Introduction

Limited fossil fuel resources and their environmental concerns have diverted human interest toward producing sustainable and renewable fuels (Bumrungtham et al. 2022). Today, ethanol is the most commonly produced biofuel worldwide, obtained from cane sugar or corn starch via fermentation technology (Aziz et al. 2023; Coimbra et al. 2023). Despite being a renewable resource, ethanol is not considered an advanced biofuel because of its subpar fuel properties. These include a low energy density, a high vapor pressure, and an incompatibility with the existing fuel infrastructure, the latter of which can be attributed in large part to ethanol’s high hygroscopicity and corrosiveness (Ni et al. 2023). This situation provides an opportunity to find alternative fuels. As a result, there has been a lot of interest in producing advanced biofuels that would possess properties similar to those fuels generated from petroleum, such as higher alcohols (Choi et al. 2014; Gupta et al. 2022). Isobutanol and 1-butanol are the two examples of these higher alcohols, which have high energy densities and low vapor pressure than ethanol (Jawed et al. 2020; Zhang et al. 2019b). They can be mixed with gasoline in any ratio and can also be used as replacements for gasoline because they can perform better in conventional gasoline engines (Yousif and Saleh 2023). Isobutanol and its derivatives also have numerous applications in various chemical industries; for instance, these can be used as solvents, additives in paints, ink ingredients, and extractants for organic compounds (de Lima et al. 2023; Wang et al. 2012). In addition, isobutanol is comparatively less toxic and has superior energy density than its straight-chain counterparts (Siripong et al. 2018; Veettil et al. 2016). Thus, these properties make isobutanol more attractive than other butanol isomers.

Isobutanol can be produced naturally by branched-chain amino acid biosynthesis through five enzymatic steps (Kobayashi et al. 2022). In the isobutanol pathway, the two molecules of pyruvate are first converted into 2-acetolactate by the acetolactate synthase (AHAS). In the second reaction, 2-acetolactate is reduced to 2,3-dihydroxy isovalerate; this reaction is catalyzed by acetohydroxyacid reductoisomerase (AHAIR). In the third reaction, 2,3-dihydroxyisovalerate is converted into 2-ketoisovalerate catalyzed by dihydroxyacid dehydratase (DHAD). Finally, isobutanol can be synthesized via the Ehrlich pathway from 2-ketoisovalerate using two more reaction steps; the enzymatic steps are catalyzed by ketoacid decarboxylases (KDCs) and alcohol dehydrogenase (ADH), respectively (Fig. 1) (Felpeto-Santero et al. 2015; Hasegawa et al. 2020; Hazelwood et al. 2008; Novak et al. 2020). Only a few microorganisms can naturally produce isobutanol in a meager amount, such as Lactococcus lactis, Saccharomyces cerevisiae, Pichia pastoris, and Candida sp. (Derrick and Large 1993; Liu et al. 2021b; Smit et al. 2004). With rapidly evolving techniques and quick access to metabolic engineering tools and taking advantage of the increasing genomic information, considerable progress has been made to improve isobutanol production in native microbial hosts (Siripong et al. 2018; Wess et al. 2019). However, the titer of isobutanol produced by the native microbial host is still considerably below the levels needed for industrial purposes. For this reason, various non-native microbial hosts, such as Escherichia coli, Bacillus subtilis, G. thermoglucosidasius, Corynebacterium glutamicum, Clostridium cellulolyticum, Ralstonia eutropha, and Synechococcus elongatus, have been engineered for isobutanol production using a diverse range of feedstocks (Fig. 2) (Atsumi et al. 2009; Higashide et al. 2011; Huo et al. 2018; La et al. 2017; Li et al. 2011; Lin et al. 2014; Lu et al. 2012; Smith et al. 2010). Several approaches have been successfully employed to enhance isobutanol production in these microorganisms, such as deleting the competitive pathways, overexpressing the key enzymes of the isobutanol synthesis pathway, cofactor engineering, and improved robustness of microbial hosts (Acedos et al. 2021; Lu et al. 2012; Matsuda et al. 2013). All the native and non-native microbial-hosts used for isobutanol production are compiled in Table 1.

Fig. 1
figure 1

Pathway used for production of isobutanol in microbial hosts. Figure modified from Felpeto-Santero et al. (2015)

Fig. 2
figure 2

Schematic representation of the microbial production of isobutanol from various feedstocks

Table 1 Summary of the engineered microbial hosts used for isobutanol production

A wide range of reviews addressed biofuel production in microorganisms (Mainguet and Liao 2010; Meadows et al. 2018; Nawab et al. 2020; Xin et al. 2019); nevertheless, a comprehensive review enlisting all genetically engineered microbial hosts for the production of isobutanol is still not available. Therefore, a comprehensive review showing a full picture of engineered microbial hosts for isobutanol production is highly required. Here, we have summarized different engineered microbial hosts for isobutanol production and discussed different engineering strategies for enhancing the production of isobutanol. This review also explains the problems plaguing modern isobutanol manufacturing and suggests some approaches for addressing them. So, we conclude that this study will assist researchers in overcoming the problems associated with engineering microbial hosts for producing isobutanol and developing efficiently engineered strains to fulfil the world’s growing energy demands.

Engineering native microbial hosts for isobutanol production

Few microbial hosts are capable of producing isobutanol naturally, i.e., S. cerevisiae, Pichia pastoris, and Lactococcus lactis (Gambacorta et al. 2022; Priyadharshini et al. 2015; Siripong et al. 2018). However, these microorganisms produce a very minute amount of isobutanol. The production range of isobutanol varies from 0.01 to 0.44 g/L (Kurylenko et al. 2020; Priyadharshini et al. 2015). In order to increase isobutanol synthesis, scientists have designed a number of metabolic engineering strategies to modify the native-host organisms’ isobutanol biosynthetic pathways.

Engineering Saccharomyces cerevisiae for isobutanol production

The S. cerevisiae (baker’s yeast) is used to produce bioethanol and is considered a biofuel industry’s workhorse (Coyle Diane Publishing 2010). S. cerevisiae has been modified for enhanced production of several advanced biofuels, including isobutanol and the sesquiterpenoids farnesene, bisabolene, and amorphadiene (Peralta-Yahya et al. 2011; Westfall et al. 2012). Due to the nature of this review, we shall restrict our attention to the isobutanol manufacturing process. The modified pathways constructed for the production of isobutanol in S. cerevisiae are depicted in Fig. 3 (Brat et al. 2012; Lee et al. 2012; Matsuda et al. 2013).

Fig. 3
figure 3

Engineered pathways used for production of isobutanol in yeast cells. Figure modified from Lee et al. (2012) and Wess et al. (2019). Native genes are shown in black, and heterologous genes in red

As mentioned before, yeast naturally produces a minute quantity of isobutanol; therefore, to enhance isobutanol production in S. cerevisiae, the genes involved in valine metabolism (ILV2, ILV5, and ILV3) were overexpressed to direct the metabolic flux toward 2-ketoisovalerate. The modified strain produced 0.97 mg of isobutanol for every gram of glucose. Isobutanol production was further increased twofold by the additional overexpression of the BAT2 gene. Moreover, the combined over-expression of the ILV2, ILV5, ILV3, and BAT2 genes in S. cerevisiae resulted in a 13-fold increased isobutanol production than the parental strain (Chen et al. 2011). In another study, isobutanol production was improved 13-fold by deleting PDC1 and overexpression of ILV2, kivd, and ADH6. Deleting the PDC1 gene shifted the metabolic flux from ethanol towards isobutanol synthesis (Kondo et al. 2012). To improve the production of isobutanol in a genetically modified strain of S. cerevisiae, an additional improvement was achieved by employing multiple strategies, such as the deletion of the lpd1 gene, overexpression of ILV2, ILV5, ILV3, ADH6, kivd, and overexpression of some genes encoding transhydrogenase-like shunts such as PYC2, MDH2, and MAE1 to solve the cofactor deficiency. After several engineering rounds, the recombinant strain accumulated 1.6 g/L of isobutanol in a batch fermentation from 100 g/L of glucose after 24 h (Matsuda et al. 2013).

Moreover, in S. cerevisiae, the enzyme for the isobutanol pathway is localized into two cellular components: mitochondria and cytosol. Thus, shifting the entire Ehrlich pathway to one component could improve the titer of isobutanol. For example, Lee et al. (2012) overexpressed three mitochondrial enzymes (ILV2, ILV3, and ILV5) in the cytosol by eliminating mitochondrial targeting sequences and introducing the kivd gene from L. lactis. The engineered strain accumulated almost sevenfold (151 mg/L) more isobutanol than its parent strain without deleting three native mitochondrial genes (Lee et al. 2012). Brat et al. (2012) employed a similar strategy by relocating the three mitochondrial enzymes (ILV2, ILV3, and ILV5) into the cytosol by eliminating mitochondrial targeting sequences and expressing ARO10 and ADH2 genes. The engineered strain with a combination of codon-optimized truncated ILV gene expression and ILV2 gene deletion in mitochondria increased isobutanol titers to 630 mg/L. The titer of isobutanol was approximately 32-fold higher than the original strain (Brat et al. 2012).

Recently, in another approach to increasing the metabolic flux towards isobutanol production in engineered S. cerevisiae, non-essential competing metabolic pathways were deleted by the elimination of multiple (bdh1, bdh2, leu4, leu9, ecm31, ilv1, adh1, gpd1, gpd2, and ald6) genes (Wess et al. 2019). The engineered S. cerevisiae produced 200-fold more isobutanol (2.09 g/L) than its parent strain after the expression of codon-optimized three endogenous enzymes (IlV2, ILV3, and ILV5) in the cytosol and deletion of the ilv2 gene to disrupt mitochondrial valine biosynthesis. This is the highest reported titer of isobutanol in S. cerevisiae (Wess et al. 2019). In another report, Park and Hahn (2019) engineered the strain of S. cerevisiae by expressing alsS from B. subtilis utilizing a copper-inducible CUP1 promoter and overexpressing two endogenous mitochondrial genes, ILV5 and ILV3, in the cytosol. The recombinant strain of S. cerevisiae accumulated 263.2 mg/L of isobutanol, which was almost 3.3 times higher than the control strain expressing the three native genes ILV2, ILV5, and ILV3 on a plasmid utilizing a strong constitutive promoter (Park and Hahn 2019). These results demonstrated that blocking the competitive pathways, overexpressing the endogenous genes, and increasing cofactor availability effectively enhanced isobutanol production. However, the production of isobutanol by S. cerevisiae is still very low. Therefore, further engineering efforts are needed to identify novel enzymes for the isobutanol pathway with higher catalytic activities.

Engineering Pichia pastoris for isobutanol production

P. pastoris, a methylotrophic yeast, has been used as a production host for various high-value chemicals. The wild strain of P. pastoris accumulates a very low amount of isobutanol (Kharkwal et al. 2009; Nor and Roshanida 2015). Therefore, to improve isobutanol production in P. pastoris KM71, Siripong et al. (2018) overexpressed KDC and ADH genes from different candidates. Three candidates for KDC genes have been tested: ARO10 and THI3 from S. cerevisiae, kivd from L. lactis, and ADH, ADH6, and ADH7 from S. cerevisiae. Different combinations of KDC-ADH genes were employed to analyze isobutanol levels under the use of the additive 2KIV (4 g/L), the isobutanol precursor. The highest titer of 48 mg/L of isobutanol was accumulated with the expression of kivd (L. lactis) and ADH7 (S. cerevisiae) (Siripong et al. 2018). To improve the titer, endogenous L-valine biosynthetic pathway genes (Ilv6, Ilv2, Ilv5, and Ilv3) were overexpressed. As a result, the engineered strain (PP300) produced nearly 0.90 g/L of isobutanol from glucose directly. Additional improvement was achieved by the introduction of an additional copy of the ILV6 and ILV2 genes by placing them on the episomal plasmid; the engineered strain (PP403) accumulated a titer of 2.22 g/L of isobutanol, which was almost a 43-fold improvement over the original strain (Siripong et al. 2018).

In another study, Kurylenko et al. (2020) reported that the wild-type strain of multinuclear yeast M. magnusii produces 0.44 g/L of isobutanol (10 to 20-fold higher than S. cerevisiae). To further enhance the production of isobutanol in M. magnusii, the exogenous ILV2 gene of S. cerevisiae was overexpressed in M. magnusii under the control of the constitutive PTEF1 promoter. The recombinant strain accumulates 0.62 g/L of isobutanol, a 1.4-fold increase over the wild-type strain (Kurylenko et al. 2020). This result showed that the wild strain of M. magnusii performs better than the wild strains of S. cerevisiae and P. pastoris regarding isobutanol production. However, the genetic toolbox available for this species is not as sophisticated as that for S. cerevisiae. Therefore, implementing a better genetic toolbox would be a promising strategy to accelerate the process of pathway design in M. magnusii for isobutanol production.

Engineering Klebsiella pneumoniae for isobutanol production

K. pneumoniae is an indispensable industrial bacterium that can accumulate 2,3-butanediol naturally by utilizing glucose or other sugars as a substrate (Chen et al. 2015). Gu et al. (2017) demonstrated the presence of an isobutanol production pathway in wild K. pneumonia (Gu et al. 2017). However, this pathway did not lead to isobutanol production in the wild-type strain, while a very high amount of 2,3-butanediol (17.3 g/L) was detected in the wild-type strain (Gu et al. 2017). Thus, deletions of budA (encoded by α-acetolactate decarboxylase) and ldhA genes diverted the carbon flux from 2,3-butanediol and lactic acid to the valine pathway, respectively. As a result, the engineered strain accumulated 157 mg/L of isobutanol and a high amount of 2-ketoisovalerate (12.7 g/L), a precursor for isobutanol. Titers were further increased to 2.45 g/L in ipdC (encoding an indole-3-pyruvate decarboxylase) complementation strain after 24 h of fermentation, which is an almost 15.5-fold increase compared to K. pneumoniae ΔbudAldhA strain (Gu et al. 2017). It was found that the ipdC gene is responsible for making isobutyraldehyde from 2-ketoisovalerate in K. pneumoniae. However, a high amount of 2-ketoisovalerate was detected in the fermentation broth of K. pneumoniae, which showed that the main bottleneck in the isobutanol pathway is the decarboxylation of 2-ketoisovalerate. Another bottleneck for isobutanol production in K. pneumoniae is the catalytic activity of the ipdC gene towards pyruvate, which reduces the available precursor pyruvate for isobutanol production. Therefore, to further improve isobutanol production, the ipdC gene of K. pneumoniae was engineered through site-specific mutations to generate different variants of Kp-IpdC. Among the different variants of Kp-IpdC, one of the variants T290L accumulated 5.5 g/L of isobutanol, which was a 1.2-fold improvement over the control strain K. pneumoniae ΔbudA–ΔldhA–ΔipdCipdC (IpdC) (Shu et al. 2022). These results showed that the decarboxylation of 2-ketoisovalerate is the main bottleneck in the isobutanol pathway in K. pneumoniae. Future studies should focus on modifying the existing enzymes through protein engineering or identifying alternative enzymes to enhance isobutanol production.

Engineering Lactococcus lactis for isobutanol production

L. lactis is a lactic acid-producing bacterium found in dairy products, vegetables, leaves, and animal skin (Li et al. 2020; Maślak et al. 2022; Perveen et al. 2023; Suzuki and Suzuki 2021). KivD and kdc are the native genes in L. lactis that convert 2-ketoisovalrate into isobutyraldehyde, and kivd is the most widely employed gene in the isobutanol synthetic pathway in microbial hosts. Additionally, L. lactis contains all the essential genes for the synthesis of isobutanol, including aminotransferase, decarboxylase, and alcohol dehydrogenase. Priyadharshini et al. (2015) examined the viability of isobutanol production by L. lactis for the first time (Priyadharshini et al. 2015). After culturing the wild-type strain of L. lactis using an M17 medium, isobutanol was detected in the supernatant after 5 days of fermentation (0.01 g/L). However, the titer of isobutanol was lower than that of E. coli and the co-culture system of L. lactis and E. coli. This result showed L. lactis produces a very low amount of isobutanol. One reason for the low yield is that there are multiple competing pathways that utilize the same substrate. Therefore, blocking competing pathways and overexpressing the isobutanol synthesis genes could be a promising strategy for improving isobutanol production in L. lactis.

Engineering non-native microbial hosts for isobutanol production and its advantages

Engineering of lignocellulolytic and thermophile microorganisms

Heterologous expression of cellulolytic enzymes has been found challenging; therefore, engineering native lignocellulolytic organisms is of great interest to researchers for the production of biofuels. Higashide et al. (2011) constructed a hybrid valine biosynthesis pathway in C. cellulolyticum by expressing the kivd, alsS, ilvCD, and yqhD genes from L. lactis, B. subtilis, and E. coli, respectively. The engineered strain produced 0.66 g/L isobutanol growing on cellulose after 192 h and 0.346 g/L isobutanol from cellobiose in 90 h of fermentation (Higashide et al. 2011).

In industrial biotechnology applications, thermophilic organisms offer various benefits over mesophilic species. High temperature facilitates the removal of volatile products and reduces the risk of contamination (Cann et al. 2020; DiGiacomo et al. 2022; Irdawati et al. 2023; Vavitsas et al. 2022). G. thermoglucosidasius is an important facultative anaerobic, rod-shaped, endospore-producing thermophile bacteria; its metabolic products are lactate, formate, acetate, and ethanol (Cripps et al. 2009), and it was genetically engineered for the production of non-native products (i.e., isobutanol) at elevated temperatures. Genes were selected from two microorganisms to engineer the thermophile G. thermoglucosidasius toward isobutanol production. The kivd gene was selected from the L. lactis, which converts 2-ketoisovalerate into iso-butyraldehyde. The alsS gene was introduced from B. subtilis along with the expression of its native ilvC gene. The engineered strain produced 3.3 g/L of isobutanol after 2 days of fermentation at 50 °C using glucose as a substrate (Lin et al. 2014). Another cellulolytic thermophilic bacteria (C. thermocellum) has been engineered to produce isobutanol. The engineered thermophile accumulated 5.4 g/L isobutanol using cellulose at 50 °C after 75 h of fermentation (Lin et al. 2015).

Engineering autotrophic microorganisms for isobutanol production

One of the advantages of autotrophic organisms is that they can utilize CO2 directly for biofuel production, eliminating the requirement for solid feedstock (i.e., glucose, glycerol, cellulose) (Purdy et al. 2022). To fix CO2 into biofuels, the energy is provided by renewable resources, either sunlight or electricity (Wang et al. 2020b). In autotrophs, cyanobacteria is an excellent microbial cell factory that converts atmospheric CO2 into useful products utilizing sunlight as an energy source (Fig. 2). Moreover, cyanobacteria can grow quickly, and its genetic modification is simple (Fathima et al. 2018; Gao et al. 2021; Lehmann et al. 2021). Owing to their advantages over heterotrophs, cyanobacteria are endowed with an isobutanol pathway to accumulate sustainable and environmentally friendly biofuels.

Atsumi et al. (2009) engineered S. elongatus for the first time to produce autotrophic isobutanol. The engineered strain was developed by expressing B. subtilis alsS, E. coli ilvCD, L. lactis kivd, and three different S. cerevisiae ADHs. Six days of fermentation resulted in 450 mg/L of isobutanol from the engineered strain (Atsumi et al. 2009). Further improvement in isobutanol production was achieved by deleting the glgC gene (encoding glucose-1-phosphate adenylyltransferase); the resulting ΔglgC strain accumulated 550 mg/L of isobutanol during 8 days of fermentation (Li et al. 2014). Another study reported the accumulation of 90 mg/L isobutanol in Synechocystis sp. strain PCC 6803 by overexpressing the two genes kivd (L. lactis) and adhA (L. lactis) without the addition of any inducers or antibiotics (Varman et al. 2013); the titers were fivefold higher compared to the previously engineered strain Synechococcus 7942 expressing the same enzymes (Atsumi et al. 2009).

In recent years, another facultative autotrophic bacterium R. eutropha has drawn the attention of researchers for its ability to fix CO2 into value-added products utilizing non-photosynthetic energy sources such as H2 or formate (Fig. 2) (Kim et al. 2022; Nangle et al. 2020; Wu et al. 2022). It can also use a wide range of feedstocks, such as glycerol, fructose, organic acids, and plant oil (Ingram and Winterburn 2022; Kajla et al. 2022; Strittmatter et al. 2022). Therefore, R. eutropha H16 was engineered to produce 270 mg/L of isobutanol and 40 mg/L of 3-methyl-1-butanol (at 48 h fermentation) from fructose. This was achieved after the deletion of carbon-consuming pathways (phaCAB, ilvE, bkdAB, and aceE) and by overexpression of the native valine biosynthetic pathway genes and alcohol dehydrogenase (adh) in combination with keto-isovalerate decarboxylase (kivd) from L. lactis (Lu et al. 2012). In another study, the recombinant R. eutropha H16 strain accumulated 536 mg/L of isobutanol from CO2 using electrolysis-generated H2 as the sole energy source; the recombinant strain was established by deleting polyhydroxybutyrate synthesis (PHB) genes and expressing isobutanol biosynthesis genes (Li et al. 2012a). Moreover, to improve isobutanol production in R. eutropha, an isobutanol tolerant strain (Re2445) was developed by deleting the acrA gene. The recombinant strain (Re2445) expressing isobutanol biosynthesis genes produced 260 mg/L isobutanol within 72 h of fermentation. The titers were almost 1.3 times higher than their parental strain (Bernardi et al. 2016). Another recent study described the production of 4 mg/L isobutanol utilizing fructose as the sole carbon source in a minimal media, using the CoA-dependent pathway along with the overexpression of native isobutyryl-CoA mutase (sbm1) in a recombinant strain of R. eutropha H16. Isobutanol production was further enhanced to 32 mg/L after supplementation with vitamin B12, an almost eightfold increase compared to no vitamin B12 supplementation (Black et al. 2018). These results demonstrated that deleting carbon-consuming pathways and overexpressing native genes significantly enhanced the productivity of isobutanol. However, the titers accumulated by the autotrophic microorganisms are still very low, so additional strain engineering would be necessary to improve the isobutanol productivity.

Engineering solvent-tolerant microorganisms for isobutanol production

The researchers have been interested in engineering those organisms to produce isobutanol, which has excellent potential to tolerate higher isobutanol concentrations (Goyal et al. 2019; Tian et al. 2022; Ullah et al. 2023). One major limiting variable that prevents microbial biofuel production is solvent toxicity (Liu et al. 2021a). Therefore, considering the solvent tolerance factor of B. subtilis, it was engineered to produce isobutanol. In unbaffled shake-flask fed-batch fermentation, the engineered strain BSUL03 accumulated 2.62 g/L of isobutanol by expressing its native alsS and ilvCD genes of C. glutamicum in combination with the kivD gene of L. lactics and the adh2 gene of S. cerevisiae (Li et al. 2011). Further, improvements were achieved by the inactivation of ldh and pdhC genes. The ldh- and pdhC-deficient strain BSUL05 resulted in 2.3-fold more isobutanol (5.5 g/L) than its parental strain BSUL03 in fed-batch fermentations (Li et al. 2012b). Further optimization of the B. subtilis strain was done via deactivating glucose-6-phosphate isomerase (pgi), overexpressing glucose-6-phosphate dehydrogenase (encoding zwf), and transhydrogenase (udhA) in the ldh- and pdhC-deficient strain BSUL05 to adjust the intracellular redox state; the resulting strain BSULO8 accumulated 11% more isobutanol (6.12 g/L) than its parental strain BSULO5 (Qi et al. 2014). Recently, Tian et al. (2022) constructed a mutant strain of B. subtilis with deletion of 134.4 kb non-essential region from the genome. The deletion mutant (GI12) accumulated 201.7 mg/L isobutanol in a shaking flask, 2.4-fold higher than the wild-type strain (Tian et al. 2022).

Similarly, another solvent-tolerant strain, P. putida, was engineered for isobutanol production. The overexpression of a single gene (kivd) of L. lactis in P. putida enabled the engineered strain to produce isobutanol (43.1 μM) directly from glucose via the valine synthesis route (Lang et al. 2014). In another study, additional improvement in isobutanol from P. putida was achieved by applying multiple strategies such as avoiding product and precursor degradation, deactivation of the sthA gene, overexpressing the native ilvC and ilvD genes, and the introduction of alsS, kivd, and yqhD from B. subtilis, L. lactis, and E. coli, respectively. P. putida-engineered strain Iso2 produced 22 mg/gglc isobutanol under aerobic conditions (Nitschel et al. 2020).

Engineering Escherichia coli for isobutanol production

E. coli is considered a suitable host for producing various high-value chemicals due to its several advantages over other microbial hosts, such as fast growth, the availability of well-established genetic tools, and utilizing different carbon sources under both anaerobic and aerobic conditions (Olavarria et al. 2019; Pasini et al. 2022; Tong et al. 2021; Wang et al. 2011). Owing to these distinct advantages, E. coli has been engineered for various biofuels, including isobutanol (Abdelaal et al. 2019; Hosseini et al. 2020; Sathesh-Prabu et al. 2020). For example, to engineer E. coli for isobutanol production, the metabolic flux was directed to 2-ketoisovalerate by introducing the alsS gene from B. subtilis and overexpressing the endogenous ilvC and ilvD genes. The accumulated ketoacid was then converted into isobutanol with the expression of the kivD and adh2 genes from L. lactis and S. cerevisiae, respectively. The engineered strain accumulated the highest titer of isobutanol at 22 g/L, increasing the theoretical yield to 86% (Atsumi et al. 2008). Additional optimization of isobutanol production in laboratory-scale fermenters with in situ product removal improved isobutanol titer to 50 g/L in E. coli (Baez et al. 2011).

It was found that high isobutanol concentration is toxic to microbial hosts and impairs microbial growth (Wilbanks and Trinh 2017). Chong et al. (2014) constructed an isobutanol-tolerant strain to alleviate this issue by modifying the global regulator cAMP receptor protein (CRP) via error-prone PCR. The recombinant strain produced 12 g/L of isobutanol without appreciable cell loss (Chong et al. 2014). Further improvements were achieved by engineering the ED pathway by deactivating the pgi and gnd genes, along with the deletions in the Embden-Meyerhof pathway (EMP), to divert the metabolic flux towards isobutanol. The engineered strain accumulated 15 g/L of isobutanol as the final titer using glucose as a substrate (Noda et al. 2019).

E. coli expressing β-glucosidase could accumulate isobutanol from cellobiose at a titer of 7.64 g/L, representing 28% of the theoretical yield (Desai et al. 2014). Additionally, a strain of E. coli has been engineered to grow on cellobionic acid, which produced 2.7 g/L isobutanol after 24 h of fermentation, yielding productivity of 0.11 g/L/h and 65% of the maximum theoretical yield (Desai et al. 2015). In another study, a promoter engineering strategy was utilized to boost the production of isobutanol by changing the upstream sequence of the promoter. The recombinant strain accumulated 5.9 g/L of isobutanol, almost tenfold higher than its parental strain after 72 h of fermentation (Jin et al. 2019). In another approach, Wang et al. (2020a, b) developed a robustly engineered strain JW128, which outperforms the original E. coli MG1655 strain regarding transformation efficiency and plasmid stability. The engineered strain JW128 was generated by substituting araBAD with tetr, eliminating mcrBC-hsdSMR-mrr, and deactivating the mcrA, endA, and recA genes. After introducing the isobutanol synthetic pathway into engineered strain JW128, the engineered strain JW128 accumulated 5.76 g/L isobutanol, which was 40-fold greater than the original strain of E. coli K12-MG1655 (Wang et al. 2020a).

Engineering Corynebacterium glutamicum for isobutanol production

The amino-acid-producing bacterium C. glutamicum was engineered to produce isobutanol owing to its potential for using amino-acid precursors for alcohol production and its greater tolerance against alcohol toxicity than E. coli (Smith et al. 2010). A titer up to 4 g/L of isobutanol was successfully obtained after the overexpression of native alcohol dehydrogenase (adhA), ilvC, and ilvD from C. glutamicum, kivd from L. lactis, and alsS from B. subtilis. Further optimization was carried out by knocking out the pyc and ldh genes to direct the metabolic flux toward the isobutanol. As a result, the recombinant strain produced 1.2-fold more isobutanol (4.9 g/L) than the strain without any gene deletion (Smith et al. 2010). In another study, to improve isobutanol production from C. glutamicum, the carbon flux was directed toward KIV by overexpression of ilvBNCD genes and elimination of ilvE, aceE, and pqo genes. Subsequently, under oxygen deprivation, efficient conversion of KIV to isobutanol was achieved by deleting the ldh and mdh genes, heterologous expression of the kivd gene from L. lactis and the pntAB genes from E. coli, and overexpression of the native adhA gene. In fed-batch fermentation, the engineered strain accumulated 12.97 g/L of isobutanol (Blombach et al. 2011). In another study, a titer up to 25 g/L of isobutanol was achieved under oxygen deprivation conditions in an IBU5 strain (C. glutamicum deficient in the ldh gene) expressing the kivd and adhP genes from L. lactis and E. coli under the control of the ldhA and gapA promoters, respectively, and overexpressing its native ilvBNCD genes (Yamamoto et al. 2013).

The expression levels of isobutanol-producing enzymes (encoded by ilvBN, ilvC™, ilvD, kivd, and adhA) have been recently adjusted by Hasegawa et al. (2020) by using several promoters in different combinations, inactivating alternative carbon-consuming pathways by the elimination of the pckA, ppc, ldhA, and ilvE genes and overexpressing glycolytic genes encoded by gapA, pgk, tpi, pfkA, and pgi to enhance glycolysis and glucose uptake. As a result, the final engineered strain accumulated a titer of 20.8 g/L of isobutanol during 24 h of fermentation under oxygen-limited conditions (Hasegawa et al. 2020).

Engineering Zymomonas mobilis for isobutanol production

Z. mobilis is a natural ethanologenic, facultative anaerobic α-proteobacterium bacterium. This bacterium has attracted the attention of researchers as a potential platform microorganism, well suited for producing chemicals and fuels from lignocellulosic sugars (Wang et al. 2018). Therefore, the Z. mobilis Z4 strain was engineered for isobutanol production by introducing the kivd and adhA genes from L. lactis. The engineered strain accumulated a very low titer of isobutanol (He et al. 2014). In another study, isobutanol production was improved by introducing a codon-optimized kdcA gene regulated by the tetracycline-inducible Ptet promoter. As a result, the engineered strain of Z. mobilis accumulated 150 mg/L isobutanol. The production of isobutanol was further improved by diverting the metabolic flux from ethanol to isobutanol synthesis through the expression of its native genes (ilvC and ilvD) and the exogenous alsS gene. This engineering led the recombinant strain to accumulate more than 1 g/L of isobutanol. Additional optimization was done by overexpressing the synthetic operon, alsS-ilvC-ilvD, and kdcA genes driven by Ptet and a strong constitutive Pgap promoter. The engineered strain accumulated the highest isobutanol titer of 4.01 g/L after 49 h of fermentation using glucose as a substrate (Qiu et al. 2020).

Engineering Enterobacter aerogenes for isobutanol production

E. aerogenes naturally synthesizes 1,3-propanediol and 2,3-butanediol (de Oliveira Paranhos and Silva 2020; Wu et al. 2021). It can consume numerous carbon sources, including glucose, xylose, sucrose, fructose, arabinose, and glycerol, and can survive at low pH (de Oliveira Paranhos and Silva 2020; Ewing et al. 2022; Sunarno et al. 2020). Acetolactate, an intermediate of the valine and 2,3-butanediol pathways, can also serve as a precursor for isobutanol production. For this reason, this bacterium has attracted researcher’s attention for isobutanol production through genetic engineering.

For example, Jung et al. (2017) developed an isobutanol-producing strain EMY-01DA by introducing kivd and adhA from L. lactis into EMY-01 (a ldhA-deficient strain of E. aerogenes). The recombinant strain EMY-01DA accumulated approximately 44 mg/L of isobutanol. Further, to enhance isobutanol production, the metabolic flux was directed to the valine synthesis pathway by knocking out the budA gene. As a result, the recombinant strain EMY-33DA produced 16-fold more isobutanol. Further optimization was carried out by knocking out the pflB gene and the expression of ilvC, ilvD, and budB genes from K. pneumoniae to direct the metabolic flux towards isobutanol. Finally, the recombinant strain EMY-333DCBDA accumulated almost 4.3 g/L isobutanol under semianaerobic conditions, including nitrate and formate in the medium (Jung et al. 2017). In addition, isobutanol production needs to be boosted in E. aerogenes; the ptsG gene was deleted in a parent strain to eliminate the preference for carbon utilization and to consume a cheaper carbon source like glucose and xylose derived from lignocellulosic biomass at the same time. After deleting the ptsG gene, strain EMY-340 was constructed, which simultaneously utilized glucose and xylose. Via the introduction of isobutanol pathway enzymes into strain EMY-340, strain EHM02 was constructed. The EHM02 strain could accumulate 5.6 g/L isobutanol after 24 h of fermentation, approximately 1.3-fold higher than its previously constructed strain EMY-333DCBDA (Jung et al. 2018). These accomplishments highlight the potential of E. aerogenes as a promising host for isobutanol production. Further improvements in isobutanol production via more novel and efficient synthetic techniques offer hope of achieving high isobutanol productivity in E. aerogenes.

Current problems and possible solutions for isobutanol production

Feedback inhibition of the enzymes

Isobutanol synthesis in microorganisms is limited by feedback inhibition of the enzymes by their products (Fig. 4a) (Zhang et al. 2022). A possible solution to eliminating the feedback inhibition created by substrates or products is using alternative enzymes or their engineering. Thus, alternative or engineered enzymes should be employed for high productivity to mitigate feedback inhibition problems. In the isobutanol synthetic pathway, an acetohydroxy acid synthase (AHAS) or an acetolactate synthase (ALS) enzyme used in branched-chain amino acid biosynthesis can catalyze acetolactate synthesis from pyruvate. AHAS enzyme consists of a large catalytic subunit and a small regulatory subunit and is normally regulated by the end products. However, the enzyme ALS is a single subunit and is highly specific for acetolactate synthase activity. Unlike most AHAS enzymes, ALS is not inhibited by the end products. For example, Atsumi et al. (2008) expressed the alsS gene of B. subtilis instead of its native biosynthetic enzymes (encoded by the ilvIH gene), which resulted in an increased isobutanol titer of 1.7-fold in E. coli (Atsumi et al. 2008).

Fig. 4
figure 4

Schematic illustration of challenges and possible solutions to improve isobutanol production in microbial hosts: a Feedback inhibition, b solvent toxicity, and c, d pathway imbalance

Additionally, Smith et al. (2010) also confirmed that the expression of B. subtilis alsS is beneficial for the production of isobutanol in C. glutamicum, by confirming its activity in the crude extract with alsS (6,700 U/mg) and without alsS expression (100 U/mg) (Smith et al. 2010). Moreover, the S. cerevisiae AHAS enzyme is composed of two subunits, the catalytic subunit ILV2 and the regulatory subunit ILV6, thus deleting the ILV6 gene improved isobutanol titer up to 2.2-fold as compared to the wild strain (Hammer and Avalos 2017). These investigations demonstrated that removing feedback inhibition of AHAS enzymes could be a promising and viable technique for increasing isobutanol synthesis in microbial hosts.

Screening isobutanol overproducers

Various engineering strategies have been used to increase biofuel production in microbial hosts. Microbial mutation breeding was a basic method used for genome modification, including random mutagenesis by alkylating agents, ethyl methanesulfonate (EMS), ultraviolet (UV), atmospheric and room temperature plasma (ARTP), and nitrosoguanidine (NTG) (Elshobary et al. 2022; Roy and Dahman 2023; Su et al. 2021). These artificial mutagenesis were successfully employed to improve the mutation rate and screen the beneficial strains. For example, a high biobutanol-producing mutant, Clostridial fusants, was constructed using different mutagens, e.g., UV and EMS; the mutated strain accumulated 5.8% more biobutanol than the wild-type strain (Roy and Dahman 2023). The methods for screening the large strain libraries after mutagenesis for beneficial varieties are based on traditional screening methods, and traditional screening methods rely on high-performance liquid chromatography (HPLC) or gas chromatography (GC), which are expensive and laborious. Therefore, high throughput screening methods are required to screen large strain libraries, which can rapidly identify the strains that produce the target compounds in high amounts. Biosensor-based screening strategies have drawn much attention, as they are cost-efficient and quicker to run than traditional chemical analysis for biofuels based on HPLC and GC methods (Bahls et al. 2022). For example, Yu et al. (2019) developed a biosensor based on the transcription factor BmoR, which responded to intracellular isobutanol in E. coli. They used this biosensor to isolate an isobutanol-producing strain from an ARTP mutagenesis library. The isolated mutant accumulated twofold more isobutanol (1597.6 mg/L) than the wild-type strain (Yu et al. 2019).

Solvent toxicity

High-value-added chemicals and biofuels (butanol and isobutanol) are harmful to microbial hosts; these chemicals disrupt the properties of the membrane, causing the leakage of cellular metabolites. As a result, cells cannot maintain cellular function (Kanno et al. 2013). Solvent toxicity is a significant limiting factor that impedes biofuel production during fermentation (Xu et al. 2021). Several engineering techniques have been employed to enhance the tolerance of microbial hosts to biofuels, such as the overexpression of well-characterized native or exogenous resistance genes, artificial mutagenesis, long-term adaptive laboratory evolution (ALE), and the introduction of engineered transcription factors (Fig. 4b). For example, Zhang et al. (2019a, b) successfully enhanced the tolerance of S. cerevisiae towards higher glucose and isobutanol concentrations by EMS mutagenesis followed by ALE. The evolved strain EMS39 performed better in higher concentrations of isobutanol (16 g/L) and glucose (100 g/L). After the expression of ILV2, ILV3, ILV5, and ARO10 genes in the S. cerevisiae evolved strain EMS39, it produced 30% more isobutanol than the control strain (Zhang et al. 2019a).

In another approach, the tolerance of E. coli for isobutanol was enhanced by modifying its transcription factor cAMP protein (CRP). The modified strain exhibited a better growth rate (0.18 h−1) compared to the control strain (0.05 h−1) in the presence of 1.2% isobutanol (9.6 g/L) (Chong et al. 2014). Recently, another study found that the overexpression of PHB synthesis genes improved the tolerance of E. coli towards isobutanol and efficiently enhanced the isobutanol titer. The titer of isobutanol was almost 1.6-fold higher than the control strain after 72 h of fermentation (Song et al. 2020). Thus, these overall results demonstrated that reducing the toxicity of the solvent should be a promising strategy to enhance the production capability of the host strain.

Pathway imbalance

The metabolic pathway for isobutanol production has been engineered for native and non-native microbial hosts. The non-natural metabolic pathway for isobutanol production has been constructed by combining multiple genes from various species. Nevertheless, imbalances in the metabolic pathways are usually observed due to cofactor imbalance and imbalanced gene expression, limiting the titer and productivity of the target compounds. Therefore, to maximize the production of the desired compound and maintain optimal cell growth, cofactor balancing and balanced expression of genes are indispensable strategies (Fig. 4c, d) (Park and Hahn 2019; Shi et al. 2013). For example, NADPH is an essential cofactor for the reaction catalyzed by ILV5 and ADH6 in the S. cerevisiae isobutanol pathway (Matsuda et al. 2013). In contrast, ethanol synthesis requires the utilization of NADH, which is generated during glycolysis. Therefore, an imbalance between NADH and NADPH occurs due to the stimulation of isobutanol production. The cofactor imbalance was adjusted by overexpressing some genes encoding transhydrogenase-like shunts, such as MDH2, PYC2, and MAE1. Therefore, the activation of transhydrogenase-like shunts improved NADPH supply and thus increased the titer of isobutanol up to 1.84-fold compared with its base strain BSW4 after 48 h of fermentation (Matsuda et al. 2013).

Another way to increase the titer of the target compounds is through balanced gene expression. Promoter engineering, RBS engineering, and gene copy number are engineering strategies that control balanced gene expression in endogenous and heterologous pathways (Fig. 4d) (Park and Hahn 2019). Promoters are the significant genetic elements that control gene expression. Thus, selecting the appropriate promoter is essential for achieving high productivity of the target compounds and ensuring smooth growth. For example, Lin et al. (2015) expressed the isobutanol synthetic pathway genes under different promoters to construct several recombinant strains. The amount of isobutanol among the different recombinant strains of C. thermocellum varied approximately from 0.05 to 0.6 g/L. The best recombinant strain accumulated 0.6 g/L of isobutanol after 24 h of fermentation, which was almost a 12-fold increase compared to the inferior strains (Lin et al. 2015). In another approach, isobutanol production was increased to 2.85-fold by balancing isobutanol pathway enzymes expressing the B. subtilis alsS gene under the control of a copper-inducible CUP1 promoter and improving the translational efficiency of the ILV5 and ILV3 genes by the addition of Kozak sequence (Park and Hahn 2019). These studies demonstrated that cofactor balancing and balanced gene expression in the isobutanol pathway play an indispensable role in obtaining a high titer of biofuels.

Future perspectives

There are several promising strategies that can be used to enhance the production of isobutanol. First, addressing the challenges mentioned earlier, such as solvent toxicity, feedback inhibition of the enzymes, and pathway imbalance, can be potential approaches to improve the isobutanol titer. Solvent toxicity is a major issue hindering the productivity of isobutanol in engineered microbial hosts. This issue can be addressed by either using a solvent-tolerant strain or enhancing the tolerance of engineered microbial hosts through genetic engineering. Another approach is continuously removing the accumulated product. Redox imbalance is another factor leading to lower isobutanol production. Therefore, achieving a balanced redox state in engineered microbial hosts by expressing heterologous cofactor regeneration systems is a promising strategy for improving isobutanol production. Improving the catalytic activity of key metabolic enzymes through protein engineering and enhancing cellular robustness through transcription factor engineering are also promising approaches to increase the titer of isobutanol. Moreover, extensive research is required to identify novel strains that efficiently utilize cost-effective carbon sources and accumulate substantial quantities of isobutanol. Expanding the substrate spectrum for isobutanol production is another key challenge. Future research should focus on engineering microbial hosts that can utilize a diverse range of renewable feedstocks for isobutanol production.

Conclusions

Isobutanol, a valuable biofuel that has gained global interest in various industries, has been the focus of extensive research in recent years. The goal of these efforts has been to enhance the efficiency of microorganisms in producing isobutanol. In this review, we provided a comprehensive review of different engineered microbial hosts that can be used for isobutanol production. In the past decade, significant advancements in synthetic biology and metabolic engineering have led to a noticeable improvement in the biological production of isobutanol. However, the amount of isobutanol that can be accumulated by native microbial hosts is still insufficient to meet the demands of industrial applications. Another drawback is that these native hosts rely on costly sugars like lactose and glucose as carbon sources. To overcome these limitations, researchers have turned to non-native microbial hosts such as E. coli and C. glutamicum, which have been genetically engineered to accumulate higher amounts of isobutanol compared to their native counterparts. However, these non-native hosts also require expensive sugars as carbon sources, and isobutanol itself poses safety hazards. In an effort to address these challenges, researchers have worked on engineering cellulolytic, photosynthetic, chemolithoautotrophic, and solvent-tolerant microorganisms for isobutanol production. Despite the potential of these non-native microbial hosts, their current production of isobutanol still falls short of that required for commercial purposes.