtaA possesses two heme-binding conserved regions (CR1 and CR2) and can acquire heme from Hb [93]. Additionally, ChtA and ChtC, which contain a single CR domain, require HtaA for the uptake of heme from HP–Hb complexes [92]. HtaB also harbors a single CR domain and can bind free heme or accept heme attached to HtaA and transport it to HmuTUV membrane transporters, which internalize heme into the cytoplasm [94].

The general model of heme uptake in fungal pathogens is quite different, including Candida albicans, Schizosaccharomyces pombe, and Cryptococcus neoformans. The mechanism is that the cell wall surface-anchored binding proteins of heme or hemoprotein (Rbt5/Pga7/Csa2 in C. albicans, Shu1 in S. pombe, and Cig1 in C. neoformans) capture free heme or extract heme from hemoproteins and steer it across the cell wall to the plasma membrane, where it is endocytosed into the cytoplasm with the aid of ubiquitinated proteins and the endosomal sorting complex required for transport (ESCRT) [95–97].

The newly discovered heme importers

With the development of genomics and protein engineering, more and more heme importers have been identified, especially in pathogens (Table 3). In Gram-negative bacteria, it was found that a HMP-like protein HmuY is essential for acquiring heme and increasing virulence in the main periodontopathogen (Porphyromonas gingivalis). The heterologous co-expression of HmuY and HmuR (TBDR) in a hemA mutant E. coli strain where the endogenous heme biosynthesis pathway is blocked, can promote the utilization of exogenous heme and growth recovery [98]. The homologous proteins of HumY have been recently identified in other periodontopathogens, such as Tfo in Tannerella forsythia [63], PinO/PinA in Prevotella intermedia [64], and Bvu in Bacteroides vulgatus [65], which display different modes of heme coordination with HmuY (heme ligands: His¹³⁴/His¹⁶⁶ in HmuY, Met¹²⁰/Met¹⁴⁹ in Tfo, Met¹¹⁹ in PinO, Met¹²⁹/Met¹⁵⁸ in PinA, and Met¹⁴⁵/Met¹⁷² in Bvu). In addition, a recent study found that the outer membrane HBP RhuA can transfer heme to the RhuR (TBDR) to facilitate heme utilization in Riemerella anatipestifer. The heme binding ability of RhuA is stronger than that of HemR (TBDR) derived from Serratia marcescens in ΔhemA mutant E. coli strain [66].

In Gram-positive bacteria, the discovery of new energy coupling factor (ECF) transporters attracted more attention. It is a noncanonical ABC transporter comprised of a substrate-specific component (EcfS), a transmembrane protein (EcfT), and a pair of cytosolic ATPases (EcfA) [76]. It can acquire heme from free heme and hemoproteins and independently functions without the Isd system. The latest identified ECF transporters include LhaSTA and its homologs SLUG_19150/19160/19170 in Staphylococcus lugdunensis [74,75], SiaFGH in group A Streptococcus [76], and Lsa_1836-1840 in Lactobacillus sakei [77]. Among them, the LhaS (EcfS) could be heterologously expressed in E. coli and absorb heme from the growth media [74]. In eukaryotes, previous studies have shown that the S. cerevisiae mutant (Δhem1) can uptake heme and the growth can be restored by introducing these heme importers derived from parasites, including: TbHRG in Trypanosoma brucei [99], BmHRG-1 in Brugia malayi [100], CeHRG1 in Caenorhabditis elegans [101], LHR1 in Leishmania [102], and TcHTE in Trypanosoma cruzi [103]. The latest discovered heme importers include: Str3 in S. pombe [78], IrHRG in Ixodes ricinus [79], LmFLVCRb in Leishmania [80], and SLCO2B1 and SLC46A1 in mammals [81,82]. Among them, besides Shu1, Str3 is the second cell-membrane surface heme importer in S. pombe. The increasing expressional level of Str3 restored the growth of S. pombe Δhem1Δshu1 mutant when hemin was supplemented [78]. Moreover, the heterologous expression of Str3 enhanced the ability of S. cerevisiae Δhem1 mutant to absorb exogenous heme [78].

The latest applications of heme importers

The application of heme importers has shown great potential in synthesizing hemoproteins and heme derivatives, assembling artificial metalloenzymes (ArMs), and developing vaccines and antibodies (Figure 2(b)). Regarding the synthesis of hemoproteins and heme derivatives, a previous study showed that the introduction of the entire HUS hug operon derived from Plesiomonas shigelloides into E. coli resulted in a 30-fold increase in the expression of human-Hb compared to the control strain without the transport system when supplemented with 2.3 mM hemin [27]. Later, it was proposed that the E. coli strain Nissle 1917 (EcN) strain bearing the ChuA (TBDR) might be an ideal host for the production of hemoproteins. It was found that the EcN strain produced a higher titer (over 82.36%) and heme-binding ratio of the protein sGAF2 than the control strain E. coli BL21(DE3) when 10 μM hemin was added [29]. However, the EcN strain lacks the T7 RNA polymerase, which limits the use of common commercial vectors based on the T7 promoter. Therefore, the gene encoding T7 RNA polymerase was integrated into the genome of EcN strain to construct the EcN(T7) strain, which exhibited more efficient expressional or synthetic efficiency of hemoprotein, BV, and PCB when hemin was supplemented [28]. In addition, the recent result showed that the co-expression of ChuA in E. coli C41(DE3) increased the whole-cell catalytic efficiency of mutated P450-BM3 and P450-sca-2 by 16.1% and 29.4%, respectively [23].

In terms of ArMs assembly, the EcN strain showed a two-fold increase in Ir content in the supernatant of cell lysates compared to the control BL21 (DE3) strain, indicating that EcN is more efficient in exogenously utilizing the cofactor Ir(Me)MPIX (porphyrin compounds) of Ir-CYP119 [104]. However, the EcN strain may not be the optimal choice. When Ir(Me)MPIX was supplemented, the co-expression of the hug operon in BL21(DE3) resulted in a significant increase in the assembly and catalytic efficiency of Ir-CYP119 compared to the co-expression of ChuA and the blank control [105]. Therefore, whether to introduce the heme receptor in the out membrane alone or the entire HUS depends on the homology of HUS components between the expression host and the original source strain, which has also been demonstrated in our recent study [23]. In addition, HUS components are crucial for the survival of pathogens that cannot synthesize heme as an internal iron source, which can be targeted for developing vaccines and antibody-based immunotherapies (Table 3), including: IsdA for S. aureus [83], NEAT proteins (IsdX1, IsdX2, and Bslk) for Bacillus anthracis [84], and Shr and HtsA for group A Streptococcus [87,88].

Dynamically regulating heme synthetic pathway by biosensor

The excess intracellular heme is cytotoxic because it can cause lipid peroxidation and trigger ROS formation [23]. To control the concentration of heme at the lower level, there are heme-responsive regulators existing in various organisms. Heme can bind to the heme-responsive regulators (such as HrtR in Lactococcus lactis), altering their conformation and the ability of DNA binding to regulate the transcription of downstream genes involved in heme synthesis [106]. Based on the special property of heme-binding, many heme-sensing regulators have been employed to develop fine-tuning systems to dynamically control the moderate supply of heme for the sustainable production of hemoproteins and heme-derived compounds. In this section, the new discovered heme-responsive regulators and several fluorescently labeled heme sensors were introduced, providing a foundation to develop further effective heme-regulating systems.

The latest discovered heme-responsive regulators include: HatR in Clostridium difficile [107], FhtR in Enterococcus faecalis [108], ChuP in Chromobacterium violaceum [72], and PhuS in Pseudomonas aeruginosa [109] (Table 4). HatR and FhtR belong to the TetR family of transcriptional regulators and function as transcriptional repressors of the hatRT and hrtBA_Ef operons, respectively [107,108]. When intracellular free heme is abundant, HatR and FhtR sense and bind to heme, inhibiting their function as repressors and increasing the transcription of hatRT and hrtBA_Ef, respectively. Next, HatT and HrtBA_Ef control the efflux of intracellular heme, thereby reducing heme toxicity in the cells. Unlike HatR and FhtR, ChuP can bind heme and activate the expression of the ChuR (TBDR) and siderophore VbuA by the regulation of post-transcription under the limitation of iron, thereby increasing the uptake of iron [72]. In addition, PhuS is a special regulator that controls the intracellular homeostasis of iron through its dual function. It can transfer heme to HO and repress the transcription of PrrH sRNA [109] that is involved in the degradation of mRNA encoding non-essential iron-containing proteins such as pyochelin siderophore [110]. Apo-PhuS binds to the PrrH promoter, but this binding is inhibited in the presence of heme. In iron-limited conditions, the equilibrium shifts toward apo-PhuS, which downregulates the relative level of PrrH; while under heme-replete conditions, the equilibrium shifts toward holo-PhuS, leading to the increased expression of PrrH.

Table 4. The new discovered heme-responsive regulators in recent years.

Strain	Regulators	Heme-binding motif	Function	Downstream genes/or sRNA	DNA-binding region	Conditions	Ref.
C. difficile	HatR	His99	Transcriptional repressor	Heme exporter hatT	Not binding, as the hatR gene is located upstream of the hatT gene and shares the same promoter	Heme-limited	[107]
E. faecalis	FhtR	Tyr132	Transcriptional repressor	Heme exporter hrtBAEf	P1 (TTATCAATCGATAA) and P2 (TTATCGATTGATAA), located at hrtBAEf promoter	Heme-limited	[108]
C. violaceum	ChuP	ND	Post-transcriptional activator	The TBDR chuR and siderophore vbuA	HPRE-S1 (CCCGCAAGCCAGCCGACAGCCAGCCAGCG), 5′-UTR of the chuR gene; HPRE-S2 (GCCAGCCAGACGACGCCGCCG), 5′-UTR of the vbuA gene	Iron-limited	[72]
P. aeruginosa	PhuS	H212	Transcriptional repressor	PrrH sRNA	Within 50 bp upstream of the prrF1 sRNA, located at prrF1 promoter	Iron-limited	[109,110]

HPRE: HmuP-responsive element; P1 and P2 are palindromic repeat sequences (5′–3′); S1/2: DNA sequence 1/2 (5′–3′); PrrH sRNA formed by the non-coding PrrF1 sRNA in tandem with PrrF2 sRNA. ND: not determined.

Based on the heme-sensing regulators, various regulatory systems have been developed to synthesize heme and its precursors, and hemoproteins. HrrSA, a two-component heme-responsive system in C. glutamicum, comprises a histidine kinase HrrS and a regulator HrrA [111]. HrrS responds to the extracellular heme by phosphorylating HrrA, which binds to the promoter of the hmuO gene encoding HO (P_hmuO) and upregulates its expression. In a previous study, the titer of ALA in C. glutamicum increased by 32.8% using P_hmuO to express the ALA exporter RhtA [45]. In addition, another study used a regulator HrtR derived from L. lactis, a hybrid promoter containing HrtR DNA-binding site (hrtO_L) and CRISPRi technique to construct a heme-responsive regulatory system in E. coli [31]. The expression of ALAD, PBGD, and FECH was dynamically downregulated using this system, respectively, resulting in the increase of 54.0%, 428.4%, and 65.0% in the titers of ALA, PBG, and porphyrins. Considering the long regulatory period (>6 h) of this CRISPRi-based system for the fine-tuning of heme synthetic pathway and the potential off-target effect of dCas9, a modified heme biosensor (regulatory period of 2–3 h) was developed in E. coli using a mutated HrtR regulator and its binding site hrtO_L, and sRNA technique [23]. Using this sensitive system to regulate the enhanced heme biosynthetic pathway, the whole-cell enzymic activity of mutated P450-BM3 increased by 1.4-fold.

The design of pathways to efficiently synthesize heme-derived compounds

As for the production of heme derivatives, although BV and BR, as well as PCB can be extracted from mammalian bile and Spirulina, respectively [24,25,112], these methods suffer from: the shortage of raw materials, unsustainability, difficulty in purification, etc. Hence, the effective biosynthetic approach is selected based on the adequate intracellular supply of heme. In the recent years, with the development of systems and synthetic biology, some novel strategies have been used to enhance the synthetic level of heme-derived compounds (Figure 3).

Figure 3. The strategies used for enhancing the production of heme derivatives (BV, BR, and PCB). (a) Metabolic engineering for predicting and designing the rate-limiting steps in the BV biosynthetic pathway [24]. MetaCyC: a website (https://metacyc.org/) used for the thermodynamic analysis. (b) Bioconversion of heme into BV and BV into BR in engineered EcN(T7) and E. coli BL21(DE3) strains, respectively [25, 30]. ChuA: an outer membrane heme receptor. (c) Modular optimization for the expression of the ho1 and pcyA genes in recombinant E. coli [26, 113]. T: T. elongatus; S: Synechocystis. (d) Spatial assembly of HO1 and PcyA by the short peptide tags and DNA scaffolds to optimize substrate trafficking and facilitate metabolic flux [26, 113]. RIDD/RIAD: the short peptide tags; ADB1/2: the artificial DNA binding domains. (e) Response surface methodology for optimizing the fermentation conditions of PCB in engineered E. coli strains. (f) Enhancement of PCB biosynthesis for optogenetic control of cell signal pathways in mammalian and B. subtilis [114–116]. Pr: red-absorbing ground state; Pfr: (far-red)-absorbing ground state; Pg: green-absorbing ground state.

Regarding the production of BV, a previous study identified the rate-limiting enzymes (GluTR and GSAM) in the heme biosynthetic pathway using thermodynamic analysis and revealed a novel Fe-coproporphyrin III decarboxylase (HemQ)-mediated coproporphyrin dependent pathway in C. glutamicum (Figure 3(a)) [24]. Subsequently, the expression of: GluTR^fbr (the feedback-resistant version of GluTR, derived from Salmonella typhimurium), GSAM (derived from E. coli), HemQ, and HmuO (hereafter HO1) was optimized in different plasmids, resulting in the highest reported titer of BV at 68.74 mg/L during 5 L fed-batch fermentation. Next, another study used the engineered EcN(T7) strain overexpressing the ho1 gene to produce BV (BVIXα) with the supplementation of 10.0 μM hemin (Figure 3(b)) [30]. It displayed that the yield of BVIXβ and -δ isomers (26% and 24%, respectively) were improved by over 700-fold and 400-fold than the yield obtained by the chemical coupled oxidation method (0.035% and 0.054%, respectively).

As for the microbial production of BR in recent years, only one research reported the biotransformation of BV using the engineered E. coli strain overexpressing the bvdR gene from cyanobacterium Synechocystis (Figure 3(b)) [25]. Using 450 mg/L of BV as the substrate, it exhibited a conversion rate of 72.3% and a yield of 20.8 mg/g (DCW), which is 35-fold higher than the traditional method extracted from pig bile.

Along with the growing demand for PCB in the: food, pharmaceutical, cosmetic, and optogenetic fields [117], the synthesis of PCB in different organisms has become a hot topic, including: E. coli, Bacillus subtilis, and mammalian cells. In E. coli, a previous study found that monocistronic expression of the ho1 and pcyA genes (derived from Synechocystis sp. PCC6803) could produce 6.64 mg/L PCB, but the upregulation of: GluTR, GSAM, PBGS, PPO, and FECH and the increased copy number of ho1 and pcyA could not significantly further improve the titer of PCB [118]. Another study achieved the PCB titer of 13 mg/L in the engineered E. coli overexpressing ho1 and pcyA (derived from Synechocystis sp. PCC6803) at the: optimized concentration of lactose (4 mmol/L), induction temperature (24.69 °C), induction time (4.6 h), and induction duration (13.57 h) obtained by response surface methodology (Figure 3(e)) [119]. Nevertheless, the PCB titers remain low due to the low expressional level of ho1 and pcyA from Synechocystis sp. and the insufficient supply of the cofactors (NADPH and Fd) for PCB biosynthesis. Therefore, our recent study obtained a higher PCB titer of 28.32 mg/L during 5 L fed-batch fermentation through the spatial assembly of the highly active ho1 and pcyA (ho1 from Thermosynechococcus elongatus and pcyA from Synechocystis sp. PCC6803) using proper DNA scaffolds, the supplementation of 200 mg/L ALA, 20 mg/L FeSO₄·7H₂O, and 5 g/L vitamin C (reducing agent), and the overexpression of PBGS, FECH, and NadK (NAD kinase) (Figure 3(c,d)) [26]. Subsequently, the PCB titer was increased to the highest reported record of 147.0 mg/L in our latest study by selecting the better chassis E. coli BL21(DE3), enhancing heme supply using the overexpression of all genes involved in heme biosynthesis, assembling HO1 and PcyA (derived from Synechocystis sp. PCC6803) using the short peptide tags RIAD-RIDD (Figure 3(c,d)) [113].

Furthermore, the biosynthesis of PCB in B. subtilis and mammalian cells was mainly used as the optogenetic control of cell signaling. The genetically encoded PhyB (phytochrome B)-PIF (phytochrome-interacting factor) LID system for manipulating cellular signaling in mammalian cells was developed based on the optical properties of PCB. PhyB can ligate with PCB to form a photoabsorbing chromophore and can rapidly bind to PIF when exposed to red light, while the PhyB–PIF complex promptly dissociates when exposed to far-red light (Figure 3(f)). In the previous reported PhyB–PIF LID system, the co-expression of: MTS-PcyA-P2A, MTS-HO1-P2A, MTS-Fd-P2A, and MTS-Fnr (MTS, mitochondria-targeting sequence; P2A, self-cleaving 2A peptide; pcyA and ho1 derived from T. elongatus BP-1; Fd and Fnr derived from Synechocystis sp. PCC 6803) produced a red fluorescent intensity equivalent to 2.5 μM PCB, which was further enhanced by knocking out/or down for BvdR [114]. Subsequently, the PhyB–PIF LID system was improved by the polycistronic co-expression of: MTS-tFnr-Fd-P2A, MTS-PcyA-P2A, and MTS-HO1 (tFnr-Fd, a CpcD-like domain-truncated version of Fnr fused with Fd; pcyA, ho1, Fd, and Fnr from T. elongatus BP-1), increasing the synthesis of PCB by 3.9-fold [115]. In B. subtilis, the binding of Cph1 (a phytochrome from Synechocystis sp. PCC 6803) with PCB resulted in weak red fluorescence. Furthermore, the fusion expression of ho1*-pcyA** (ho1*, codon-optimized initial 15 codons of ho1; pcyA**, codon-optimized complete pcyA with a transcriptional terminator to eliminate potential mRNA instability) led to a significant 80-fold increase in red fluorescence [116]. The enhancement of the PCB synthesis could support the robust function of CcaSR (a transcriptional regulatory pathway driven by optogenetic) in B. subtilis (Figure 3(f)).

Conclusions and future perspectives

The bioproduction of heme derivatives and hemoproteins depends on the availability of intracellular heme. Over the past five years, various approaches, including systems metabolic engineering, genome-scale metabolic modeling, and pathway reconstruction, have been used to increase intracellular heme levels in different microorganisms, resulting in the successful synthesis of heme derivatives and hemoproteins. However, the titer, yield, and productivity of these products in the recombinant strains are still low for industrial applications, and several challenges are required to be overcome. First, the regulation of heme synthesis varied a lot in different microorganisms, and the effective strategies cannot be universally applied [55]. Second, the heme biosynthetic pathway is long and complex, and the overexpression of all necessary genes can result in the accumulation of heme intermediates rather than heme [61]. Third, the ability of heme uptake is too weak in the commercial microbial hosts and are required to be further enhanced [23]. Fourth, the novel regulatory elements are limited to balance the heme supply and the production of heme derivatives/or hemoproteins in different microorganisms [23, 31]. Fifth, the activities of several enzymes related to the synthesis of heme derivatives are relatively low [26, 113].

To address these issues, five potential directions can be attempted. At first, the metabolic flux analysis can be used to identify the rate-limiting steps of heme synthesis in diverse microbial hosts [120]. Then, the variants of rate-limiting enzymes with high activities can be obtained through de novo design based on the predictive tools such as AlphaFold or the insertion of nonstandard amino acids [121,122]. In addition, the directed evolution method [123] or the saturation mutagenesis for the key residues of heme‐binding can be used to enhance the capacity of heme importers to absorb exogenous heme. Furthermore, the transcriptomic analysis can be utilized to analyze the transcriptional changes under the different levels of intracellular heme to develop an increasing number of regulatory elements for heme biosynthesis, such as hybrid promoters and DNA adapters. Finally, highly active enzymes for the synthesis of heme derivatives (HO, etc.) can be obtained from algae, microorganisms, plants, and animals by the combination of machine learning and high-throughput screening [124]. By employing these strategies, the production of heme derivatives and hemoproteins can be significantly enhanced through the fine-tuning the enhanced heme biosynthetic pathway.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Key Research and Development Program of China (2021YFC2101400) and the National First-class Discipline Program of Light Industry Technology and Engineering (LITE2018-08).