New Per Ox i Some Complete

7/30/2019 New Per Ox i Some Complete

1/41

Biochemistry of MammalianPeroxisomes Revisited

Ronald J.A. Wanders and Hans R. Waterham

Department of Clinical Chemistry and Pediatrics, Laboratory Genetic MetabolicDisease, Academic Medical Center, University of Amsterdam, 1105 AZ Amsterdam

The Netherlands; email: [email protected], [email protected]

Annu. Rev. Biochem.2006. 75:295332

First published online as aReview in Advance on

March 22, 2006

The Annual Review of

Biochemistry is online atbiochem.annualreviews.org

doi: 10.1146/annurev.biochem.74.082803.133329

Copyright c 2006 byAnnual Reviews. All rightsreserved

0066-4154/06/0707-0295$20.00

Key Words

fatty acid oxidation, plasmalogens, reactive oxygen species, genediseases

Abstract

In this review, we describe the current state of knowledge abthe biochemistry of mammalian peroxisomes, especially human poxisomes. The identification and characterization of yeast muta

defective either in the biogenesis of peroxisomes or in one of

metabolic functions, notably fatty acid beta-oxidation, combinwith the recognition of a group of genetic diseases in man, wher

these processes are also defective, have provided new insightsall aspects of peroxisomes. As a result of these and other stud

the indispensable role of peroxisomes in multiple metabolic paways has been clarified, and many of the enzymes involved in th

pathways have been characterized, purified, and cloned. One asp

of peroxisomes, which has remained ill defined, is the transportmetabolites across the peroxisomal membrane. Although it is cl

that mammalian peroxisomes under in vivo conditions are clostructures, which require the active presence of metabolite tran

porter proteins, much remains to be learned about the permeabiproperties of mammalian peroxisomes and the role of the four h

ATP-binding cassette (ABC) transporters therein.

295


2/41

Contents

INTRODUCTION.. . . . . . . . . . . . . . . . 296

PEROXISOMAL PROTEINS . . . . . . 297When to Call a Protein a

Peroxisomal Protein? . . . . . . . . . . 297

Strategies to Identify Putative

Peroxisomal Proteins . . . . . . . . . . 300Strategies to Demonstrate the

Peroxisomal Localization of

Proteins . . . . . . . . . . . . . . . . . . . . . . 300SELECTED METABOLIC

PATHWAYS . . . . . . . . . . . . . . . . . . . . . 301

Oxygen Metabolism, ReactiveOxygen Species, and Reactive

Nitrogen Species Metabolism . . 301Ether-Phospholipid Biosynthesis . 302

Peroxisomal Fatty Acid

Beta-Oxidation... . . . . . . . . . . . . . 305Peroxisomal Fatty Acid

Alpha-Oxidation . . . . . . . . . . . . . . 311Glyoxylate Metabolism . . . . . . . . . . . 312

Amino Acid Catabolism. . . . . . . . . . . 313Pentose Phosphate Pathway. . . . . . . 313

Polyamine Oxidation . . . . . . . . . . . . . 313Miscellaneous Peroxisomal

Enzyme Activities . . . . . . . . . . . . . 314

Isoprenoid and CholesterolMetabolism . . . . . . . . . . . . . . . . . . . 314

BIOCHEMISTRY OF HUMANPEROXISOMAL DISORDERS. . 315

MOUSE MODELS FOR

PEROXISOMAL DISORDERS. . 316PEROXISOMAL METABOLITE

TRANSPORT . . . . . . . . . . . . . . . . . . 319Permeability Properties of

Peroxisomes . . . . . . . . . . . . . . . . . . 319The Intraperoxisomal pH . . . . . . . . . 322

Peroxisomal ABC Transporters. . . . 322

Peroxisomal ATP Transporter . . . . . 324Other Putative Peroxisomal

Transporters . . . . . . . . . . . . . . . . . . 325CONCLUDING REMARKS . . . . . . . 325

INTRODUCTION

Peroxisomes belong to the microbody fam-

ily, with glyoxysomes and glycosomes as theother members, and represent a class of ubiq-

uitous and essential cell organelles characterized by the presence of a proteinaceous matrix

surrounded by a single membrane. Since this

topic was last reviewed in this journal in 1992(1), our knowledge about the biochemistry o

peroxisomes has increased substantially for anumber of different reasons. First, the identi-

fication and characterization of yeast mutantsdefective in peroxisome biogenesis have al-

lowed the resolution of the principal features

of peroxisome biogenesis, which includes thetargeting of peroxisomal matrix proteins via

one of two distinct peroxisomal targetingsignals (PTS1 and PTS2). This knowledge

has been used to perform computer-basedsearches for proteins that contain a PTS1 or a

PTS2 notably in the yeastSaccharomyces cere-

visiae and to identify the corresponding mam-malian orthologues, using homology probing

This strategy, together with more classical ap-proaches, such as protein purification, has led

to the identification of a series of peroxisomaproteins. Second, the recognition of a large

class of genetic diseases in man, in which ei-

ther peroxisome biogenesis per se or a certainperoxisomal function is defective, has pro-

vided new insights into the metabolic role ofperoxisomes in humans. As a result of these

combined studies, it is nowclear that fatty acid(FA) beta-oxidation is a general feature of vir-

tually all types of peroxisomes. In addition

peroxisomes in higher eukaryotes, includinghumans, catalyze a number of additional per-

oxisomal functions not shared by peroxisomesin lowereukaryotes, including etherphospho

lipid biosynthesis, FA alpha-oxidation, andglyoxylate detoxification.

Another major step forward has been

the discovery that, in contrast to the long-held view that the peroxisomal membrane

is freely permeable to low-molecular-weightsubstances, peroxisomes are in fact closed

structures under in vivo conditions. This

296 WandersWaterham


3/41

requires the presence of peroxisomal mem-

brane proteins, which allow the specific en-trance and exit of metabolites. Some progress

has been made in this respect with the iden-tification of a set of half ATP-binding cas-

sette (ABC) transporters, possibly involved intransmembrane FA transport.

In this review, we present the current stateof knowledge about (a) the biochemistry ofperoxisomes, with special emphasis on the en-

zymology and metabolic functions of mam-mals, notably in humans; (b) the transport

properties of mammalian peroxisomal mem-branes; and (c) the biochemistry in patients

suffering from different peroxisomal disor-

ders and corresponding mouse models.

PEROXISOMAL PROTEINS

It has become clear that peroxisomes are in-volved in a variety of metabolic pathways,

which implies the presence of a large number

of proteins in the peroxisomal matrix. Belowwe define criteria for the peroxisomal localiza-

tion of proteins and discuss some commonlyused strategies to demonstrate the peroxiso-

mal localization for a given protein.

When to Call a Protein aPeroxisomal Protein?

It is estimated that mammalian peroxisomescontain some 50 different enzyme activities,

many of which have been attributed to es-

tablished peroxisomal proteins (Table 1). Inaddition, a number of peroxisomal proteins

have been reported without a known catalyticactivity. This includes the peroxisomal pro-

tein PeP, encoded byFNDC5, with no signif-icant homology to any known protein except

for a short stretch of amino acids containingthe fingerprint of the fibronectin type III su-perfamily (2). In addition, a number of per-

oxisomal proteins have been identified witha catalytic activity of unknown function, as,

for example, the peroxisomal nudix hydrolaseNudt7 (3). Furthermore, some enzyme activ-

ities are not yet linked to a specific peroxiso-

FA: fatty acid

mal protein. These activities may be catalyzed

by new, yet unidentified peroxisomal proteinsor may be a side reaction of a known pro-

tein, as shown, for instance, for 2,4-dienoyl-coenzyme A (CoA) reductase, which also cat-

alyzes the NADPH-dependent reduction ofretinal to retinol (4), at least in vitro, and the

peroxisomal enzyme D-bifunctional protein,which has abundant 17 beta-estradiol dehy-drogenase activity under in vitro, but not in

vivo, conditions. The latter data further im-ply that the identification of a certain enzyme

activity in peroxisomes under in vitro condi-tions does not necessarily imply that peroxi-

somes also catalyze this activity under in vivo

conditions.Most enzyme activities listed in Table 1

are unique to peroxisomes, but some are

shared with other subcellular compartments,including the mitochondria and cytosol. Sucha multiple subcellular localization may be due

to the existence of different isoforms of a pro-

tein targeted to different subcellular sites, as,for example, is the case for the enzymes in-

volved in FA beta-oxidation in mammals (per-oxisomal and mitochondrial), or the presence

of multiple targeting signals within the sameprotein, as, for example, shown for 3-hydroxy-

3-methylglutaryl-CoA lyase (5) and alpha-

methylacyl-CoA racemase (AMACR) (68).These enzymes are located in both peroxi-

somes and mitochondria.Conclusive evidence for the peroxiso-

mal localization of a certain protein and/orits activity requires the actual identification

and characterization of the protein and the

demonstration of its physical presence insidethe organelles. In this respect, it should be

noted that the peroxisomal localization of acertain protein may be species dependent.

Furthermore, the finding of a peroxisomal lo-calization in one species does not necessarily

mean that its homologue is also peroxisomal

in other species. A well-documented exampleof this is alanine:glyoxylate aminotransferase

(AGT), which is peroxisomal in humans, rab-bits, and guinea pigs, has a dual peroxiso-

mal and mitochondrial location in rats and

www.annualreviews.org Biochemistry of Peroxisomes 297


4/41

Table1

Listofperoxisomalenzymesandotherperoxisomalproteinsfromhumansandtheirperoxisomaltargetingsequences(PT

S1orPTS2)

Peroxisomal(enzyme)protein

Gene

symbol

Enzyme

symbol

EC

number

Humanlocus

Targeting

signal

PTS1

or

PTS

2

Targeting

sequence

Peroxisomalbeta-oxidation

Acyl-CoAoxidase1(palmitoyl-CoAoxidase)

AC

OX1

ACOX1

1.3.3.6

17q25

PTS1

-SKL

Acyl-CoAoxidase2(branched-chainacyl-CoAoxidase)

AC

OX2

ACOX2

1.3.3.6

3p14.3

PTS1

-SKL

Acyl-CoAoxidase3(pristanoyl-CoAoxidase)

AC

OX3

ACOX3

1.3.3.6

4p15.3

PTS1

-SKL

L-bifunctionalprotein(

peroxisomalmultifunctional

enzyme1)

EH

HADH

LBP/MFP1

1.1.1.35;

5.3

.3.8;

4.2

.1.17

3q26.33q28

PTS1

-SKL

D-bifunctionalprotein(peroxisomalmultifunctional

enzyme2)

HSD17B4

DBP/MFP2

4.2.1.-;

1.1

.1.35

5q2

PTS1

-AKL

Peroxisomalbeta-ketoth

iolase1(straight-chainthiolase)

AC

AA1

2.3.1.16

3p23-p22

PTS2

-RLQVVLGHL

Peroxisomalbeta-ketoth

iolase2(branched-chain

thiolase)

SC

P2

SCP2

1p32

PTS1

-AKL

Alpha-methylacyl-CoAracemase

AM

ACR

AMACR

5.1.99.4

5p13.2q11.1

PTS1

-(K)ASL

Carnitineacetyltransferase

CR

AT

CAT

2.3.1.7

9q34.1

PTS1

-AKL

Carnitineoctanoyltransferase

CR

OT

COT

7q21.1

PTS1

-THL

Delta3,5-,delta2,4-dien

oyl-CoAisomerase

EC

HI

19q13.1

PTS1

-SKL

Peroxisomal2,4-dienoyl-CoAreductase2

DECR2

16p13.3

PTS1

-AKL

Peroxisomal3,2-trans-e

noyl-CoAisomerase

PE

C1

6p24.3

PTS1

-SKL

Very-long-chainacyl-CoAsynthetase

SL

C27A2

VLCS

6.2.1.-

15q21.2

PTS1

-LKL

Acyl-CoAthioesterase2

PT

E1

3.1.1.2

20q12q13.1

PTS1

-SKL

Acyl-CoAthioesterase1

B

PT

E2

3.1.1.2

14q24.3

PTS1

-SKV

Peroxisomaltrans-2-eno

yl-CoAreductase(NADPH)

PE

CR

2q35

PTS1

-AKL

Peroxisomalalpha-oxida

tion

Phytanoyl-CoA2-hydro

xylase

PH

YH/PAHX

PHYH/PAHX

1.14

.11.18

10pterp11.2

PTS2

-RLQIVLGHL

2-Hydroxyphytanoyl-CoAlyase

HPCL2

HPCL2

3p25.1

PTS1

-(R)SNM

Plasmalogenbiosynthesis

Dihydroxyacetonephosphateacyltransferase

GNPAT

DHAPAT

2.3.1.42

1q42.1142.3

PTS1

-AKL

Alkyldihydroxyacetonephosphatesynthase

AG

PS

ADHAPS

2.5.1.26

2q31

PTS2

-RLRVLSGHL

Fattyacyl-CoAreductase1

MLSTD2

FAR1

11p15.2

Fattyacyl-CoAreductase2

MLSTD1

FAR2

12p11.22

298 WandersWaterham


5/41

Glyoxylatemetabolism

Alanine:glyoxylateamin

otransferase

AG

XT

AGT

2.6.1.44;

2.6

.1.55

2q36q37

PTS1

-KKL

Lysinemetabolism

Peroxisomalsarcosineo

xidase/L-pipecolateoxidase

PIPOX

PIPOX

17q11.2

PTS1

-AHL

Oxygenmetabolism

Catalase

CA

T

CAT

1.11

.1.6

11p13

PTS1

-(K)ANL

PeroxiredoxinV(PMP2

0)

PR

DX5

PROX5/PMP20

1.11

.1.7

11q13

PTS1

-SQL

d-aminoacidoxidase

DAO

DAOX

1.4.3.3

12q24

PTS1

-SHL

d-aspartateoxidase

DDO

DASPOX

1.4.3.1

6q21

PTS1

-(K)SNL

Glycolateoxidase(hydroxyacidoxidase1)

HAO1

GOX/HAO1

1.1.3.15

20p12

PTS1

-SKI

Hydroxyacidoxidase2

HAO2

HAO2

1.1.3.15

1p13.3p13.1

PTS1

-SRL

Hydroxyacidoxidase3

HAO3

HAO3

1.1.3.15

-

PTS1

-SRL

Epoxidehydrolase

EP

HX2

EPH2

3.3.2.3

8p21-p12

PTS1

-SKM

GlutathioneS-transferaseclassKappa

GSTK1

GSTK1

7

PTS1

-ARL

Polyaminemetabolism

N1-acetylspermine/sper

midineoxidase

PA

OX

PAO

10q26.3

PTS1

-(R)PRL

Additional(enzyme)pro

teins

Malonyl-CoAdecarboxylase

MLYCD

4.1.1.9

16q12

PTS1

-SKL

3-Hydroxy-3-methylglu

taryl-CoAlyase

HMGCL

HL

4.1.3.4

1p36.1p35

PTS1(+

MTS)

-CKL

Isocitratedehydrogenase(NADP+-linked)

ID

H1

IDH1

1.1.1.42

2q33.3

PTS1

-AKL

NudixhydrolasespecificforCoA

NUDT7

NUDT7

16q23.1

PTS1

-SRL

Insulin-degradingenzym

e

ID

E

IDE

3.4.24.56

10q23q25

PTS1

-AKL

Serinehydrolaselike

SE

RHL

22q13.2

q13.31

Lonprotease

LO

NP

16q12.1

PTS1

-SKL

Nudix-typemotif19(Roswell-Parkcomplex2)

D7RP2e

RP2p

19q13.11

PTS1

-SHL

Trim37

TR

IM37

6p21.3

PeP

FN

DC5

FNDC5/PeP

1p35.1

PTS1

-SKI

PMP22

PX

MP2

PMP22

12q24.33



6/41

marmosets, and is mitochondrial in cats (9).

Another example is the localization of the FAbeta-oxidation pathway, which is exclusively

peroxisomal in yeast but shows a dual peroxi-somaland mitochondrial localization in mam-

mals and plants (10).

Strategies to Identify PutativePeroxisomal Proteins

Both forward and reverse genetics approacheshave been employed to identify putative per-

oxisomal proteins. In short, the forward ge-netics approach involves the conventional

purification of a protein, which exhibits a cer-

tain enzyme activity assumed to be perox-isomal, followed by the generation of spe-

cific antibodies against the purified protein,

which can be used to demonstrate its actualsubcellular localization. Furthermore, the en-coding cDNA and corresponding gene can

be identified by a degenerated polymerase

chain reaction approach that is based on apartial amino acid sequence of the purified

protein. The reverse genetics approach be-came popular because of the availability of the

genomic sequences of various organisms, in-cluding yeasts, mice, and humans, in conjunc-

tion with the increased knowledge of func-

tional domains within amino acid sequences(e.g., catalytic sites and targeting signals), al-

lowing selective database searches for genesencoding putative peroxisomal proteins with

certain activities. This in silico strategy hasbeen very successful in identifying novel puta-

tive peroxisomal proteins [e.g., proteins with

consensus peroxisomal targeting signals (11)]as well as orthologues (i.e., functional homo-

logues identified on the basis of significant se-quence similarity) of proteins determined to

be peroxisomal in other species, a strategy alsoreferred to as homology probing (12, 13).

Strategies to Demonstrate thePeroxisomal Localization of Proteins

The most important criterion for the per-

oxisomal localization of a certain protein

remains that the protein should be shownas physically present inside peroxisomes

This is true particularly for candidate per-

oxisomal proteins identified by the reversegenetics approach, because the presence of a

consensus PTS in the amino acid sequencei.e, a PTS1-consensus sequence: (S/A/C)

(K/R/H)-(L/M) or a PTS2-consensus sequence: (R/K)-(L/I/V)-X5-(Q/H)-(L/I/V)does not necessarily mean that the protein

is truly peroxisomal. For example, althoughboth pristanoyl-CoA oxidase and the bile acid

conjugating (BACAT) enzyme harbor thesame C-terminal PTS1-like SQL tripeptide

only pristanoyl-CoA oxidase is peroxisoma

and BACAT cytosolic (14, 15). Also, despitethe presence of a consensus PTS1 (-SRL)

phosphomevalonate kinase was recently

demonstrated to be a cytosolic protein (16).Several approaches can be employed

to demonstrate a peroxisomal localization

These include conventional subcellular frac-

tionation studies by which tissues or cells arefirst homogenized and separated by differ-

ential centrifugation into organelle-enrichedfractions that are further fractionated by den-

sity gradient centrifugation. The fractions ofthese gradients are then analyzed by specific

enzyme activity measurements and/or im-

munoblotanalysis,using specific antibodies todeterminethe profileof the protein of interes

in the gradient in comparison to the profileof established marker proteins/enzymes for

the various subcellular organelles. As activitymeasurements may not always be specific be-

cause enzymes often can handle multiple sub-

strates, the combined approach with specificantibodies is highly recommended.

The peroxisomal localization of solubleproteins can also be studied through con-

trolled permeabilization of cellular mem-branes with digitonin. Because digitonin per-meabilizes cellular membranes forming a

complex with cholesterol and organelle mem-branes contain lower levels of cholestero

than the plasma membrane, this treatmentleads to differential leakage of proteins. This

leakage can be assessed by determining the

300 WandersWaterham


7/41

releaseoftheproteinofinterestincomparison

to that of established marker proteins usingspecific enzyme activity measurements and/or

immunoblot analyses using specific antisera(see, for examples, References 16, 17, and 18).

Conclusive evidence for the peroxisomallocalization of a protein can be obtained by

in situ (immuno) microscopical techniques,including immunoelectron microscopy, im-munofluorescence microscopy, and immuno-

histochemistry. Because the outcome of thesetechniques relies heavily on the quality of an-

tisera used to detect the protein, it is essen-tial to obtain antisera that are highly spe-

cific and exclusively recognize the protein

of interest under native conditions. The im-portance of performing immunolocalization

studies to establish the peroxisomal localiza-

tion of a certain protein using specific anti-bodies has been shown on numerous occa-sions. This is exemplified by recent studies

by Yokota and coworkers (19), showing that

the NADP-linked isocitrate dehydrogenaseencoded by IDH1 is predominantly, if not

exclusively, peroxisomal, whereas subcellu-lar fractionation studies, based on differen-

tial centrifugation of homogenates, revealedthat the enzyme is predominantly cytosolic

and only partially peroxisomal (19).

The functionality of putative PTS1 orPTS2 motifs in the amino acid sequence of

a protein can be tested by reporter studiesin which the protein or portions thereof are

fused to a reporter protein such as green fluo-rescent protein or specific epitopes (e.g., myc,

HA). This allows easy detection of the protein

constructs upon expression in cells. It shouldbe noted, however, that the observation that a

certain (truncated) amino acid sequence is ca-pable of targeting a reporter to peroxisomes

does not necessarily imply that it also func-tions as a true PTS in the authentic protein if

the latter has not been shown to actually re-

side in peroxisomes by other means. Indeed,a putative C-terminal PTS1 may not be func-

tional when, for example, the PTS1 is hiddenwithin the three-dimensional structure of the

protein or when, in addition, an N-terminal

mitochondrial targeting sequence is presentin the same protein. Furthermore, great care

should be taken with the interpretation of re-

sults obtained with reporter studies becausein most cases these involve overexpression,

which may introduce unpredictable artifacts(16, 20). It should be noted that the specific

context of a putative PTS is important fortargeting, by increasing the affinity between

the PTS-containing peptide and its receptor

(2124). For instance, it has been shown forcatalase, which ends in ANL, a weak PTS1,

that the lysine present in the (4) position ofcatalase (-KANL) greatly stimulates binding

to the PTS1-receptor (22). Optimized aminoacid residues at positions (4) and (5) can en-

hance affinities by at least two orders of mag-

nitude (23).

SELECTED METABOLICPATHWAYS

Below we describe the role of peroxisomes inselected metabolic pathways.

Oxygen Metabolism, ReactiveOxygen Species, and Reactive

Nitrogen Species Metabolism

Peroxisomes harbor a number of oxidases thatreduce O2 to H2O2 (25). The H2O2 produced

can be disposed of via several enzymes, in-cluding catalase, glutathione peroxidase, and

peroxiredoxin V (PMP20). The decomposi-

tion of H2O2 by catalase may occur catalyt-ically (2H2O2 O2 + 2H2O) or peroxi-

datically (H2O2 + AH2 A+ 2H2O), inwhich the conversion of one molecule H2O2to two molecules of H2O is coupled to the ox-idation of different hydrogen donors (AH2),

such as ethanol, methanol, formaldehyde, for-mate, and nitrite. In addition to catalase, per-oxisomes also contain glutathione peroxidase

activity (26). A third peroxisomal enzyme thatremoves H2O2 is PMP20 (27), which exhibits

thiol-specific antioxidant activity.Apart from H2O2, peroxisomal en-

zymes also generate other reactive species,



8/41

including superoxide anions. One source of

superoxide anions is xanthine oxidoreductase,which can exist in two forms, including

a dehydrogenase and oxidase form (thelatter form generates superoxide anions).

Angermuller et al. (28) were the first toidentify xanthine oxidase activity in the core,

but not in the matrix, of peroxisomes. Recentstudies in rat liver in which use was made ofimproved methods applied to unfixed cryostat

sections have shown that xanthine oxidase isnot only present in the core of peroxisomes

but also in the peroxisomal matrix (29).Furthermore, xanthine oxidase appears to be

the predominant, if not exclusive, form of

xanthine oxidase in peroxisomes, whereas inthe cytosol, the reverse is true with xanthine

dehydrogenase predominating over xanthine

oxidase.Inactivation of superoxide anions is

brought about by superoxide dismutases. Sev-

eral reports have shown the presence of

Cu/Zn-SOD (30, 31) and Mn-SOD activi-ties (32) in peroxisomes, although theproteins

responsible for these activities remain to beidentified. It has also been claimed recently

that peroxisomes contain inducible nitric ox-ide (NO) synthase activity (33), which, if true,

would be an important intraperoxisomal gen-

erator of NO species. Together with super-oxide anions, NO would generate peroxyni-

trite, a highly reactive species. Interestingly,peroxiredoxin V, which has been localized to

peroxisomes (see Table 1) as well as to mito-chondria and cytosol, was recently shown to

exhibit potent peroxynitrite reductase activity

(34).Peroxisomes also contain epoxide hydro-

lase activity (3537). Epoxides are a group ofhighly reactive molecules of both exogenous

and endogenous origin. Some of the most po-tent carcinogenic and mutagenic compounds

only become active when transformed into

their epoxides. Because they are very elec-trophilic, they easily react with nucleophilic

groups such as lipids containing unsaturatedFAs, DNA, RNA, and proteins. Epoxides,

which can be synthesized endogenously, in-

clude epoxides of prostaglandins, leukotriensarachidonic acid, cholesterol, and unsaturated

FAs.Onesinglegene coding fora protein with

a weak PTS1 signal (SKI) has been identifiedwhich gives rise to a bicompartmental distri-

butionof epoxide hydrolasein both theperox-isomes and cytosol (36). According to others

however, the peroxisomal epoxide hydrolaseis different from that present in other com-

partments (37).

Finally, peroxisomes also contain glu-tathione S-transferase (GST) activity (38)

GSTs catalyzethe conjugation of electrophilicsubstrates to glutathione but, in addition, have

reduced glutathione-dependent peroxidaseand isomerase activities. The GST identified

in peroxisomes belongs to the kappa family

The true function of this GSTK1 remains to

be identified, however. The enzyme shows reactivity with 1-chloro-2,4-dinitrobenzene as

well as with cumene hydroperoxide and 15-S-

hydroperoxy-5,8,11,13-eicosatetraenoic acid(38).

Ether-Phospholipid Biosynthesis

Ether phospholipids may occur in two

forms, including (a) plasmanyl-phospholipidand (b) plasmenyl-phospholipids (plasmalo-

gens) with a 1-O-alkyl and 1-O-alk-11-enyether bond, respectively, and usually con

tain ethanolamine or choline as head groupwith ethanolamine predominating in humans

(fourfold). In humans, plasmalogens make up

some 18% of total phospholipid mass andshow a cell- and tissue-specific distribution

High levels of plasmenyl ethanolamine occurin brain, heart, lung, kidney, spleen, skele-

tal muscle, and testis, whereas high levels oplasmenyl choline occur in heart and skele-

tal muscle with very low levels in all othertissues. Macrophages and neutrophils contain not only high plasmenyl ethanolamine

levels, but also significant levels of the satu-rated ether-phospholipid plasmanyl choline

which is used by these cells for the produc-tion of platelet-activating factor (1-O-alkyl-2

acyl-sn-glycero-3-phosphocholine). The sn-1

302 WandersWaterham


9/41

position of plasmalogens is occupied predom-

inantly by C16:0, C18:0, and C18:1 fatty al-cohols, whereas the sn-2 position of ether

phospholipids usually contains polyunsatu-rated FAs.

Peroxisomes and the enzymology of ether-

phospholipid biosynthesis. The first com-mitted step in the biosynthesis of ether-linkedglycerolipids is the formation of the ether

linkage by the enzyme alkyldihydroxyace-tone phosphate synthase (alkyl-DHAP syn-

thase or ADHAPS) (Figure 1a) encoded bythe AGPS gene. In this reaction, the ester-

linked FA of acyl-DHAP is replaced by a fatty

alcohol with an ether bond, forming alkyl-DHAP. The reaction proceeds by a ping-pong

mechanism: acyl-DHAP first binds to the en-

zyme, followed by release of the FA, result-ing in an activated enzyme-DHAP complex,

which then reacts with a fatty alcohol to pro-

duce alkyl-DHAP. Alkyl-DHAP synthase is

an established peroxisomal enzyme (3941),which can react with a range of fatty alco-

hols, including saturated (C10:0 to C18:0) aswell as mono- (C18:1) and polyunsaturated

(C18:2 and C18:3) alcohols, and this contrastsmarkedly to the fatty alcohols found in plas-

malogens, which are C16:0, C18:0, and C18:1

only (see below). Alkyl-DHAP synthases havebeen identified in various eukaryotic species

and represent one of the few peroxisomal en-zymes equipped with a PTS2-targeting se-

quence in all organisms exceptCaenorhabditis

elegans. In this organism, the enzyme contains

a PTS1, which is in line with the notion that

thePTS2 pathway is missing in C. elegans(42).The two substrates of alkyl-DHAP syn-

thase, i.e., acyl-DHAP and a long-chainfatty alcohol, are also generated by peroxi-

somes (Figure 1b). Acyl-DHAP is synthe-sized from DHAP and an acyl-CoA esterby the peroxisomal enzyme dihydroxyace-

tone phosphate acyltransferase (DHAPAT)encoded by GNPAT. The enzyme can han-

dle only a small range of acyl-CoAs, includingsaturated (C14:0 and C16:0) and unsaturated

(C18:1) acyl-CoAs (43), shows a broad pH

Figure 1

(a) Schematic representation of the steps involved in the biosynthesis of (phosphatidylcholine) and PE (phosphatidylethanolamine) plasmalogensand (b) the topology of the enzymes involved in the biosynthesis of

plasmalogens. Abbreviations used: AADHAPR, acylalkyl-dihydroxyacetophosphate reductase; ADHAPS, alkyl-DHAP synthase; DHAPAT,dihydroxyacetone phosphate acyltransferase; FAR, fatty acyl-CoAreductase; G3PDH, glycerol-3-phosphate dehydrogenase; and VLCS,

very-long-chain acyl-CoA synthetase.

CHO: Chinesehamster ovary

optimum between 7.0 and 9.0, and is mem-

brane associated, with its catalyticsite exposedto the peroxisome interior. All DHAPAT

amino acid sequences known from differenteukaryotic species contain a PTS1 sequence

(44, 45). DHAPAT is crucial for plasmalo-gen synthesis because DHAPAT-deficient hu-man and Chinese hamster ovary (CHO) cell

lines are unable to synthesize plasmalogens.Interestingly, acyl-DHAP can also be syn-

thesized outside peroxisomes by other acyl-transferases including microsomal G3PAT

(46), but this acyl-DHAP is not available for



10/41

Figure 1

(Continued)

peroxisomal alkyl-DHAP synthase, mostlikely because acyl-DHAP synthesized out-

side peroxisomes is unable to traverse the

peroxisomal membrane (Figure 1b). Alkyl-DHAP synthase and DHAPAT form a 210-

kDa heterotrimeric complex within peroxi-

somes (44, 47) (Figure 1b). DHAPAT is onlystable inside peroxisomes and when present inthe 210-kDa complex, whereas alkyl-DHAP

synthase is stable in peroxisomes and active

even in the absence of DHAPAT (48).Although some of the long-chain fatty

alcohols required for the alkyl-DHAP syn-thase reaction maycome from dietary sources,

the bulk is synthesized from acyl-CoAs. Thetwo consecutive reductions required to trans-

form acyl-CoAs into the corresponding al-

cohols (acyl-CoA aldehyde alcoholare catalyzed by the same fatty acyl-CoA re-

ductase (FAR), which does not release the

intermediate aldehyde and has NADPH asthe preferred cosubstrate. In developing ratbrain, a long-chain acyl-CoA reductase ac-

tivity was identified as reactive with C16:0-

C18:0-, and C18:1-CoA only (49). The en-zyme was specifically localized in the perox-

isomal membrane with its catalytic site ex-posed to the cytosol (49) (see Figure 1b)

304 WandersWaterham


11/41

On the basis of a specific substrate spec-

trum, it was speculated that this enzyme is re-sponsible for the virtually exclusive presence

of C16:0, C18:0, and C18:1-alk(en)yl chainsat the sn-1 position of ether phospholipids.

Recently, the identification of two acyl-CoAreductases, called FAR1 and FAR2, with dif-

ferent substrate specificities and tissue distri-butions was reported (50). Both FAR1 andFAR2 are peroxisomal membrane proteins,

but they lack clear transmembrane-spanningregions as well as PTS1 or PTS2 sequences.

FAR1 is reactive with saturated and unsatu-rated acyl-CoAs of 1618 carbons, whereas

FAR2 prefers saturated acyl-CoAs of 16 or 18

carbon atomsonly. FAR1 expression was iden-tified in many mouse tissues, with the highest

level in the preputial gland, a modified seba-

ceous gland. FAR2 expression was more re-stricted in distribution and most abundant inthe eyelid, which contains wax-laden meibo-

mian glands. Both FAR1andFAR2expression

was observed in the brain, a tissue rich in etherlipids. These findings suggest that fatty alco-

hol synthesis in mammals is accomplished bytwo FAR enzymes (FAR1and FAR2), encoded

byMLSTD2 and MLSTD1, respectively, andexpressed at high levels in tissues known to

synthesize wax esters and ether lipids (50).

The last contribution of peroxisomes toether-phospholipid biosynthesis is the re-

duction of alkyl-DHAP, generated by alkyl-DHAP synthase, to alkyl-G3P. The respon-

sible enzyme acyl/alkyl-DHAP reductase ismembrane bound, faces the cytosol both in

peroxisomes and in the endoplasmatic reticu-

lum, and preferentially reacts with NADPHrather than NADH (51). All subsequent

steps occur in the endoplasmic reticulum(Figure 1b) (see References 46 and 52 for re-

cent reviews).The physiological role of ether phos-

pholipids, including plasmalogens, has not

been established with certainty. They havebeen implicated in membrane dynamics, in-

tracellular signaling, cholesterol transport andmetabolism, oxidative stress, and polyunsat-

urated FA metabolism (46, 5358). The fact

that isolated deficiencies of DHAPAT andalkyl-DHAP synthase in humans are associ-

ated with severe clinical abnormalities and

early death (see the section on peroxisomaldisorders) indicates that ether phospholipids

are essential for life.

Peroxisomal Fatty AcidBeta-Oxidation

In contrast to most other functions of per-oxisomes, which may vary between different

species and within specific cell types in a sin-gle species, FA beta-oxidation is a universal

property of peroxisomes in most, if not all, or-

ganisms. In yeast and plants, peroxisomes arethe sole site of FA beta-oxidation, whereas in

higher eukaryotes beta-oxidation may occur

in both mitochondria and peroxisomes, fol-lowing a mechanism involving dehydrogena-tion, hydration, dehydrogenation again, and

thiolytic cleavage as depicted in the panels of

Figure 2a,b for mitochondrial beta-oxidationand peroxisomal beta-oxidation, respectively.

Although similar in mechanism, mitochon-drial and peroxisomal beta-oxidation fulfill

different functions, as concluded from theusually severe but different clinical signs and

symptoms associated with inherited defects

in either mitochondrial (59) or peroxisomalbeta-oxidation (60).

FAs destined for beta-oxidation may orig-inate from outside the cell or result from

intracellular breakdown of lipids, for in-stance in lysosomes. Extracellular FAs prob-

ably enter cells by a saturable mechanism

mediated by candidate proteins, such as theplasma membrane FABPPM, the FA translo-

case (FAT/CD36), as well as one or moremembers of the FA transport protein (FATP)

family of molecules, which have been hypoth-esized to harbor both FA transport as wellas acyl-CoA synthetase activity (6163). FAs

generated within the cell may be activatedby one of the acyl-CoA synthetase enzymes

(Acs1-6), which activate different FAs withunique efficiencies (64). Once activated, FAs

cannot repartition back into the membrane



12/41

Figure 2

Schematic representation of the mitochondrial and peroxisomalbeta-oxidation systems in humans. (a) In mitochondria, the FADH2 andNADH, generated in the first and third steps of beta-oxidation, aredirectly reoxidized by the respiratory chain (RC), (b) whereas inperoxisomes, molecular oxygen is the electron acceptor in the first step ofbeta-oxidation, resulting in the formation of H2O2, which is reconvertedinto O2 by catalase. The NADH generated in the third step ofperoxisomal beta-oxidation is reoxidized via a NAD(H)-redox shuttle,

involving the cytosolic and peroxisomal isoforms of malate dehydrogenasein S. cerevisiae and lactate dehydrogenase in higher eukaryotes.

VLCFA:very-long-chain fattyacid

because of their decreased hydrophobicity.The activation also ensures low unesterified

FA levels in the cell, thereby maintaining aconcentration gradient that is favorable for

the entry of more unesterified FAs into the

cytosol. The major differences between per-oxisomal and mitochondrial beta-oxidation

include different substrate specificities and

transport of substrates and products of beta-oxidation across the membrane (see Refer-

ences 65 and 66 for reviews).

Different substrate specificities. Short-and medium-chain FAs are exclusively andlong-chain FAs are predominantly beta-

oxidized in mitochondria, whereas very-long-chain FAs (VLCFAs), notably 26:0

can only be handled by peroxisomesOther substrates handled only by per-

oxisomes are (a) pristanic acid (2,4,6,10-

tetramethylpentadecanoic acid), derived fromdietary sources such as pristanic acid itself

or its precursor phytanic acid, which is con-

verted to pristanic acid by alpha-oxidation; (b)di- and trihydroxycholestanoic acid (DHCAand THCA), produced from cholesterol in

the liver and converted to chenodeoxycholic

and cholic acid, respectively, after one cy-cle of beta-oxidation in the peroxisome

(c) long-chain dicarboxylic acids, producedby omega-oxidation of long-chain monocar-

boxylic acids; (d) certain polyunsaturated FAsincluding tetracosahexaenoic acid (C24:6)

which undergoes one cycle of beta-oxidation

in peroxisomes to produce docosahexaenoicacid (C22:6); (e) certain prostaglandins and

leukotrienes; ( f) some xenobiotics; and (g)vitamins E and K.

Transport of substrates and products of

beta-oxidation across the membrane. In

the case of mitochondria, long-chain FAs(LCFAs) enterthe mitochondrialspacevia the

carnitine cycle (Figure 2a), whereas short-and medium-chain FAs enter directly in their

protonated form. For peroxisomes, the situa-tion is less clear, but a carnitine-mediated im-port mechanism has been ruled out (66). As

discussed below, the FAs destined for beta-oxidation in peroxisomes probably enter per-

oxisomes as acyl-CoA esters. Oxidation oFAs in peroxisomes generates a number of

acyl-CoA esters, including (a) medium-chain

306 WandersWaterham


13/41

Figure 2

(Continued)

acyl-CoAs, e.g., 4,8-dimethylnonanoyl-CoAin the case of pristanoyl-CoA beta-oxidation;

(b) proprionyl-CoA, and (c) acetyl-CoA. The

fate of each of these products may vary amongdifferent organs and cell types. In principle,

LCFA: long-chaifatty acid

there are different ways in which these CoAesters canbe metabolized further (Figure 2b).

First, different acyl-CoAs can be con-

verted into the corresponding carnitine es-ters via the peroxisomal enzymes carnitine



14/41

acetyltransferase and carnitine octanoyltrans-

ferase, as encoded by CRAT and CROT, re-spectively, followed by export from the per-

oxisomes and uptake into the mitochondrionvia the carnitine acylcarnitine carrier as was

shown for acetyl-CoA and propionyl-CoA(67) and 4,8-dimethylnonanoyl-CoA (68) in

cultured skin fibroblasts. Furthermore, acyl-CoA esters may be hydrolyzed within theperoxisome by one of the peroxisomal acyl-

CoA thioesterases (69), yielding the free acidand CoA (Figure 2b). In hepatocytes, the

thioesterase route is very active, with acetateas a major product of the acetyl-CoA units

produced in peroxisomes (70). Recent studies

on the peroxisomal and mitochondrial oxida-tion of FAs have shown that there is no de-

tectable transfer of peroxisomal acetyl-CoA

units to the mitochondrion for oxidation toCO2 and H2O, at least in perfused rat hearts(71). This implies that, in contrast to the

liver, peroxisomal FA oxidation in the heart is

not accompanied by the hydrolysis of acetyl-CoA and release of acetate. The exact fate of

the acetyl-CoA units produced in heart per-oxisomes remains to be determined. Recent

studies in HepG2 cells have shown that theacetyl-CoA units generated in liver peroxi-

somes are not only converted into acetate (70)

but are also used for chain elongation (72)(see Figure 2b).

Enzymology of the peroxisomal beta-

oxidation system. Saturated unbranched

and 2-methyl-branched FAs are the onlyFAs that can undergo direct beta-oxidation.

In contrast, other FAs, such as mono- andpolyunsaturated FAs, 3-methyl branched-

chain FAs, and 2-hydroxy FAs, first needto undergo remodeling before they become

substrate for peroxisomal beta-oxidation(Figure 3). The first step of beta-oxidation inmammalian peroxisomes is catalyzed by dif-

ferent acyl-CoA oxidases, with important dif-ferences between the rat and human. Extra-

hepatic peroxisomes in the rat contain twoacyl-CoA oxidases, including palmitoyl-CoA

oxidase (ACOX1) and pristanoyl-CoA oxi-

dase (ACOX3), whereas liver peroxisomescontain an additional cholestanoyl-CoA ox-

idase (ACOX2), specifically reacting with the

CoA esters of the bile acid intermediatesDHCA and THCA (73). Rat ACOX1 is ac-

tive with CoA esters of straight-chain mono-and dicarboxylic FAs, prostaglandins, VLC

FAs, and xenobiotics, whereas rat ACOX3is active with 2-methyl-branched-chain acyl-

CoAs, such as pristanoyl-CoA, but also

handles long and very-long straight-chainacyl-CoAs (73). In the rat, only ACOX1 is in-

ducible by peroxisome proliferators. Interest-ingly, human peroxisomes contain only two

oxidases; the first one is palmitoyl-CoA ox-idase, the counterpart of rat ACOX1, with

similar substrate spectrum and molecular

characteristics (74). The second human per-

oxisomal oxidase is the branched-chain acyl-CoA oxidase, active with 2-methyl-branchedcompounds, such as pristanoyl-CoA and the

CoA esters of DHCA and THCA, as well

as straight-chain acyl-CoAs, including theCoA esters of VLCFAs and dicarboxylic

acids (74). Cloning of the cDNA for hu-man branched-chain acyl-CoA oxidase re-

vealed that it is the homologue of the raliver-specific cholestanoyl-CoA oxidase (75)

Incontrasttotheratenzyme,however,human

branched-chain acyl-CoA oxidase is ubiqui-tously present in all tissues. Remarkably, the

gene coding for the homologue of rat ACOX3was also identified in the human genome, bu

both immunoblot and northern blot analysesfailed to identify its expression in human tis-

sues (76). It was therefore speculated that the

gene was only expressed under certain, forinstance developmental, conditions. Interest-

ingly, abundant expression of this pristanoyl-CoA oxidase was found recently in human

prostate tissue as well as in some prostate can-cer cell lines (77).

Human, rat, and mouse peroxisomes con-

tain two distinct bifunctional proteins thatdisplay both enoyl-CoA hydratase and 3-

hydroxy-acyl-CoA dehydrogenase activitiesand catalyze the conversion of 2-trans-enoyl-

CoAs to 3-ketoacyl-CoAs. l-bifunctiona

308 WandersWaterham


15/41

Figure 3

Schematic description of the enzymes required for the conversion of different FAs into thecorresponding acyl-CoA esters. Abbreviations used: ALDH, aldehyde dehydrogenase; HPCL2,2-hydroxyphytanoyl-CoA lyase.

protein (LBP) forms and dehydrogenates l-

3-hydroxyacyl-CoAs, andd-bifunctional pro-tein (DBP) forms and dehydrogenates d-

3-hydroxyacyl-CoAs. Alternative names forLBP and DBP are multifunctional enzymes

I and II (perMFE-I and -II) (78), multifunc-tional proteins 1 and 2 (MFP-1 and -2) (79),

and l- and d-peroxisomal bifunctional en-

zyme(l-PBE andd-PBE)(80).Itiswellestab-lished now that DBP is the main, if not exclu-

sive, enzyme involved in the beta-oxidation ofVLCFAs, pristanic acid, as well as DHCA and

THCA (Figure 4). Substrate specificity stud-ies have shown that both LBP and DBP re-

act with straight-chain enoyl-CoAs, whereasonly DBP reacts with the enoyl-CoA estersof pristanic acid, DHCA, and THCA (78, 79,

8184). The importance of DBP in the beta-oxidation of all these compounds has become

clear through the identification and charac-terization of patients with a deficiency of DBP

(85) and by the generation of a DBP (MFP-2)

knockout mouse (86). The physiological roleof LBP remains unclear, although recent stud-

ies indicated that it may be the primary en-

zyme involved in dicarboxylic acid oxidation(87).

Both bifunctional proteins show very littlesequence homology and are structurally very

different. The N-terminal part of LBP con-

tains theenoyl-CoAhydratase activity andtheC-terminal part, the 3-hydroxyacyl-CoA de-

hydrogenase activity. In addition, LBP alsoharbors 3,2-enoyl-CoA isomerase activ-

ity. In contrast, the N-terminal domain ofDBP is responsible for the 3-hydroxyacyl-

CoA dehydrogenase activity, the central partcontains the enoyl-CoA hydratase activity,and the C-terminal domain, sterol carrier pro-

tein (SCP) 2 activity (88).Mammalian peroxisomes also contain

multiple peroxisomal thiolases. Mouse andrat liver peroxisomes contain three dif-

ferent 3-oxoacyl-CoA thiolases, including



16/41

Figure 4

Overview depicting the involvement of the different peroxisomal beta-oxidation enzymes in theperoxisomal beta-oxidation of VLCFAs, pristanic acid (PRIS), DHCA, THCA, tetracosahexaenoic acid(C24:6), and long-chain dicarboxylic acids (DCA). Abbreviation: CoASH, free unesterified coenzyme A.

(a) 3-oxoacyl-CoA thiolase A, (b) 3-oxoacyl-

CoA thiolase B, and (c) SCP-2/3-oxoacyl-CoA thiolase (SCPx). The constitutively ex-

pressed thiolase A and inducible thiolase

B have a virtually identical substrate spec-trum, which is active toward short-, medium-,

long-, and very-long-chain 3-oxoacyl-CoAs,

and are involved in the peroxisomal beta-oxidation of straight-chain FAs (89). SCPxis active toward medium-, long-, and very-

long-chain 3-oxoacyl-CoAs, and it is also re-

active toward the 3-oxoacyl-CoA species of2-methyl branched-chain FAs, such as pris-

tanic acid and the bile acid intermediatesDHCA and THCA (9093). Human per-

oxisomes only contain two thiolases; theseinclude a straight-chain 3-oxoacyl-CoA thi-

olase, encoded by ACAA1, and the SCPx,

encoded by SCP2, which is essential forthe oxidation of 2-methyl branched-chain

FAs, i.e., pristanic acid, DHCA, and THCA(94). The involvement of the different beta-

oxidation enzymes in the oxidation of VL-CFAs, pristanic acid, DHCA, THCA, C24:6,

and long-chain dicarboxylic acids is shown in

Figure 4.

The enzymes described above are neces-

sary and sufficient for the beta-oxidation ofstraight-chain saturated FAs as well as alpha-

methyl branched-chain FAs with the methygroup in the (2S)-configuration. However

auxiliary enzymes are needed for the beta-

oxidation of (2R)-methyl branched-chain FAs

and unsaturated FAs (Figure 3). Oxidationof (2R)-methyl branched-chain FAs requiresthe active participation of the peroxisoma

enzyme 2-methylacyl-CoA racemase, whichis capable of converting (2R)- into (2S)-

branched-chain acyl-CoAs (9597). Interest-

ingly, a single gene (AMACR) codes for a pro-tein equipped with both a mitochondrial and

peroxisomal targeting signal thus explainingits bicompartmental presence in both peroxi-

somes and mitochondria (68). Both the per-

oxisomal and mitochondrial AMACRs are re-quired for the oxidation of pristanic acid (7).

Peroxisomes also contain the full enzy-matic machinery to remove the double bonds

in mono- and polyunsaturated FAs. FAs witha double bond at an even-numbered posi-

tion require the subsequent action of two aux-

iliary enzymes, including 2,4-dienoyl-CoA

310 WandersWaterham


17/41

reductase and 3,2-enoyl-CoA isomerase,to produce the corresponding saturated acyl-

CoA esters. The 2,4-dienoyl-CoA reductase

and 3,2-enoyl-CoA isomerase in perox-isomes are different from their mitochon-

drial counterparts and the products of dis-tinct genes, i.e., the DECR2 (98) and PEC1

(99) genes (see Table 1). Oxidation of FAswith a double bond at an odd-numbered po-sition may proceed via two different path-

ways. Pathway one only involves the 3,2-enoyl-CoA isomerase, whereas pathway two

requires the active participation of three aux-iliary enzymes, including a distinct3,5,2,4-

dienoyl-CoA isomerase, 2,4-dienoyl-CoA

reductase, and 3,2-enoyl-CoA isomerase.After its prior identification in mitochondria,

He et al. (100) detected 3,5,2,4-dienoyl-

CoA isomerase activity in peroxisomes. Sub-sequent studies have shown that a singlegene ECH1 codes for a protein with both

a mitochondrial and peroxisomal targeting

signal, thus explaining the bicompartmen-tal distribution of3,5,2,4-dienoyl-CoA iso-

merase in both mitochondriaandperoxisomes(101).

Peroxisomal Fatty Acid

Alpha-OxidationFAs with a methyl-group at the carbon 3

position are not a substrate for beta-oxidationbut must first undergo alpha-oxidative de-

carboxylation to produce the corresponding

(n-1) FA, with the methyl-group at the 2position, which then can undergo beta-

oxidation. In the past decade, it has becomeclear that, in contrast to FA beta-oxidation,

FA alpha-oxidation is confined to perox-isomes and only accepts acyl-CoA esters

as substrate (Figure 5). Most studies onalpha-oxidation have been performed withthe physiological substrate phytanic acid

(3,7,11,15-tetramethylhexadecanoic acid),which is known to accumulate in different

genetic disorders including Refsum disease(see below). Activation of phytanic acid

to its CoA-ester can occur either outside

the peroxisome by the enzyme long-chainacyl-CoA synthetase (102), present at the

cytosolic face of the peroxisome (103), or

inside the peroxisome by the enzyme very-long-chain acyl-CoA synthetase, a peripheral

peroxisomal membrane protein equippedwith a PTS1-like signal (Figure 5) (104).

Subsequently, phytanoyl-CoA is convertedinto 2-hydroxyphytanoyl-CoA by the en-

zyme phytanoyl-CoA 2-hydroxylase, which

belongs to the family of 2-oxoglutarate-dependent oxygenases, the largest known

family of nonheme metal-dependent oxidiz-ing enzymes (see References 105 and 106

for reviews). Conversion of phytanoyl-CoAto 2-hydroxyphytanoyl-CoA is stoichiomet-

rically coupled to the decarboxylation of

2-oxoglutarate into succinate and CO2, after

which one of the oxygen atoms of the dioxy-gen (O2) molecule is incorporated into thecarboxyl group of succinate and the other in

the 2-hydroxy group of 2-hydroxyphytanoyl-

CoA. The primary sequences of a numberof phytanoyl-CoA 2-hydroxylases from

different species have become available inrecent years with little overall sequence

identity. All phytanoyl-CoA 2-hydroxylasescontain the 2-His-1-carboxylate motif and

the RXS motif responsible for the binding

of iron and 2-oxoglutarate, respectively (107,108). Site-directed mutagenesis studies have

established the identity of His-175, Asp-177,and His-264 as the iron-binding ligands in

the human phytanoyl-CoA 2-hydroxylase

(109).All hydroxylases identified so far con-

tain PTS2 sequences except for the C. ele-

gans 2-hydroxylase, which contains a PTS1-

like sequence (-RSNL) in accordance withthe absence of the PTS2 pathway in C. ele-

gans (42). The next enzyme in the pathway,i.e., 2-hydroxyphytanoyl-CoA lyase, converts2-hydroxyphytanoyl-CoA into pristanal and

formyl-CoA, is a peroxisomal matrix en-zyme of four identical 63-kDa subunits,

and contains a PTS1-like sequence (-SNM)(110). At neutral pH, formyl-CoA is split

into formate and free CoA nonenzymatically



18/41

Figure 5

Schematic

illustration andtopology of theenzymes involvedin the phytanic acidalpha-oxidationpathway inmammalianperoxisomes.

(111). Pristanal is converted into pristanic

acid via a still poorly defined peroxisomalaldehyde dehydrogenase (112). Finally, pris-

tanic acid is activated to pristanoyl-CoA,probably via the enzyme VLCS (104, 113).

Pristanoyl-CoA undergoes three cycles of

beta-oxidation in the peroxisome to produce4,8-dimethylnonanoyl-CoA, which is then

transported to the mitochondria for full ox-idation (68). Recent studies have shown that

peroxisomes also catalyze the oxidation of 2-

hydroxy FAs, which are first activated to the

respective CoA esters, followed by cleavageinto the corresponding aldehyde and formyl-

CoA by 2-hydroxyphytanoyl-CoA lyase (114)

Glyoxylate Metabolism

In humans, the enzyme alanine:glyoxylate

aminotransferase (AGT) is exclusively ex-pressed in liver peroxisomes and converts gly-

oxylate generated in peroxisomes into glycine

312 WandersWaterham


19/41

using alanine as the primary amino group

donor. This prevents the conversion of gly-oxylate into the toxic metabolite oxalate,

which can be catalyzed by various dehydro-genases and oxidases, including lactate dehy-

drogenase. In the case of AGT deficiency, asin patients with hyperoxaluria type 1, glyoxy-

late accumulates and is converted into ox-alate, which precipitates in the liver as wellas in other organs, including the kidneys, ul-

timately causing kidney failure and loss. It hasbeen postulated that glyoxylate, which accu-

mulates in peroxisomes, diffuses across theperoxisomal membrane and is converted into

oxalate by cytosolic lactate dehydrogenase.

Such a scenario is unlikely, however, becauseunder these conditions the cytosolic enzyme

glyoxylate reductase is expected to convert cy-

tosolic glyoxylate to glycolate. Therefore, itis more likely that glyoxylate, which accumu-lates in peroxisomes, is converted into oxalate

by peroxisomal enzymes, including glycolate

oxidase and lactate dehydrogenase (115). In-terestingly, hyperoxaluria type 1 may not only

be caused by a functional deficiency of per-oxisomal AGT but also by mislocalization of

AGT to mitochondria (116). Under the lat-ter conditions, the AGT, mislocalized to mi-

tochondria, is catalytically active, but glyoxy-

late generated in peroxisomes cannot reachthe AGT in mitochondria and instead is con-

verted into oxalate, giving rise to hyperox-aluria type 1 (9).

Amino Acid Catabolism

Mammalian peroxisomes contain d-aminoacid oxidase, which oxidizes the d-isomers

of neutral and basic amino acids, as well asd-aspartate oxidase (117), which oxidizes the

d-isomers of acidic amino acids. The oxida-tion of amino acids by both enzymes yieldsthe corresponding keto acids, ammonia, and

hydrogen peroxide. Peroxisomes are also in-volved in the oxidation of some l-amino acids,

e.g., l-lysine, which may be degraded to l-2-amino adipic acid either via the saccha-

ropine pathway or via the l-pipecolate path-

way. l-pipecolate oxidase, which oxidizes l-pipecolate to 1-piperideine-6-carboxylate,

is a peroxisomal enzyme identified in human

(118) and monkey liver (119), has been pu-rified and cloned (120, 121), and is a typi-

cal PTS1 protein. l-pipecolate accumulatesin tissues and body fluids of patients who lack

peroxisomes, emphasizing the importance ofthe l-pipecolate pathway in humans. Lysine,

hydroxylysine, and tryptophan can be con-

verted to glutaryl-CoA, for which a perox-isomal glutaryl-CoA oxidase has been de-

scribed in rat and man. However, because theglutaryl-CoA oxidase activity copurifies with

palmitoyl-CoA oxidase activity (73), the ex-

istence of a separate glutaryl-CoA oxidase isquestionable.

Pentose Phosphate PathwayIn the pentose phosphate pathway, NADPH

is generated when glucose 6-phosphate is ox-idized to ribose-5-phosphate. Although the

pentose phosphate pathway has always beenassumed to be cytosolic, approximately 10%

of the total activity of the two pentose phos-phatepathway enzymes, glucose-6-phosphate

dehydrogenase and 6-phosphogluconate de-

hydrogenase, is peroxisomal (122). It isproposed that these two enzymes provide

intraperoxisomal NADPH as needed, for ex-ample, for the 2,4-dienoyl-CoA reductase

reaction. As discussed below, peroxisomes

also contain a different system that pro-vides intraperoxisomal NADPH, i.e., via the

2-oxoglutarate/isocitrate NADP(H) shuttle.Because peroxisomes lack any of the sub-

sequent enzymes of the pentose phosphatepathway, the intraperoxisomal generation of

NADPH by the glucose 6-phosphate and 6-

phosphogluconate dehydrogenases would re-quire import of glucose 6-phosphate and ex-

port of ribulose 5-phosphate.

Polyamine Oxidation

Spermine (SPM) and spermidine (SPD) arerequired for numerous fundamentally im-

portant cellular processes, and the levels of



20/41

these polyamines are under tight control.

The main mechanism by which spermine andspermidine are degraded involves transforma-

tion of SPM and SPD into N1-acetyl-SPMand N1-acetyl-SPD by the cytosolic enzyme

acetyl-CoA:SPD/SPM N1-acetyltransferase,followed by oxidation of the N1-acetylated

polyamines by a peroxisomal N1

-acetylatedpolyamine oxidase. This enzyme convertsN1-

acetyl-SPM to SPD and 3-acetamidopropanal

and N1-acetyl-SPD to putrescine and 3-acetamidopropanal. The same enzyme also

oxidizes SPM to SPD and 3-aminopropanal,although very inefficiently. Most cells also

contain a cytosolic spermine oxidase, which

oxidizes SPM to SPD and 3-aminopropanalbut does not react with N1-acetyl-SPM and

N1-acetyl-SPD. The mouse, bovine, and hu-

man peroxisomal N1

-acetylated polyamineoxidases (123, 124) all have a typical PTS1sequence (Table 1). The fate of the prod-

ucts of the peroxisomal polyamine oxidase re-

action is not clear. It may well be that the3-acetamidopropanal is further metabolized

within peroxisomes because peroxisomes har-bor both aldehyde (125) and alcohol dehydro-

genase (126) activities.

Miscellaneous Peroxisomal EnzymeActivities

Peroxisomes have been said to contain a num-ber of different enzyme activities for which

the responsible enzymes and genes have re-

mained unknown as well as proteins with un-known functions (Table 1). Indeed, rat liver

peroxisomes contain a clofibrate-induciblealcohol:NAD+ oxidoreductase activity, with

tetradecanol showing the highest catalytic ef-ficiency, i.e. Vmax/Km (126). Furthermore, a

clofibrate-inducible aldehyde dehydrogenaseactivity was identified in rat liver peroxisomes

with nonanal as the substrate with the high-

est catalytic efficiency (125). Peroxisomes alsocontain pristanal dehydrogenase activity (112)

and retinal reductase activity (4). As discussedabove, the peroxisomal 2,4-dienoyl-CoA re-

ductase is responsible for the latter activ-

ity (4). Peroxisomes also contain a trans-2-enoyl-CoA reductase, which may play a role

in chain elongation (127). Because of the ob-

servation that phytol is converted into phy-tanic acid via the reduction of phytenoyl-CoA

into phytanoyl-CoA, it was recently hypothe-sized that the enoyl-CoA reductase identified

by Das et al. (127) may mediate this reductionstep (128).Recent studies have shown that one or

more members of the nudix hydrolase familyare present in peroxisomes (3). One of these

enzymes, called NUDT7, catalyzes the hy-drolysis of CoA and its derivatives, and its

function may be to eliminate oxidized CoA

from peroxisomes and/or to regulate the lev-els of CoASH and acyl-CoAs in this organelle

in response to metabolic demands. Recently

we performed proteomic analysis of mouseperoxisomes and identified several new per-oxisomal proteins, including a novel nudix

hydrolase designated RP2p and encoded by

the D7RP2e gene. RP2p is a CoA diphos-phatase with activity toward CoASH, oxidized

CoASH, and a wide range of acyl-CoA esters(R. Ofman and R.J.A. Wanders, submitted for

publication). Finally peroxisomes also containNAD-linked glycerol-3-phosphate dehydro-

genase activity (122, 129), which may play a

role in the provision of DHAP for the DHA-PAT reaction (see Figure 1b).

Isoprenoid and CholesterolMetabolism

Many studies performed from the 1950s to the

1980s showed that eight of the nine enzymesof the first part of the isoprenoid biosyn-

thesis pathway, involved in the conversion ofacetyl-CoA to farnesyl pyrophosphate, are cy-

tosolic (130). The exception is 3-hydroxy-3-methylglutaryl-CoA reductase, which hadbeen localized to the endoplasmic reticulum

(ER), as are the enzymes involved specificallyin cholesterol synthesis. Since 1985, however

a series of reports have claimed that many ofthe enzymes (or the reactions they catalyze)

of the first part of the pathway are partly

314 WandersWaterham


21/41

mainly, or even exclusively located in perox-

isomes (131). Moreover, several enzymes in-volved in cholesterol synthesis were reported

to be colocalized in peroxisomes and in theER. This has led to the rather generally ac-

cepted view that peroxisomes would be di-rectly involved in isoprenoid and cholesterol

biosynthesis.Different observations have been used tosupport the claim that peroxisomes would be

involved in isoprenoid/cholesterol biosynthe-sis. Among these arethe decreasedactivities of

several isoprenoid biosynthetic enzymes ob-served in postmortem liver homogenates of

patients, whosuffered from a fatal peroxisome

biogenesis disorder (PBD) and thus lackedfunctional peroxisomes (see References 131

and132for reviews). Recent studies, however,

showed that these decreased activities resultfrom inactivation owing to the bad conditionand/or preservation of the livers rather than

from mislocalization to the cytosol (17). This

corresponds to the finding of normal activi-ties and protein levels of these enzymes in cul-

tured primary skin fibroblasts of PBDpatientsand in liver of the pex5 knockout mouse (18,

133), which constitutes a well-defined modelfor human PBDs (134).

Conflicting data have been published on

the de novo cholesterol synthesis rate inperoxisome-deficient fibroblasts and CHO

cells. Although a few groups reported de-creased rates (135137), others found normal

or even increased rates in such cells (138140),indicating that the loss of peroxisomes per se

does not affect the enzyme activities or de

novo cholesterol biosynthesis.Also, with respect to the subcellular local-

ization of isoprenoid biosynthetic enzymes,conflicting data have been published indicat-

ing either a cytosolic or a peroxisomal local-ization of the proteins. It should be noted,

however, that in most cases the claim of a per-

oxisomal (co)localization was based on (a) thefinding of only (very) low amounts of pro-

teins in peroxisomal fractions obtained aftersubcellular fractionation of rat liver tissue, (b)

immunocytochemicalstudies using antiseraof

PBD: peroxisombiogenesis disorde

undefined specificity, and/or (c) the results of

overexpression studies with tagged proteinsor(portions of) proteins fused to reporter pro-

teins in cell lines. As discussed above, the lat-ter studies can be informative but may be full

of pitfalls and can only support, but neverfully replace, studies aimed at determining

the subcellular localization under physiolog-ical conditions. Indeed, the subcellular lo-calization under physiological conditions and

after overexpression of the three human iso-prenoid biosynthetic enzymes mevalonate ki-

nase, phosphomevalonate kinase, and meval-onate pyrophosphate decarboxylase, which

were previously claimed to be predominantly

peroxisomal, was reinvestigated in great detailusinga variety of biochemical andmicroscopi-

caltechniques.The results of these studies un-

ambiguously pointed to an exclusive cytosoliclocalization of these enzymes with no indica-tion of even a partial peroxisomal localization

(1618).

When combining all available data withemphasis on studies of the subcellular lo-

calization of authentic, nonengineered pro-teins under physiological conditions, the con-

clusion must be that there is little, if any,evidence for a direct peroxisomal involve-

ment in the biosynthesis of isoprenoids and

cholesterol.

BIOCHEMISTRY OF HUMANPEROXISOMAL DISORDERS

The importance of peroxisomes for normal

mammalian development and growth is un-derlined by the existence of a group of in-

herited diseases in humans, the peroxisomaldisorders, which can be classified into two

groups, including(a)thePBDsand(b)thesin-gle peroxisomal enzyme deficiencies. The fa-tal cerebro-hepato-renal syndrome, in short

Zellweger syndrome (ZS), is the prototypeof the first group and is characterized by

the complete absence of peroxisomes. Theunderlying basis for the inability to synthe-

size peroxisomes in ZS has been resolved in



22/41

recent years and involves mutations in at least

12 differentPEXgenes (141143).The absence of peroxisomes has major

consequences for most of the metabolic path-ways in which peroxisomes are involved.

This is evident from the biochemical ab-berrations observed in ZS patients; these

abberrations range from the accumulationof substrates normally handled by peroxi-somes, e.g., VLCFAs, pristanic acid, phytanic

acid, DHCA, THCA, and pipecolic acid, toa shortage of end products of peroxisomal

metabolism, e.g., plasmalogens, cholic andchenodeoxycholic acid, and docosahexaenoic

acid (see Table 2). From these and other ob-

servations, it hasbecome clear that theseques-tration of (or parts of ) certain metabolic path-

ways in peroxisomes is essential for efficient

substrate channeling and to protect the cellagainst toxic metabolites generated in perox-isomes, e.g., reactive oxygen species, reactive

nitrogen species, and glyoxylate. The abnor-

malities observed in ZS patients are causedby the deficiency of most, but not all, per-

oxisomal enzymes destined for peroxisomes.Indeed, most peroxisomal enzymes are unsta-

ble in the cytosol and are rapidly degraded,whereas a few peroxisomal enzymes, such as

catalase (144) and alanine:glyoxylate amino-

transferase (145), are assembled correctly intoactive multimers (tetramer and dimer, respec-

tively) and are stable in the cytosol.Most peroxisomal disorders belong to

group two, which can be subdivided fur-ther into distinct subgroups, depending

upon which peroxisomal function is impaired

(Table 2). Virtually all peroxisomal enzymedeficiencies are associated with severe clinical

aberrations. Remarkably, defects in enzymeswithin the same metabolic pathway may re-

sult in different phenotypes (Table 2). Thisis especially true for the disorders of perox-

isomal beta-oxidation (Table 2). For exam-

ple, patients with DBP deficiency are severelyaffected with clinical signs and symptoms re-

sembling those observed in ZS patients. In pa-tients with DBP deficiency, the peroxisomal

beta-oxidation of all major substrates is im-

paired, resulting in the accumulation of VL-CFAs, pristanic acid, DHCA, and THCA in

tissues and plasma (85, 146). Conversely, pa-

tients with AMACR deficiency show a mildclinical phenotype resembling Refsum dis-

ease (see below) and only accumulate pris-tanic acid, DHCA, and THCA (147, 148)

Another example is X-linked adrenoleukodys-trophy, which affects boys who develop nor-

mally for the first few years of life and then

rapidly deteriorate, followed by early deathIn X-linked adrenoleukodystrophy, only VL-

CFAs accumulate (68, 149).Other single peroxisomal enzyme de

ficiencies are rhizomelic chondrodysplasiapunctata type 2 and 3, caused by mutations

in GNPAT and AGPS, respectively, which

encode the peroxisomal enzymes DHAPAT

and ADHAPS, respectively. Patients affectedby these deficiencies show markedly loweredplasmalogen levels, which is in line with the

notion that both enzymes play an indispens-able role in ether-phospholipid biosynthesis

(52). Refsum disease also belongs to group 2

and is caused by mutations in the gene en-coding phytanoyl-CoA hydroxylase; as a re-

sult, phytanic acid gradually accumulates andreaches toxic levels later in life (150, 151). An-

other disorder belonging to group 2 is hyper-

oxaluria type 1, caused by mutations in thealanine:glyoxylate aminotransferase encoding

gene (AGXT). When glyoxylate accumulatesin peroxisomes, oxalate is formed, which pre-

cipitates as calcium oxalate in tissues, includ-ing the kidney. This explains the patients loss

of kidney function over time. Acatalasaemia isthe last single peroxisomal enzyme deficiency

caused by mutations in the catalase-encoding

gene, and is associated with an increased ten-dency to develop oral gangrene in otherwise

asymptomatic patients (152).

MOUSE MODELS FORPEROXISOMAL DISORDERS

Because the number of patients with spe-cific peroxisomal defects is rather limited and

the majority of defects lead to early death

316 WandersWaterham


23/41

Table2

Biochemistry

ofhumanperoxisomaldisordersa

Biochemicalabnormalities

Groups

Pe

roxisomaldisorder

Mutantgen

e

VLCFA

PRIS

PHYT

D

/THCA

PL

Clinicalsign

sandsymptoms

Group1

(peroxisome

biogenesis

defects)

Zellwegerspectrum

disorders(ZS,NALD,

IRD)

PEX1,2,3,5,

6,10,1

3,14,

16,1

9,26

b

b

ZS,NALD,andIRD

representaspectrum

ofdiseaseseveritywithZSbeingthemost

andIRDtheleasts

everedisorder.

CommontoZS,N

ALD,andIRDareliver

disease,variablene

uro-developmental

delay,retinopathy,andperceptivedeafness.

ZSpatientsareusu

allyhypotonicfrom

birthanddiebeforeoneyearofage,

whereasNALDpatientsshowneonatal

onsethypotoniaan

dseizures,andthey

haveprogressivewhitematterdisease,

usuallydyinginlateinfancy.IRDpatients

maysurvivebeyondinfancy,andsomemay

evenreachadulthood

RCDPtype1

PEX7

N

N

b

N

Patientshaveadisproportionallyshort

statureprimarilyaf

fectingtheproximal

partsoftheextremities.Othersymptoms

includetypicalfacialabnormalities,

congenitalcontractures,ocular

aberrations,severe

growthdeficiency,and

mentalretardation

Group2(single

peroxisomal

enzyme

deficiencies)

X-linked

adrenoleukodystrophy

(X-ALD)

ABCD1

N

N

N

N

Twomajorforms,in

cludingchildhood

cerebraladrenoleukodystrophy(CCALD)

andadrenomyeloneuropathy(AMN);in

thesevereform(CCALD),normal

developmentuntilsixyearsofage,

followedbyrapidd

eteriorationanddeath

withintwoyears

(Continued)



24/41

Table2

(Continued)

Biochemicalabnormalities

Groups

Pe

roxisomaldisorder

Mutantgen

e

VLCFA

PRIS

PHYT

D

/THCA

PL

Clinicalsign

sandsymptoms

Acyl-CoAoxidase

deficiency(ACOX1

deficiency)

ACOX1

N

N

N

N

Hypotonia,earlyonsetseizures,hearing

loss,retinopathy,neurological

abnormalities

d-Bifunctionalproteinde-

ficiency/multifunctional

pro

tein2deficiency

(DBP/MFP2deficiency)

HSD17B4

b

b

N

Craniofacialabnorm

alities;neurological

disturbances;Zellw

eger-likephenotype,

includingneuronal

migrationdefect

2-M

ethyl-acyl-CoA

rac

emasedeficiency

(AMACRdeficiency)

AMACR

N

b

b

N

Slow,progressivelossofvision;neurological

deterioration;inso

mepatientsmarked

hepatopathy

RCDP2(DHAPAT

deficiency)

GNPAT

N

N

N

N

Severegrowthretardation,mental

retardation,rhizom

elia,earlydeath

RCDP3(ADHAPS

deficiency)

AGPS

N

N

N

N

Severegrowthretardation,mental

retardation,rhizom

elia,earlydeath

Refsumdisease

(ph

ytanoyl-CoA

hydroxylasedeficiency)

PAHX/PHYH

N

N

N

N

Lossofvision,cereb

ellarataxia,anosmia,

ichtyosis,cardiacproblems

Hyp

eroxaluriatype1

(AGTdeficiency)

AGXT

N

N

N

N

N

Progressivelossofk

idneyfunction

Acatalasaemia

CAT

N

N

N

N

N

Increasedtendencytodeveloporalgangrene

aAbbreviations:ZS,Zellwege

rsyndrome;NALD,neonataladrenoleukod

ystrophy;IRD,infantileRefsumdisease;RC

DP,rhizomelicchondrodysplasiapunctata;X-ALD,X-linked

adrenoleukodystrophy;VLCFA,very-long-chainFAs;PRIS,pristanicacid;PHYT,phytanicacid;D/THCA,di-andt

rihydroxycholestanoicacid;PL,plasmalogens;N,normal;,elevated;,

decreased.

bLevelsmayvaryfromnormaltoelevatedbecausephytanicacidandpristanicacidarederivedfromexogenous(dietary)sourcesonly.

318 WandersWaterham


25/41

most biochemical studies on these defects

have been done on cells of such patients, inparticular primary skin fibroblasts. Although

valuable and informative, these studies do notshow how and why a given peroxisomal dys-

function leads to the specific pathophysiologyassociated with these defects. To obtain more

insight into these pathophysiological conse-quences,anumberofmousemodelshavebeengenerated; the biochemical and phenotypical

characteristics of these models are summa-rized in Table 3. All of these mice were gen-

erated through targeted gene disruption.Most mouse models, for which currently a

human peroxisomal disorder is known, show

biochemical and phenotypical defects similartothoseobservedinthecorrespondinghuman

disorders. Moreover, studies on these mouse

models have led to the identification of ad-ditional defects that could not be readily in-

vestigated in patients or their cells, including

ossification and neuronal migration defects.

A few mouse knockouts have been gener-ated that create peroxisomal enzyme deficien-

cies for which no human peroxisomal disorderhas been identified (see Table 3). The bio-

chemicaland phenotypical characterization ofthese mouse models may aid in the recogni-

tion of their possible human counterparts. In-

terestingly, it appears that, at least in mice, theabsence of certain peroxisomal enzymes in-

cluding l-bifunctional enzyme (80), thiolase-B (153), and catalase (154) does not result in

noticeable biochemical and phenotypical ab-normalities, making the specific physiologi-

cal function of these enzymes difficult to dis-

cern. It should be noted, however, that suchenzymes maybe required only for certain spe-

cific conditions, as exemplified by the Scp2

(/) mouse. In this mouse model, severe

biochemical and phenotypical defects becameapparentonlyafterfeedingitthephytanicacid

precursor phytol, leading to the accumulation

of pristanic acid. From this observation andadditional findings in the Amacr(/) mouse

(155), one may predict that the possible hu-man counterpart of SCP deficiency presents

onlylaterinlife,asisthecasewithRefsumdis-

ease, which presents only after gradual accu-mulation of phytanic acid to toxic levels over

time.

Another aspect that should be mentionedis the short life span of mice in comparison to

humans, which maynot allow enoughtime forthe development of certain phenotypical ab-

normalities. One example is the observationthatAbcd1 (/) mice, a mouse model for X-

linked adrenoleukodystrophy (X-ALD), only

develop mild neurological and behavioral ab-normalities later in life (> 15 months of age),

and these abnormalities resemble those ob-served in adrenomyeloneuropathy, a milder

variant of X-ALD (156).The generation of additional mouse mod-

els for selected peroxisomal proteins has been

reported forwhichno detailed information on

the biochemical and phenotypical character-istics is available yet. These include mice withtargeted disruptions of the genes encoding

phytanoyl-CoA hydroxylase (a model for Ref-sum disease), peroxisomal ABC proteins, i.e.,

the 70-kDa peroxisomal membrane protein

(PMP70), ALD-related protein (ALDRP),PMP70-related protein (PMP70R) (no hu-

man diseases known), and alanine:glyoxylateaminotransferase (a model for hyperoxaluria

type 1).

PEROXISOMAL METABOLITETRANSPORT

Correct execution of the many metabolic

functions of peroxisomes requires the trans-port of a large variety of metabolites across

the peroxisomal membrane. In recent years,much hasbeen learned about the permeability

properties of peroxisomes as discussed below.

Permeability Properties ofPeroxisomes

Because after isolation mammalian peroxi-somes are freely permeable to low-molecular-

weight compounds and peroxisomal enzymesdo not show structure-linked latency, it has

long been assumed that the peroxisomal



26/41

Table3

Mousemodelsofperoxisomebiogenesisandperoxisomefunctiona

Disrupted

gene

Deficient

(enzyme)protein

Corresponding

humandisease

Biochemicalphenotype

VLC

FA

PRIS

PHYT

D/THCA

PL

Clinicalcharacteristics

References

Pex2

Pex2

Zellweger

spectrum

disorder

(ZS/NALD/IRD)

Intrauterinegrowthretardation,

severehypotonia,neonatald

eath,

delayedneuronalmigration

inCNS,

cerebellarabnormalitieswith

reducedPurkinjecelldevelo

pment

(192,193)

Pex5

Pex5

Zellweger

spectrum

disorder

(ZS/NALD/IRD)

Lowbirthweight,hypotonia,poor

feeding,neuronalmigration

defect,

neonataldeath

(134,194)

Pex13

Pex13

Zellweger

spectrum

disorder

(ZS/NALD/IRD)

Lowbirthweight,hypotonia,poor

feeding,neuronalmigration

defect,

neonataldeath

(195)

Pex7

Pex7

RCDPtype1

N

N

N


severehypotonia,delayedossification

ofdistalboneelements,dwa

rfism,

delayedneuronalmigration

(196)

Pex11

Pex11

N

N

N

N

N

Nophenotypicabnormalities

(197)

Pex11

Pex11

Zellweger

spectrum

disorder

(ZS/NALD/IRD)

N

N

N

N

N


hypotonia,developmentaldelay,

neonataldeath,impairedneuronal

migration

(198)

Gnpat

Dhapat

RCDPtype2

N

N

N

N


hypotonia,maleinfertility,d

efectsin

eyedevelopment,cataract,o

ptic

nervehypoplasia,prenataldeathof

Dhapat(/)embryos

(199)

Acox1

Acox1

Acyl-CoAoxidase

deficiency

N

N

N

N

Viable,butinfertile;retarded

postnatalgrowth;microvesicular

steatosis;focalcelldeath;

inflammatoryreactions;livertumors

atlaterage(>15months)

(200)

320 WandersWaterham


27/41

Hsd17B4

Dbp/Mfp2

d-Bifunctional

proteindeficiency

N

Normalbirthweight,dramat

icgrowth

retardation,upto30%diebefore

postnatalday12,maleinfertility,no

neuronalmigrationdefect

(86)

Scp2

Peroxisomal

thiolase2

(Scpx)

N

N

Nophenotypicabnormalities,phytol

feedinginducesweightloss,

neurologicalabnormalities,

andearly

deathwithinthreeweeksof

birth

(94)

Abcd1

Aldp

Adrenomyel-

oneuropathy

N

N

N

N

Noapparentphenotype;how

ever,

beyondage15months,late-onset

neurologicalandbehavioral

abnormalities,axonallossin

the

spinalcord,andslowernerv

e

conduction

(156,

201203)

Ehhadh

Lbp/Mfp1

N

N

N

N

N


(204)

Slc27a2

Vlcs

N

N

N

N

N


(176)

Amacr

Amacr

AMACR

deficiency

N

N-

N

Nophenotypicabnormalities,

intolerancetophytolwithliver

diseaseandearlydeath

(155)

mThB

ThiolaseB

N

N

N

N

N


(153)

Cat

Catalase

Acatalasemia

N

N

N

N

N

Nophenotypicabnormalitiesexcept

forincreasedsusceptibilityto

trauma-induceddysfunction

ofbrain

mitochondria

(154)

aSeeTable2forabbreviations.



28/41

membrane does not constitute a permeability

barrier to small molecules, at least in mam-mals. More recent studies, notably in the yeastS. cerevisiae and partly confirmed in mam-malian cells, however, revealed that in vivo

the peroxisomal membrane is impermeable tosmall metabolites, which implies the existence

of peroxisomal metabolite carriers. Hence, itseems plausible that the in vitro permeabilityis due to the disruption of protein/membrane

structures as a consequence of the cell frac-tionation methods used for their isolation. It

should be noted that structure-linked latencyof peroxisomal enzymes has been observed

in other members of the microbody family,

including glyoxysomes and glycosomes. Onthe basis of pulse-labeling experiments in Try-

panosoma brucei, it was concluded that the gly-

cosomal membrane is poorly permeable toglycolytic intermediates with the exception ofglycerol-3-phosphate, which was postulated

to be transported via a specific translocator

(157). Recent studies by Antonenkov et al.(158, 159) have shown that under properly

controlled conditions several peroxisomal en-zymes do show structure-linked latency in

peroxisomes from rat liver.

The Intraperoxisomal pHOne of the first indications that the perox-

isomal membrane may form a closed struc-ture in vivo was the demonstration of a pH

gradient across the membrane in yeast, which

could be dissipated by uncouplers (160).Morerecently, a peroxisomal pH gradient was also

shown in human cells (161). It remains un-clear whether this gradient is due to an ac-

tive proton-translocating protein, the conse-quence of a high intraperoxisomal metabolic

activity, or both. Moreover, there is no agree-ment concerning the orientation of the pro-ton gradient because some groups reported a

lower (162), whereas others found a higheror similar intraperoxisomal pH in compari-

son to the cytosolic pH (163), despite the factthat similar methodologies and cell systems

were used. This also makes it difficult to de-

termine whether theproton gradient is used asa driving force for metabolite transport, e.g.

FA transport (163), or is a mere consequence

of metabolite transport (164). Further insightinto a possible physiological function of a per-

oxisomal proton gradient may come from invitro studies with reconstituted peroxisoma

membrane proteins aimed at determining theproperties, characteristics, and requirements

of peroxisomal metabolite carriers and shut-

tle systems,which were identified recently.

Peroxisomal ABC Transporters

Mamma

Documents

New Per Ox i Some Complete