Báo cáo khoa học: Cholesterol oxidase: biotechnological applications Loredano Pollegioni, Luciano Piubelli and Gianluca Molla pot

ChOx is produced by two types of bacteria: a nonpathogenic bacteria, which utilize cholesterol as a carbon source; and b Keywords biocatalysis; cholesterol; cholesterol determination; di

Cholesterol oxidase: biotechnological applications

Loredano Pollegioni, Luciano Piubelli and Gianluca Molla

Dipartimento di Biotecnologie e Scienze Molecolari, Universita` degli studi dell’Insubria, Varese, Italy, and Centro di Ricerca Interuniversitario

in Biotecnologie Proteiche ‘‘The Protein Factory’’, Politecnico di Milano and Universita` degli studi dell’Insubria

Introduction

In eukaryotes, cholesterol is essential for maintaining

cell membrane structure and for synthesizing a number

of important compounds Moreover, improper

mainte-nance of cholesterol concentrations can severely

affect the physiological function of an organism

Cho-lesterol oxidase (ChOx, EC 1.1.3.6), or more precisely,

3b-hydroxysterol oxidase, is a bacterial enzyme that

has proven to be very useful in biotechnological

appli-cations related to the detection (and conversion) of

cholesterol and to the disruption of

cholesterol-con-taining membranes As described in the two

accompa-nying reviews in this miniseries [1,2], ChOx catalyzes

the oxidation of the C3-OH group of cholesterol (and

other sterols) to give the corresponding D5-3-ketone

(cholest-5-en-3-one) and its isomerization to D4 -3-ke-tone (cholest-4-en-3-one, see Scheme 1 in the first mini-review [1]) In bacteria, ChOx is the first enzyme in the catalytic pathway that yields propionate and acetate as final products Importantly, there is no mammalian homolog of ChOx

Since 1973, ChOx has been used in clinical chemistry

to measure serum cholesterol ChOx research has grown largely in the past several years: a PubMed search for the term ‘cholesterol oxidase’ yielded 124,

212 and > 220 papers in the periods 1980–1989, 1990–

1999 and 2000–2008, respectively ChOx is produced

by two types of bacteria: (a) nonpathogenic bacteria, which utilize cholesterol as a carbon source; and (b)

Keywords

biocatalysis; cholesterol; cholesterol

determination; diagnostic enzyme; enzyme

biotechnology; flavoproteins; virulence

Correspondence

Loredano Pollegioni, Dipartimento di

Biotecnologie e Scienze Molecolari,

Universita` degli studi dell’Insubria,

via J.H Dunant 3, 21100 Varese, Italy

Fax: 0332 421500

Tel: 0332 421506

E-mail: loredano.pollegioni@uninsubria.it

(Received 23 July 2009, revised

2 September 2009, accepted 10

September 2009)

doi:10.1111/j.1742-4658.2009.07379.x

Cholesterol oxidase is a bacterial FAD-containing flavooxidase that catalyzes the first reaction in cholesterol catabolism Indeed, this enzyme catalyzes two reactions: the oxidation of the C3-OH group of cholesterol (and other sterols)

to give cholest-5-en-3-one; and its isomerization to cholest-4-en-3-one In the past several years, the structural and functional characterization of choles-terol oxidase has been developed together with its application as a biological tool Cholesterol oxidase has been used in biocatalysis for the production of

a number of steroids, as an insecticidal protein against boll weevil larvae and,

in particular, as a diagnostic enzyme for determining serum levels of choles-terol These applications prompted various laboratories worldwide to isolate this flavooxidase from different sources and to improve its properties by pro-tein engineering, further increasing our knowledge on its structure–function relationships These studies also discovered new physiological roles for cho-lesterol oxidase (e.g in virulence and as an antifungal sensor) We assume that the investigations of cholesterol oxidase and its applications will con-tinue to grow quickly in the near future, in particular to uncover unexpected, new areas of application

Abbreviations

BsChOx, Brevibacterium sterolicum cholesterol oxidase; ChOx, cholesterol oxidase (EC 1.1.3.6); ReChOx, Rhodococcus equi cholesterol oxidase; ShChOx, Streptomyces hygroscopicus cholesterol oxidase.

pathogenic bacteria, which require ChOx for infection

of the host macrophage because of its ability to alter

the physical structure of the lipid membrane by

con-verting cholesterol into cholest-4-en-3-one (see below)

Both pathogenic and nonpathogenic bacteria

upregu-late the expression of ChOx in the presence of

choles-terol

This review summarizes our knowledge concerning

the production of ChOx and its main uses By

under-standing the structure–function relationships of this

flavoenzyme and the mode of action by which it alters

the physical properties of lipid membranes, together

with its overexpression in recombinant systems, we

may discover new and⁄ or optimized applications of

ChOx

Sources of ChOx

A list description of ChOx-containing microorganisms

was published in 2000 [3] This enzymatic activity is

produced by a variety of microorganisms that are found

in different environments: a detailed list is reported in

the Supplementary materials section Microorganisms

such as Arthrobacter, Rhodococcus equi, Nocardia

erythropolis, Mycobacterium and Nocardia rhodochrous

produce an intracellular⁄ membrane ChOx, while

Pseudomonas, Schizophyllum commune,

Brevibacteri-um sterolicBrevibacteri-um, Streptoverticillum cholesterolicum and

Streptomyces violascens produce an extracellular

enzyme As a general rule, ChOx enzymes require the

prosthetic group FAD for their activity (both as a

cova-lently or noncovacova-lently bound cofactor) An exception

to this could be the ChOx isolated from

c-Proteobacteri-um Y-134, which was reported to contain covalently

bound FMN as a cofactor [4] This enzyme also differs

in substrate specificity compared with other known

ChOxs (e.g it is not active on pregnenolone and

ergosterol, see Table 1)

Generally, the presence of cholesterol (or of a similar

compound) in the growth medium induces the

produc-tion of ChOx Among the ChOx-producing bacteria,

R equi (previously identified as B sterolicum ATCC

21387) [5] exhibits high levels of ChOx activity, but

com-mercial use of this microorganism is limited because of

its highly pathogenic nature (see below) This enzyme is

the product of the gene choE and it is homologous to

other secreted ChOxs identified in Streptomyces spp (see

Fig S1) This protein also exhibits significant similarities

to putative ChOxs encoded by Mycobacterium

tubercu-losisand Mycobacterium leprae Rhodococcus

erythropo-lisChOx was produced (using cholesterol solubilized in

Tween-80 in the culture medium) both as an intracellular

enzyme and an extracellular enzyme: the maximal

production of both forms of ChOx was obtained after

70 h of growth, using a spray-dry method of preparation

of cholesterol (up to 365 UÆL)1 and 1.7 UÆg)1 of cells) [6] Rhodococcus (Nocardia) erythropolis IMET 7185 also produces ChOx: the addition of 1 gÆL)1 of cholesterol results in a fivefold increase in ChOx production, up to 3.3 UÆg)1of cells [7] The soil-isolated strain Rhodococ-cussp GK1 produces ChOx both as membrane-bound and extracellular forms [8] The synthesis of ChOx depends on the presence of either phytosterols or hex-anoate as the sole carbon source; under these conditions, the membrane-bound ChOx was produced up to

100 UÆg)1 of cells and 400 UÆL)1 [8] By contrast, the synthesis of ChOx from Burkholderia cepacia (wrongly classified as Pseudomonas sp strain ST-200 and corre-sponding to the choS gene, Fig S1) is not induced by cholesterol and this bacterial strain produced only up to

13 UÆL)1 [9] The highest production levels have been reported in the nonpathogenic Streptomyces sp., reach-ing up to 2500 UÆL)1 in fermentation broth in terms

of purified protein [10] (for the sake of comparison, the activity values not measured at 25C have been cor-rected for the effect of temperature)

Because of the moderate productivity of the various microorganisms used and the requirement of adding cholesterol to culture media to induce⁄ enhance produc-tion, the cost of ChOx remained relatively high [6,8] The yield of ChOx production was enhanced by alter-ing the composition of the growth medium and the physical parameters of culture In S commune, ChOx production depends on the amount of oleic acid in the culture broth but, interestingly, this applies to the level

of insoluble oleic acid rather than to the soluble com-ponent adsorbed onto the cell [3] A detailed investiga-tion of the growth condiinvestiga-tions was also reported for

R equi no 23 grown in a batch fermentor: by adjust-ing the pH (from 6.5 to 7.5), the temperature (from 39

to 37C) and the agitation speed (from 200 to

300 rpm) after 24 h of growth, maximal production of ChOx (340 UÆL)1, corresponding to 11 UÆh)1ÆL)1) was reached in 30 h of culture at an aeration rate of

5 LÆmin)1[11]

To overcome the difficulties related to overproduc-tion of this flavoenzyme in the original organism, ChOx genes from several sources have been cloned and expressed as recombinant proteins However, most of these recombinant microorganisms produce ChOx at levels close to those of the original source

An exception to this is the expression of the gene encoding Rhodococcus equi cholesterol oxidase (ReChOx) in Streptomyces lividans that resulted in a yield of protein production  85-fold higher than that from the natural organism [12] Its expression in

S lividans (and in Escherichia coli) was accomplished

by using various deletions within the 5¢-flanking

(non-coding) region When the ReChOx protein was fused

with the N-terminal part of the lacZ protein, large

amounts of inactive enzyme were produced as

‘inclu-sion bodies’ In contrast, by using a gene with less

than 256 bp at the 5¢-flanking region, enzyme could be efficiently produced in S lividans, secreting

 30000 UÆL)1 into the culture medium Recently, ReChOx was also efficiently synthesized in E coli under the control of the pT7lac promoter in the pET28 plasmid (up to 3.7 UÆmg)1 of protein was

Table 1 Relative activity of ChOx from different sources In all cases, activity on cholesterol (k cat value) was taken to be 100%.

c-Proteobacterium Y-134c

Chromobacterium DS-1 d

Cholesterol

HO

7a- or

7b-OH-Cholesterol

7-Keto-cholesterol

Ergosterol

HO

b-Sitosterol

HO

Pregnenolone

HO

O

Cholestanol

HO

a

Left column: 0.6% Triton X-100, pH 7.5 [70]; right column: 0.7% propan-2-ol, pH 7.5 [71].

b 100 m M potassium phosphate, pH 7.5, 0.4% Thesit (data in parenthesis were obtained in 500 m M potassium phosphate) [72,73].

c 100 m M potassium phosphate, pH 7.0, 0.05% Triton X-100, pH 7.5 [4,74].

d 50 m M sodium phosphate buffer, pH 7.0, 64 m M sodium cholate, 0.34% Triton X-100 [24].

produced in the crude extract and 2200 UÆL)1 was

produced in the fermentation broth) [13,14], as well

as in Chromobacterium DS-1 ChOx (1870 UÆL)1 was

produced in the fermentation broth and 2.3 UÆmg)1of

protein was produced in the crude extract) [15] The

only data concerning the production of true B

steroli-cum ChOx (i.e that containing the FAD cofactor

covalently linked to the apoprotein moiety) were

obtained for E coli using the pET24 plasmid (up to

4 UÆmg)1 of protein was produced in the crude extract

and 12000 UÆL)1 was produced in the fermentation

broth) [F Volonte`, L Pollegioni, G Molla, F

Mari-nelli and L Piubelli, personal communication]

The Streptomyces ChOx gene was also overexpressed

in S lividans by subcloning the choP-choA operon into

a multicopy shuttle vector, which produced a 70-fold

greater amount of ChOx than the original organism

(i.e up to 4400 UÆL)1 in the fermentation broth) [16]

Interestingly, more than 90% of the protein produced

is extracellular The same ChOx was then produced in

E coli JM109 under the P8 promoter, reaching a

specific activity in the crude extract of 9.7 UÆmg)1 of

protein after 3 days of growth [17] A detailed review

of the methods for extracting and purifying ChOx has

been published previously [3]

ChOx as a biological tool

Effect of detergents and solvents on ChOx

activity and stability

ChOx is an enzyme that interacts with membranes and

micelles and acts on hydrophobic substrates whose

sol-ubility is increased by using detergents and⁄ or solvents

These compounds also affect the activity⁄ stability of

the flavoprotein One of the first investigations on the

effect of detergents on ReChOx activity showed a

sig-moidal behavior versus cholesterol concentration at a

Triton X-100 concentration of ‡ 0.3% (the

corre-sponding Hill plot gave an n-value of 4) with a

con-comitant increase in the Km value [18] The transition

from a simple Michaelis–Menten behaviour to a

sig-moidal behavior was also recently observed for the

V121A mutant of Streptomyces hygroscopicus

choles-terol oxidase (ShChOx) at a fixed Triton X-100

con-centration of 0.077% (the activity decreases with an

increase in Triton X-100 concentration) [19] Val121 in

ShChOx is located in a hydrophobic loop near the

active site, which should play an important role in

extracting cholesterol from micelles: the V121A

substi-tution negatively affects this step only if cholesterol is

solubilized using nonionic detergents (no sigmoidal

behavior is observed when cholesterol is solubilized in

propan-2-ol) For example, the enzymatic activity of Nocardia ChOx is induced upon the formation of mixed micelles of cholesterol with molecules of the detergent Surfal: the enzyme is half activated at a detergent concentration corresponding to its critical micelle concentration [20]

In 1988, Cees Veeger’s group published a very important work regarding clarification of the effect of the solvent⁄ detergent on the reaction catalyzed by ChOx [21] The dependence of catalytic activity of No-cardia ChOx on the composition of a solution com-posed of n-hexane, propan-2-ol and water (a detergentless microemulsion) revealed two maxima The maximal catalytic activity in ternary systems is equal to that determined in aqueous solution, whereas the Km value increases with solvent concentration because of a preferential interaction with the surround-ing organic solvent with respect to the hydrophobic ChOx active site By contrast, the stability profile shows a single sharp maximum corresponding to the microemulsion region of the phase diagram [21]; in any case, the storage stability is higher in aqueous solution We have also systematically studied the effect

of the nonionic detergents Thesit and Triton X-100, and of propan-2-ol (used as a substrate solubilizer), on ShChOx and Brevibacterium sterolicum cholesterol oxi-dase (BsChOx) activity [22] At a low concentration of Thesit, activity increases for both enzymes, whereas at higher detergent concentrations (‡ 2–5%) the opposite effect occurs On the other hand, Triton X-100 inacti-vates both enzymes at all concentrations We thus deduced that these surfactants exert their effect by interacting with the enzymes and not by affecting micellar phenomena Analogously, the increase in con-centrations of propan-2-ol (or other organic solvents)

up to  5–10% (v ⁄ v) induces an increase in the activ-ity of both ShChOx and BsChOx and a decrease at higher solvent concentration A significant difference between the two ChOx enzymes emerges when stability

is analyzed as a function of concentration of propan-2-ol: BsChOx is rapidly inactivated, whereas for ShChOx 70% of the initial activity still remains after 5 h in the presence of 30% propan-2-ol [22] This observation is

in line with the stability to detergents reported in Fig 1 (see below)

The stability of ChOxs from different sources, in the presence of several detergents, was also compared after

1 h of incubation at 30, 45 and 60 C [4,23,24] The sta-bility at 30C was similar among the ChOxs analyzed (the only exception was observed in the presence of

‡ 0.1% SDS that fully inactivated all enzymes except for those from Burkholderia and Chromobacterium DS-1), with the most significant changes being observed

after 1 h of incubation at 60C in the presence of

nonionic detergents (see Fig 1, which reports additional

data from our experiments on covalent BsChOx and its

noncovalent variant H69A mutant) At 0.5% Triton

X-100, all ChOxs assayed showed a high stability, while

in the presence of sodium cholate, only the ChOxs from Proteobacterium Y-134, Chromobacterium DS-1 and Pseudomonaswere stable Interestingly, the ChOxs from Proteobacterium Y-134 and Chromobacterium DS-1 were also more stable to temperature than all other ChOxs: they maintained more than 90% of the original activity after 1 h of incubation at 60C The results obtained using BsChOx were significantly different from those observed with other ChOxs Both wild-type and H69A BsChOxs showed very low stability under the conditions tested: the residual activity after 60 min of incubation at 60C was £ 3% in all cases, despite the presence of the covalent FAD linkage in the wild-type BsChOx Analogously, a systematic comparison was carried out to determine the stability of ChOxs, from different sources, after 12–24 h of exposure to solvents

at 37C [23,24] Interestingly, only the enzyme from Burkholderia is not inactivated by 33% (v⁄ v) acetone and propan-2-ol

Uses in biocatalysis Many Actinobacteria are highly efficient in oxidizing the 3-OH group of D5-hydroxysteroids coupled to

D5fi D4 isomerization, thus providing valuable inter-mediates for industrial steroid drug production For example, R equi DSM 89-133 was used to convert cho-lesterol and other sterols to androst-4-ene-3,17-dione and androsta-1,4-diene-3,17-dione [25]: up to 82% of cholesterol was converted when the growth medium was supplemented with acetate The synthesis of 7b-hy-droxytestosterone was achieved by incubating a C8-ester

at the C-17 position of 5-androstene-3b,7b,17b-triol-17-caprylate with ChOx and subsequently eliminating the intermediate ester with porcine lipase [26] Analogously,

in the bioconversion of cholesterol into bile acids, ReChOx was used to convert 3b,7a-cholest-5-ene-3,7-diol to 7a-hydroxycholest-4-en-3-one, obtaining a yield higher than 90% [27] As another alternative, ShChOx was used to prepare unsaturated C21 triols as reference standards to study adrenal steroid production in Smith-Lemli-Opitz syndrome, with the ultimate aim of developing a prenatal or a postnatal diagnostic method [28] R erythropolis ChOx was employed for the preparative oxidation of cyclic allylic, bicyclic (e.g 10b-methyl-D1(9)-2b-octalol) and tricyclic alcohols, as well as of a synthon to synthesize several ergot alkaloids [29] This work also demonstrated a lack of enantiospec-ificity for the steroids Recently, ChOx from Chromo-bacterium DS-1 was reported to catalyze the oxidation

of cholesterol to hydroperoxy-cholest-4-en-3-one with the consumption of 2 mol of O2[24]

A

B

C

Fig 1 Comparison of the stability of ChOxs, from different

sources, in the absence (A) and in the presence of 0.1% (B) and

0.5% (C) Triton X-100 (black bars) or sodium cholate (white bars).

The enzyme (0.1 UÆmL)1) was incubated at 60 C for 60 min in

100 m M potassium phosphate, pH 7.0, before the enzymatic

activ-ity assay was performed: the residual activactiv-ity was calculated as a

percentage of the enzyme activity without heating and incubation.

Data for Pseudomonas (Pse), Nocardia (Noc), Rhodococcus (Rho),

Streptomyces (Str) and c-Proteobacterium Y-134 (Y-134) are from

Isobe et al [4] and for Chromobacterium sp DS-1 are from Doukyu

et al [24] Bre, BsChOx n.d., not determined.

With the ultimate goal of modifying the relative

ratios of specific plant sterols to stanols [i.e the

hydro-genated forms having the C-5 double bond reduced

(and which attracted attention because of their

benefi-cial effect in reducing serum and low-density

lipopro-tein cholesterol levels)], it is noteworthy that ShChOx

was used for engineering oil seeds from rapeseed and

soybean [30]

Clinical uses

ChOx is a useful analytic tool for determining

choles-terol in various samples: (a) total and esterified serum;

(b) from low-density lipoproteins to high-density

lipo-proteins; (c) on the cell membrane of erythrocytes (and

of other cells and cellular compartments); and (d) in

gall stones and in human bile

Normal human blood serum contains less than

5.2 mm (200 mgÆdL)1) of cholesterol; in plasma,

lipoproteins contain cholesterol and about 70% is

esterified by fatty acids Determining the concentration

of serum cholesterol is fundamental in the assessment

of a variety of diseases (e.g in atherosclerosis and

other lipid disorders) and for estimating the risk of

thrombosis, myocardial infarction, etc The risk for

Alzheimer disease is also related to

hypercholesterol-emia via mechanisms involving oxidative stress: this

disease is characterized by the accumulation of

amyloid b-peptide (a 39–43 amino acid peptide) in the

neocortex, which is connected to peroxidative damage

The amyloid b-peptide forms complexes with Cu2+

ions, which oxidize cholesterol into cholest-4-en-3-one,

thus mimicking the activity of ChOx In fact, brain

tissues from Alzheimer disease patients had a

cholest-4-en-3-one content  2-fold higher than brain tissues

from controls [31]

The chemical methods for determining serum

choles-terol (e.g Liebermann–Burchard reactions) were ousted

in 1973 by Richmond [32]: he had demonstrated that

ChOx from Nocardia can be used to measure serum

cholesterol by assaying the amount of hydrogen

perox-ide produced Hydrogen peroxperox-ide is reacted with

quadri-valent titanium and xylenol orange, yielding a colored

product measured at 550 nm This method relied on the

nonenzymatic hydrolysis (alkaline saponification) of the

cholesterol esters The same enzyme was then used for

the serum cholesterol assay by directly measuring the

absorbance of the product cholest-4-en-3-one at

240 nm, following a lengthy 2-hour incubation period

and subsequent product extraction [33] Total

choles-terol could also be determined in serum (including the

esters) by employing a totally enzymatic method in

which cholesterol hydrolase (EC 3.1.1.13) was used to

hydrolyze cholesterol esters [34] Free cholesterol is then oxidized by ChOx to produce hydrogen peroxide, which

is finally assessed enzymatically with horseradish peroxidase by the oxidative coupling of 4-aminoantipy-rine and phenol: this results in a quinoneimine dye (with

an absorption maximum at 500 nm) This method is characterized by the advantages of simplicity and little interference, and is highly reproducible: it is still routinely used in clinical laboratories The colorimetric method can also accurately determine serum cholesterol levels but, because of interference from pigments, it is difficult to assess bile cholesterol and thus an electro-chemical method based on oxygen consumption by ChOx reaction is better [35] The different methods of determining serum cholesterol are reported in Table 2 Recently, the performance of the end point versus the kinetic method for enzymatic assay of cholesterol was compared [36] The end point method is more accurate and precise; however, the kinetic method shows a lower sensitivity to interfering substances and analysis times are shorter

The method of Allain et al [34] for assaying high-density lipoprotein cholesterol was improved by employing polyethylene glycol-modified ChOx and cholesterol esterase [37] The serum, to which a-cyclo-dextran sulfate and a small amount of a-cyclo-dextran sulfate was added, was directly analyzed without precipitating lipoprotein micelles, and the results compared favour-ably with the previous method Subsequently, the same group reported on an automated method for measur-ing serum low-density lipoprotein cholesterol without ultracentrifugation separation [38] This result was achieved by using a nonionic surfactant, polyoxyethyl-ene-polyoxypropylene block copolyether, magnesium ions and a sodium salt of sulfated cyclic maltohexaose, a-cyclodextrin sulfate Polyethylene glycol-modified ChOx and cholesterol esterase were then used for the amperometric determination of high-density lipopro-tein cholesterol in 1–2 lL of serum [39]

For many years, ChOx was used in a great variety

of biosensors based on different detection systems For example, ChOx immobilized on a nylon membrane was used to build a fiberoptic biosensor based on the change in fluorescence of an oxygen-sensitive dye [40]; the analytical range is 0.2–3 mm, and the steady-state signal is achieved in 7–12 min Subsequently, a choles-terol sensor based on the electrochemical reduction of oxygen was developed using a bilayer-film coating: this sensor is less sensitive to organic interferences [41] ChOx was immobilized on nylon nets placed over the membrane of an oxygen probe or over the cellulose acetate membrane of a hydrogen peroxide detector and then covered with a polycarbonate membrane [42]

When incorporated in a flow injection system, the

hydrogen peroxide-based device was fast, reproducible

and linear over a 0.5–5 mm cholesterol concentration

range An amperometric biosensor was also obtained

by reconstituting the apoprotein of Pseudomonas

fluo-rescens ChOx with a FAD monolayer [43]: the sensor

shows high sensitivity and selectivity towards

electroac-tive interference Most recently, self-assembled

mono-layers have been used to improve response and binding

in biosensors based on immobilized ChOx The

prop-erties of biosensors for determining cholesterol using

ChOx immobilized by silane-based self-assembled

monolayers onto an indium-tin oxide surface have

been compared [44] An amperometric-rotating

biosen-sor using immobilized ChOx, cholesterol esterase and

peroxidase was also produced; this biosensor contains

a micropacked column with immobilized ascorbate

oxidase to eliminate l-ascorbic acid interference [45]

By use of this device, cholesterol can also be

deter-mined in the range of 1.2 lm–1 mm with a lifetime of

25 days of use Furthermore, ChOx was also used to

develop a sensor for determining total cholesterol in different food samples [46]

ChOx activity has also been employed to analyze cholesterol in cultured cells by means of an HPLC method At first, the cholesteryl esters were quantita-tively hydrolyzed using cholesterol esterase, and then total cholesterol was converted by Nocardia ChOx to cholest-4-en-3-one before detection at 240 nm In 3 min, free and total cholesterol can be measured in as few as

5000 cells of human monocyte-derived macrophages [47] Subsequently, the same assay was developed using

a fluorimetric method: cholesterol and related sterols have been assayed in sickle-cell and healthy erythrocytes [48] Other examples of cholesterol analysis in cells have been briefly described [3] Interestingly, the surface area

of a cell can be easily estimated by using ChOx to deter-mine cell-surface cholesterol For example, an area of

17500 lm2 was calculated for human fibroblasts from the cholesterol concentration in the plasma membrane (equivalent to 44 fmolÆcell)1) and a known choles-terol⁄ phospholipid molar ratio of 0.8 [49]

Table 2 Alternative ChOx-based methods for determining cholesterol concentration.

Cholest-4-en-3-one HPLC with detection under UV

light (240 nm)

0.15 nmol Determination of unsaponifiable

lipids and thoracic duct lymph chylomicrons

[75]

Hydrogen peroxide Amperometric assay using

immobilized ChOx and cholesterol esterase

1 lmolÆL)1 Determination of both free and

total cholesterol in blood serum

[76]

Hydrogen peroxide Chemiluminescent detection of

the reaction using lucigenin

2.6 lmolÆL)1 Requires accurate adjustment of

the pH to 11.75–11.9

[77] Hydrogen peroxide Temperature-enhanced

chemiluminescent flow system using bis-(2,4,6-trichlorophenyl) oxalate-perylene in the presence

of Triton X-100 (detection

at 500 nm)

19 lmolÆL)1 Significant increase in the

signal-to-noise ratio

[78]

Hydrogen peroxide Fluorimetric determination (excitation

320 nm; emission 400 nm) using 4-hydroxyphenylacetic acid

0.5 nmolÆL)1 Measurements can be performed

on 1 lL of serum

[79]

Hydrogen peroxide Spectrophotometric determination

of the reaction with 4-aminophenazone plus phenol (at 520 nm) using immobilized ChOx, cholesterol esterase and horseradish peroxidase on arylamine glass beads

54 lmolÆL)1 The co-immobilized enzymes did

not show loss of activity after 300 uses

[80]

Electron transfer from ChOx Electrochemical determination

using electron mediators (i.e.

1-methoxy-5-methylphenazinium, thionine, etc.)

0.25–2.5 mmolÆL)1 The measurement is altered by the

dissolved oxygen and requires

‡ 10 min to obtain a steady-state response No effect of classical interference was observed in the colorimetric assays.

[81]

ChOx in agriculture

Genetically modified plants that produce insecticidal

proteins (e.g the Bacillus thuringiensis toxin) are now

available to control insect pests of several major crops

In 1993, from a random screening of > 10000 filtrates

from microbial fermentations, Monsanto Co (St Louis,

MO, USA) discovered a highly efficient protein in

cul-ture filtrates that killed boll weevil (Anthonomus grandis

Boheman) larvae [50] Histological and biochemical

studies identified the protein as ChOx: purified ChOx is

active against boll weevil larvae at a 50% lethal

con-centration (LC50) of 20.9 lgÆmL)1, which is comparable

to the bioactivity of B thuringiensis proteins against

other insect pests The ChOx gene (Monsanto named it

choM, see Fig S1), from the Streptomyces sp strain

A19249, exhibits 85% DNA sequence identity and

89% amino acid sequence similarity with known

Strep-tomyces sp SA-COO (choA gene) In addition to boll

weevil, several lepidopterans were negatively affected

by the presence of ChOx at a dietary concentration of

0.001% Adding the products of cholesterol oxidation

by ChOx (i.e cholest-en-3-one and hydrogen peroxide)

to the diet, and pretreating the diet with the enzyme,

excluded insecticidal effects caused by the ingestion of

toxic compounds However, the boll weevil larvae are

acutely sensitive to ingested ChOx because it induces

lysis at the midgut epithelium Boll weevil adults are

insensitive to ingested ChOx, although the fecundity of

adult females was greatly reduced if 50 lgÆmL)1of the

enzyme was present in the diet [51] ChOx reduced

sub-sequent oviposition (up to 83% in eggs laid) and larval

survival (97% reduction as compared to controls)

because of poorly developed ovaries and few

develop-ing oocytes ChOx was expressed in transformed

tobacco plants, and the synthesis levels in leaf tissues

routinely ranged from approximately 5–50 lg of

enzyme per g fresh weight In the absence of a

chloro-plast-targeting sequence, ChOx production resulted in

severe abnormalities in plant development and fertility

When produced as a fusion with a chloroplast-targeting

peptide, synthesis of the mature and the full-length

enzyme did not cause the deleterious phenotypic effect

observed with untargeted ChOx [52] Transgenic leaf

tissues expressing ChOx exerted insecticidal activity

against boll weevil larvae When produced in the

cyto-sol, or when targeted to chloroplasts, ChOx

metabo-lizes phytosterols in vivo Transgenic plants expressing

ChOx in cytosol accumulated low levels of saturated

sterols (stanols), while the transgenic plants expressing

chloroplast-targeted ChOx maintained a greater

accu-mulation of stanols and appeared phenotypically and

developmentally normal It was proposed that ChOx

could modify sterol ratios, thus influencing cell division, or could affect brassinosteroid biosynthesis in steroid hormones

ChOx in virulence ChOx is an interesting pharmaceutical target for treat-ing bacterial infections R equi is a Gram-positive coc-cobacillus that resides within macrophages of the host

It is a common soil organism that frequently infects young horses; the most common manifestation is a chronic suppurative bronchopneumonia with abscess formation and cavitary pneumonia (for a review, see [53]) Since 1967, this organism has also been reported

to infect humans, being frequently diagnosed as an opportunistic pathogen in immunocompromised patients,  70% of whom are infected with HIV [54] The clinical manifestations of R equi infections are different: the most frequent form is severe pyrogranul-omatous pneumonia Among the candidate virulence factors of this pathogenic actinomycete, in vitro data suggested that during R equi infection of the host cell, membrane lysis is facilitated by the induction of extra-cellular ChOx [55] Mutational analysis indicated that ChOx is the membrane-damaging factor responsible for the synergistic hemolytic reaction elicited by

R equi in the presence of sphingomyelinase C-produc-ing bacteria, such as Listeria ivanovii, Bacillus cereus and Staphylococcus aureus [55] The membrane-damag-ing activity of R equi requires the presence of bacterial sphingomyelinase C, thus indicating that the ChOx substrate is not directly accessible to the enzyme in intact membranes A detailed description of the role of ChOx in the virulence of R equi, as well as recently reported criticisms [56], has been reported previously [2]

ChOx as antifungal sensor Streptomyces natalensis ChOx (encoded by the pimE gene; Fig S1) plays a main role in the biosynthesis of the polyene macrolide pimaricin [57] (the details of ChOx involvement in this biosynthetic process have been reported previously [2]) This 26-member tetraene macrolide antifungal antibiotic is widely used in the food industry (to prevent contamination of cheese and other nonsterile food with mold) and in the treatment of fungal keratitis because it interacts with membrane ster-ols (ergosterol is the major sterol found in fungal mem-branes), altering the membrane structure and causing cell leakage Interestingly, putative ChOx-encoding genes are present in other known biosynthetic gene clus-ters of antifungal polyketides, such as filipin (pteG) and

rimodicin⁄ CE-108 (rimD) All these polyketides are

elicited by soil bacteria against their fungal competitors,

whose membranes contain ergosterol, representing a

selective advantage for the producing organism

Protein engineering studies

Improvement in thermal stability

The temperature sensitivity of ChOxs from a number

of sources has been compared recently [23] The

ther-mal stability (as well as the stability at alkaline pH) of

ShChOx is not optimal for use as a diagnostic enzyme

(discussed above) In order to improve its properties, a

random mutagenesis approach by error-prone PCR of

the choA gene was used [58] Following screening of the

mutant libraries at 50C, four single point mutants

(S103T, V121A, R135H and V145E ShChOx, see the

sequence numbering reported in Fig S1) were isolated

whose kinetic properties resemble those of the wild-type

enzyme but whose thermal stability is increased In

par-ticular, the V145E ShChOx shows a marked increase in

thermal stability and an enlarged range of optimal pH (from acid to alkali).The single mutations identified by random mutagenesis were combined All the mutations, except for R135H, had an additive effect: the double mutant S103T⁄ V145E was the most improved ShChOx because it shows unaltered activity after 4 weeks at

40C Concerning the rationale of the observed changes, the Cc atom of the threonine residue in S103T ShChOx seems to be accommodated in a relatively large cavity, thus strengthening the atomic packing (Fig 2A) The V145E substitution introduces an addi-tional hydrogen bond and a salt bridge with the side chain of R417, a residue close to the FAD cofactor These new interactions could help to stabilize the native conformation, thus increasing the thermal resistance and yielding a marked change in the optimal pH BsChOx is a monomeric flavoenzyme containing one molecule of FAD covalently linked to H69 (mature protein numbering, corresponding to H121 on the full-length polypeptide; Fig S1) The covalent link

is eliminated by the H69A substitution: wild-type and H69A BsChOx do not show significant structural

A

B

Fig 2 (A) Details of the substitutions

intro-duced in ShChOx (Protein Data Bank, PDB

code: 1mxt) by random mutagenesis and

resulting in a higher stability to temperature

and to pH [58] Mutated residues are

col-ored according to the reporting authors as in

Fig S1 FAD (yellow) and

3-beta-hydroxy-5-androsten-17-one (AND, gray) are shown in

CPK representation Left: details of the

regions surrounding two selected mutations,

as reported by Nishiya et al [58] The

molecular surface is shown in light brown.

(B) Comparison of the active site of ReChOx

(PDB code: 1coy) and the active site of

ShChOx (PDB code: 1mxt), highlighting the

position important for substrate specificity.

Residues of the ReChOx active site are

numbered as reported by Li et al [68];

residues of the ShChOx active site are

numbered as reported in Lario et al [69]

(representation licorice) to which were

added residues mutated as described in

Xiang & Sampson [64] (representation

ball-and-stick) Proposed hydrogen bonds are

represented in pink The ligand AND was

placed in the ShChOx active site according

to the position occupied in the active site of

ReChOx.

differences [59] The mutant enzyme retains the

oxida-tion and the isomerizaoxida-tion catalytic activities

How-ever, the H69A BsChOx shows a 35-fold decrease in

maximal oxidation activity and a flavin midpoint

reduction potential of 100 mV lower ()204 mV

ver-sus )101 mV for wild-type ChOx) [60] We

demon-strated that the covalent bond of the flavin in ChOx

represents a structural device for stabilizing the protein

tertiary structure [61] In fact, the urea-induced

unfold-ing of H69A BsChOx occurred at significantly lower

urea concentrations ( 2 m lower), as did the

tempera-ture-induced unfolding ( 10–15 C lower), than for

the wild-type enzyme

Alteration of the substrate specificity

Table 1 describes the substrate preference of some

ChOxs (expressed as relative activity with respect to

cholesterol) In vivo random mutagenesis of the choA

gene decreased the Kmvalue for cholesterol by 10-fold

without changing maximal activity: the evolved

ShChOx mutant contained the V145E⁄ G405S

substitu-tions [62] Based on sequence comparison between

Streptomyces and Rhodococcus ChOxs, six positions

were mutated in ShChOx [63] Among the mutants

generated, the S379T variant showed a twofold higher

activity towards pregnenolone than cholesterol as a

substrate (interestingly, a threonine is present in the

same position in ReChOx, i.e., T387; see Fig 2B) The

catalytic efficiency (kcat⁄ Km) of this mutant enzyme for

cholesterol and pregnenolone was two- and sixfold

higher, respectively, than that of the wild-type enzyme

[63] Noteworthy is that the cavity around the side

chain of S379 in ShChOx is larger than that

surround-ing T387 in ReChOx (Fig 2B) Structure-based

rational design of ShChOx was also used to identify

the residues involved in substrate binding [64]: M58,

L82, V85, M365, and F433 were identified as being in

direct contact with the C17 sterol tail (these residues

correspond to M95, L119, V122, M402 and F470 in

the full-length sequence reported in Fig S1) Each

position was degenerated using NYS (N = A, C, G,

T; Y = C, T; S = C, G) codons to limit amino acids

to predominantly hydrophobic residues The

L82A⁄ V85T ⁄ F433L ShChOx mutant showed a twofold

increase in the kcat⁄ Kmratio with b-sitosterol and

stig-masterol, which differ from cholesterol for the

extended C17 tail

By deleting the five residues belonging to the

active-site loop in ShChOx (mutant D79-83

corre-sponding to sequence 115-119 in Fig S1), the kinetic

efficiency (kcat⁄ Km ratio) was increased 5.6-fold for

dehydroepiandrosterone compared with the wild-type

enzyme, while the same ratio decreased 170-fold with micellar cholesterol and up to 2800-fold with choles-terol in a vesicle [2] These results also suggest that the tip of the active-site loop is necessary for the packing with the C17-tail, being mainly responsible for the substrate specificity at this position of ChOx [65]

The oxidation and isomerization reactions catalyzed

by ShChOx could be separated by site-directed muta-genesis The N480A⁄ Q mutants (corresponding to position N522 in the full-length sequence reported in Fig S1) showed no oxidative activity but retained the ability to isomerize cholest-5-en-3-one into cholest-4-en-3-one [66] However, the E361D ShChOx mutant (corresponding to the E356D mutant in [66] and to position 398 in the full-length sequence; Fig S1) retained the oxidative activity, but not the isomeriza-tion activity [67] Furthermore, the kcat⁄ Km ratio for the oxidation reaction was increased sixfold in the E356D ShChOx mutant and the kcat⁄ Km ratio for the isomerization reaction was increased threefold in the N480A mutant

Conclusions

Certainly, future molecular biology, biochemical and structural investigations will enable us to clarify the role that ChOx activity plays in different microorgan-isms and also to identify new functional roles for this flavooxidase Furthermore, in order to comprehend why this enzyme exists as a flavoprotein containing noncovalently or covalently linked FAD, we will also need to understand its physiological role These inves-tigations will be carried out whilst optimizing known applications of the ChOx reaction and developing new, and more sophisticated, biotechnological uses of this flavoenzyme The latter hinges both on the discovery

of novel ChOx activities (see the recent identification

of ChOx from thermophilic bacteria) and on the engi-neering of enzyme variants We assume that the num-ber of investigations of ChOx will continue to grow quickly in the near future, in particular to identify unexpected areas of application

Acknowledgements

This work was supported by grants from Fondo di Ateneo per la Ricerca We are grateful for the support from Consorzio Interuniversitario per le Biotecnologie (CIB) and from the Research Center of Biotecnologie per la Salute Umana We would like to apologize to many colleagues whose work we could not discuss in detail because of space limitations