CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES

CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES CHAPTER 14 – INVENTORY AND EVOLUTION OF FUNGAL ABC PROTEIN GENES

I NTRODUCTION

The baker’s yeast Saccharomyces cerevisiae was

the first eukaryotic organism to have its

com-plete genome sequence determined, revealing

30 distinct genes encoding ATP-binding

cassette (ABC) proteins (Bauer et al., 1999;

Decottignies and Goffeau, 1997; Taglicht and

Michaelis, 1998) ABC proteins are ubiquitous

and form one of the largest gene families

known with more than 2000 distinct ABC genes

present in various current databases, e.g

Interpro (www.ebi.ac.at/interpro/) or Prosite

(www.expasy.ch/Prosite) All known ABC

pro-teins share a common hallmark domain, the

highly conserved ABC domain, also known as

the nucleotide-binding domain (NBD) The NBD

contains signature motifs found in all ABC

pro-teins operating from bacteria to man (Higgins,

1992) Membrane-bound ABC proteins also

con-tain variable numbers of membrane-spanning

domains arranged in certain membrane

archi-tectures Many ABC proteins transport a

vari-ety of compounds across cellular membranes

by an active process that is coupled to ATP

hydrolysis These ABC proteins are therefore

referred to as ABC transporters or pumps

While some pumps seem to transport various

xenobiotics, others exhibit a rather narrow

sub-strate spectrum Notably, for many ABC

pro-teins no defined substrates or even physiological

roles are known Interestingly, ABC proteins not

only function as simple membrane translocators

for molecules, they can also act as receptors, sensors, proteases, channels, channel regula-tors and even signal-ing components (Higgins, 1995) The question of how the highly con-served molecular architecture of ABC proteins entertains such a functional diversity remains elusive Hence, the functions of many ABC pro-teins may hold surprises and many important issues remain to be discovered In this chapter,

we will discuss the structure, function and properties of fungal ABC proteins, focusing on

the inventory of ABC genes in S cerevisiae.

Because the functional annotation of the yeast genome is fairly advanced, we will also compare the yeast ABC inventory to those

from fungal pathogens (Candida albicans and

Aspergillus fumigatus) whose genomes have

been sequenced or are close to being sequenced

Based on their molecular architecture, one can distinguish two types of yeast ABC proteins

The first type contains at least one transmem-brane domain (TMD), while the second type

lacks any obvious MSD (Figure 14.1) The

ABC Proteins: From Bacteria to Man ISBN 0-12-352551-9

Copyright 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

14

I NVENTORY AND E VOLUTION OF

F UNGAL ABC P ROTEIN G ENES

CHAPTER

architecture of yeast ABC proteins also includes

one or two highly conserved ABC domains

or NBDs, encompassing roughly 200 amino

acids The most conserved features found in

any given NBD are the Walker A and B motifs

[AG]-X(4)-G-K-[ST] and

[RK]-X(3)-G-X(3)-L-hydrophobic(4)-D, which are present in all

ATP-binding proteins (Walker et al., 1982), and

the ABC signature motif

[LIVMFYC]-[SA]-

[SAPGLVFYKQH]-G-[DENQMW]-[KRQASP-CLIMFW]-[KRNQSTAVM]-[KRACLVM]-[LIV

MFYPAN]-{PHY}-[LIVMFW]-[SAGCLIVP]-{FYWHP}-{KRHP}-[LIVMFYWSTA] – (Prosite

PS00211) Moreover, two additional regions

pro-vide diagnostic sequences for ABC proteins –

the center motif located between Walker A and

B, and the sequences found downstream of

Walker B (Michaelis and Berkower, 1995) The

molecular architecture of eukaryotic ABC

pro-teins arranges NBDs with TMDs in two

possi-ble ways Yeast ABC proteins come in a

duplicated TMD1-NBD1-TMD2-NBD2 forward

or a mirror image NBD1-TMD1-NBD2-TMD2

reverse topology The reverse architecture of

such full-size transporters is found mainly in

the PDR subfamily (Table 14.1), while the

for-ward orientation is similar to the one present in

mammalian P-glycoproteins (Gros et al., 1986).

However, so-called half-size transporters of

both the TMD-NBD and NBD-TMD topologies

are also known (Figure 14.1) Half-size ABC

transporters are believed to dimerize to form functional transporter molecules The recent elucidation of a high-resolution 3-D crystal

struc-ture of the Escherichia coli MsbA protein nicely

illustrates this interaction (Chang and Roth, 2001) In bacteria, each domain of a given ABC protein is encoded by a single gene, although many variations on this theme also exist

in prokaryotes (Young and Holland, 1999; Chapter 8)

Each TMD usually contains six predicted

␣-helical transmembrane-spanning segments (TMSs), although in some cases four to eight predicted TMSs per TMD are also known In sharp contrast to NBDs encompassing the hall-mark domains, only limited homology can be found within the TMDs of different ABC pro-teins (Decottignies and Goffeau, 1997; Michaelis and Berkower, 1995) The NBDs serve to bind and hydrolyze ATP or other NTPs, thereby fuel-ing transport processes However, numerous studies and genetic analyses have shown that NBDs not only serve as the fueling domains, but they appear intimately linked to the func-tion and/or structure of individual ABC pro-teins Importantly, the functions of N-terminal and C-terminal NBDs are not necessarily equivalent and thus each NBD of a eukaryotic ABC protein is indispensable The analysis of the evolutionary sequence relationships between individual NBDs of yeast ABC proteins rev-ealed five distinct clusters of homology Hence, the yeast ABC gene inventory comprises

30 genes subdivided into the PDR, MDR, ALDP, MRP/CFTR, and YEF3/RLi families (Bauer

et al., 1999; Decottignies and Goffeau, 1997;

Michaelis and Berkower, 1995)

This subfamily includes the Pdr5p, Pdr10p, Pdr15p, Pdr11p, Pdr12p, Snq2p, Ynr070p, Adp1p and Aus1p/YOR011w ABC proteins Their func-tion might be linked to cellular detoxificafunc-tion, although in several cases no substrates have been identified The overexpression of Pdr5p, Snq2p and Yor1p confers pleiotropic drug resistance (PDR) phenotypes These genes confer resistance to hundreds of chemically unrelated

Figure 14.1 Molecular architecture and predicted

membrane topology of yeast ABC proteins The

cartoon depicts the predicted membrane topologies

and architecture present in distinct subfamilies of

yeast ABC proteins.

drugs, including agricultural fungicides,

benz-imidazoles, dithiocarbamates, azoles,

myco-toxins, herbicides, cycloheximide, sulfometuron,

nigericin and anticancer drugs (Balzi et al., 1987;

Bissinger and Kuchler, 1994; Cui et al., 1996;

Hirata et al., 1994; Katzmann et al., 1995; Kralli

et al., 1995; Servos et al., 1993) These ABC genes

and their regulation are described in great

detail in Chapter 15

The Pdr12p pump seems to have a distinct physiological role, as it does not transport hydrophobic drugs, but confers resistance to weak organic acids Pdr12p mediates the energy-dependent extrusion of carboxylate anions

(Piper et al., 1998), such as those used as food

preservatives, including benzoate, sorbate and propionate, as well as C1–C7 weak organic acids, some of which are produced during normal

TABLE14.1 THE INVENTORY OFABC PROTEINS INSACCHAROMYCES CEREVISIAE

MDR family

Ste6p a-factor pheromone 1290 (TMS6-ABC)2 PM, GV, ESM Atm1p Fe/S proteins 694 TMS6-ABC Mito IM Mdl1p ? 696 TMS6-ABC Mito IM Mdl2p ? 820 TMS6-ABC Mito IM

PDR family

Pdr5p Drugs, steroids, antifungals, PL 1511 (ABC-TMS6)2 PM Pdr10p ? 1564 (ABC-TMS6)2 PM Pdr15p ? 1529 (ABC-TMS6)2 PM Snq2p Mutagens, drugs 1501 (ABC-TMS6)2 PM Pdr12p Weak organic acids 1511 (ABC-TMS6)2 PM Pdr11p ? 1411 (ABC-TMS6)2 PM Aus1p/YOR011c ? 1394 (ABC-TMS6)2 ? Adp1p ? 1049 TMS2-ABC-TMS7 ? YNR070w ? 1333 (ABC-TMS6)2 ? YOL075c ? 1095 (ABC-TMS6)2 ?

MRP/CFTR family

Yor1p Oligo, revero, PL 1477 TMD0(TMS6-ABC)2 PM Ycf1p GS-conjugates, Cd 2+ , UCB, BA 1515 TMD0(TMS6-R-ABC)2 Vacuole Ybt1p BA 1661 TMD0(TMS6-ABC)2 Vacuole Bpt1p UCB 1559 TMD0(TMS6-ABC)2 Vacuole, ERM?

YHL035c ? 1592 TMD0(TMS6-ABC)2 ? YKR103w/YKR104c 1524 (TMS6-ABC)2 ?

ALDp family

Pxa1p LCFA 870 TMS6-ABC Peroxisomes Pxa2p LCFA 853 TMS6-ABC Peroxisomes

YEF3/RLI family

Yef3p Hygromycin, paro 1044 ABC2 Ribo?, Cyt?

Gcn20p 752 ABC2 Polysomes Hef3p 1044 ABC2 Cytosol?

New1p/YPL226w 1196 TMS3-ABC2 ? Kre30p/YER036c 610 ABC2 ? Rli1p/YDR091c 608 ABC2 ?

Non-classified

Caf16p/YFL028c 289 ABC ? ABC, ATP-binding cassette; TMD0, transmembrane domain; TMS, transmembrane segment; GS, glutathione S;

UCB, unconjugated bilirubin; BA, bile acids; PL, phospholipids; oligo, oligomycin; revero, reveromycin A;

paro, paromomycin; LCFA, long chain fatty acids; PM, plasma membrane; ERM, endoplasmic reticulum membrane;

ESM, endosomal membranes; Cyt, cytoplasm; Ribo, ribosome; Mito IM, mitochondrial inner membrane.

cellular metabolism Notably, PDR12 mRNA

synthesis is dramatically induced by sorbic acid

stress and by exposure of yeast cells to low pH

stress (Piper et al., 1998), demonstrating that

Pdr12 in fact represents a stress response gene

Aus1p (YOR011w) is closely related to Pdr11p, sharing more than 65% sequence

iden-tity Non-essential Aus1p appears to be involved

in the uptake of sterols, as a ⌬aus1 deletion

mutant exhibits a reduced accumulation of

cholesterol, while no obvious phenotypes

are discernible under standard growth

condi-tions (SGD http://genomewww.stanford.edu/

Saccharomyces/) The function of other

mem-bers of the PDR subfamily such as Pdr11p,

Pdr10p, Pdr15p, Adp1p and Ynr070p remains

unknown and no data are currently available

regarding their substrates or physiological roles

Because Pdr10p and Pdr15p are tightly regulated

by adverse conditions such as high osmolarity

and heat shock, respectively, their functions

might also be linked to a cellular response

(Wolfger et al., in preparation).

With the exception of Adp1p, all members

of this group display a predicted

NBD1-TMD1-NBD2-TMD2 structure with usually 12

pre-dicted TMSs Adp1p exhibits a slightly different

architecture, replacing the first NBD with a large

soluble domain, followed by a

TMD1-NBD2-TMD2 topology (Figure 14.1) It is noteworthy

that all PDR members localize to the plasma

membrane as shown in Figure 14.2 This cell

surface localization further supports their

pur-ported function in cellular detoxification and

cellular stress responses, although their precise

roles and cellular substrates remain an enigma

Finally, a substantial number of PDR subfam-ily members have been identified in other fungal

species, including fungal pathogens All PDR

homologues linked to multidrug resistance are

extensively discussed in Chapter 15 These ABC

proteins currently total almost 50 fungal PDR

genes For many related PDR family members,

a cellular function has not been established

beyond the one known for the corresponding

counterpart in baker’s yeast Examples for

mem-bers of this impressive and growing group of

fungal detoxification proteins are Candida krusei

Abc1p, Schizosaccharomyces pombe bfr1+/Hba2p

(Turi and Rose, 1995), Candida glabrata Cgr1p/

Pdh1p (Miyazaki et al., 1998), Penicillium

digita-tum Pmr1p (Nakaune et al., 1998), Emericella

nidu-lans AtrAp/ANPGP1p (Del Sorbo et al., 1997) and

AtrBp/ANPGP2p (Andrade et al., 2000), A

fumi-gatus AtrFp, C albicans Cdr1p (Prasad et al.,

1995), Cdr2p (Sanglard et al., 1997), Cdr3p (Balan

et al., 1997) and Cdr4p (Franz et al., 1998), Botrytis cinerea BCPGP1p, Cryptococcus neofor-mans eCdr1p and Magnaporte grisea Abc1p

(Urban et al., 1999) This incomplete list

illus-trates the diversity of this ABC transporter fam-ily and hence underscores its importance, with more members surfacing at a rapid pace The interested reader is referred to publicly accessible databases such as Swissprot (www.expasy.ch) or Interpro (www.ebi.ac.uk/Interpro) to obtain continuously updated information

Members of this class exhibit a membrane topol-ogy such as TMD0-TMD1-NBD1-TMD2-NBD2

(Tusnady et al., 1997) The C-terminal TMD

comprises 11 predicted TMSs, interrupted by a small cytoplasmic domain Yeast MRP/CFTR-like pumps include Yor1p, Ycf1p, Bpt1p, Ybt1p, YHL035w and YKR103/YKR104w The YKR103/YKR104w open reading frames (ORFs) include a stop codon between MSD2 and NBD2 and thus represent perhaps a pseudogene or a sequencing error

Figure 14.2 Subcellular localization of yeast ABC proteins The cartoon depicts the subcellular

localization of yeast ABC proteins in various cellular membranes or compartments For a list of ABC proteins and further details see text and

Table 14.1.

Yor1p is probably among the best-studied members of the MRP family The gene was

initially isolated in a genetic screen for genes

conferring resistance to oligomycin (Katzmann

et al., 1995) Yor1p is localized to the plasma

membrane and has overlapping functions with

PDR pumps such as Pdr5p, Snq2p and even

Pdr12p, although Yor1p exhibits quite unique

substrate specificities (Table 14.1) The ⌬yor1

null mutant is viable, but displays increased

sensitivity to a variety of compounds,

includ-ing azoles, antibiotics such as tetracycline,

erythromycin and oligomycin, as well as

anti-cancer drugs like daunorubicin and

doxo-rubicin, carboxylic acids such as acetic, propionic

and benzoic acids, and heavy metals such as

cadmium (Cui et al., 1996) In contrast, Yor1p

overproduction confers resistance to many of

these compounds (Ogawa et al., 1998) The

function of Yor1p and its regulation is also

extensively discussed in Chapter 15

In contrast to Yor1p, Ycf1p is localized to the

vacuolar membrane (Figure 14.2)

Neverthe-less, like Yor1p, Ycf1p confers resistance to

cad-mium (Szczypka et al., 1994) Besides vacuolar

Cd2⫹ sequestration, Ycf1p is also involved in

vacuolar transport of reduced glutathione and

glutathione S-conjugates such as

glutathione-conjugated arsenite A homologue of Ycf1p,

Bpt1p, mediates transport of unconjugated

bilirubin into the vacuole A⌬ycf1 ⌬bpt1 double

mutant is blocked for vacuolar transport of

unconjugated bilirubin Ycf1p is related to the

human multidrug resistance proteins MRP1

and MRP2, and has 45% overall similarity to

human CFTR (cystic fibrosis transmembrane

conductance regulator) based on a ClustalW 1.4

alignment It is interesting to note that yeast

sequesters heavy metals to the vacuole, rather

than extruding them Such a ‘social’ behavior of

a unicellular organism might be explained by a

beneficial effect on immediate neighbors Finally,

Ybt1p, the yeast bile transporter (formerly Bat1p)

mediates vacuolar uptake of bile acids such as

taurocholate (Ortiz et al., 1997) Another close

homologue of Ybt1p, the YHL035w gene

prod-uct, has not been studied and its physiological

cargo and cellular localization has not been

elucidated as yet

ABC proteins of the MRP/CFTR family have also been identified in other fungi However,

in contrast to the large PDR family,

substan-tially less information is available on fungal

genes of this family In S pombe, YAWB (also

SPAC3F10.11C) and ABC1 (Christensen et al.,

1997b) have been identified as MRP/CFTR

family members, as well as a gene from

Neurospora crassa (B7A16.190) and a Yor1p

homo-logue in C albicans (Ogawa et al., 1998).

This small subfamily contains only two half-size transporters, Pxa1p and Pxa2p, displaying a TMD-NBD membrane topology Pxa1p/ Pxa2p are yeast orthologues of human Pmp70/ABCD3/

PXMP1, ALD/ALDR/ABCD2 and ABCD4/

PXMP1L/PMP69 peroxisomal disease genes associated with neurodegenerative diseases such as adrenoleukodystrophy and Zellweger

syndrome (Gartner and Valle, 1993; Holzinger et

al., 1997, 1999; Kamijo et al., 1992) Indeed, both

Pxa1p and Pxa2p localize to the peroxisomal membrane and might function as heterodimers

(Shani et al., 1996; Swartzman, et al., 1996) They

are thought to mediate peroxisomal uptake of very long chain fatty acids to undergo degrada-tion through ␤-oxidadegrada-tion (Watkins et al., 2000), which is consistent with the presence of a fatty acid-binding domain in Pxa1p/Pxa2p The null mutants fail to grow on fatty acids such as palmi-tate or oleate as the sole carbon source Although the Pxa1p/Pxa2p complex is required for perox-isome function, it is dispensable for peroxperox-isome biogenesis or for import of peroxisomal matrix

proteins While the PXA1 gene is only expressed when cells grow on oleate, the PXA1 and PXA2

promoters lack any consensus oleate-response

elements, yet PXA1, but not PXA2, is

oleate-induced and transcription is

Oaf1p/Pip2p-dependent (Bossier et al., 1994; Swartzman et al.,

1996) The regulators Oaf1p and Pip2p represent the two key transcription factors for peroxisome biogenesis in yeast In contrast to the situation with the PDR family, only a few ALDP homo-logues have been described in other fungi, mostly from genomic sequencing approaches of

other fungal genomes (Figure 14.3 A, B).

This subfamily contains the ABC proteins Mdl1p, Mdl2p, Atm1p and Ste6p The Ste6p

a-factor pheromone transporter is a full-size

transporter displaying the duplicated

(TMD-NBD)2 topology Ste6p is localized in Golgi

vesicles, the plasma membrane and perhaps

endocytic vesicles (Kölling and Hollenberg,

1994; Kuchler, 1993; Michaelis, 1993) Ste6p is

a haploid-specific transporter required for the

export of farnesylated a-factor, a pheromone

absolutely required for mating in yeast Ste6p

was the first ABC transporter identified in yeast

(Kuchler et al., 1989; McGrath and Varshavsky,

1989), closing an evolutionary gap between the

E coli hemolysin transport system (Wang et al.,

1991) and mammalian Mdr1p P-glycoprotein

mediating multidrug resistance (Chen et al.,

1986; Gros et al., 1986) Interestingly, the steady

state concentration of Ste6p was found to be

highest in the Golgi vesicles, although its

func-tion is clearly required in the plasma membrane

(Berkower et al., 1994; Kuchler et al., 1993).

Because Ste6p travels through all exo- and

endo-cytic compartments, it serves as a useful model

membrane protein for intracellular trafficking, proteolytic degradation, endocytosis, and even

vacuolar sorting studies (Berkower et al., 1994;

Kölling and Hollenberg, 1994; Kuchler, 1993;

Kuchler et al., 1989) Moreover, Ste6p has been

subjected to extensive molecular studies to unravel the molecular mechanisms of ABC transporter-mediated peptide transport Ste6p function can be easily tested through convenient assays such as mating (Kuchler and Egner, 1997) Notably Ste6p, although an MDR family member, does not confer typical multidrug

resistance phenotypes Extracellular a-factor

pheromone is essential for the sexual reproduc-tion cycle of haploid yeast cells Ste6p funcreproduc-tions

at the plasma membrane, providing the

rate-limiting step in a-factor export After

phero-mone extrusion, Ste6p is rapidly removed from the cell surface through ubiquitin-mediated endocytosis, and delivered to the vacuole for

terminal degradation (Egner et al., 1995; Kölling

and Hollenberg, 1994) Pheromone export

MRP/

CFTR MDR

Figure 14.3 Similarity relationships of fungal ABC proteins A, A dendrogram in which the entire yeast

inventory is compared with sequences from the Aspergillus genome project, including the apparent

classification into yeast subfamilies Preliminary sequence data was obtained from The Institute for

Genomic Research website at http://www.tigr.org B, The same dendrogram for the Candida albicans

genome Sequence data for C albicans was obtained from the Stanford Genome Technology Center website

at http://www-sequence.stanford.edu/group/candida Sequencing of C albicans was accomplished with the support of the NIDR and the Burroughs Welcome Fund.

occurs through a non-classical route, bypassing

the vesicular secretory pathway (Kuchler, 1993)

Interestingly, severing experiments demon-strate that both Ste6p halves, when coexpressed

as individual half-size transporters, mediate

pheromone export (Berkower et al., 1996) This

indicates that both Ste6p halves are required for

function and that they can interact in vivo to

form a functional a-factor transporter The

trans-port substrate, a-factor, is extremely

hydropho-bic due to its C-terminal lipid modification and

carboxy-methylation While mutations in the

structural gene encoding a-factor do not

dra-matically affect its secretion, a lack of a-factor

farnesylation or methylation debilitates its

release (Sapperstein et al., 1994) Hence, the lipid

moiety or its hydrophobicity may represent an

essential recognition determinant for Ste6p As

with many other eukaryotic ABC transporters,

Ste6p is powered by ATP hydrolysis, because

many NBD mutations destroy function (Browne

et al., 1996), and because Ste6p binds

photo-activatable ATP analogues (Kuchler et al., 1993).

Interestingly, Ste6p might also play a role in cell

fusion, since ste6 mutants were isolated that still

mediate a-factor export, but fail to complete

fusion of haploid mating partners (Elia and

Marsh, 1996) Taken together, the precise

mecha-nism by which the Ste6p ABC transporter

medi-ates the actual pheromone translocation across

the plasma membrane is somewhat

mysteri-ous, but it appears as if intracellular a-factor

pre-cursor processing and translocation across the

plasma membrane are tightly coupled (Kuchler

and Egner, 1997; Michaelis, 1993)

The half-size molecules Mdl1p, Mdl2p and Atm1p display a similar TMD-NBD topology

and localize to the mitochondrial inner

mem-brane (Figure 14.2) Mdl1p is related to

mam-malian P-glycoproteins and to a greater extent

to the mammalian peptide transporter of

anti-gen presentation, TAP (Dean and Allikmets,

1995) It is required for efficient mitochondrial

export of rather long peptides of 2100–2600 Da

molecular mass These peptides are proteolytic

degradation products of inner membrane

pro-teins generated by mAAA proteases Afg3p and

Yta12p However, Mdl1p fails to transport

short peptides or free methionine (Young et al.,

2001) Notably, Mdl2p seems to play a different

role in mitochondrial function, since it has not

been implicated in peptide transport processes

It is therefore likely that Mdl1p and Mdl2p

may form functional homodimers, which

con-trasts with the situation of peroxisomal Pxa1p

and Pxa2p Furthermore, Mdl1p and Mdl2p

co-purify at molecular masses of approximately

200 kDa and 300 kDa, respectively, suggesting that they are part of distinct oligomeric protein

complexes (Young et al., 2001).

The third member of the yeast MDR group, Atm1p, is related to the human ABCB7/ABC7 protein, which is implicated in the mitochon-drial X-linked sideroblastic anemia and ataxia

(Allikmets et al., 1999) Atm1p is required for

mitochondrial DNA maintenance or stability, but this function might be an indirect pheno-typic effect observed in the ⌬atm1 mutant The

atm1-1 mutant displays a high level of damage

and even loss of mitochondrial DNA during

growth on rich medium Interestingly, the ATM1

mRNA localizes in close proximity to mito-chondria in living cells, as demonstrated using

a GFP fusion protein that binds to a

heterolo-gous sequence in a reporter ATM1 mRNA (Corral-Debrinski et al., 2000) Atm1p is also

required for the assembly of iron–sulfur clus-ters of cytoplasmic iron–sulfur-containing pro-teins, and thus may be involved in the export of mitochondrial heme required for cluster

assem-bly (Pelzer et al., 2000).

ABC proteins of the MDR family have also been identified in other fungal species For

example S pombe Mam1p is similar in length

and domain structure to Ste6p and shares about 30% sequence identity, thus representing the Ste6p orthologue in fission yeast (Christensen

et al., 1997a) The C albicans Hst6p transporter

can also functionally complement a ⌬ste6 mutant

(Raymond et al., 1998) Further, MDR family homologues have been identified in A fumigatus (Mdr2p) (Tobin et al., 1997), C albicans Mdl1p (Swissprot ID: P97998), S pombe (YFX9 C9B6.09c) and Rhizomucor racemosus (Trembl ID: Q9C163/

Pgy1p)

This S cerevisiae subfamily includes Yef3p,

Hef3p, Rli1p, Gcn20p, Kre30p, Caf16p and New1p Except for New1p, these ABC proteins lack any predicted TMSs normally present in other ABC transporters Surprisingly, three ABC

proteins from this class, Yef3p, Kre30p and

Rli1p, are essential for viability under standard

growth conditions These proteins are involved

in cellular functions that appear unrelated to

transport events, and the functional role of the

NBDs is in most cases not clear

Yef3p perhaps localizes to the cytoplasm or even to ribosomes It is also known as

transla-tion elongatransla-tion factor EF-3A, which has a

func-tion in tRNA binding and dissociafunc-tion from the

ribosome (Chakraburtty, 1999) Yef3p displays

basal ATPase activity, which is stimulated by

the presence of ribosomes by two orders of

magnitude, suggesting that Yef3p might at

least interact with ribosomes or in fact localize

to ribosomes (Gontarek et al., 1998) A

ribosome-binding site and a putative tRNA-ribosome-binding

domain is located near the C-terminus of Yef3p

ATP hydrolysis facilitates EF-3 dissociation

from the ribosome In eukaryotes only fungal

homologues are known, suggesting that Yef3p

is a unique fungal translation elongation factor

(Sarthy et al., 1998) Whole-genome

transcrip-tome profiling of a conditional null-mutant

indicates a gene expression pattern that

resem-bles that of wild-type cells treated with

cyclo-heximide, suggesting a role for Yef3p in

blocking ribosomes in vivo (Hughes et al., 2000).

The YEF3 mRNA levels are modulated by a

variety of conditions It is repressed by

rapamycin and peroxide or heat shock stress

conditions (Causton et al., 2001), while

hyper-expressed in high density cultures and during

diauxic shift Notably, overexpression of Yef3p

renders cells hypersensitive to paromomycin

and hygromycin B, two translational inhibitors

(Sandbaken et al., 1990).

Hef3p (also known as Yef3Bp) shares 84%

overall identity with Yef3p, implying a similar

or overlapping function Indeed, Hef3p can

rescue a yef3 null mutant when expressed from

the YEF3 or ADH1 promoter (Sarthy et al.,

1998) In striking contrast to loss of Yef3p,

how-ever, a ⌬hef3 null mutant has no obvious growth

defect This might be explained by the fact that

Hef3p is not expressed under normal culture

conditions and its promoter is therefore inactive

(Maurice et al., 1998) Interestingly, HEF3 mRNA

levels are highly upregulated by limiting zinc

concentration in the growth medium (Yuan,

2000) The HEF3 mRNA abundance increases

during nitrogen starvation and during

station-ary phase, but is repressed by a shift to high

osmolarity (Causton et al., 2001) It will be

inter-esting to uncover the role of Hef3p under these

conditions

Like Yef3p, and perhaps Hef3p, Gcn20p has

a functional role in translation Gcn20p is a component of a protein complex required for the response to amino acid starvation, glucose

limitation and osmotic stress (Marton et al.,

1997) Together with Gcn1p, Gcn20p is proba-bly involved in detection of uncharged tRNA and transmission of this signal to Gcn2p, a pro-tein kinase which phosphorylates eIF2alpha Gcn1p, Gcn2p and Gcn20p form a complex and the apparent role of Gcn20p is to activate Gcn2p, through the stabilization of the interac-tion between Gcn1p and Gcn2p (Garcia-Barrio

et al., 2000) The gcn20 mutant phenotype is

similar to a gcn1 mutant, in that the null mutant

is viable under normal conditions and inviable under starvation conditions (Vazquez de

Aldana et al., 1995) The C-terminal region of

Gcn20p containing the ABC domain is dispen-sable for complex formation with Gcn1p and for the stimulation of Gcn2p kinase activity

(Marton et al., 1997), and the role of the Gcn20p

NBD remains obscure

The physiological roles of the following non-transporter ABC proteins are largely unknown and they may therefore provide some surprises

in the future Kre30p is required for viability and was initially isolated in a genetic screen for Killer toxin-resistant mutants The cellular function of Kre30p is not known, but it seems

to interact with other proteins as determined

by two-hybrid assays Interactions with several proteins, including Sma1p (spore membrane assembly) and Cbk1p (an S/T kinase required for sporulation) were discovered, but the phys-iological relevance of these interactions, if any, remains to be established

The N-terminal domain of New1p, which is especially rich in glutamine and asparagine residues, is able to support prion

inheri-tance when fused to SUP35 Sup35p is a

trans-lational release factor, eRF3, which interacts with Sup45p (eRF1) to form a translational release factor complex Moreover, Sup35p is also a prion-like molecule responsible for the [PSI⫹] determinant (Tuite et al., 1981) Although the cellular function of New1p remains elu-sive, it may behave as an epigenetic switch

(Santoso et al., 2000) The New1p sequence

also includes three predicted TMSs, although they are not clustered within a classical

TMD Finally, NEW1 mRNA levels are

repres-sed under stress conditions such as changes

in temperature, oxidation, nutrients, pH and

osmolarity (Causton et al., 2001; Jelinsky et al.,

2000)

The Caf16p and Ydr061w ABC proteins con-tain only a single NBD While nothing is

known about YDR061w, Caf16p seems to have

a role in PoIII-dependent transcription of some,

but not all, promoters Caf16p forms a dimer

and interacts with the RNA polymerase II

holoenzyme components Srb9p, Ssn3p, and

Ssn8p Finally, Rli1p is similar to the human

RNase L inhibitor (RLI) Its precise function

has not been established, although a ⌬rli1 null

mutation in yeast is lethal Human RLI is

prob-ably a regulator of the

2⬘,5⬘-oligoadenylate-dependent RNase L, which is involved in the

antiviral activity of interferons Some viruses

developed strategies to bypass the antiviral

activity of RNase L by virus-induced expression

of RLI (Martinand et al., 1999) Interestingly, the

C-terminal tail domain of yeast Ire1p

displays sequence similarity to mammalian

RNase L (Sidrauski and Walter, 1997) Ire1p is a

regulator of the unfolded protein response

pathway (UPR), which signals from the ER

to the nucleus (Cox et al., 1993) A direct role for

Rli1p in the UPR is possible but untested

as yet

Numerous homologues of yeast ABC genes

have also been identified in other fungal

species through functional complementation

approaches More importantly, genome

sequenc-ing of fungal pathogens such as C albicans and

A fumigatus provided complete sequence

datasets from their genomes and the data are

publicly available (TIGR: http://www.tigr.org/;

Stanford Genome Technology Center: http://

sequence-www.stanford.edu/) Although

func-tional annotation of ABC genes in these fungal

pathogens has been a difficult task, the

com-parison of various fungal ABC inventories has

become possible Because a global picture of

the evolutionary relationships of ABC genes

from various fungi has not been reported, we

have compared the inventory of baker’s yeast

ABC genes to various fungal genomes Yeast

ABC genes guided a search to detect and

iden-tify homologous sequences in other fungal

species, including C albicans and A fumigatus

(Figure 14.3 A, B) Previous work demonstrated

that all S cerevisiae NBDs generate clusters of

five subfamilies (Table 14.1) In a first round of

comparison, NBDs were identified using a translated pattern search against the nucleotide sequence databases The patterns were gener-ated by alignment of the respective subgroups

of S cerevisiae NBD sequences The

compari-son of amino acid sequence patterns with a translated nucleotide sequence minimizes the effect of sequencing errors causing trunca-tions or frameshift mutatrunca-tions In addition, the sequences were searched with the Prosite pat-terns for ABC proteins In the next step, the regions surrounding hits were analyzed in detail by extracting putative NBDs In cases where truncations due to frameshift mutations had occurred, ORFs were appropriately edited

to allow for the generation of meaningful den-drograms Next we generated an alignment

using the entire set of NBDs including S

cere-visiae NBDs The A fumigatus candidate genes

were first identified through a tblast at TIGR (http://www.tigr.org/) The dendrograms

shown in Figure 14.3 represent a graphical

dis-play of the sequence homologies as detected through the alignment, although it should be noted that this is not a phylogenetic tree

Furthermore, we intended to include a

dendro-gram showing the relationships to C

neofor-mans ABC genes, the sequence data of which

can be publicly accessed at http://www-sequence.stanford.edu/group/C.neoformans/

However, because of the confidentiality poli-cies of the sequencing consortium, we were prohibited from doing so

As shown in Figure 14.3, each subfamily from

baker’s yeast has an equivalent family in other fungi Thus, ABC proteins from other fungi form similar evolutionary relationships, and can thus

be classified into similar subfamilies The PDR

subfamily contains five Candida homologues

(Cdr1p, Cdr2p, Cdr3p, Cdr4p and Cdr99p) of Pdr5p, all of which are more similar to each other than to other yeast members of the Pdr5p-family

(Figure 14.3A) As in yeast, not all CDR genes are

implicated in drug resistance While Cdr1p and Cdr2p mediate clinical antifungal resistance, the function of Cdr3p and Cdr4p has not been linked

to drug efflux Homozygous deletion of CDR4

did not confer hypersensitivity to fluconazole

(Franz et al., 1998) Interestingly, the CDR3 gene

is regulated in a cell-type-specific manner, as it appears important in morphology switching, and it is not expressed in the standard labora-tory strain CAI4 However, in a WO-1 genetic strain background that switches between two

morphological states, white and opaque, the

CDR3 mRNA is only present in the opaque form.

Here, overexpression of Cdr3p did not result in

increased resistance to known drug substrates

of the PDR family (Balan et al., 1997)

Substan-tially less information is available on the PDR

family homologues in A fumigatus (Figure 14.3),

although in general, a clear species-specific

clustering becomes immediately apparent in

this family

As expected, the yeast ALDP subfamily has equivalent orthologues in all other fungal

pathogens In C albicans a homologue to

both Pxa1p and Pxa2p is detectable, while in

A fumigatus only one close match could be

identified Concerning the MDR subfamily,

single nearest matches to each Atm1p, Mdl1p,

Mdl2p and Ste6p were found in C albicans The

situation seems to be somewhat more

compli-cated in A fumigatus The A fumigatus MDR

members cluster together and do not allow even

a tentative assignment The C albicans

ortho-logue of Ste6p has been previously described

as Hst6p (Raymond et al., 1992) Surprisingly,

despite the diploid nature of C albicans, Hst6p

is able to functionally complement a ste6 null

mutant for a-factor transport in S cerevisiae.

The MRP subfamily indicates some differences

between S cerevisiae and C albicans No close

homologue to Ste6 was identified in A fumigatus.

Furthermore, we do not find a close neighbor

of Ybt1p and YHL035c, a fact that could also

be the consequence of incomplete databases

While single orthologues to Ycf1p and Yor1p

are present, two Candida ORFs are similar to

Bpt1p Thus, further experimental evidence

will be necessary to establish the roles of the two

Candida Bpt1ps, whether or not one of them

represents a functional homologue of Ybt1p

In the A fumigatus alignments we find a

Yor1p and Ycf1p homologue but several

can-didates for Ybt1p remain Finally, the

non-transporter ABC genes from the YEF3/RLI

subfamilies, as well as non-classified ABC genes,

all have corresponding genes in other fungal

species For instance, both Yef3p and Hef3p

cluster with the C albicans homologue Tef3p.

The Candida Eif3p, however, appears more

sim-ilar to New1p than to Hef3p Both Gcn20p and

Kre30p also have close homologues in Candida,

and Caf16p and Rli1p also have a single

coun-terpart in Candida Taken together, the

inven-tory of ABC proteins from fungal pathogens is

quite similar to the one present in baker’s

yeast, with similar subfamilies of close

evolu-tionary relationships

The completion of the entire yeast genome, and the availability of genomic tools such as whole-genome DNA microarrays, permitted the transcriptional profiling of many metabolic pathways It is therefore not surprising that expression regulation of yeast ABC genes was observed in numerous studies that investigated genome-wide expression of yeast genes For

example, PDR5, SNQ2, YOR1, PDR10, PDR11 and PDR15 share common transcriptional

regu-lators, such as the zinc-finger proteins Pdr1p,

Pdr3p or Yrr1p (Del Sorbo et al., 1997) These

regulators, also instrumental for PDR develop-ment, control a number of genes of both the

ABC family and non-ABC genes (DeRisi et al., 2000; Wolfger et al., 2001) A detailed

transcrip-tome analysis revealed the identification of numerous potential Pdr1p/Pdr3p target genes

(DeRisi et al., 2000) Moreover, PDR target genes

were also identified simply by the presence of

potential PDRE cis-acting motifs in yeast gene

promoters However, a Pdr1p/Pdr3p-dependent regulation has only been experimentally veri-fied for certain ABC genes and two MFS

perme-ases (Wolfger et al., 1997) It should be

emphasized that the molecular signals, includ-ing the transduction pathways affectinclud-ing tran-scriptional activities of Pdr1p, Pdr3p or Yrr1p, remain elusive A specific activation of these fac-tors by drugs has not been reported It is tempt-ing to speculate that PDR could evolve through increased mutation rate upon drug challenge or other adverse conditions Apart from other reg-ulatory influences, the mRNA levels of several ABC genes show dependencies on carbon and/or nitrogen source, stress regulation as well

as cell cycle-dependent fluctuations A closer inspection of the available literature on yeast ABC gene expression leads to the conclusion that individual mRNAs display a distinctive expression pattern Even closely related proteins such as the PDR group display striking dif-ferences under various conditions In many cases, the transcription factors involved remain unknown but a functional link between stress response and drug resistance is evident

Whole genome transcriptome analysis sug-gested that Snq2p is induced by heat shock,

H2O2and rapamycin, whereas PDR5 mRNA is