Cholesterol is a steroid commonly found in nature with a great relevance in biology, medicine and chemistry, playing an essential role as a structural component of animal cell membranes. The ubiquity of cholesterol in the environment has made it a reference biomarker for environmental pollution analysis and a common carbon source for different microorganisms, some of them being important pathogens such as Mycobacterium tuberculosis. This work revises the accumulated biochemical and genetic knowledge on the bacterial pathways that degrade or transform this molecule, given that the characterization of cholesterol metabolism would contribute not only to understand its role in tuberculosis but also to develop new biotechnological processes that use this and other related molecules as starting or target materials.
Cholesterol is a steroid, i.e. a class of terpenoid lipids, with a carbon skeleton virtually flat and relatively rigid formed by four fused alicyclic rings (Fig. 1). This polycyclic sterol has a great importance in biology, medicine and chemistry, being one of the most common molecules in nature, as an essential structural component of animal cell membranes (Staytor and Bloch, 1965), although it has been also exceptionally found in some prokaryotes (Hayami et al., 1979).
The metabolism of cholesterol is critical in eukaryotes as precursor of vitamins, steroid hormones and bile acids (Miller, 1988; Björkhem and Eggertsen, 2001) (Fig. 1), but according to its physiological relevance, an imbalance in its blood level causes serious diseases in humans.
Cholesterol is frequently found in the biosphere, not only because of its natural abundance, but also due to its high resistance to microbial degradation. Cholesterol is a recalcitrant molecule to biodegradation because of its low number of functional groups (one CC double bond and a single hydroxyl group), its low solubility in water (3 × 10−8 M) and its complex spatial conformation constituted by four alicyclic rings and two quaternary carbon atoms. The high hydrophobicity and low volatility of cholesterol leads to a high absorption to solid phases. Because of their high rate of persistence, cholesterol and some derived compounds, such as coprostanol, have been used as reference biomarkers for environmental pollution analysis (Veiga et al., 2005).
Steroids, some of them derived from cholesterol, constitute a new class of pollutants discharged into the environment as a result of human activity (Gagnéet al., 2006). The potent metabolic activities of these compounds affect a large number of cellular processes and thus, their presence and accumulation in water wastes and in certain ecological niches can affect the endocrine system of animals and humans (Hotchkiss et al., 2008). Moreover, cholesterol is the main component of lanolin, and this and other related sterols are natural contaminants fairly resistant to the anaerobic treatment that is carried out on the effluents from wool-processing industry (Poole and Cord-Ruwisch, 2004).
The ubiquity of cholesterol and related sterols in the environment has made them a common carbon source for many different microorganisms, some of them being important pathogens as Mycobacterium tuberculosis. On the other hand, the bacterial metabolism of steroids has attracted also considerable attention as low-cost natural steroids like cholesterol can be used as starting materials for many bioactive synthetic steroids (Fernandes et al., 2003). Therefore, the characterization of the bacterial metabolism of cholesterol could be useful not only to understand its influence in pathological processes like tuberculosis, but also to develop new organisms by metabolic engineering with potential use as biotechnological tools. In a recent work we reviewed the accumulated biochemical and genetic information on the bacterial pathways that degrade or transform biochemically different steroids (García et al., 2011). This review is mainly focused in cholesterol catabolism and its biotechnological applications including the development of cholesterol biosensors for clinical diagnosis and food industry, microbial steroid biotransformations for the production of steroid drugs and hormones, probiotics with cholesterol-lowering effects, and enzymes for insecticide and fungicide purposes.
Catabolism of cholesterol
The search for microorganisms capable of degrading cholesterol was started over 70 years ago. In the early 20th century, it was observed that several species of Mycobacterium could use cholesterol as the sole source of carbon and energy (Söhngen, 1913; Tak, 1942). Later on, it was reported that microorganisms belonging to the genus Proactinomyces, like Rhodococcus erythropolis (formerly Nocardia erythropolis), have the ability to degrade cholesterol and a potential capacity for the synthesis of new steroid derivatives (Turfitt, 1944; 1948) Moreover, some species of Azotobacter transform cholesterol in cholest-4-en-3-one (cholestenone) or 7-dehydrocholesterol, and have the ability to hydrolyse the side-chain of cholesterol generating methylheptanone (Horvath and Kramli, 1947). Thereafter, other bacteria belonging to the genera Nocardia, Arthrobacter, Bacillus, Brevibacterium, Corynebacterium, Streptomyces, Microbacterium, Serratia, Achromobacter, Pseudomonas or Protaminobacter, among others, were reported to accomplish partial or complete cholesterol degradation (Whitmarsh, 1964; Brown and Peterson, 1966; Arima et al., 1969; Nagasawa et al., 1969; Chipley et al., 1975; Martin, 1977; Ferreira and Tracey, 1984; Drzyzga et al., 2009; 2011; Fernández de Las Heras et al., 2009; Ge et al., 2011). Nevertheless, in spite of the large number of described bacteria able to use cholesterol as a sole source of carbon and energy only the genomes of few of them have been completely sequenced (Table 1). When we analysed in the genome of these microorganisms the presence of some of the most representative gene loci involved in cholesterol metabolism [e.g. igr (side-chain degradation), hsa (central pathway) and mce (transport)], we have observed that some of them do not contain these genes. This suggests that although these bacteria could transform cholesterol they may not be able to completely mineralize this compound. The relevance of these genes in cholesterol catabolism has been also confirmed recently by high-density mutagenesis and deep-sequencing analysis in M. tuberculosis (Griffin et al., 2011). Finally, it is worth to mention that most of these genes are randomly located in different clusters in the bacterial chromosome with the exception of M. tuberculosis where it has been assumed that the catabolic predicted genes are mainly located in a single cluster (80 genes in region of about 50 kb) (Van der Geize et al., 2007). However, the recent work of Griffin and colleagues (2011) have shown that while the genes found to be required for growth in cholesterol were clustered in this region, the majority (60%) were distributed elsewhere in the chromosome.
Table 1. Sequenced genomes of described organisms able to growth in cholesterol as a sole source of carbon and energy.
It is generally assumed that the first reaction in the aerobic metabolism of cholesterol is its oxidation to cholestenone through two sequential reactions. That is, the oxidation of cholesterol to cholest-5-en-3-one [Fig. 2 (2)] followed by its isomerization to cholestenone [Fig. 2 (3)]. The enzymes responsible for this activity are cholesterol oxidases (ChOx), or 3β-hydroxysteroid dehydrogenase/isomerases (HSDs).
ChOx is a monomeric flavoenzyme that catalyses the oxidation and isomerization of cholesterol to cholestenone with a concomitant reduction of FAD. Regeneration of the oxidized cofactor is achieved by the reduction of O2 to hydrogen peroxide. The enzyme is extracellular and occurs as secreted and/or cell-surface-associated forms, depending on the microorganism and growth conditions (Kreit and Sampson, 2009). The ChOx from Streptomyces sp. SA-COO (Yue et al., 1999; Lario et al., 2003) and the BOC1 form of ChOx from Brevibacterium sterolicum (Vrielink et al., 1991; Li et al., 1993) have been classified as type I oxidases and belong to the glucose-methanol-choline (GMC) oxidoreductase family of flavoenzymes where the FAD cofactor is non-covalently bound. However, the BOC2 form of ChOx from B. sterolicum (Coulombe et al., 2001) has been classified as a type II oxidase belonging to the family of vanillyl-alcohol oxidase enzymes, which contain a covalently bound FAD molecule (Motteran et al., 2001). Both classes of ChOxs share the same catalytic activity but show significant differences in their redox and kinetic properties. In general, they exhibit a broad range of steroid specificities and can oxidize a number of hydroxysterols, including sterols, steroid hormones and bile acids, but the presence of a 3β-hydroxyl group is an important requirement for activity (Pollegioni et al., 1999). Remarkably, some ChOxs, like those from Burkholderia cepacia ST-200, Pseudomonas spp. and Chromobacterium sp. DS-1, oxidize cholesterol mainly to 6β-hydroperoxy-cholest-4-en-3-one and hydrogen peroxide instead of cholestenone (Doukyu and Aono, 1999; Doukyu et al., 2008; Doukyu, 2009) but this intermediate has not been associated so far to any cholesterol degradation pathway. Recently, a cholesterol inducible ChoX encoding gene (choG) has been identified in Rhodococcus sp. strain CECT3014 within a gene cluster involved in the metabolism of cholesterol (Fernández de Las Heras et al., 2011).
The 3β-hydroxy-Δ5-steroid HSD belongs to a family of enzymes that catalyse the conversion of 3β-hydroxy-5-ene-steroids to 3-oxo-4-ene-steroids through consecutive oxidation and isomerization reactions, using NAD or NADP as cofactors. Gene database analysis revealed the existence of a protein superfamily that in addition to the bacterial HSDs includes mammalian 3β-hydroxysteroid dehydrogenases, plant dihydroflavonol reductases, bacterial UDP-galactose-4-epimerases and viral 3β-hydroxysteroid dehydrogenase (Baker and Blasco, 1992). All of them present a glycine-rich sequence (Gly-X-X-Gly-X-X-Gly) similar to the Rossman fold sequence and an atypical active site motif (Tyr-X-X-X-Lys). The NAD(P)-dependent HSD from Nocardia sp. has been well characterized and appears to be responsible for the transformation of cholesterol to cholestenone in this cholesterol degrader organism (Horinouchi et al., 1991). Recent genetic and biochemical studies have also demonstrated the direct involvement of a NAD-dependent HSD in mycobacterial cholesterol catabolism (Yang et al., 2007; Griffin et al., 2011; Thomas et al., 2011; Uhía et al., 2011a). In fact, it has been demonstrated that the ChOx analogue (ChoD) found in Mycobacteria does not seem to play a role in cholesterol degradation (Griffin et al., 2011; Uhía et al., 2011a).
Cholesterol side-chain degradation
It has been assumed that simultaneously and independently of the first oxidative process, the long aliphatic chain of cholesterol is removed via a process similar to the β-oxidation of fatty acids, leading to the formation of a steroid C22-oic acid intermediate with the concomitant release of two molecules of propionyl-CoA and one acetyl-CoA (Fig. 2). This process has been elucidated at the biochemical level in strains of Mycobacterium and Nocardia and proceeds differently for cholesterol and β-sitosterol (Sih, 1968; Martin, 1976; Wilbrink et al., 2011).
Recent studies have pointed out that the Cyp125 cytochrome from Rhodococcus jostii RHA1 (ro04679 coding gene) and M. tuberculosis (Rv3545c coding gene) initiates the side-chain degradation of cholesterol (Capyk et al., 2009a; Rosłoniec et al., 2009). This cytochrome belongs to the P450 family and catalyses the hydroxylation at carbon C-26 yielding 26-hydroxycholest-4-en-3-one. Cyp125 from M. tuberculosis CDC1551 was proved to be involved not only in the hydroxylation of cholesterol but also in its further oxidation to cholest-4-en-3-one-27 oic acid (Ouellet et al., 2010). The crystallographic structure of Cyp125 from M. tuberculosis has been recently determined (McLean et al., 2009).
Driscoll and colleagues (2010) and Johnston and colleagues (2010) have demonstrated that Cyp142 is a cytochrome that is able to replace the activity of Cyp125 in M. tuberculosis H37Rv, explaining why a Cyp125 deletion mutant of this strain is still able to grow using cholesterol as a carbon source. Furthermore, Δcyp125 mutant strains of Mycobacterium bovis and M. bovis BCG (Driscoll et al., 2010) and M. tuberculosis CDC1551 (Johnston et al., 2010) could not grow on cholesterol because they have a defective cyp142 gene. In addition, its crystal structure was very similar to that of Cyp125 (Driscoll et al., 2010). Although Griffin and colleagues (2011) have identified Cyp125 as the sole monooxygenase significantly required for growth on cholesterol the involvement of Cyp142 cannot be completely discarded.
After the formation of the carboxylic acid intermediate, the cholesterol side-chain is shortened by a β-oxidation-like process initiated by an ATP-dependent sterol/steroid CoA ligase catalysing the CoA activation of the C27 carboxylic acid intermediate (Sih, 1968; Sih et al., 1968). Subsequent cleavage of the side-chain to 17-ketosteroid takes place stepwise in three consecutive cycles and two molecules of propionyl-CoA and one molecule of acetyl-CoA are formed from the side-chain (Fig. 3). This process involves two initial steps of typical β-oxidation with the participation of one or two different sets of four enzymes, i.e. acyl-CoA dehydrogenases, enoyl-CoA-hydratases, hydroxyacyl-CoA dehydrogenases and thiolases. These steps will release one propionyl-CoA and one acetyl-CoA. The additional propionyl-CoA cannot be released by a conventional β-oxidation due to the presence of the cyclopentane ring D. Although the two following steps would proceed by the action of an acyl-CoA dehydrogenase and an enoyl-CoA-hydratase, as occurs in a typical β-oxidation, they will render a tertiary alcohol, which is not susceptible to dehydrogenation to form the keto group. Therefore, the last step will require the participation of a hydroxyacyl-CoA lyase similar to other CoA-lyases (e.g. malyl-CoA lyase, citramalyl-CoA lyase or 3-hydroxy-3-methylglutaryl-CoA lyase) that will render androst-4-ene-3,17-dione (AD) and propionyl-CoA. A similar mechanism has been proposed for the metabolism of cyclohexaneacetate in Arthrobacter sp. CA1 (Ougham and Trudgill, 1982). A number of genes related to β-oxidation have been recently identified in M. tuberculosis by transposon mutagenesis and suggested to be involved in cholesterol side-chain degradation (Griffin et al., 2011). Some of these genes were located inside the large cholesterol degradation gene cluster (hsd4A, ltp2, fadE29, fadE28, fadA5, fadE30, fadE32, fadE33, fadE34 and hsd4B), but others were found outside of the cluster (fadE5, echA9, fadD36 and fadE25).
Our knowledge of genes involved in microbial degradation of cholesterol side-chain is still very limited, but a catabolic gene cluster was identified in R. jostii RHA1 and Mycobacterium smegmatis, encoding several enzymes predicted to be involved in the β-oxidation process (van der Geize et al., 2007; Uhía et al., 2012) (Fig. 4). Recently, the fadD19 gene product of this cluster has been identified as a steroid-CoA ligase with an essential role in the degradation of C-24 branched-chain-sterols (Wilbrink et al., 2011). In addition, the fadA5 (Rv3546) gene from M. tuberculosis H37Rv has been proposed to encode a β-ketoacyl-CoA thiolase that functions in cholesterol side-chain β-oxidation (Nesbitt et al., 2010). This gene is upregulated by cholesterol and repressed by KstR (see below), and is required for utilization of cholesterol as a sole carbon source in vitro and for full virulence of M. tuberculosis in the chronic stage of mouse lung infection (Nesbitt et al., 2010).
It is worth to mention that cholesterol catabolism would follow an obligatory order for side-chain and ring structure modifications depending of the microorganism. In this sense, while the Δcyp125 mutants of BCG (Bacillus Calmette-Guerin, an inactivated form of M. bovis) and Rhodococcus rhodochrous RG32 transformed cholesterol to cholestenone, the corresponding mutant of R. jostii RHA1 did not (Capyk et al., 2009a). Furthermore, ring degradation cannot proceed with substrates having a full-length aliphatic side-chain in the BCG mutant as it is not able to grow or further transform cholestenone, while R. jostii RHA1 mutant grows in this compound (Capyk et al., 2009a).
The combination of these two first enzymatic steps, i.e. the oxidation of cholesterol together with the degradation of its lateral chain, renders the central intermediate AD [Fig. 2 (5A)] that is further metabolized by the central catabolic pathway.
Subsequently 3-HSA is hydroxylated by enzymes similar to the two-component oxygenase TesA1A2 from C. testosteroni (Horinouchi et al., 2004) leading to the production of 3,4-dihydroxy-9,10-secoandrosta-1,3,5(10)-trien-9,17-dione (3,4-DHSA) [Fig. 2 (8)]. Mycobacterium tuberculosis and R. jostii RHA1 have a similar enzyme named HsaAB (Dresen et al., 2010). The crystal structure of HsaA from M. tuberculosis revealed that the enzyme is very similar to p-hydroxyphenylacetate hydroxylase (Dresen et al., 2010). The last catechol derivative is cleaved by a meta extradiol dioxygenase similar to TesB in C. testosteroni (Horinouchi et al., 2001), which was named HsaC in R. jostii RHA1 (van der Geize et al., 2007) or M. tuberculosis (Yam et al., 2009) yielding 4,5,9,10-diseco-3-hydroxy-5,9,17-trioxoandrosta-1(10),2-diene-4-oic acid (4,9-DSHA) [Fig. 2 (9)]. This compound is then hydrolysed by enzymes similar to TesD from C. testosteroni (Horinouchi et al., 2003b) named HsaD in M. tuberculosis (Lack et al., 2008; 2010) or R. jostii RHA1 (van der Geize et al., 2007) yielding 2-hydroxyhexa-2,4-dienoic [Fig. 2 (10)] and 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic (DOHNAA) acids [Fig. 2 (11)].
Lower catabolic pathway
The lower metabolic steps of the cholesterol catabolic pathway should degrade the 2-hydroxyhexa-2,4-dienoic and DOHNAA acids leading to metabolites able to enter the central metabolic pathways (Kieslich, 1985), but the enzymes involved in this downward pathway have not been yet investigated in detail.
Most likely the catabolism of the 2-hydroxyhexa-2,4-dienoic acid involves genes similar to the tesE, tesF and tesG testosterone catabolic genes from C. testosteroni (Horinouchi et al., 2005). In this sense, the 2-hydroxyhexa-2,4-dienoic acid can be metabolized to 4-hydroxy-2-oxo-hexanoic acid by the TesE-like hydratase [Fig. 2 (12)] that eventually will be transformed by the action of TesG-like aldolase into pyruvic acid and propionaldehyde that is further transformed into propionic acid by the action of TesF-like aldehyde dehydrogenase (Horinouchi et al., 2005). Although it has been suggested that TesF could render propionyl-CoA, the similarity of TesF with CoA acylating aldehyde dehydrogenases is very low and therefore, the most probably end-product should be propionic acid.
DOHNAA would be further transformed to succinic acid and it is assumed that the first step in the degradation of this compound could be the elimination of a propionyl moiety, to produce 9,17-dioxo-1,2,3,4,5,6,10,19-octanorandrostan-7-oic acid [Fig. 2 (13)] through a typical β-oxidation (Kieslich, 1985). This reaction is carried out in two steps, i.e. the activation of DOHNAA by ATP and coenzyme A, followed by a reduction of DOHNAA-CoA by an NADPH-dependent dehydrogenase (Miclo and Germain, 1990). The enzymes specifically involved in these catabolic steps have not been precisely identified nor characterized so far.
A denitrifying β-proteobacterium, strain 72Chol, was the first bacterium isolated and cultured on cholesterol under strictly anoxic conditions able to mineralize cholesterol to carbon dioxide (Harder and Probian, 1997; Hylemon and Harder, 1998). However, the anoxic metabolism of cholesterol has been mainly studied at the biochemical level in the denitrifying β-proteobacterium Sterolibacterium denitrificans, now used as a model organism (Tarlera and Denner, 2003). Interestingly, this bacterium was also able to catabolize anaerobically testosterone (Chiang et al., 2010).
The first steps of the anoxic catabolism of cholesterol are similar to those described for the aerobic process and thus, first the hydroxyl group at C-3 is oxidized into the keto group leading to cholest-5-en-3-one followed by a subsequent isomerization to yield cholestenone (Chiang et al., 2007). These two reactions are catalysed by the bifunctional enzyme AcmA that belongs to the short-chain dehydrogenase reductase (SDR) superfamily (Chiang et al., 2008a). The oxidation of cholestenone to cholesta-1,4-diene-3-one, is catalysed by AcmB, a cholest-4-en-3-one-Δ1-dehydrogenase (Chiang et al., 2008a,b). The enzyme is very similar to the flavin adenine dinucleotide-dependent 3-ketosteroid-Δ1-dehydrogenase from Nocardia corallina (Itagaki et al., 1990).
In contrast with the aerobic pathway where cholestenone is hydroxylated by a monooxygenase at C-27 to produce the corresponding primary alcohol, the anaerobic hydroxylation of the cholesterol side-chain is oxygen-independent (Chiang et al., 2007). This step is catalysed by an enzyme belonging to the molybdenum-containing hydroxylase family and involves the hydroxylation of C-25 of cholestenone or cholesta-1,4-diene-3-one using water as a source of the oxygen atoms (Hille, 2005).
The cholesterol uptake system
Cholesterol as other steroids does not appear to be able to diffuse to the bacterial cytoplasm and needs to be transported by specific uptake systems before being metabolized. A pioneering work in steroid uptake systems was carried by Watanabe and Po (1974) demonstrating the presence of an inducible transport system for testosterone in Pseudomonas testosteroni.
Mce-family proteins are closely involved in the invasion and prolonged existence of M. tuberculosis in host macrophagues. The first mce gene to be discovered was found to promote the uptake of bacteria into nonphagocytic cells and was thus designated to function in mycobacterial cell entry (‘mce’) (Arruda et al., 1993). The genome of M. tuberculosis contains four mce operons (mce1-4) each including 9–13 genes with a similar arrangement encoding two transmembrane proteins with homology to the permease subunits of ABC transporters, along with several putative secreted or cell-surface proteins (Pandey and Sassetti, 2008). Recently, it was reported that the mycobacterial gene cluster mce4 encodes a cholesterol import system that enables M. tuberculosis to derive both carbon and energy from this component of host membranes (Pandey and Sassetti, 2008). The energy for the substrate translocation by all four Mce systems in M. tuberculosis is generated by a common ATPase, named MceG, belonging to the Mlk family (Joshi et al., 2006). Interestingly, the mceG gene is located far away from the four mce operons in M. tuberculosis (Mohn et al., 2008). Genome sequence analyses revealed that diverse actinobacteria, including members of the Mycobacterium, Nocardia, Rhodococcus, and Streptomyces genera, harbour mce transporter operons involved in steroid uptake (Casali and Riley, 2007; García et al., 2011). The reason why these Mce systems require many more proteins than classical ABC transporters remains unclear. It has been suggested that these proteins might form a large complex necessary for the movement of the substrate across cell walls (Mohn et al., 2008).
The mce4 locus also encodes a system involved in steroid uptake in R. jostii RHA1 (Mohn et al., 2008) and in Rhodococcus equi (van der Geize et al., 2008b). The mce4 locus of R. jostii RHA1 is upregulated by cholesterol, and mutants with deletions in this operon lost their ability to grow in this compound (van der Geize et al., 2007). A deletion of supAB genes, which form part of the mce4 operon in R. equi, confirmed that these genes are essential for growth of this organism on cholesterol (van der Geize et al., 2008b). Therefore, up to now, the mce4 locus has been associated with steroid or cholesterol metabolism in all annotated bacterial genomes.
On the other hand, the Mce4 uptake system is also essential for growth of R. jostii RHA1 in cholesterol related compounds such as 5α-cholestanol, 5α-cholestanone and β-sitosterol (Mohn et al., 2008). Nevertheless, mutants of this uptake system were unaffected for growth on steroid compounds having shorter polar side-chains (Mohn et al., 2008). These data indicate that only side-chains of at least eight carbons are recognized by Mce4 steroid uptake system.
Regulation of the cholesterol catabolism
From the very beginning it was recognized that expressions of several genes/enzymes involved in steroid degradation were induced by their respective steroid substrates (Marcus and Talalay, 1956; Möbus et al., 1997; Skowasch et al., 2002), but very few data are still available upon the regulation of genes involved in cholesterol degradation. Kendall and colleagues (2007) have recently identified a repressor, named KstR, belonging to the TetR family of transcriptional regulators, which is involved in the regulation of 83 genes in M. smegmatis related with cholesterol and lipid metabolism (Fig. 4). The operator motif of the KstR dimer was predicted in silico as a 14 bp conserved sequence (TnnAACnnGTTnnA, where n is any nucleotide) located in the upstream region of the KstR-regulated genes (Kendall et al., 2007; Uhía et al., 2011b). The KstR-dependent promoter of the MSMEG_5228 gene of M. smegmatis, which encodes an HSD responsible for the first step in the cholesterol degradative pathway, has been characterized in more detail (Uhía et al., 2011a). DNaseI footprint assays have experimentally demonstrated that KstR specifically binds to an operator region of 31 nucleotides containing the predicted palindromic sequence AACTGgAACgtGTTtCAGTT overlapping the −10 and −35 boxes of the P5228 promoter, suggesting that KstR represses the MSMEG_5228 transcription by preventing the binding of RNA polymerase. The repression mediated by KstR is alleviated by the binding of an inducer molecule that still remains unknown.
Using the microarray data from M. smegmatis, a KstR regulon was also deduced for the human pathogen M. tuberculosis due the high degree of sequence similarity. In this bacterium, KstR is involved in the transcriptional repression of 74 genes (Kendall et al., 2007). The KstR regulon in M. tuberculosis contains a large number of genes that have been found to be essential for the survival of this pathogen in both macrophages and mice (Schnappinger et al., 2003). The 3D-structure of KstR from M. tuberculosis has been already determined (PDB ID: 3MNL).
More recently, another transcriptional repressor has been described as KstR2, which controls the expression of other additional 15 genes related to the lower cholesterol catabolic pathway in M. tuberculosis and M. smegmatis, suggesting that this complete pathway is controlled by at least two regulatory proteins (Fig. 4) (Kendall et al., 2010).
The transcriptome of M. smegmatis growing in cholesterol has been compared with that of cells growing in glycerol as the sole carbon and energy sources (Uhía et al., 2012). The results supported the role of KstR and KstR2 as autoregulated repressors of cholesterol catabolism. Microarray analyses revealed that 89 genes were upregulated at least threefold during growth on cholesterol. These upregulated genes are scattered throughout the M. smegmatis genome and likely reflect a general physiological adaptation of the bacterium to grow on this highly hydrophobic polycyclic compound. The catabolic genes are mainly organized in three large clusters that are putatively controlled by KstR and KstR2. Most of the cholesterol upregulated genes within these clusters appear to be responsible for the uptake of cholesterol, the β-oxidation of the branched side-chain at C17, and the catabolism of the rings to central metabolites. However, the transcriptome analysis also revealed the presence of other cholesterol upregulated genes that were not clustered. Within these genes, some are putatively controlled by the KstR repressor whereas others do not contain binding motifs for KstR or KstR2 in their promoter regions. This finding suggests the putative existence of additional regulatory systems different to KstR and KstR2.
The KstR binding motif identified in mycobacteria is also present in the genome of R. RHA1 (van der Geize et al., 2007; Kendall et al., 2007). Moreover, the kstR and kstR2 orthologues in R. jostii RHA1, i.e. ro04482 and ro04598, respectively, are induced by cholesterol (van der Geize et al., 2007). In R. jostii RHA1 several cholesterol-induced genes do not seem to be regulated by KstR or KstR2, which could also indicate the presence of additional regulators of the catabolic pathway (Uhía et al., 2012).
Pathogenic implications of cholesterol metabolism in mycobacteria
In contrast to previous thoughts (Av-Gay and Sobouti, 2000), it has been recently demonstrated that M. tuberculosis can grow using cholesterol as a sole carbon and energy source (Pandey and Sassetti, 2008). In this sense, cholesterol has recently been identified as an important lipid for mycobacterial infection (de Chastellier and Thilo, 2006; Ouellet et al., 2011). The relatively abundant cholesterol present in host cells is an important growth substrate for these bacteria in different infection stages (e.g. intracellular growth or intracellular persistence).
Mycobacterium tuberculosis growing in the human cells appears to obtain the energy from host lipids rather than carbohydrates, supported by the finding that numerous genes involved in lipid metabolism are upregulated during growth in the mammalian macrophage and the mouse lung (Dubnau et al., 2005). Some genes that are critical for survival of M. tuberculosis in macrophages are involved in cholesterol degradation (Rengarajan et al., 2005; van der Geize et al., 2007). The Mce4 transport system that, at stated above, is required for cholesterol import into bacterial cells, plays an important role in pathogenesis because a mutant lacking this transporter fails to persist in the lungs of chronically infected mice and cannot grow in (IFN-γ)-activated macrophages (Pandey and Sassetti, 2008). Furthermore, it has been proposed that the ability of Mce4 proteins to bind cholesterol could act as a signal allowing the pathogen to interact with the host (Mohn et al., 2008).
The igr locus, which was identified because it is necessary for intracellular growth of M. tuberculosis in macrophages and mice, is absolutely required for cholesterol metabolism, because igr-deficient bacteria cannot grow using cholesterol as a primary carbon source (Chang et al., 2007; Miner et al., 2009). This locus consists of a single operon containing genes for a putative cytochrome P450 (igrA), two acyl coenzyme A dehydrogenases (igrBC), a conserved hypothetical protein (igrD), a putative enoyl coenzyme A hydratase (igrE) and a lipid carrier protein (igrF) (Chang et al., 2007). A homologous operon is present in R. jostii RHA1 and its expression is shown to be induced during growth in cholesterol (van der Geize et al., 2007). The igr cluster is encompassing the mce4 genes, which encode the uptake system for cholesterol and related steroids (Mohn et al., 2008). Interestingly, the igrA gene is homologous to the Rv3545 and ro04679 genes from M. tuberculosis and R. jostii RHA1, respectively, and encode the Cyp125 cytochrome involved in cholesterol C27 (or C26) hydroxylation described above (Capyk et al., 2009a; Rosłoniec et al., 2009).
Mycobacterium tuberculosis lacking the kshA and kshB genes, which encode the two-component system catalysing the NADH dependent 9α-hydroxylation of 3-ketosteroids (Petrusma et al., 2009), failed to use cholesterol and 4-androstene-3,17-dione as carbon sources and its growth and survival is attenuated in macrophages and mice (Hu et al., 2009).
On the other hand, ChOxs have been considered as potential pathogenic determinants. In this sense, the ChOX (ChoE) from R. equi, a primary pathogen of horses and an opportunistic pathogen of humans, has been described as the major membrane-damaging factor (Navas et al., 2001), although its role in pathogenesis is controversial (Pei et al., 2006). The ChOx (ChoD) from M. tuberculosis also appears to be important for the growth of the pathogen in peritoneal macrophages and lungs of mice (Brzostek et al., 2007). Mycobacterium leprae, an obligate intracellular pathogen that shows a dramatic reduction of functional genes when compared with M. tuberculosis, still conserves and expresses the choD orthologue (Marques et al., 2008) and although this pathogen cannot catabolize cholesterol, it is a necessary factor for growth in macrophages (Kato, 1978). Nevertheless, the relevance of ChoD in cholesterol metabolism has been recently questioned (Griffin et al., 2011; Uhía et al., 2011a).
Interestingly, a mutation in the HSD (Rv1106c) of M. tuberculosis that is required for growing in cholesterol does not reduce its growth in infected macrophages or guinea pigs suggesting that cholesterol is not a sole nutrition source in vivo (Yang et al., 2011), consistently with the idea that there are multiple lipids in the host that M. tuberculosis would be able to catabolize for surviving (de Carvalho et al., 2010).
As mentioned above, the FadA5 thiolase involved in the side-chain metabolism of cholesterol is required for full virulence of M. tuberculosis as a fadA5 mutant showed attenuated phenotype and at 8 weeks after infection, the number of cfu in the mouse lungs decreased about 10-fold compared with the level in the wild-type strain (Nesbitt et al., 2010).
Finally, the development of chemical inhibitors targeting cholesterol catabolic enzymes has been proposed as a novel route for designing new therapeutics to deal with the re-emergent pathogen M. tuberculosis (Ouellet et al., 2011).
Biotechnological applications of cholesterol degrading bacteria
Microorganisms capable of degrading and/or transforming cholesterol have been widely used for many years in biotechnological applications. These applications include the development of cholesterol biosensors for clinical diagnosis, microbial steroid biotransformations for the production of steroid drugs and hormones, probiotics with cholesterol-lowering effects, and enzymes for insecticide and fungicide purposes.
Biosensors for cholesterol analysis in biological samples
ChOxs are widely utilized for the enzymatic analyses of cholesterol. The most common source of ChOxs are from Streptomyces hygroscopicus, B. sterolicum and Pseudomonas fluorescens (Arya et al., 2008; Nien et al., 2009; Singh et al., 2009). Because 70% of cholesterol exists in ester form and 30% as free form in blood samples, cholesterol esterases are also included in biosensors together with ChOxs (Arya et al., 2008; Shih et al., 2009). In addition, horseradish peroxidase has been used in the fabrication of photometry based biosensors (Arya et al., 2008). This enzyme can be also coupled to 4-aminoantipyrine and phenol to determine the hydrogen peroxide released in the ChOx reaction, a method routinely used in clinical laboratories (Pollegioni et al., 2009). To avoid the interference of blood pigments in colorimetric assays some electrochemical methods based on oxygen consumption by ChOx reaction have been also developed (Cheillan et al., 1989; Pollegioni et al., 2009).
Microbial ChDHs, which catalyse the conversion of 3β-hydroxy-5-ene-steroids to 3-oxo-4-ene-steroids using NAD or NADP as cofactors, are also utilized in cholesterol biosensors (Ishige et al., 2009; Wallace-Davis et al., 2011).
Several authors have also shown that the P450scc cytochrome, an isoform of the P450 (CYP) cytochrome superfamily, which catalyses the side-chain cleavage of cholesterol to pregnenolone in the presence of molecular oxygen and NADPH-ferrihaemoprotein reductase (Carrara et al., 2008), can be utilized for determination of cholesterol with high sensitivity (Antonini et al., 2004; Paternolli et al., 2004; Shumyantseva et al., 2004; 2005). The highly specific conversion of steroid hormones by cythocromes makes feasible a specific substrate determination in complex media by using CYP based sensors (Bistolasa et al., 2005)
The synthesis of the complex steroid molecules requires very specific reactions, and the utilization of microorganisms should enable the performance of reactions of high regio- and stereo-selectivity. Moreover, the mild conditions required for bioconversions as well as the more ecological processes compared with chemical synthesis make the high yield biological production an interesting field to develop.
The importance of microbial biotransformations in the production of steroid drugs and hormones became a reality when the 11α-hydroxylation of progesterone by a Rhizopus species was patented in 1952 (Hogg, 1992). Since then, microbial reactions for the transformation of steroids have proliferated and specific transformation steps have been incorporated into numerous partial syntheses of steroids with therapeutic use and commercial value (Mahato and Garai, 1997). Although some of these bioconversions are well-established, increasing the efficiency of the existing processes as well as identifying new potentially useful bioconversions, are two current objectives of the steroid industry (Fernandes et al., 2003).
Microbial bioconversion has been focused mainly in steroid hydroxylation, Δ1-dehydrogenation and sterol side-chain cleavage that associated to chemical synthesis steps, have enhanced the large-scale production of natural and modified steroid analogues. The use of whole cells instead of enzymes is preferred as the production costs are lower and it is possible to perform multi-steps conversions with a single biocatalyst (Bortolini et al., 1997). The manufactured steroid compounds have a wide range of therapeutic applications as anti-inflammatory, immunosuppressive, diuretic, anabolic, contraceptive agents, breast and prostate cancer and anti-obesity agents, among others. The therapeutic action of steroid hormones has been traditionally associated to their binding to the respective intracellular receptors, which act as transcription factors in the regulation of gene expression. Some steroid molecules are also called neurosteroids due to their role as memory enhancers, inducers of endocrine response to stress, anxiolytic agents, anticonvulsants, antidepressives and neuroprotective effect (Fernandes et al., 2003).
Microbial transformation of steroids can also be an environmentally beneficial process. Dias and colleagues (2002) studied the composition of different industrial waste materials for their steroid/cholesterol content in order to find out if these industrial by-products were suitable as steroid sources for the microbiological production of AD and ADD. They were used as a substrate for microbial degradation by a Mycobacterium sp. strain and proved to be easily converted to AD and ADD. Thus, a profitable application of these wastes could be the production of sterol-rich fractions, reducing the pollution loads.
Probiotics are living microorganisms, which upon ingestion in certain numbers exert health benefits on the host beyond inherent basic nutrition (Guarner and Schaafsma, 1998). Although many in vitro and in vivo studies have shown that the administration of probiotics reduces serum/plasma total cholesterol, LDL-cholesterol and triglycerides or increases HDL-cholesterol, their hypocholesterolemic effects remain controversial (Ooi and Liong, 2010). The species with cholesterol-lowering effects studied include genera as Lactobacillus, Lactococcus, Enterococcus, Streptococcus or Bifidobacterium (Pereira and Gibson, 2002).
Probiotics have been suggested to reduce cholesterol via various mechanisms that do not necessarily imply the transformation of cholesterol by the probiotic strains, e.g. enzymatic deconjugation of bile acids by bile-salt hydrolases of probiotics (Lambert et al., 2008), assimilation of cholesterol (Pereira and Gibson, 2002), co-precipitation of cholesterol with deconjugated bile (Liong and Shah, 2006), cholesterol binding to cell walls of probiotics (Liong and Shah, 2005), incorporation of cholesterol into the cellular membranes of probiotics during growth (Lye et al., 2010a), production of short-chain fatty acids upon fermentation by probiotics in the presence of prebiotics (De Preter et al., 2007) and conversion into coprostanol (Lye et al., 2010b).
As stated above, the most frequent transformation of cholesterol in the intestinal anoxic habitats is a biogenic reduction of its double bond yielding coprostanol, which is directly excreted in faeces (Li et al., 1995; Harder and Probian, 1997). Many intestinal fermenting bacteria are able to catalyse this reaction (Groh et al., 1993; Freier et al., 1994) decreasing the amount of absorbed cholesterol and leading to a reduced concentration in the physiological cholesterol pool (Ooi and Liong, 2010). Lye and colleagues (2010b) evaluated the conversion of cholesterol to coprostanol by strains of lactobacilli such as L. acidophilus, L. bulgaricus and Lactobacillus casei detecting both intracellular and extracellular cholesterol reductases in these strains, indicating possible intracellular and extracellular conversions of cholesterol to coprostanol. Cholesterol reductase has been also directly administered to humans to convert cholesterol to coprostanol in the small intestines to reach a bloodstream cholesterol-lowering effect (Ooi and Liong, 2010).
One of the best-studied mechanisms proposed for the cholesterol-lowering effects of probiotics is the enzymatic deconjugation of bile acids by bile salt hydrolases (Begley et al., 2006). Bile, a water-soluble end-product of cholesterol in the liver that consists of cholesterol, phospholipids, conjugated bile acids, bile pigments and electrolytes, is released into the duodenum upon ingestion of food. When it is deconjugated by bile salt hydrolases-producing strains, bile acids are less soluble, being no absorbed by intestines and eliminated in the faeces. Cholesterol is then used to synthesize new bile acids as a homeostatic response, resulting in a hypocholesterolemic effect of probiotics (Begley et al., 2006; Ooi and Liong, 2010).
The cholesterol-lowering effects are also attributed to the ability of probiotic strains to bind cholesterol, a process that is growth-dependent and strain-specific. However, although growing cells remove more cholesterol than dead cells, the heat-killed cells can still remove cholesterol from media, indicating that some cholesterol is bound to the cellular surface (Usman and Hosono, 1999). During the probiotic growth, the presence of cholesterol increases the concentration of saturated and unsaturated fatty acids, leading to increased membrane strength and subsequently higher cellular resistance towards lysis (Lye et al., 2010a; Ooi and Liong, 2010).
Microbial enzymes for insecticides and fungicides
Besides the nutritional role of ChOx, other roles as biological weapons have been ascribed to these enzymes (Ghoshroy et al., 1997; Martín and Aparicio, 2009). Purcell and colleagues (1993) discovered a highly efficient protein that killed boll weevil (Anthonomus grandis grandis Boheman) larvae in Streptomyces culture filtrates and identified the protein as ChOx. Purified enzyme was active against boll weevil larvae at a concentration comparable with the bioactivity of Bacillus thuringiensis proteins against other insect pests. The enzyme is involved in the lysis of the midgut epithelial cells of the larvae and also exhibits insecticidal activity against lepidopteran cotton insect pests, including tobacco budworm (Heliothis virescens), corn earworm (Helicoverpa zea) and pink bollworm (Pectinophora gossypiella) (Corbin et al., 2001). Recently, it was reported that Chromobacterium subtsugae has insecticidal properties (Martin et al., 2007). ChOx might be involved in this insecticidal activity because it was recently found that Chromobacterium strains produce this enzyme (Doukyu et al., 2008). Some insecticide proteins are vital for pest control strategies employing transgenic crops. The ChOx encoding gene has been expressed in tobacco cells (Corbin et al., 1994; Cho et al., 1995; Doukyu, 2009).
Despite the large number of works carried out during the last century, the bacterial catabolism of cholesterol and other steroids is still far for being completely understood. For instance, the enzymes involved in the degradation of the steroid side-chain as well as those responsible for the last steps of the catabolic pathway remain to be precisely established. Moreover, the physiological role played by bacterial ChOxs is still unclear. Very few data are available about the proteins responsible for the bacterial uptake of steroids and this field will require supplementary attention as the possibility to create by metabolic engineering heterologous recombinant bacteria able to transform steroids for biotechnological purposes will depend on the possibility of conferring these bacteria-specific steroid transport capacities. As stated above, only a limited number of regulatory proteins involved in the metabolism of cholesterol have been identified and partially characterized and the real inducers of these regulators remain to be elucidated as it has not been determined yet whether the cholesterol or their catabolic intermediates are their effectors. Taking into account that the availability of commercial pathway intermediates is very limited the identification of such putative inducers appears to be a difficult task.
Although cholesterol is equally disseminated both in oxic and anoxic environments only two bacterial strains were reported so far to mineralize cholesterol to carbon dioxide under anoxic conditions. Therefore, the anoxic catabolism of cholesterol remains challenging because compared with the oxic metabolism of cholesterol, the metabolism of this substrate in the absence of oxygen is still very poorly understood (Chiang et al., 2008a).
While powerful technological approaches based on modern omic technologies (genomics, metagenomics, transcriptomics, etc.) will be useful to solve many of the questions raised here, the complexity of the pathways and their intermediates still demand to carry out huge efforts by using conventional biochemical and genetic tools to be able to decipher all the steps and processes involved in steroid catabolism.
In spite of the fact that the increment of cardiovascular diseases makes the availability of cholesterol biosensors of outstanding interest, few have been successfully commercialized. Optimization of critical parameters, such as enzyme stabilization, quality control and instrumentation design is needed. Increased understanding of the immobilized bioreagents, improved techniques for immobilization and technological advances in the microelectronics will lead to the rapid development of more accurate cholesterol biosensors (Arya et al., 2008).
Expected developments in biological production of steroid compounds range from the identification of novel biocatalysts or the improvement of existing ones to the enhancement of the biotransformation processes. The application of genetic engineering and systems biology tools for the upgrading of steroid-transforming microorganisms; solubility improvement of substrates that are sparingly soluble in water; immobilization of enzymes or whole cells in a suitable matrix for repetitive economic utilization of enzymes and development of a continuous process for economic product recovery are some of the biotransformation fields of research.
Lamb and colleagues (1998) have proposed the existence of a sterol biosynthetic pathway in M. smegmatis, after ascertaining a significant number of homologies between the genome sequence of M. tuberculosis and the sterol biosynthetic enzymes of Saccharomyces cerevisiae. Therefore, the development of a steroid cell-factory in M. smegmatis based on the approach followed by Duport and colleagues (1998), who engineered six genes to construct a yeast strain that produces pregnenolone and progesterone from galactose, could be envisaged (Fernandes et al., 2003).
The knowledge of microbial cholesterol catabolic pathways could lead to the introduction of several changes in probiotic organisms thus transforming them in strains capable of reducing cholesterol levels in foods and intestines. In this regard, Park and colleagues (2008) constructed a constitutive high-level expression vector for the genus Bifidobacterium and used it to express ChOx from Streptomyces coelicolor. In vivo studies of the cholesterol-lowering effects of this genetically modified organism as well as of the hypocholesterolemic properties of other modified strains that could be constructed in the future would be of outstanding interest.
We thank Dr E. Díaz and Dr M.A. Prieto for the critical reading of the manuscript and helpful discussions. This work was supported by grants from the Ministry of Science and Innovation (BFU2006-15214-C03-01, BFU2009-11545-C03-03), and an I3P predoctoral fellowship from the Consejo Superior de Investigaciones Científicas.