Gut microbial-host isozymes: A novel perspective on gut microbiota-host interactions

Summary

The interaction between gut microbiota and host health has garnered significant attention since its initial discovery. Dysbiosis of the gut microbiota is implicated in various diseases, particularly metabolic disorders such as diabetes, as well as neuro-, cardiovascular-, and hepatic-metabolic diseases. In recent years, the concept of microbial-host isozymes (MHIs) has emerged as a novel research area within the microbiome field. These enzymes, encoded by intestinal microbiota, can replicate the functions of host enzymes and contribute to disease development, presenting potential new therapeutic targets. In this review, we examine the current understanding of the discovery, function, and potential applications of MHIs. We summarize the distribution and functional enrichment of identified MHIs, provide examples of MHI-targeted interventions aimed at optimizing diabetes treatment, and discuss existing challenges and future research directions in this area.

Graphical abstract

Gastroenterology; Microbiome; Endocrinology

Introduction

The gut microbiota plays a crucial role in maintaining host health and influencing the onset and progression of various diseases. Beyond its well-established function in sustaining intestinal homeostasis, this complex microbial community is actively involved in numerous physiological processes, including host immune regulation, nutrient digestion, and drug metabolism.¹ These intricate interactions are facilitated through various media and messengers, primarily metabolites and proteins, which serve as essential communication tools between the microbiota and the host.^2,3

Recently, researchers have identified a novel interaction mechanism that enhances our understanding of the gut microbiota’s impact on host biology. It has been discovered that gut microbiota can produce enzymes that mimic host functions, referred to as gut microbial-host isozymes (MHIs).⁴ This groundbreaking finding not only expands our perspective on microbial-host interactions but also provides new opportunities for therapeutic strategies aimed at modifying the gut microbiota to treat diseases.

In this review, we explore the discovery and identification of MHIs, offering an overview of the methodologies employed to uncover these intriguing enzymes. We analyze the diverse characteristics and functions of MHIs, emphasizing their potential to modulate host physiological processes. Additionally, we summarize the current understanding of the roles that MHIs play in gut microbiota-based treatments, discussing their potential as therapeutic targets for a variety of diseases (Figure 1). This review aims to provide a comprehensive understanding of MHIs and their implications for future research and therapeutic interventions, ultimately contributing to improved host health outcomes through the manipulation of gut microbiota.

Figure 1.

Overview of the discovery, identification, classification, distribution, and functions of microbial host isozymes (MHIs)

The discovery and identification of MHIs

The concept of MHIs was recently introduced in a previous study by Jiang et al.’s team, who discovered that microbial-derived protein can perform functions similar to those of host antidiabetic protein.⁴ Unlike traditional microbiome and associated metabolome analysis strategies, the authors established an innovative activity-based protein screening platform. This platform combined a stable human gut microbiota culture system with enzyme activity testing assays of 110 host enzymes, selected from US Food and Drug Administration (FDA)-approved drug targets and human key drug-target database SuperTarget.⁵ These 110 host enzymes were selected based on their established roles in human disease and therapy, thereby supporting the identification of MHIs with potential pathophysiological relevance. Through the functional protein screening framework, 71 enzymes were identified as MHIs—microbial enzymes that can replicate the functions of host enzymes.

The function-based screening strategy provides a promising new approach for identifying MHIs. In the study by Jiang et al., enzyme activities were assessed by extracting proteins from human stool-derived ex vivo microbial communities, followed by measuring absorbance or fluorescence intensity, depending on the specific enzyme catalytic reaction types and established detection methods. Among the 71 MHIs, enzyme activities were further compared between feces from germ-free (GF) and conventionally raised (CONV-R) mice. The results revealed that 56 of these MHIs exhibited higher activity in the feces from CONV-R mice compared to those from GF mice (Figure 2A).

Figure 2.

Microbial-host-isozyme (MHI) screening and verification system

(A) The flowchart of the MHIs screening and verification.

(B) The Venn diagram illustrates the MHI number screened at different stages. CONV-R mice: conventionally raised mice.

The screening and validation process in the study by Jiang et al. was designed with rigorous controls to ensure specificity and reproducibility. Enzyme activity was measured using assays specific to each enzyme’s catalytic function. For example, oxidoreductase activity was assayed by monitoring NADH absorbance at 340 nm, while hydrolase activity was measured via the release of p-nitroaniline (pNA) at 405 nm. To confirm that the observed enzymatic activity originated from the gut microbiota, GF mice were used as an in vivo control for CONV-R mice. To effectively control for background signals from host cells or diet, unfermented brain-heart infusion (BHI) medium extracting solution was used as an in vitro blank control for protein extracts from mouse fecal samples. Furthermore, a statistical threshold (Z-factor >0.5) was applied to define positive activity, ensuring a high-quality screen.

Based on previous foundational discoveries, we hereby propose a functional definition for MHIs.^4,6 An enzyme should be classified as an MHI if it fulfills two core criteria: (1) functional equivalence, demonstrated by its ability to catalyze the same biochemical reaction as a specific host enzyme, thereby implying a direct functional counterpart in the host; (2) pathophysiological relevance, evidenced by its impact on host physiology. Table 1 summarizes the detailed criteria for MHIs. This activity-based screening strategy serves as an example of establishing functional equivalence, as it directly measures the catalytic activity of gut microbial enzymes toward substrates specific to host enzymes.⁴ Moreover, this functional definition prioritizes catalytic mimicry over sequence or structural homology.⁷ MHIs are distinct from traditional gut microbial enzymes that produce host-active metabolites (e.g., short-chain fatty acids and vitamins) but do not mimic the function of a specific host enzyme.⁶ The unique consequence of MHI’s activity is that it can directly compete with, compensate for, or interfere with the native host enzyme’s role in a physiological pathway, thereby introducing a microbial regulation over host processes.^4,6,7

Table 1.

Definition criteria for microbial-host isozymes (MHIs)

Criteria Category	Specific Requirements	Validation Methods	Status of Criterion
Sequence Similarity	•amino acid sequence homology in catalytic domains •conservation of active site residues	•BLASTp alignment •multiple sequence alignment •phylogenetic analysis	sufficient but not necessary
Structural Homology	•similar tertiary structure •conserved active site architecture •comparable cofactor binding sites	•homology modeling •X-ray crystallography (when available) •molecular dynamics simulations	sufficient but not necessary
Functional Equivalence	•catalyzes an identical biochemical reaction •similar substrate specificity •comparable kinetic parameters	•enzymatic activity assays •substrate specificity profiling •kinetic parameter determination (Km, kcat)	necessary but not sufficient
Physiological Relevance	•expressed under physiological conditions •contributes to host-microbe metabolic interactions •impacts host phenotype when modulated	•metatranscriptomics •metaproteomics •gnotobiotic animal models •human cohort studies	necessary but not sufficient
Distinction from Analogous Enzymes	•not merely functionally analogous but evolutionarily or structurally related •distinguish from convergent evolution cases	•deep evolutionary analysis •structural comparison with non-homologous enzymes	necessary but not sufficient

With ongoing advancements in host intestinal flora acquisition and diversity retention, improvements in enzyme activity detection technologies, and the continued expansion of key protein databases, it is anticipated that more MHIs will be discovered in the future. These advancements will help to deepen our understanding of the complex interactions between the gut microbiota and the host, offering new insights into the functional roles of these microbial-host enzymes.

The classification and distribution of MHIs

Based on Jiang et al.’s work, a total of 56 enzymes have been finally identified as MHIs for further investigation through in vitro and in vivo validation in a mouse model. We reanalyzed these MHIs utilizing the enzyme, Kyoto Encyclopedia of Genes and Genomes (KEGG), STRING databases, and TissueEnrich application.^8,9,10,11 As some enzymes have different isoforms or homologs in the same or different species, an enzyme with a specific EC number may correspond to multiple KEGG Orthology (KO) numbers due to slight sequence variations or differences in regulatory mechanisms.⁸ We searched the KO numbers of these MHIs and their corresponding proteins in Homo sapiens and found that the 56 enzymes were derived from 97 proteins in different human organs and tissues (Figure 2B; Table S1).

To identify the distribution of the MHI proteins, we performed the TISSUE and COMPARTMENTS enrichment analyses. The TISSUE enrichment indicated that these MHI proteins were highly expressed in the nervous system, liver, and digestive gland (Figure 3A). At the subcellular level, these proteins were mostly localized within secretory granules, lysosomes, and vesicles, suggesting their involvement in transport between the cytoplasm and extracellular space (Figure 3B). These findings implied that MHI proteins may function as signaling molecules involved in both neural and humoral regulation, which is consistent with previous reports on the role of gut microbiota in influencing host health and diseases.^12,13

Figure 3.

Distribution enrichment analyses and category of microbial-host-isozyme (MHI) proteins

(A) Enrichment analysis of tissue expression of MHI proteins.

(B) Enrichment analysis of subcellular localization of MHI proteins.

(C and D) The enzyme type perception of the MHI proteins.

To better understand the features of MHI proteins in a structured way, we classified them by enzyme function. The major types identified were hydrolases and oxidoreductases, which together comprised 39 different enzymes and accounted for more than three-quarters of the total MHI proteins (Figures 3C and 3D; Table 2). Hydrolases are enzymes that catalyze the hydrolysis of macromolecular substances, thereby facilitating their absorption and utilization by the host.¹⁴ For MHIs, the hydrolysis profiles are closely linked to the dietary habits and tissue microenvironments of the host, and the ecological environments to which distinct microbial species have adapted. For example, Lactobacillus and Bifidobacterium have been shown to enhance the activity of arylesterase, one of the identified MHIs, which in turn reduces lipid absorption and increases high-density lipoprotein levels, exhibiting cardio- and neuro-protective properties.^15,16 Oxidoreductases represent another significant group of enzymes involved in microbial-host interaction. These enzymes could participate in host metabolism and immune regulation by modifying various processes such as short-chain fatty acids synthesis, bile acids conversion, intestinal inflammation, drug metabolism, and so on.^17,18,19 Notably, sucrase isomaltase and dipeptidylpeptidase 4 (DPP4) were two important oxidoreductases involved in the oxidative utilization of polysaccharides and glucose,^20,21 consistent with the observed enrichment of the starch utilization system complex in subcellular location analysis. This structured classification underscores the functional importance of these enzymes in facilitating host-microbe interactions, driven by both metabolic and environmental factors. Other enzyme types include transferase and lyase. The primary function of transferases is to facilitate the transfer of functional groups, such as methyl (-CH₃), amino (-NH₂), and glycoside (-O-glycosyl) groups. Hydrolases are enzymes that catalyze hydrolysis reactions, converting complex biomolecules (e.g., proteins, fats, and nucleic acids) into simpler forms that can be utilized by cells. These enzymes play crucial roles in biological processes, signal transduction, and cellular regulation.

Table 2.

The enzyme type of the host key enzymes

Enzyme Type	Database (Total = 110)	Ex vivo screened MHIs (Total = 71)	In vivo verified MHIs (total = 56)
Hydrolase	40	29	23
Oxidoreductase	38	24	16
Transferase	18	13	13
Lyases	6	4	4
Ligases	5	1	0
Isomerases	3	0	0

MHIs, Microbial-host-isozymes.

The functions of MHIs

To elucidate the working pattern of the MHIs, we performed further functional analysis by searching the Gene Ontology (GO), KEGG and Human Phenotype Ontology (Monarch) databases.^22,23,24 Notably, both the GO and KEGG analyses suggested that these MHI proteins were highly enriched in multiple metabolic pathways (Figures 4A–4C). Genetic variations in these proteins were primarily associated with patient phenotypes with metabolic/homeostatic and digestive abnormalities. (Figure 4D). Collectively, specific bacteria could produce distinct proteins, most of which were metabolic enzymes, mimic the function of host’s own proteins, regulate specific pathways, and induce abnormal phenotypes. Table 3 summarizes 11 key MHIs.

Figure 4.

Function enrichment analyses of microbial-host-isozyme (MHI) proteins

(A and B). Gene Ontology (GO) functional annotation of MHI proteins.

(C) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the MHI proteins.

(D) Human phenotype (Monarch) enrichment analysis of the MHI proteins.

Table 3.

Comprehensive MHIs analyses

Enzyme Name	Microbial Source(s)	Host Counterpart	Key Functions	Associated Diseases	Therapeutic Potential	Potential Risks	Research Status	References (PMID)
Dipeptidyl peptidase 4 (DPP4)	Bacteroides thetaiotaomicron, B. fragilis, B. ovatus, P. copri	Human DPP4 (CD26)	•Cleaves N-terminal dipeptides from GLP-1, GIP •Regulates incretin hormone activity •Modulates glucose homeostasis	Type 2 diabetes, metabolic syndrome, cardiovascular disease, Crohn’s disease	•Dau-d4 (selective mDPP4 inhibitor) improves glucose tolerance •Explains variable response to sitagliptin •DPP4 inhibitor enhances intestinal fibrosis in Crohn’s disease	Risk of off-target effects on beneficial bacteria, ecological disturbance in gut microbiota	In vivo validated (mouse models, human fecal samples)	37815552; 41334589; 41043862
Alcohol dehydrogenase	Escherichia coli, Lactobacillus spp., Enterococcus faecalis	Human ADH1/7, ADH4, ADH6	•Ethanol oxidation to acetaldehyde •Metabolism of aliphatic alcohols •Redox balance maintenance	Alcohol use disorder, alcoholic liver disease, microbiome dysbiosis	•Potential target for reducing acetaldehyde production in gut •Could complement host-targeted ADH inhibitors	Competition with host enzyme, potential acetaldehyde accumulation	In vivo/animal model	41326333; 41272311; 41154505
GTP cyclohydrolase 1	Bacteroides spp., Prevotella spp., Bifidobacterium spp.	Human GCH1	•Catalyzes first step in tetrahydrobiopterin (BH4) synthesis •BH4 is essential cofactor for NO synthesis •Modulates vascular tone and immune function	Endothelial dysfunction, hypertension, autoimmune disorders	•Could affect NO bioavailability and vascular health •Potential target for cardiovascular protection	Potential interference with host folate metabolism	Ex vivo screened (activity detected); preclinical	28079055
Thioredoxin-dependent peroxiredoxin	Bacteroides spp., Faecalibacterium prausnitzii, Roseburia intestinalis	Human peroxiredoxin 1/2/4/5	•Reduction of hydrogen peroxide and organic hydroperoxides •Protection against oxidative stress •Redox signaling modulation	Atherosclerosis, hypertension, heart failure	•Potential probiotic strains engineered to overexpress protective peroxiredoxins •Biomarker for cardiovascular risk assessment	Potential interference with host redox signaling; Strain-dependent expression levels; Complex regulation in polymicrobial communities	Ex vivo screened (activity detected in anaerobic culture)	39428740
UDP-glucose 6-dehydrogenase	Bifidobacterium spp., Roseburia spp., Akkermansia muciniphila	Human UGDH	•Converts UDP-glucose to UDP-glucuronic acid •Provides precursors for glycosaminoglycan synthesis •Involved in detoxification pathways	Atherosclerosis, vascular calcification	•Potential target for modulating vascular extracellular matrix •Could influence host glucuronidation capacity	Potential interference with host hyaluronan synthesis	Ex vivo screened (activity detected)	40005761
α-Glucosidase	Bacteroides spp., Prevotella spp., Ruminococcus spp.	Human MGAM, SI	•Hydrolyzes α-1,4 glycosidic bonds in oligosaccharides •Releases glucose from complex carbohydrates •Affects postprandial glucose levels	Type 2 diabetes, obesity, metabolic syndrome	•Microbial α-glucosidase inhibitors may complement acarbose therapy •Explains variable response to acarbose treatment		In vivo validated (enzyme activity resistant to acarbose)	41278090; 41151127
Cystathionine β-synthase (CBS)	Fusobacterium nucleatum, Clostridium spp., Bilophila wadsworthia	Human CBS	•Produces hydrogen sulfide (H2S) from cysteine •Modulates vascular tone and inflammation •Affects mitochondrial function	Homocystinuria, cardiovascular disease, CKD	•Microbial CBS inhibitors may reduce pathogenic H2S production •Potential target for cardiovascular and kidney protection	Potential off-target effects on sulfur amino acid metabolism	animal/ex vivo screened	39777464; 35856606
Inosine-5′-monophosphate dehydrogenase (IMPDH)	Clostridium spp., Enterococcus spp., Streptococcus spp.	Human IMPDH1/2	•Rate-limiting enzyme in de novo guanine nucleotide synthesis •Essential for rapidly proliferating cells (e.g., T cells) •Affects immune cell function	Autoimmune disorders, transplant rejection, cancer	•Microbial IMPDH may affect host response to mycophenolate drugs •Potential target for selective immunomodulation	Potential impact on rapidly dividing immune cells, Competition with host IMPDH inhibitors	Ex vivo screened (activity confirmed)/animal	41110544; 39841148
Purine-nucleoside phosphorylase (PNP)	Bacteroides spp., Lactobacillus spp., Bifidobacterium spp.	Human PNP	•Catalyzes phosphorolysis of inosine/guanosine •Part of purine salvage pathway •Affects T cell development and function	Immunodeficiency disorders, autoimmune diseases	•Microbial PNP may affect host response to PNP inhibitors •Potential target for immunomodulation	Risk of nucleotide pool imbalance, Potential impact on host DNA/RNA synthesis	Ex vivo screened (activity detected)	38234794; 33878133
γ-Glutamyl transpeptidase (GGT)	Helicobacter pylori, Brucella spp., Bacillus spp.	Human GGT1/5/6/7	•Transfers γ-glutamyl group from glutathione to amino acids •Modulates glutathione metabolism and redox balance •Affects xenobiotic detoxification	Gastric cancer, liver disease, oxidative stress disorders	•H. pylori GGT inhibitors as potential therapeutics •Biomarker for gastric pathogen virulence	–	In vivo validated (H. pylori models)	40996235; 40205659
Tyrosine decarboxylase	Enterococcus faecalis, Lactobacillus spp., Streptococcus spp.	Human AADC (broad specificity)	•Converts tyrosine to tyramine •Tyramine affects gut motility, vascular tone, CNS function •Modulates catecholamine signaling	Migraines, hypertension, IBS, neuropsychiatric disorders	•Low-tyramine diets for migraine prevention •Microbial tyrosine decarboxylase inhibitors as potential therapeutics	–	Ex vivo screened (activity confirmed)	4,550,085; 31196984; 24928805

MHIs and glucose metabolism

Human DPP4 (hDPP4) is normally secreted by intestinal enteroendocrine cells. It could serve as an antidiabetic target because of its capacity to degrade glucagon-like peptide-1 (GLP-1) and regulate blood glucose homeostasis.²⁵ The study by Jiang et al. found that microbial DPP4 (mDPP4), mainly secreted by Bacteroides spp,²⁶ could also reduce active GLP-1 when the intestinal barrier is compromised and subsequently affect the host’s glucose level.⁴ The key mechanism is that a compromised intestinal barrier (“leaky gut”) allows mDPP4 to translocate from the gut lumen into the intestinal tissue.⁴ Once there, mDPP4 degrades locally secreted GLP-1, thereby impairing its ability to stimulate insulin secretion and control postprandial blood glucose.⁴ Furthermore, they also found that sitagliptin, a clinically used DPP4 inhibitor for treating type 2 diabetes (T2D), exhibited compromised efficacy against mDPP4, as compared to hDPP4.⁴

This difference is attributed to the non-conserved active site architecture of mDPP4, as revealed by co-crystal structures (detailed in section the potential applications of MHIs in disease treatment).^4,7 This may account for the low response to sitagliptin observed in some patients with T2D.⁴ Clinically, low responder patients exhibited higher fecal DPP4 activity (primarily of microbial origin) and an enriched abundance of Bacteroides spp (which produce mDPP4), providing a direct link between this MHI and heterogeneous drug response.⁴

To address this limitation, high-throughput screening (HTS) identified daurisoline-d4 (Dau-d4) as a selective mDPP4 inhibitor.^4,27 Dau-d4 was shown to improve glucose metabolism in diabetic mice, and its co-administration with sitagliptin yielded significant benefits, underscoring the therapeutic potential of targeting both host and microbial enzymes.^4,27 This represents a dual-targeting strategy to block the DPP4-GLP-1 axis in both host and microbe.^4,27

Alpha (α)-glucosidase is a type of hydrolase that digests dietary carbohydrates. Microbial α-glucosidase and human α-glucosidase (specifically, human maltose-glucoamylase and sucrase-isomaltase) share homology in both sequence and structure.^28,29 Specifically, it has been reported that genes encoding α-glucosidase from human gut tissue and Ruminococcus obeum share a highly conserved catalytic domain.³⁰ In vitro maltose hydrolysis assays and X-ray diffraction and structure determination analyses have demonstrated that bacterial α-glucosidase from Ruminococcus obeum may interact with anti-diabetic drugs (α-glucosidase inhibitors: voglibose, miglitol, and acarbose).³¹ These findings suggest that this potential unintended in vivo cross-reaction, mediated by a gut bacterial enzyme, may trap a portion of the oral drug before it reaches the host’s small intestine, thereby reducing the bioavailability of α-glucosidase inhibitors.^27,31 This may compromise the therapeutic efficacy of these α-glucosidase inhibitors in patients with T2D and contribute to individual variation in drug response.^27,31 Furthermore, due to the inhibition of bacterial enzymes, the drug may influence the ecological balance of the host gut microbiota, which could be linked to the drug-associated gastrointestinal side effects.^27,31

Beta (β)-galactosidase is a glycoside hydrolase enzyme that catalyzes the hydrolysis of terminal non-reducing β-D-galactose residues in β-D-galactosides.³² This reaction could hydrolyze the disaccharide lactose into glucose and galactose, hydrolyze allolactose into monosaccharides, and convert lactose into allolactose. A recent study revealed that S. thermophilus could produce β-galactosidase to inhibit colorectal tumorigenesis in mice.³³ Furthermore, β-galactosidase secreted by L. vaginalis also showed a protective effect on acetaminophen-induced hepatotoxicity through glucose metabolism regulation and ferroptosis prevention.³⁴

MHIs and L-DOPA metabolism

Specific gut microbiotas also produce enzymes that can promote the synthesis of neurotransmitters or their precursors, which are usually catalyzed by enzymes in the host’s brain.³⁵ The conversion of L-DOPA to dopamine through aromatic amino acid decarboxylase (AADC) in the brain is the core process of Parkinson’s disease treatment.³⁶ Another enzyme, tyrosine decarboxylase (TyrDC), has a similar function and was found to exist in Enterococcus faecalis in the gut as well. However, TyrDC in the microbiota breaks down L-DOPA into dopamine and subsequently rapidly into m-tyramine, which could potentially reduce the availability of the medication (L-DOPA) for patients with Parkinson’s disease .³⁷ This constitutes a direct pharmacokinetic competition: Gut bacterial TyrDC peripherally metabolizes L-DOPA in the intestine, preventing the drug from crossing the blood-brain barrier to reach the brain.³⁷ This unintended process irreversibly attenuates L-DOPA’s therapeutic effect in Parkinson’s disease.³⁷ AADC inhibitors such as carbidopa are used to reduce peripheral L-DOPA metabolism, but up to 56% of the drug still fails to reach the brain, and patient responses remain highly variable.³⁸ This limited efficacy is likely due to carbidopa’s poor inhibition of gut bacterial TyrDC.

The L-tyrosine analog (S)-α-fluoromethyltyrosine (AFMT) was designed to selectively target the MHI TyrDC. In vitro, AFMT potently inhibited bacterial TyrDC (IC₅₀ = 4.7 μM) with minimal effect on the human enzyme AADC.³⁷ AFMT completely inhibited L-DOPA metabolism in gut microbiota samples from patients with Parkinson’s disease and neurologically healthy control volunteers, indicating its preliminary effectiveness in complex human microbial environments.³⁷ In gnotobiotic mice colonized with E. faecalis, co-administration of AFMT with L-DOPA and carbidopa significantly increased the peak serum concentration of L-DOPA, confirming its ability to block microbial metabolism in vivo and improve drug bioavailability.³⁷

MHIs and lipid metabolism

Gut microbiota can both synthesize and break down lipids with host modulatory properties. Among the MHI proteins, there were 30 proteins enriched in the lipid metabolic process (Figure 4A). Research on microbial enzymes for the regulation of host lipid homeostasis has been performed, and several enzymes were identified. For instance, conjugated linoleic acid (CLA) synthesis enzymes secreted by Bifidobacterium spp., Lactobacillus spp. and Streptococcus spp. were involved in PUFA synthesis^39,40; IsmA secreted by Eubacterium spp. participated in cholesterol metabolism.⁴¹ IsmA catalyzes the conversion of cholesterol to coprostanol, a non-absorbable sterol, thereby reducing intestinal cholesterol absorption and lowering host serum cholesterol levels.^6,41 Phospholipase A2, another screened MHI, plays a crucial role in phospholipid metabolism and immune regulation. Research has demonstrated that Akkermansia muciniphila phospholipid could active Toll-like receptor 2 and induce the release of inflammatory factors.⁴² However, most isoenzymes still lack direct evidence demonstrating their impact on host function when derived from microorganisms. The existing studies primarily offer indirect clues that may be relevant.

MHIs and purine metabolism

Gut microbiota participated in purine synthesis and degradation. It was reported that multiple phyla of gut bacterial taxa, including Bacillota, Fusobacteriota, and Pseudomonadota, could anaerobically utilize purines.²³ The screening system also identified purine-nucleoside phosphorylase as an MHI.⁴ Future studies are needed to establish the direct relationship between MHI in specific microbes and host purine metabolism.

MHIs and amino acid metabolism

Tryptophan is the precursor of serotonin. A series of studies has revealed that gut microbes influence the host’s tryptophan metabolism and serotonin synthesis. Specific microbes such as Lactococcus, Lactobacillus, Streptococcus, Escherichia coli, and Klebsiella are able to express tryptophan synthetase for serotonin production.^43,44,45 Furthermore, microbial colonization/transplantation could help normalize tryptophan levels and improve depressive behaviors in mice.⁴⁶ Memapsin-2, also known as beta-site amyloid precursor protein (APP) cleaving enzyme 1 (BACE1), beta-secretase, or Asp2, is a type of hydrolase enzyme encoded by the BACE1 gene in humans.⁴⁷ It is primarily expressed in neurons and oligodendrocytes.⁴⁸ Memapsin-2 is a membrane-anchored aspartic protease that catalyzes the initial cleavage of beta-APP, leading to the formation of amyloid β (Aβ) in the brain—a key factor in the pathogenesis of Alzheimer’s disease (AD).⁴⁹ As a result, β-secretase has emerged as an important therapeutic target for the development of inhibitory drugs aimed at treating AD.⁵⁰

Arginine decarboxylase (ADC) catalyzes arginine decarboxylation to agmatine and is widely distributed in human tissues and diverse organisms, including bacteria. Recent studies have shown that Bacteroides vulgatus, a common gut bacterium, uses its ADC (encoded by speA) to convert intestinal arginine into agmatine.⁵¹ An in vivo study demonstrated that agmatine treatment in apo ^E−/− mice, a model of diet-induced fatty liver and atherosclerosis, alleviated both atherosclerotic lesions and hepatic steatosis by suppressing the hepatic de novo lipogenesis pathway.⁵² Together, these findings suggest that gut bacteria harboring the MHI ADC may, through the production of agmatine, provide the host with an endogenous form of protection, aiding in the regulation of hepatic lipid metabolism and potentially playing a beneficial role in counteracting metabolic dysfunction-associated steatohepatitis (MASH).^6,52

MHIs and heme metabolism

Biliverdin reductase (BVR) is an oxidoreductase enzyme in the heme metabolism pathway, responsible for reducing biliverdin, a product of heme degradation, to bilirubin, which has strong antioxidative properties.⁵³ BVR is co-expressed with heme oxygenases (HO-1, HO-2) in various tissues, such as the stomach and liver, and regulates HO activity to maintain the balance of heme metabolism.⁵⁴ Although BVR is primarily cytoplasmic, it can localize to subcellular compartments such as the endoplasmic reticulum, mitochondria, and cell membrane under pro-oxidative conditions. Its membrane localization is associated with tyrosine kinase activity, enabling the regulation of intracellular signaling pathways through the activation of downstream signaling molecules.⁵⁵ BVR activity is dependent on cofactors NADH/NADPH, the pH environment, and Ser149 autophosphorylation.⁵⁶ Additionally, BVR is involved in insulin signaling, protein kinase cascades, and anti-inflammatory pathways, suggesting its potential role in the pathophysiology of diabetes and obesity.

MHIs and oxidative stress

Superoxide dismutase (SOD) is ubiquitous in living organisms, and plays a crucial role in catalyzing the conversion of superoxide radicals (O²⁻) into oxygen (O₂) and hydrogen peroxide (H₂O₂).⁵⁷ This enzyme is essential for maintaining intracellular reactive O₂ species (ROS) balance and protecting cells from oxidative damage. Najmuldeen et al. discovered 3 SOD isoenzymes—Mn-, Fe-, and CuZn-auxiliary SOD—in Klebsiella pneumoniae.⁵⁸ Beyond scavenging O²⁻, microbial SOD also plays significant roles in metabolism, bacterial virulence, and host colonization.⁵⁸

Nicotinamide mononucleotide adenylyltransferase (NMNAT) is a transferase enzyme of the nucleotidyltransferase alpha/beta phosphodiesterase superfamily that catalyzes the synthesis of NAD from NMN and ATP, representing the final step in NAD biosynthesis.⁵⁹ NAD serves as an essential cofactor in cellular redox reactions.⁵⁹ Mammals express three isoforms of NMNAT, each of which supplies NAD to specific NAD-dependent enzymes and influences various biological processes, such as DNA repair, protein homeostasis, cell differentiation, and neuronal function maintenance.⁶⁰ NMNAT1, the nuclear isoform, is the most abundant and efficient, with a 4-fold preference for NMN over NAMN and is widely expressed.⁶⁰ NMNAT2, which is associated with the Golgi apparatus, exhibits lower efficiency and is found in tissues such as the brain, heart, and pancreas.⁶⁰ NMNAT3, localized in the cytoplasm, mitochondria, and lysosomes, is less abundant in most tissues compared to NMNAT1, but it is the predominant isoform in erythrocytes.⁶⁰

The potential applications of MHIs in disease treatment

MHIs represent a promising frontier in biomedical research and application, particularly in the areas of target discovery, drug screening, and therapeutic efficacy prediction. Besides, the properties of MHIs also provide opportunities to design isoform-selective inhibitors or activators. HTS platforms, coupled with structure-based drug design and computational modeling, can be used to screen large libraries of compounds for their ability to specifically bind and modulate the activity of the target isoenzyme. Furthermore, machine learning algorithms can be employed to analyze MHI datasets generated from preclinical studies, enabling the prediction of therapeutic efficacy and patient stratification based on individual genetic and phenotypic characteristics.

The distinct sequence and structural features of microbial enzymes, despite their functional similarity to host counterparts, present a rich source for identifying novel therapeutic targets. By targeting these microbial enzymes, known as the “drug the bug” strategy, we can modulate host physiology with high specificity.⁶¹ Prior HTS against purified MHI targets, coupled with structure-based drug design, is a powerful methodology to identify such selective modulators.⁴ For example, crystal structures have revealed that the trifluoromethyl triazolopyrazinyl ()-stabilizing residues (F357, S209, and R358) in human DPP4 (hDPP4) are not conserved in microbial DPP4 (mDPP4).⁴ Furthermore, a protruding residue (E342) in mDPP4 prevents the inhibitor sitagliptin from binding effectively.⁴ This difference in active site architecture between hDPP4 and mDPP4 provides a molecular basis for the selective targeting of microbial enzymes over host ones.^4,27 Dau-d4, a small-molecule selective inhibitor of mDPP4 identified by HTS and structure-based drug design, can improve glucose metabolism in diabetic mice without significantly affecting the host enzyme.^4,27 The discovery of Dau-d4 serves as a definitive proof-of-concept for this approach.^4,27

Beyond traditional small-molecule inhibitors, the use of engineered probiotics designed to lack detrimental MHIs or to express beneficial ones is another promising approach to target MHIs.²⁷ This strategy is exemplified by the work of Jiang et al., who have constructed a Dpp4-deficient B. thetaiotaomicron to explore the effect of btDPP4 in B. thetaiotaomicron on glucose metabolism.^4,27 In contrast to the wild-type strain, this engineered bacterium does not impair host glucose levels, demonstrating the potential of using engineered bacteria to selectively remove detrimental MHI activities.^4,27

There are several technical challenges in targeting MHIs. Firstly, to avoid protein inactivation during the oral route, engineered delivery systems (e.g., L100-55 beads) and gut-restricted compounds with low systemic bioavailability were developed to effectively deliver MHIs modulators to the gut.⁴ Another challenge in targeting MHIs is the risk of potential off-target effects, as a single MHI is often produced by a range of bacteria, including beneficial ones. This means that an inhibitor designed to hit a detrimental MHI in one species may inadvertently affect beneficial commensals within the same genus that share the same enzyme, leading to unforeseen ecological consequences.^27,61 Several strategies can be used to mitigate the challenge of off-target effects. These include using highly specific small-molecule inhibitors or bacteriophages to precisely target detrimental bacterial species without broadly impacting the commensal community.^4,27 Furthermore, compounds with low bioavailability (such as Dau-d4), which have low systemic drug exposure, tend to concentrate primarily in the gut, thereby minimizing the impact on the host’s own enzymes.⁴ Lastly, by screening patient fecal samples to identify individuals with high-activity, harmful target MHIs, we can ensure that only these patients receive treatment, thereby guaranteeing the effectiveness and safety of the therapy.²⁷

Targeting MHIs to optimize diabetes treatment is emerging, moving from concept to preclinical validation. As previously described, Dau-d4 was identified as a highly selective inhibitor of mDPP4, without affecting the host enzyme (hDPP4).⁴ In diabetic mouse models, Dau-d4 alone improved glucose tolerance, and its co-administration with sitagliptin, an hDPP4 inhibitor, yielded additive benefits.^4,27 This demonstrates a dual-targeting strategy that concurrently regulates both the host and microbial arms of the DPP4-GLP-1 axis.^4,27 The efficacy of anti-diabetic drugs, α-glucosidase inhibitors, may be diminished due to the impact of gut bacterial α-glucosidase before reaching the small intestine.^31,62 This phenomenon represents a previously overlooked and microbial-mediated mechanism for reduced drug bioavailability and variable patient response.^31,62 Therefore, next-generation α-glucosidase inhibitors should be designed to enhance selectivity toward human enzymes over microbial ones, thereby maximizing therapeutic efficacy.

Future perspectives

While several studies have explored the impact of MHIs on the host, the key unanswered question of how the host regulates MHI production in gut bacteria is still largely unknown.⁷ We hypothesize that the expression of MHIs is dynamically regulated by host physiology and environment. A diet rich in fiber, including whole grains, traditional Chinese medicinal foods, and prebiotics, has been shown to significantly alleviate T2D.⁶³ This beneficial effect is linked to a decline in Bacteroides species that produce mDPP4, an MHI with an adverse impact on glucose homeostasis.^61,64 This evidence suggests that host dietary habits may regulate MHI activity by reshaping the composition of the gut microbial community.

The current activity-based screen platform in the study by Jiang et al. is powerful but limited to: 1) the use of non-native substrates (fluorogenic or other mimics), 2) the inability to trace activities back to their encoding genes, 3) a focus on fecal microbiota.⁴ Using metagenomics, we can expand the MHI library and get a more comprehensive MHI repertoire.²⁷ There are two possible technological strategies: (1) Single-cell metagenomics combined with functional screens. Single-cell sequencing of human fecal samples can obtain vast genetic information from uncultured microbes. The predicted MHI genes can be synthesized and cloned into expression vectors to produce recombinant proteins, which are then subjected to functional screening. (2) Artificial intelligence and machine learning can be used to predict MHI function from genetic sequences and to identify candidate MHIs in large-scale metagenomic databases, even in the absence of significant sequence homology to known host enzymes.²⁷

To better understand the role of MHI in human diseases, future research needs to integrate more techniques from different fields, such as clustered regularly interspaced short palindromic repeats (CRISPR)-based screens. Specifically, CRISPR interference (CRISPRi) can be applied to screen bacterial genes on a large-scale.⁶⁵ By using guide RNA libraries to knock down bacterial genes (including putative MHIs) using a CRISPRi-based platform, and then colonizing GF mice, we can identify which genes, when suppressed, affect host phenotypes (e.g., glucose levels).^66,67 This method provides direct and high-throughput functional evidence for the role of specific MHIs in vivo.

In Jiang et al.’s study, a broad-spectrum antibiotic cocktail was orally administered to mice for one day to deplete the gut microbiota, suggesting the powerful impact of the antibiotics on the gut microbiota and enzyme activity.⁴ The use of antibiotics is a rather strong intervention method, which is only used for specific pathological conditions (e.g., pathogen infection or inflammatory bowel disease), and their misuse risks inducing antibiotic resistance.²⁷ Consequently, antibiotic perturbation is a suboptimal strategy for MHI-targeted therapy due to its broad ecological impact and potential for unpredictable consequences. Therefore, the application of small-molecular MHI inhibitors has been proposed as an alternative and effective strategy to regulate gut microbiota, which can specifically modulate microbial function without changing global compositions.²⁷

The study of MHIs offers a novel perspective for understanding the interaction between intestinal microbiota and the host, while providing opportunities for the development of new therapeutic strategies. Future research should integrate multidisciplinary methods and technologies, expand the library of MHIs, accumulate more direct evidence, and explore the applications of MHIs in disease treatment.

Acknowledgments

We thank Ruixuan Wang and Yu Zhang for their help with figure design. This work was supported by the National Natural Science Foundation of China (no. 82202071), the Health Human Resources Development Center, the National Health Commission (RCLX2315027), and the Natural Science Foundation of Sichuan Province (2024NSFSC1608).

Author contributions

X.L., S.Z., and P.Z. contributed to the concept and design of this review. X.L., C.W., S.G., Y.H., Y.X., C.C., S.Z., and P.Z. performed the literature search and selected the papers for inclusion. X.L., S.G., Y.H., and S.Z. created the table and figures. X.L., C.W., S.G., Y.H., Y.X., C.C., S.Z., and P.Z. wrote and edited the manuscript. All authors read and approved the final version of the manuscript.

Declaration of interests

The authors declare no competing interests.