Báo cáo khoa học: New roles of flavoproteins in molecular cell biology: An unexpected role for quinone reductases as regulators of proteasomal degradation pptx

In recent years, an additional role of quinone reductases as regulators of proteaso-mal degradation of transcription factors and possibly intrinsically unstruc-tured protein has emerged.

New roles of flavoproteins in molecular cell biology: An unexpected role for quinone reductases as regulators of proteasomal degradation

Sonja Sollner and Peter Macheroux

Technische Universita¨t Graz, Institut fu¨r Biochemie, Austria

Introduction

Quinones are abundant cyclic organic compounds

present in the environment as well as in pro- and

eukaryotic cells They can be reduced by two- or

one-electron reduction to either the hydroquinone or the

semiquinone form A number of organisms express

enzymes that afford strict two-electron reduction to

the hydroquinone form in an attempt to avoid the

generation of semiquinones This species is known to

cause oxidative stress by reacting with molecular

oxy-gen, eventually leading to the generation of superoxide

radicals (redox cycling) Hence, quinone reductases

(QRs) from pro- and eukaryotes have a protective

effect against quinone-related oxidative cell damage

Consequently, QRs have been identified in bacteria, fungi, plants and mammals

Originally, QRs were classified as ‘DT-diaphorases’ to express the fact that they utilize both DNPH (reduced diphosphopyridine nucleotide, NADH) and TPNH (reduced triphosphopyridine nucleotide, NADPH)

as a source of reducing equivalents [1] At the time, the term ‘diaphorase’ was used to describe an enzyme (preferentially a flavoprotein) capable of transferring electrons from reduced pyrimidine nucleotides to electron acceptors [2] This nomenclature led to con-fusion because ‘diaphorase activity’ could be detected

in numerous biological systems The first

‘DT-diapho-Keywords

flavin; NAD(P)H; ornithine decarboxylase;

oxidative stress; peptide flip; proteasome;

redox state; reduction; transcription factors;

ubiquitination

Correspondence

P Macheroux, Graz University of

Technology, Institute of Biochemistry,

Petersgasse 12⁄ II, A-8010 Graz, Austria

Fax: +43 316 873 6952

Tel: +43 316 873 6450

E-mail: peter.macheroux@tugraz.at

(Received 9 December 2008, revised

29 April 2009, accepted 4 May 2009)

doi:10.1111/j.1742-4658.2009.07143.x

Quinone reductases are ubiquitous soluble enzymes found in bacteria, fungi, plants and animals These enzymes utilize a reduced nicotinamide such as NADH or NADPH to reduce the flavin cofactor (either FMN or FAD), which then affords two-electron reduction of cellular quinones Although the chemical nature of the quinone substrate is still a matter of debate, the reaction appears to play a pivotal role in quinone detoxification

by preventing the generation of potentially harmful semiquinones In recent years, an additional role of quinone reductases as regulators of proteaso-mal degradation of transcription factors and possibly intrinsically unstruc-tured protein has emerged To fulfil this role, quinone reductase binds to the core particle of the proteasome and recruits certain transcription fac-tors such as p53 and p73a to the complex The latter process appears to be governed by the redox state of the flavin cofactor of the quinone reductase, thus linking the stability of transcription factors to cellular events such as oxidative stress Here, we review the current evidence for protein complex formation between quinone reductase and the 20S proteasome in eukaryotic cells and describe the regulatory role of this complex in stabili-zing transcription factors by acting as inhibitors of their proteasomal degradation

Abbreviations

NQO, mammalian NAD(P)H:quinone oxidoreductase; ODC, ornithine decarboxylase; QR, quinone reductase; ROS, reactive oxygen species.

rase’, reported by Ernster & Navazio [3], is now

known as mammalian NAD(P)H:quinone

oxidoreduc-tase (NQO1, isozyme 1) However, the acronym NQO

has traditionally been confined to QRs from

mamma-lian sources

Although the first successful crystallization of QR

was reported in the late 1980s [4,5], it was another

couple of years before Li and coworkers eventually

solved the structure of rat liver NAD(P)H:quinone

oxi-doreductase [6] The crystal structure revealed that the

fold of the N-terminal portion resembles that of

flavo-doxin, a bacterial electron-transfer protein involved in

a variety of photosynthetic and nonphotosynthetic

reactions [7] The biological unit for NQO1, as for

most QRs studied to date, is a dimer The overall fold

of the flavodoxin-like catalytic domain consists of a

twisted, central five-stranded parallel b sheet

sur-rounded by helices The FAD cofactor is

noncovalent-ly bound at the interface of the monomers, with the

redox-active isoalloxazine ring positioned at one side

of two equivalent crevices, thereby forming two

identi-cal, independent active sites [6] Structure

determi-nation of NQO2 confirmed the close structural

relationship between NQO1 and NQO2 Similar to

NQO1, NQO2 self-associates as a homodimer and

contains two identical catalytic sites located at

oppo-site ends of the dimer interface Each catalytic oppo-site

forms a large cavity, lined by residues from both

pep-tide chains with the flavin isoalloxazine ring forming

the bottom [8] Interestingly, the flavodoxin-like

struc-ture of QRs is not restricted to mammalian enzymes,

but is conserved down to yeast Lot6p, a homologous

QR from the unicellular model organism

Saccharo-myces cerevisiae, also adopts a dimeric flavodoxin-like

fold but binds one FMN cofactor per protomer

instead of FAD [9]

The unique property of QRs is their ability to

trans-fer two electrons to a quinone, thereby catalyzing the

formation of a two-electron reduced hydroquinone

without the generation of a one-electron reduced

semiquinone [10] This feature seems to be crucial for

understanding the physiological role of QRs as a

cellu-lar device to avoid the formation of semiquinones

and hence the generation of harmful reactive oxygen

species (ROS)

Although some QRs can utilize both NADH and

NADPH as a source of electrons (e.g NQO1, Lot6p),

others have developed a strong preference for either

NADH (e.g AzoA from Enterococcus faecalis [11])

or NADPH (e.g YhdA from Bacillus subtilis [12])

By contrast, NQO2 is unable to employ NADH or

NADPH as a source of electrons, but instead uses

reduced N-ribosyl- and N-alkyl-dihydronicotinamide

However, the issue of oxidizing substrates seems to be far more complicated It is generally assumed that enzymes involved in the detoxification mechanisms of xenobiotics do not possess endobiotic substrates but have evolved in such a way that a broad range of chemical structures can be processed In fact, the size and shape of the catalytic sites of NQO1, NQO2 and Lot6p suggest that these enzymes have evolved to accept a variety of ring-containing substrates Never-theless, a number of naturally occurring quinones com-prising vitamin K derivatives (menaquinone and phylloquinone), coenzyme Q (ubiquinone) and dopa-quinone have also been shown to be substrates for mammalian QRs [13]

The functional importance of QRs has been a matter

of discussion since their discovery The discovery that

a vitamin K reductase, described by Maerki & Martius [14], and the DT-diaphorase first described by Ernster

& Navazio [3] is actually the same enzyme added to speculation revolving around the physiological role of QRs The opinion that mammalian QRs are primarily involved in xenobiotic metabolism and in preventing the carcinogenicity and toxicity of highly reactive com-pounds is a more recent development An explicit mechanism by which NQO1 might protect cells was first provided by Iyanagi & Yamazaki [10] by distin-guishing between flavoproteins that catalyze one-electron reductions and those, like NQO1, that catalyze strict two-electron reductions Accordingly, two-electron reduction of quinones can avert: (a) one-electron redox cycling, which generates highly ROS; and (b) depletion of cellular glutathione by decreasing the levels of quinones, which react easily with thiol groups Furthermore, in contrast to semiquinone products generated by one-electron reducing flavopro-teins, the hydroquinone products of the NQO1 path-way are not only more stable, but can also be further metabolized to glucuronide and sulfate conjugation products, thereby facilitating their excretion Thus, the possibility of forming reactive semiquinone radicals, potential mediators of oxidative stress, is highly reduced [15], and consequently, QRs have been proposed to serve as a cellular control device against quinone toxicity

A new role for an old enzyme: QR and the proteasome

Many proteins have a dual biological role and partici-pate in regulatory cellular functions in addition to their metabolic function as an enzyme [16] For exam-ple, glutathione S-transferase associates with c-Jun N-terminal kinase leading to inhibition of kinase

acti-vity and modulation of signalling and cellular

prolifer-ation [17] Similarly, recent studies have demonstrated

that eukaryotic QRs bind to the 20S proteasome and

affect the lifetime of several transcription factors and

ornithine decarboxylase by inhibiting their

ubitquitin-dependent and -inubitquitin-dependent proteasomal degradation

This role of QRs is the focus of this minireview Before

we shed further light on this recently discovered

func-tion of a historically old enzyme family, we provide a

brief introduction to the role of the proteasome A

more detailed description of the structure and function

of the proteasome is given in several recent review

arti-cles [18–20]

The bulk of cellular proteins in eukaryotic cells are

degraded by the 26S proteasome This 2.5 MDa

proteo-lytic machinery consists of a 20S barrel-structured core

that provides a catalytic chamber and a 19S regulatory

particle This latter protein complex binds to the edges

of the core particle and regulates access to the catalytic

chamber The process that leads to proteasomal

degra-dation is initiated by selective polyubiquitination

fol-lowed by recognition of the condemned protein

through the 19S regulatory caps Ubiquitin consists of

76 amino acids and is covalently attached in a highly

regulated multistep process to the substrate protein

[21–23] The 19S caps are involved in recognition of

the polyubiquitinated protein substrates, unfolding of

the condemned protein [24], removing ubiquitin chains

for recycling [25,26] and opening an axial gate into the

20S catalytic chamber [27] Whereas 26S proteasomal

degradation requires ubiquitination of substrate

pro-teins, the 20S proteasome degrades structurally

abnor-mal, misfolded or highly oxidized proteins in a

ubiquitin-independent manner under conditions of

cel-lular stress [28] In mammalian cells (cytoplasm and

nucleus), most of the 20S core particles are present in

their free uncapped form with only a smaller fraction

being capped with the 19S regulatory protein complex

[29]

The association of mammalian QR with the 20S proteasome was first described by Shaul and cowork-ers in the course of their investigation into the deg-radation of transcription factors They found that the vast majority of QR 1 (NQO1) from mouse liver extract is bound to the 20S, but not the 26S, protea-some [30] Similar results were obtained with extracts from human red blood cells and with different com-mercially available proteasome preparations [31] This link between NQO1 and the 20S proteasome at the protein level is also reflected at the transcriptional level: Nrf2, a transcription factor that is activated upon oxidative stress, is a major transcriptional acti-vator of NQO1 [32] Furthermore, activation of the Nrf2 pathway by treatment with 3H-1,2-dithiole-3-thione also leads to enhanced expression of most of the 20S proteasome subunits (a1, a2, a4–a7, b1–b6) [33]

The demonstration that a complex between QR and the 20S proteasome exists in mammalian cells prompted further experiments in the unicellular model organism S cerevisiae (baker’s yeast) It could be shown that Lot6p, a QR and ortholog of human NQO1, is physically associated with the 20S protea-some [34] Using the 20S proteaprotea-some and recombi-nant Lot6p, several biochemical issues revolving around the stoichiometry and importance of the flavin cofactor could be resolved Fluorescence titration studies exploiting the intrinsic fluorescence of the fla-vin cofactor demonstrated that two QR dimers bind

to one 20S proteasome core particle (Fig 1) [34] Furthermore, QR lacking both flavin cofactors (apo-Lot6p) was unable to bind to the proteasome, indi-cating that the presence of the flavin cofactor is required for complex formation Not surprisingly, the enzymatic activity of the QR is also compromised in the presence of proteasome, supporting the direct or indirect involvement of the flavin cofactor in protein complex formation [34]

α

β

β

α

α

β

β

α

NQO1 (reduced)

Stabilization Degradation

20S proteasome

26S proteasome 19S cap

Mdm2 dependent polyubiquitination

Ubiquitin

19 S cap

p53 p53

p53 Fig 1 Proposed mechanisms of

ubiquitin-dependent (right) and ubiquitin-inubiquitin-dependent

(left) degradation pathways of p53 Free p53

(red) binds to the preformed 20S

protea-some–QR (green and pale green,

respec-tively) complex after reduction of the QR.

After reoxidation of the QR flavin cofactor,

p53 is released from the complex, becomes

ubiquitinated (blue) and is eventually

degraded by the 26S proteasome.

The impact of the 20S proteasome on QR activity

raised the question of whether a reciprocal effect on

proteasomal activity occurs or, in other words, does

the QR act as a gatekeeper for the proteasome, as

recently suggested [35]? Detailed analysis of the

trypsin-like, chymotrypsin-like and

peptidyl-glutamyl-protein-hydrolysing or caspase-like activity [36]

demonstrated up to 10 times lower proteolytic activity

in the presence of Lot6p [34] Because substrate access

through the gated entry port in the outer a-rings of

the core particle is considered to be the rate-limiting

step in catalysis, it appears likely that the QR binds to

or near the a-rings of the proteasome, thereby

affect-ing access to the catalytic chamber, leadaffect-ing to reduced

proteasomal activity [37] In this context, it is

impor-tant to emphasize that it is not at all clear whether this

effect on proteasomal activity is part of a regulatory

mechanism because 20S proteasome core particles

out-number QR molecules by a factor of 10 [38] Thus, it

appears that the majority of 20S core particles exist in

their ‘free’ form and only some are associated with

Lot6p However, if the number of QR molecules

increases in the cells, for example, by overexpression

during oxidative stress, it is conceivable that 20S

pro-teasomal activity is severely downregulated by binding

of QR

Contradictory results concerning the physical

associ-ation of the mammalian NQO1 and 20S proteasome

were recently reported by Jaiswal’s group: although

copurification of the 20S proteasome and QR from

mouse liver cytosol was observed, they were unable to

detect complex formation by immunoprecipitation

using a 20S proteasome antibody Therefore, these

authors conclude that mammalian QR and the 20S

proteasome do not form a protein complex, as

pro-posed by others [39] Unfortunately, no explanation

for the copurification of QR and the 20S proteasome

is provided in this report and the failure to detect the

protein complex by immunoprecipitation was not

confirmed by an independent method Hence, the

relevance of their findings concerning binding of

mam-malian QR to the 20S proteasome remains unclear at

present

Protection of transcription factors

from proteasomal degradation through

association with the QR–proteasome

complex

Protein degradation is a key cellular process involved

in almost every aspect of the living cell [21,22] The

prevailing concept assumes that proteins are not

proteolysed unless marked by polyubiquitination, a

prerequisite for recognition and degradation by the 26S proteasome [31] For example, the transcription factor and tumour suppressor p53 is subject to ubiquitin-dependent proteasomal degradation [41,42] and only accumulates following various types of stress, leading to growth arrest and apoptosis Ubiquitination of p53 is carried out by Mdm2, an E3 ubiquitin ligase, which binds to the N-terminus of the transcription factor Stabilization of p53 towards proteolysis can be achieved by disruption of the p53– Mdm2 interaction, thereby preventing Mdm2 from ubiquitinating p53 The main process that leads to disruption of the Mdm2 recognition of p53 involves post-translational modifications Accumulation of p53 then results in the expression of a variety of genes necessary to cope with DNA damage and other forms of cellular stress [43]

The ubiquitin-dependent degradation of p53 and other proteins appears to be the main pathway for regulating proteolysis by the 26S proteasome The discovery by Shaul and coworkers that p53 (and other proteins as well) is degraded more rapidly when human QR (NQO1) is inhibited by dicoumarol, a potent and specific inhibitor of QRs, was the first hint that another and different regulatory pathway may exist in eukaryotic cells As a result of enhanced degra-dation of p53 and hence lower levels of the trans-cription factor, p53-dependent apoptosis in both c-irradiated normal thymocytes and in myeloid leukae-mic cells was suppressed These effects could be pre-vented by overexpression of NQO1, supporting the idea that it might be involved in regulating cellular p53 levels [44] These findings raised questions concerning the role of NQO1 in ubiquitin-dependent proteasomal degradation Does NQO1 affect the ubiquitination process directly or is it involved in an alternative path-way? Further studies addressing this question revealed that regulation of p53 degradation by NQO1 does not require ubiquitination by Mdm2 Instead, a variant of p53, which is resistant to Mdm2-mediated degradation, was shown to be susceptible to dicoumarol-induced degradation, indicating that NQO1-regulated proteaso-mal p53 degradation is Mdm2 independent Accord-ingly, two alternative pathways for p53 proteasomal degradation have been proposed: one is ubiquitin dependent and regulated by Mdm2, whereas the other

is ubiquitin independent and regulated by the QR NQO1, implying that p53 stabilization is not solely dependent on inhibition of the p53–Mdm2 interaction but also requires physical association with NQO1 (Fig 1) [45]

Accumulation of p53 also leads to expression of the PIG3 (QR homologue) and FDXR (ferredoxin

reduc-tase) genes and stabilizes p66Shc (the 66 kDa isoform

of ShcA, an adaptor protein that relays extracellular

signals downstream of receptor tyrosine kinases) The

latter protein gives rise to increased levels of

intracellu-lar ROS, thereby promoting apoptosis of damaged

cells [46,47] This generation of ROS increases

expres-sion of NQO1 [48] which then further stabilizes p53

This sequence of events is consistent with the proposed

feed-forward loop for p53 stabilization by ROS [46]

The association of human QR with transcription

factors is not restricted to p53 Following the discovery

of p53 as an interaction partner of NQO1, p73 was

also reported to be regulated by a

ubiquitin-indepen-dent process [30] p73, also known as tumour protein

73 (TP73), was the first identified homologue of the

tumour suppressor p53 Overexpression, and thus

accumulation, of p73 in cultured cells promotes growth

arrest and⁄ or apoptosis similarly to p53 [49,50] The

p73 gene encodes a protein with significant sequence

and functional similarity to p53 Like p53, p73 has

several variants It is expressed as distinct forms

differ-ing either at the C- or the N-terminus Currently, six

different C-terminal splicing variants have been found

in normal cells The a-splice variant of p73 (p73a)

con-tains an additional structural domain near its

C-termi-nus known as the sterile a-motif that is probably

responsible for regulating the p53-like functions of

p73 [51] This motif appears to be essential for

inter-action with NQO1 and subsequent stabilization as the

p73b isoform lacking the C-terminal sterile a-motif

domain was not protected against 20S proteasomal

degradation

Recently, levels of the tumour suppressor p33ING1b

have also been found to be regulated by NQO1 The

ING1gene was originally identified through subtractive

hybridization between normal human mammary

epithelial cells and seven breast cancer cell lines, and

subsequent in vivo selection of genetic suppressor

ele-ments that displayed oncogenic characteristics [52]

Three alternatively spliced transcripts of the ING1 gene

have been found, encoding protein variants with a

pre-dicted size of 47, 33 and 24 kDa p33ING1b (ING1b for

inhibitor of growth family, member 1b) has been

reported to be downregulated in several carcinomas

The protein was shown to be a major player in cellular

stress responses, including cell-cycle arrest, apoptosis,

chromatin remodelling and DNA repair [53]

Phos-phorylation of p33ING1b at Ser126 was reported to be

important for proliferation in malinoma cells, as well

as modulation of its degradation [54] Garate et al

recently detected this tumour suppressor in purified

fractions of 20S proteasome and provided evidence

that p33ING1b is degraded by the 20S proteasome

Further results indicate that NQO1 inhibits the degra-dation of p33ING1b and that ultraviolet irradiation stabilizes p33ING1b by inducing phosphorylation at Ser126, thereby enhancing its interaction with NQO1 [55]

Protection of transcription factors through associa-tion with a QR is not restricted to NQO1 NQO2, a homologue of NQO1, was shown to prevent transcrip-tion factors from being degraded as well [39] Human

QR 2 was described for the first time in 1961 as an unknown mammalian cytosolic FAD-dependent flavo-protein catalyzing the oxidation of reduced N-ribosyl-and N-alkyldihydronicotinamides by menadione N-ribosyl-and other quinones, but not the oxidation of NADH, NADPH or NMNH (reduced nicotinamide mono-nucleotide) The enzyme was extensively characterized, but was completely forgotten for several decades In the early 1990s, Jaiswal and coworkers isolated and described a second NAD(P)H:quinone oxidoreductase, which they discovered in the course of cloning human NQO1, and named it NQO2 [39] In 1997, Zhao et al demonstrated that NQO2 was indeed the flavoprotein discovered more than 30 years before [56] Jaiswal’s group then developed NQO2-null mice models to investigate the role of the second human QR in regula-tion of p53 and found that loss of NQO2 significantly decreases the level of p53 [39] Co-immunoprecipita-tion studies revealed a physical interacCo-immunoprecipita-tion of NQO2 with p53, indicating that an increased amount of cytosolic NQO2 protects p53 from 20S proteasomal degradation through physical association [39]

Not just transcription factors

Although transcription factors appear to be prime tar-gets for QR-regulated degradation, recent studies have also identified an enzyme – ornithine decarboxylase (ODC) Catalysing the first and rate-limiting step in the polyamine biosynthesis pathway, ODC is one of the most labile cellular proteins [57] The polyamines spermidine, spermine and their precursor putrescine are abundant polycations that are present in all living cells Polyamines are essential for cellular proliferation and are involved in regulating additional fundamental cellular processes such as cellular transformation and differentiation [35] In its active form, ODC is a homo-dimer with two enzymatic active sites catalyzing the decarboxylation of ornithine to putrescine [58] The cellular level of ODC and its activity need to be strictly controlled because polyamines act as a double-edged sword On the one hand, they are absolutely required for maintaining growth, whereas, on the other hand, excessive levels are cytotoxic ODC degradation is

mediated by interaction with a polyamine-induced

pro-tein termed antizyme [57] Association of antizyme

with ODC subunits triggers disruption of ODC

homo-dimers and the formation of enzymatically inactive

ODC–antizyme heterodimers [59] Both in vivo and

in vitro studies have indicated that ODC degradation

by the 26S proteasome requires interaction with

anti-zyme, but not ubiquitination However, recent studies

revealed that there is a second, ubiquitin-independent

degradation pathway for ODC that is regulated by

NQO1 The QR was shown to protect ODC against

proteasomal degradation both in vivo and in vitro [60]

Disruption of NQO1 binding under several conditions

such as oxidative stress or upon exposure to

dicou-marol, a competitive inhibitor of NQO1, results in

ubiquitin-independent 20S proteasomal degradation of

ODC Notably, only ODC monomers are degraded by

this pathway Thus, the role of antizyme in this

pro-cess, if any, is confined to the ODC monomerization

step ODC monomerization is also obligatory for the

antizyme-independent degradation of ODC in vitro

An ODC mutant that fails to dimerize is susceptible to

20S proteasomal degradation, but not to degradation

by the 26S proteasome Interaction with NQO1

pro-tects monomeric ODC from this degradation pathway,

whereas inhibition of NQO1 dissociates the complex

and promotes ODC degradation [35,61]

Although specific mechanisms mediate the

recogni-tion of proteins destined for degradarecogni-tion by the 26S

proteasome, it is not yet clear how proteins are

recog-nized for degradation by the 20S core particle Recent

studies have suggested that unstructured proteins such

as a-synuclein and p21cip are intrinsically unstable

because of their capacity to enter the 20S proteasome

pore [62,63] Even a segment of unstructured region

within a protein might be sufficient to direct a protein

to 20S proteasomal degradation Analysis of the ODC

sequence using different computational prediction

algorithms indicates that ODC contains several

unstructured regions Similarly, > 80% of

transcrip-tion factors have been reported to possess extended

regions of intrinsic disorder [64]

From mammalian cells to yeast:

a homologous system in a unicellular

organism

All of the initial studies indicating a role for QR in

stabilizing transcription factors and tumour suppressors

were performed with cells from a narrow range of

multicellular eukaryotic organisms, i.e mammalian

cells of human or murine origin Until recently, it was

unclear whether the QR-operated regulation of protein

degradation discovered in mammalian cells was con-served in all eukaryotes or is a recent addition to the arsenal of regulatory mechanisms found in higher multicellular eukaryotes such as mammals This issue could be partially resolved because studies with baker’s yeast (S cerevisiae) have unambiguously demonstrated that Lot6p, a homologue of mammalian NQO1, binds

to the 20S proteasome and forms a ternary complex with Yap4p, a member of the yeast activator protein family of transcription factors [34] Yap4p (CIN5p, Hal6p, YOR028Cp) has been shown to increase sodium and lithium tolerance upon overexpression [65] and to confer resistance to cisplatin, a chemotherapy drug [66] In the yeast system, binding of Yap4p to the proteasome–QR complex depends on the redox state

of the flavin cofactor present in the active site of Lot6p Recruitment of the transcription factor occurs when the flavin is in its fully reduced form This pro-cess is independent of the mode of reduction – either

by addition of NADH or by light – and suggests that the change in redox state governs the association of Yap4p and Lot6p Although it appears likely that the redox state of the FAD cofactor of NQO1 and NQO2 plays a similar role in the mammalian system, unequivocal experimental evidence is not available at present However, the observation that recruitment of transcription factors occurs only in the presence of NADH is consistent with a redox-controlled process Further biochemical studies with Lot6p have shown that the native dimeric quaternary structure is a prere-quisite for formation of the ternary complex: in its monomeric form, Lot6p still binds to the proteasome, but is no longer able to recruit the transcription factor

to the complex Degradation of Yap4p by the 20S pro-teasome is prevented by the formation of a ternary protein complex consisting of the 20S core particle, reduced Lot6p and Yap4p Interestingly, formation of this ternary protein complex not only prevents degra-dation of Yap4p, but also influences the localization of the transcription factor In normal, unstressed yeast cells, Yap4p is present in the cytosol, whereas under oxidative stress it relocates to the nucleus Apparently, oxidation of the flavin cofactor of Lot6p results in the dissociation and concomitant relocalization of the released transcription factor to the nucleus where expression of stress related genes then occurs

Taken together, several studies investigating tran-scription factors from mammalian to yeast cells, as well as regulatory proteins such as ODC, suggest that short-lived proteins are intrinsically prone to degrada-tion by the 20S proteasome The associadegrada-tion of a QR (NQO1, NQO2 or Lot6p) with the 20S proteasome, together with its ability to regulate the stability, and in

the case of the yeast system, even the localization, of

those short-lived proteins suggests that QRs might play

a general and central role in regulating the metabolic

stability of a subset of cellular proteins

Molecular mechanism of interaction

Protection against 20S proteasomal degradation relies

on the recognition of transcription factors and perhaps

other intrinsically unstable proteins such as ODC by

the QR The active site of the enzyme with its flavin

cofactor – FAD in the case of the mammalian enzymes

and FMN in the case of the homologous Lot6p – is

clearly essential for the interaction, as documented by

the effect of dicoumarol, a potent competitive inhibitor

that pi-stacks on top of the isoalloxazine ring system,

preventing association with p53 [44,67] Similarly,

removal of the flavin in Lot6p disables the interaction

with the yeast transcription factor Yap4p and the 20S

proteasome [34] Moreover, both the presence of the

cofactor and its redox state appear to be relevant In

the yeast system, only reduced QR is able to recruit

transcription factor Yap4p Although clear evidence is

currently available only for the yeast system, studies

with the mammalian system have also indicated the

necessity to reduce the flavin cofactor in order to

enable interaction with target proteins such as

tran-scription factors p53 and p73a [44,45] Thus, it can be

concluded that transcription factors (and probably

intrinsically unstructured proteins such as ODC) bind

to or near the flavin site of the QR What structural

changes occur upon reduction of the flavin cofactor

and how are these sensed by potential interaction

part-ners? In principle, reduction of the isoalloxazine ring

system converts N(5) from a hydrogen-bond acceptor

to a hydrogen-bond donor (Fig 2) In other words,

reduction of the isoalloxazine ring system may cause

reorganization of hydrogen-bond interactions with

neighbouring amino acids, which in turn may lead to

local structural changes in the protein An instructive

example for such a restructuring is given by the X-ray

analysis of oxidized and reduced flavodoxin from

Clos-tridium beijerinckii [68] In the oxidized state, C@O of

Gly57 points away from N(5) of the isoalloxazine ring

system Upon reduction, the C@O turns towards the

N(5) position to form a hydrogen bond As a result,

Gly57 adopts a different conformation and this

‘pep-tide flip’ also causes the movement of some amino acid

side chains (Fig 2) [68] As mentioned in the

introduc-tion, QRs also adopt a flavodoxin-fold and the

isoal-loxazine ring engages in a similar interaction with a

peptide chain In the reported structure of oxidized

Lot6p, the backbone amide group of Asn96 forms a

hydrogen bond to N(5) Upon reduction, N(5) will be protonated and hence this interaction is no longer feasible, and it is conceivable that this leads to a con-formational change similar to that observed in flavo-doxins Interestingly, structural comparison of QRs (human NQO1, human NQO2, mouse NQO1 and yeast Lot6p; Fig 3) shows the conservation of large hydrophobic amino acid residues (i.e conservation of

a Trp residue; Fig 3E) in the peptide segment that runs along the N(5)–C(4)=O edge of the isoalloxazine ring system It is conceivable that a similar peptide flip occurs in QRs upon reduction, which then results in a repositioning of these large hydrophobic side chains such that interaction with transcription factors is enabled Interestingly, utilization of hydrophobic

resi-A

B

Fig 2 Structural changes occurring in flavodoxin upon reduction of the flavin cofactor (A) Overall structure of oxidized flavodoxin from Clostridium beijerinckii (PDB code 5nll) The FMN cofactor is shown

as a colour-coded stick model (carbon, yellow; nitrogen, blue; oxy-gen, red) The box indicates the area where structural changes occur upon reduction of the flavin (B) Close-up of the flavin cofac-tor and the loop close to the flavin N(5) for both the oxidized (PDB code 5nll) and reduced (PDB code 5ull) form of flavodoxin Amino acid residues are depicted as colour-coded stick model with carbons from the oxidized form coloured grey and carbons from the reduced form coloured green In the oxidized state, C=O of Gly57 points away from N(5) of the isoalloxazine ring system Upon reduction, the C=O turns towards the N(5) position to form a hydrogen bond Consequently, Gly57 adopts a different conforma-tion and this ‘peptide flip’ also causes movement of some amino acid side chains.

dues for the recognition and binding of unfolded

protein substrates is a very common mechanism

employed by chaperones [69] For example,

a-crystal-lin, a prominent member of the small heat shock

pro-tein family, was proposed to suppress the aggregation

of other proteins through an interaction between

hydrophobic patches on its surface and exposed

hydro-phobic sites of partially unfolded protein substrates

[70] Thus, it is conceivable that a similar mechanism

is used by QRs for sensing and binding of intrinsically

unfolded proteins such as transcription factors This

mechanism could be probed by combined structure–

function analysis of oxidized and reduced QR and

mutagenesis of conserved residues

As far as the transcription factor p53 is concerned,

several amino acid residues have been implicated in

the interaction with NQO1 This tumour suppressor

is mutated in > 50% of human cancers [43], with

Arg175His, Arg248His and Arg273His being the most

frequent ‘hot-spot’ mutants [71] Mutations in the p53

gene often result in the accumulation of p53 protein

variants in human cancer cells [72] Asher and

cowork-ers investigated whether those common mutations may

have an effect on binding of p53 to NQO1 They

showed that the most frequent p53 variants in human

cancer mentioned above were resistant to

dicoumarol-induced degradation (unlike wild-type p53), probably

E

Fig 3 Comparison of the loops close to the flavin N(5) of several quinone reducta-ses (A) NQO1 from Homo sapiens (PDB code 1d4a) (B) NQO2 from H sapiens (PDB code 1qr2) (C) NQO1 from Mus musculus (PDB code 1dxq) (D) Lot6p from Saccharomyces cerevisiae (PDB code 1t0i) Both the flavin cofactor (carbon, yellow; nitrogen, blue; oxygen, red) and the amino acid residues (carbon, grey; nitrogen, blue; oxygen, red) are depicted as colour-coded stick model (E) Sequence alignment of the loop regions shown in Fig 3A–D h NQO1, human NQO1; h NQO2, human NQO2; m NQO1, mouse NQO1; y Lot6p, yeast Lot6p.

Fig 4 Transcription factor p53 in complex with DNA The p53 monomer is depicted as a cartoon (light blue), DNA is shown as purple stick model (PDB code 2ahi) Residues probably involved in binding to NQO1 (Arg248, Arg273) are shown as colour-coded sticks (carbon, cyan; nitrogen, blue; oxygen, red).

because of increased binding to NQO1, but remained

sensitive to Mdm2–ubiquitin-mediated degradation

Hence they concluded that NQO1 plays a major role

in stabilizing p53 hot-spot mutants in human cancer

cells [73] However, it needs to be taken into

consider-ation that some variants of p53 not only lose their

function, but also adopt a fold different to wild-type

protein [74] In addition, p53 possesses extensive

unstructured regions in its native state [75] Thus, it

cannot be ruled out that the observed effect of

p53 hot-spot variants on association with NQO1 is

attributable to an altered conformation of the p53

variant that is different from the wild-type one

However, their finding implies that the two

alterna-tive pathways for p53 degradation, the NQO1

depen-dent and the Mdm2 dependepen-dent, must have different

p53 structural requirements Crystal structures of

com-plexes between the core domain of human p53 and

DNA half-sites reveal that two of the three residues

mentioned above (Arg248, Arg273) are located at the

interface of p53 and the DNA helix [76,77], indicating

that the same residues that are involved in DNA

recognition and binding are actually responsible for

association of p53 with NQO1 (Fig 4)

Conclusions and open questions

The last decade has witnessed accumulating evidence

for a role of eukaryotic QRs in regulating the 20S

pro-teasomal degradation of certain transcription factors

(e.g p53, Yap4p) and possibly proteins possessing a

high degree of unstructured segments (e.g ODC) The

body of information available clearly indicates that

this pathway is relevant for the cell and complements

other pathways such as ubiquitin-dependent 26S

prote-asomal degradation mediated by Mdm2 The concept

of protecting a protein by ‘hiding it near the lion’s

den’ (the catalytic chamber of the proteasome) is at

first unexpected However, the proteasome represents

an enormous surface (213 210 A˚2) [78] that offers itself

for extensive protein–protein interactions and perhaps

the interaction of QR and the 20S proteasome is just

one example of many others still to be discovered

Several issues remain unclear As discussed above,

structural and biochemical information on how the

recognition and binding between the various players

occurs is lacking Although NAD(P)H is the most

likely reducing agent for QR, it is not clear how the

flavin is reoxidized, or in other words by which

chemi-cal messenger (a quinone?) transcription factors are

released from their protecting protein complex

Fol-lowing this event, the transcription factor must rapidly

relocate to the nucleus or else be degraded by the

20S or 26S proteasome Again, the mechanism of relocalization remains obscure Finally, we do not know whether our list of proteins that are subject to protection by binding to QRs is complete Are there other transcription factors and proteins in eukaryotes? And what does that tell us about their cellular function?

Acknowledgements

This work was supported by the Austrian Fonds zur Fo¨rderung der wissenschaftlichen Forschung (FWF) through the Doktoratskolleg ‘Molecular Enzymology’ W901-B05 to PM

References

1 Lind C, Cadenas E, Hochstein P & Ernster L (1990) DT-diaphorase: purification, properties, and function Methods Enzymol 186, 287–301

2 Vasiliou V, Ross D & Nebert DW (2006) Update of the NAD(P)H:quinone oxidoreductase (NQO) gene family Hum Genomics 2, 329–335

3 Ernster L & Navazio F (1958) Soluble diaphorase in animal tissue Acta Chem Scand 12, 595–602

4 Amzel LM, Bryant SH, Prochaska HJ & Talalay P (1986) Preliminary crystallographic X-ray data for an NAD(P)H:quinone reductase from mouse liver J Biol Chem 261, 1379

5 Ysern X & Prochaska HJ (1989) X-ray diffraction anal-yses of crystals of rat liver NAD(P)H:(quinone-accep-tor) oxidoreductase containing cibacron blue J Biol Chem 264, 7765–7767

6 Li R, Bianchet MA, Talalay P & Amzel LM (1995) The three-dimensional structure of NAD(P)H:quinone reductase, a flavoprotein involved in cancer chemopro-tection and chemotherapy: mechanism of the two-elec-tron reduction Proc Natl Acad Sci USA 92, 8846–8850

7 Sancho J (2006) Flavodoxins: sequence, folding, binding, function and beyond Cell Mol Life Sci 63, 855–864

8 Foster CE, Bianchet MA, Talalay P, Zhao Q & Amzel

LM (1999) Crystal structure of human quinone reduc-tase type 2, a metalloflavoprotein Biochemistry 38, 9881–9886

9 Liger D, Graille M, Zhou CZ, Leulliot N, Quevillon-Cheruel S, Blondeau K, Janin J & van TH (2004) Crystal structure and functional characterization of yeast YLR011wp, an enzyme with NAD(P)H-FMN and ferric iron reductase activities J Biol Chem 279, 34890–34897

10 Iyanagi T & Yamazaki I (1970) One-electron-transfer reactions in biochemical systems V Difference in the mechanism of quinone reduction by the NADH dehydrogenase and the NAD(P)H dehydrogenase (DT-diaphorase) Biochim Biophys Acta 216, 282–294

11 Chen H, Wang RF & Cerniglia CE (2004) Molecular

cloning, overexpression, purification, and

characteriza-tion of an aerobic FMN-dependent azoreductase from

Enterococcus faecalis Protein Expr Purif 34, 302–310

12 Deller S, Sollner S, Trenker-El-Toukhy R, Jelesarov I,

Gubitz GM & Macheroux P (2006) Characterization of

a thermostable NADPH:FMN oxidoreductase from the

mesophilic bacterium Bacillus subtilis Biochemistry 45,

7083–7091

13 Vella F, Ferry G, Delagrange P & Boutin JA (2005)

NRH:quinone reductase 2: an enzyme of surprises and

mysteries Biochem Pharmacol 71, 1–12

14 Maerki F & Martius C (1960) Vitamin K reductase,

preparation and properties Biochem Z 333, 111–135

15 Deller S, Macheroux P & Sollner S (2008)

Flavin-dependent quinone reductases Cell Mol Life Sci 65,

141–160

16 Ross D & Siegel D (2004) NAD(P)H:quinone

oxidoreductase 1 (NQO1, DT-diaphorase), functions

and pharmacogenetics Methods Enzymol 382,

115–144

17 Wang T, Arifoglu P, Ronai Z & Tew KD (2001)

Gluta-thione S-transferase P1-1 (GSTP1-1) inhibits c-Jun

N-terminal kinase (JNK1) signaling through interaction

with the C-terminus J Biol Chem 276, 20999–21003

18 Groll M, Bochtler M, Brandstetter H, Clausen T &

Huber R (2005) Molecular machines for protein

degra-dation Chembiochem 6, 222–256

19 Groll M & Clausen T (2003) Molecular shredders: how

proteasomes fulfill their role Curr Opin Struct Biol 13,

665–673

20 Groll M & Huber R (2003) Substrate access and

pro-cessing by the 20S proteasome core particle Int J

Bio-chem Cell Biol 35, 606–616

21 Hershko A & Ciechanover A (1998) The ubiquitin

sys-tem Annu Rev Biochem 67, 425–479

22 Glickman MH & Ciechanover A (2002) The ubiquitin–

proteasome proteolytic pathway: destruction for the

sake of construction Physiol Rev 82, 373–428

23 Asher G & Shaul Y (2006) Ubiquitin-independent

deg-radation: lessons from the p53 model Isr Med Assoc J

8, 229–232

24 Pickart CM & Cohen RE (2004) Proteasomes and their

kin: proteases in the machine age Nat Rev Mol Cell

Biol 5, 177–187

25 Verma R, Aravind L, Oania R, McDonald WH,

Yates JR III, Koonin EV & Deshaies RJ (2002) Role

of Rpn11 metalloprotease in deubiquitination and

degradation by the 26S proteasome Science 298, 611–

615

26 Yao T & Cohen RE (2002) A cryptic protease couples

deubiquitination and degradation by the proteasome

Nature 419, 403–407

27 Kohler A, Cascio P, Leggett DS, Woo KM, Goldberg

AL & Finley D (2001) The axial channel of the

protea-some core particle is gated by the Rpt2 ATPase and controls both substrate entry and product release Mol Cell 7, 1143–1152

28 Shringarpure R, Grune T, Mehlhase J & Davies KJ (2003) Ubiquitin conjugation is not required for the degradation of oxidized proteins by proteasome J Biol Chem 278, 311–318

29 Brooks P, Fuertes G, Murray RZ, Bose S, Knecht E, Rechsteiner MC, Hendil KB, Tanaka K, Dyson J & Rivett J (2000) Subcellular localization of proteasomes and their regulatory complexes in mammalian cells Bio-chem J 346 Pt 1, 155–161

30 Asher G, Tsvetkov P, Kahana C & Shaul Y (2005) A mechanism of ubiquitin-independent proteasomal degra-dation of the tumor suppressors p53 and p73 Genes Dev 19, 316–321

31 Asher G & Shaul Y (2005) p53 proteasomal degrada-tion: poly-ubiquitination is not the whole story Cell Cycle 4, 1015–1018

32 Nioi P & Hayes JD (2004) Contribution of NAD(P)H:quinone oxidoreductase 1 to protection against carcinogenesis, and regulation of its gene by the Nrf2 basic-region leucine zipper and the arylhydrocar-bon receptor basic helix–loop–helix transcription fac-tors Mutat Res 555, 149–171

33 Anderson SP, Howroyd P, Liu J, Qian X, Bahnemann

R, Swanson C, Kwak MK, Kensler TW & Corton JC (2004) The transcriptional response to a peroxisome proliferator-activated receptor alpha agonist includes increased expression of proteome maintenance genes

J Biol Chem 279, 52390–52398

34 Sollner S, Schober M, Wagner A, Prem A, Lorkova L, Palfey BA, Groll M & Macheroux P (2009) Quinone reductase acts as a redox switch of the 20S yeast pro-teasome EMBO Rep 10, 65–70

35 Kahana C, Asher G & Shaul Y (2005) Mechanisms of protein degradation: an odyssey with ODC Cell Cycle

4, 1461–1464

36 Orlowski M, Cardozo C & Michaud C (1993) Evidence for the presence of five distinct proteolytic components

in the pituitary multicatalytic proteinase complex Properties of two components cleaving bonds on the carboxyl side of branched chain and small neutral amino acids Biochemistry 32, 1563–1572

37 Groll M, Ditzel L, Lowe J, Stock D, Bochtler M, Bartunik HD & Huber R (1997) Structure of 20S proteasome from yeast at 2.4 A˚ resolution Nature 386, 463–471

38 Ghaemmaghami S, Huh WK, Bower K, Howson RW, Belle A, Dephoure N, O’Shea EK & Weissman JS (2003) Global analysis of protein expression in yeast Nature 425, 737–741

39 Gong X, Kole L, Iskander K & Jaiswal AK (2007) NRH:quinone oxidoreductase 2 and NAD(P)H:quinone oxidoreductase 1 protect tumor suppressor p53 against