Abbreviations FCCP, carbonyl cyanide p-trifluoromethoxy-phenylhydrazone; Flx1p, mitochondrial FAD transporter; HA, hemagglutinin; PGI, phosphoglucoisomerase; RR-MADD, riboflavin-responsi
Trang 1in Saccharomyces cerevisiae – involvement of the
mitochondrial FAD transporter, Flx1p
Teresa A Giancaspero1, Robin Wait2, Eckhard Boles3and Maria Barile1
1 Dipartimento di Biochimica e Biologia Molecolare ‘‘E Quagliariello’’, Universita` degli Studi di Bari, Italy
2 Kennedy Institute of Rheumatology Division, Faculty of Medicine, Imperial College London, UK
3 Institut fu¨r Molekulare Biowissenschaften, J.W Goethe-Universita¨t, Frankfurt am Main, Germany
Several mitochondrial dehydrogenases and oxidases
require FMN and FAD for their activity [1,2] Thus,
intramitochondrial flavin cofactor availability is
poten-tially a crucial regulator of oxidative terminal
metabo-lism Consistent with this, some patients suffering from
riboflavin-responsive multiple acyl-CoA dehydrogenase
deficiency (RR-MADD) exhibit profound disorders in
mitochondrial biochemistry that are reversed by
treat-ment with high doses of riboflavin [3]
Mammals obtain flavin cofactors from dietary
ribo-flavin, which enters their cells via plasma membrane
riboflavin transporters, although these have not yet
been characterized at the molecular level [1,4] In
Sac-charomyces cerevisiae, the product of the MCH5 gene
was recently identified as a plasma membrane riboflavin
transporter [5], although this organism, in common with
other yeasts and plants, is able to synthesize riboflavin
de novoand export it into the culture medium [3–9]
In previous publications, we proposed that mainte-nance of flavin cofactor levels inside mitochondria requires the activity of mitochondrial riboflavin trans-port system(s) and two enzymes, riboflavin kinase (EC 2.7.1.26) and FAD synthetase (EC 2.7.7.2), which catalyze the synthesis of FMN and FAD respectively [10–13] In this scenario, the lumiflavin-sensitive flavin transporter, Flx1p, is responsible for FAD export from
S cerevisiae mitochondria (SCM) [13] Alternatively,
on the basis of the cytosolic localization of the FAD synthetase, encoded by FAD1 [14], other authors sug-gested that Flx1p is involved in mitochondrial FAD import in exchange with FMN [15]
FLX1 deletion or mutation results in a respiration-deficient phenotype, in which the activities of the mitochondrial FAD dependent-enzymes, lipoamide dehydrogenase and succinate dehydrogenase (SDH), are reduced [13,15] Measurement of the mitochondrial
Keywords
flavin; Flx1p; mitochondrial FAD transporter;
post-transcriptional control; succinate
dehydrogenase flavoprotein subunit
Correspondence
M Barile, Via Orabona, 4, 70126 Bari, Italy
Fax: +39 0805443317
Tel: +39 0805443604
E-mail: m.barile@biologia.uniba.it
(Received 30 July 2007, revised 27
December 2007, accepted 4 January 2008)
doi:10.1111/j.1742-4658.2008.06270.x
The mitochondrial FAD transporter, Flx1p, is a member of the mitochon-drial carrier family responsible for FAD transport in Saccharomyces cerevi-siae It has also been suggested that it has a role in maintaining the normal activity of mitochondrial FAD-binding enzymes, including lipoamide dehydrogenase and succinate dehydrogenase flavoprotein subunit Sdh1p A decrease in the amount of Sdh1p in the flx1D mutant strain has been deter-mined here to be due to a post-transcriptional control that involves regula-tory sequences located upstream of the SDH1 coding sequence The SDH1 coding sequence and the regulatory sequences located downstream of the SDH1 coding region, as well as protein import and cofactor attachment, seem to be not involved in the decrease in the amount of protein
Abbreviations
FCCP, carbonyl cyanide p-(trifluoromethoxy)-phenylhydrazone; Flx1p, mitochondrial FAD transporter; HA, hemagglutinin; PGI,
phosphoglucoisomerase; RR-MADD, riboflavin-responsive multiple acyl-CoA dehydrogenase deficiency; SCM, Saccharomyces cerevisiae mitochondria; SDH, succinate dehydrogenase; Sdh1p, succinate dehydrogenase flavoprotein subunit; WT, wild-type; a-FAD, polyclonal antibody against FAD covalently bound to protein; a-HA, monoclonal antibody against hemagglutinin epitope; b-Gal, b-galactosidase.
Trang 2flavin content in wild-type (WT) and flx1D mutant
yeast strains suggested that the impairment in
flavo-enzyme activity was not strictly correlated with flavin
cofactor availability, but seemed to be associated with
a significant decrease in levels of the SDH flavoprotein
subunit (Sdh1p) [13] These data thus imply a role for
Flx1p in the control of Sdh1p levels Whether this
reg-ulation is achieved via modreg-ulation of rates of protein
expression or degradation is, however, unclear
We have therefore investigated Sdh1p biogenesis by
using both epitope tagging and lacZ reporter
strate-gies, and have demonstrated that Flx1p controls Sdh1p
expression, presumably at the post-transcriptional
level
Results
FLX1p controls SDH activity by regulating the
amount of flavinylated Sdh1p
We previously showed that deletion of FLX1, the
mitochondrial FAD transporter gene, results in a
res-piration-deficient phenotype in which cells are unable
to form colonies on glycerol-containing or pyruvate-containing agar, and exhibit reduced growth rates in YEP liquid media with these carbon sources [13] Polarographic measurements of oxygen consumption induced by addition of succinate to WT and flx1D mutant mitochondria are reported in Fig 1 WT SCM utilized succinate with a rate equal to 100 ngatoms OÆmin)1Æmg protein)1 (Fig 1A) Respiration was com-pletely inhibited by malonate, an inhibitor of SDH [16,17], and this inhibition was reversed by exogenous NADH, with a rate equal to 135 ngatoms OÆmin)1Æmg protein)1, but was blocked by the complex III inhibi-tor antimycin A Succinate respiration in flx1D SCM was reduced by 40% (to 59 ngatoms OÆmin)1Æmg pro-tein)1), but NADH oxidase activity (121 ngatoms OÆmin)1Æmg protein)1) was similar to that in WT SCM (Fig 1B) As both succinate and NADH oxidation involve common electron carriers downstream of ubi-quinone reduction, the defect in succinate metabolism
in the flx1D mutant could be located either in com-plex II (SDH) or in the succinate transporter To exclude the possibility that succinate transport limits the rate of the overall process of succinate
mitochon-Fig 1 Polarographic measurements of the succinate-dependent oxygen uptake rate in SCM SCM (0.1 mg) isolated from WT (A) and flx1D (B) cells grown until the stationary phase in YEP liquid medium supplemented with glycerol were incubated in respiration medium as reported in Experimental procedures The arrows indicate when the additions were made The numbers along the trace refer to the oxygen uptake rate expressed as ngatoms OÆmin)1Æmg protein)1 In the table, the mean (± SD) of the oxygen uptake rates induced by succinate and NADH and the normalization of the succinate versus NADH-dependent oxygen uptake rate, determined in three experiments performed with different mitochondrial preparations, are reported Statistical evaluation was carried out according to Student’s t-test (*P < 0.05) In (C),
1 min before succinate addition, either phenylsuccinate (s) or malonate (d) were added at the reported concentrations.
Trang 3drial metabolism, we applied control strength analysis,
essentially as described in Pastore et al [18] and
refer-ences therein, using the impermeable inhibitor
phenyl-succinate (Fig 1C) Over the concentration range
0.1–0.5 mm, the overall process of succinate respiration
was not reduced By increasing the phenylsuccinate
concentration, and therefore reducing the succinate
transporter activity, we obtained a significant reduction
in succinate respiration Conversely, the SDH inhibitor
malonate reduced the oxygen consumption rate at
con-centrations below 0.25 mm Thus, we conclude that in
SCM isolated from glycerol-grown WT cells, the
rate-limiting step of respiration was the SDH complex
To prove the specificity of SDH impairment, use
was made of glycerol-3-phosphate (5 mm) and
d-lac-tate (5 mm), which yield electrons to the respiratory
chain via other two flavoenzymes, i.e
glycerol-3-phos-phate–ubiquinone oxidoreductase and
d-lactate–cyto-chrome c oxidoreductase, encoded by the genes GUT2
and DLD1, respectively [19,20] Glycerol-3-phosphate
and d-lactate respiration rates measured in WT SCM
were found to be equal to 88 ± 24 and 63 ± 10
nga-toms OÆmin)1Æmg protein)1 Similar respiration rate
values were determined in flx1D SCM (107 ± 40 and
63 ± 20 ngatoms OÆmin)1Æmg protein)1, respectively, for glycerol-3-phosphate and d-lactate)
We also measured SDH activity directly in both sol-ubilized SCM [13] and cellular extracts, and showed that SDH activity was eight-fold to 10-fold higher in cells grown on glycerol or ethanol than in cells grown
on glucose (Fig 2A) However, no change in the activ-ity of the constitutive enzyme phosphoglucoisomerase (PGI) [21] was observed (Fig 2A) A statistically sig-nificant reduction of SDH activity was found in the flx1D mutant as compared to the wild-type, ranging from about 30% (P < 0.05) in early exponential phase
in ethanol to about 70% (P £ 0.01) in glycerol (Fig 2A) No change in the enzymatic activities of the mitochondrial flavoenzymes Gut2p and Dld1p was found (data not shown)
The lower SDH activity observed in flx1D SCM is hypothesized to be due to decreased levels of the flavo-protein subunit Sdh1p [13] This was confirmed by probing cellular extracts with an antibody against the flavin moiety of covalently flavinylated proteins (a-FAD) Following western blotting analysis, a band
Fig 2 (A) Succinate dehydrogenase (SDH)
activity in cellular extracts WT (a) and flx1D
(b) cells were grown for up to 3 h in YEP
liquid medium supplemented with different
carbon sources SDH (black bars) and PGI
(white bars) enzymatic activities were
mea-sured in cellular extracts as described in
Experimental procedures (B) Level of
fla-vinylated Sdh1p Proteins from WT (a) and
flx1D (b) cellular extracts were separated by
SDS ⁄ PAGE and transferred onto
nitrocellu-lose membrane Covalently flavinylated
Sdh1p (FAD-Sdh1p, black bars) was
detected with a-FAD, and its amount was
densitometrically evaluated The
a-FAD-reac-tive band migrating at the same molecular
mass as the ESI-MS ⁄ MS-identified
chaper-one Hsc82p (i.e 83 kDa) was used as an
internal standard (FAD-83p, white bars).
The values are the mean (± SD) of four (A)
and three (B) experiments performed with
different cellular extract preparations
Statis-tical evaluation was carried out according to
Student’s t-test (* P < 0.05; ** P £ 0.01).
Trang 4migrating at 69 kDa (theoretical molecular mass
67 kDa) was revealed, corresponding to flavinylated
Sdh1p (FAD-Sdh1p); an aspecific a-FAD-crossreactive
band (FAD-83p) was observed at 83 kDa, identified
by ESI-MS⁄ MS as the constitutive molecular
chaper-one Hsc82p (theoretical molecular mass 80.7 kDa)
Densitometric analysis of these a-FAD-crossreactive
bands (Fig 2B) revealed a significant reduction in
FAD-Sdh1p that paralleled the reduction in enzymatic
activity No change was observed in the amount of the
internal standard FAD-83p
Biogenesis and mitochondrial import of
HA-tagged Sdh1p in a WT-HA strain and
an flx1D-HA strain
The level of the flavinylated Sdh1p in functional
com-plex II could potentially be regulated at several points
between transcription and cofactor addition inside mitochondria [22–26] To investigate these processes,
we constructed a novel yeast strain, WT-HA, in which Sdh1p was fused to three consecutive copies of an influenza HA epitope (YPYDVPDYA) The HA-tag was inserted at the C-terminal end of Sdh1p, so as not
to disrupt the N-terminal mitochondrial targeting sequence Both the NCBI tool orf finder (http:// www.ncbi.nlm.nih.gov/gorf/gorf.html) and the bestorf gene prediction program from Softberry Inc (http:// www.softberry.com) predicted a single 680 amino acid translation product from the recombinant SDH1-HA gene sequence Its theoretical molecular mass is 74.4 kDa The growth properties on YEP plates of the novel strain are shown in Fig 3A WT-HA cells exhib-ited a respiration-deficient phenotype, as they were able to grow well on a fermentable carbon source (glu-cose), more slowly on ethanol, and not at all on
A
Fig 3 (A) Growth properties of the WT-HA strain and detection of Sdh1-HAp The 3xHA-loxP-kanMX-loxP cassette (1669 bp) was genomi-cally fused in frame to the 3¢-end of the SDH1 ORF of a WT strain (first line) to obtain a new strain (WT-HA, second line), as described in Experimental procedures In (A) WT-HA, flx1D-HA, WT and flx1D strains were streaked on YEP solid medium supplemented with different carbon sources The plates were incubated at 30 C for up to 2 days In (B), proteins from cellular extracts (EC), mitochondria (SCM) and postmitochondrial supernatant (SN postSCM) (1.5 lg each) prepared from WT-HA cells grown for up to 3 h in YEP liquid medium supple-mented with glycerol were separated by SDS ⁄ PAGE and transferred onto a poly(vinylidene difluoride) membrane Sdh1-HA proteins were detected with a-HA In (C), proteins from SCM and EC and rat liver mitochondria (RLM) (15 lg each) were separated by SDS ⁄ PAGE and transferred onto a nitrocellulose membrane Covalently flavinylated proteins were detected with a-FAD.
Trang 5glycerol In YEP liquid medium supplemented with
these nonfermentable carbon sources, they exhibited a
reduced growth rate (data not shown)
In cellular lysates of glucose-grown cells, Sdh1-HAp
was detected after SDS⁄ PAGE as a single band of
about 70 kDa, which increased in abundance about
10-fold when glycerol or ethanol was the carbon
source (data not shown) Two additional
a-HA-reactive bands were detected under these growth
conditions, with molecular masses of 74 and 66 kDa
(Fig 3B)
The correct delivery of the recombinant protein to
mitochondria (Fig 3B) was indicated by the
observa-tion that HA-tagged proteins were fourfold to
eight-fold enriched in the mitochondrial fraction as
compared to cellular extracts and were absent in
postmitochondrial supernatants
As it has been reported that cofactor attachment
requires correctly folded Sdh1p [23], it is possible that
the C-terminal HA-tag may inhibit flavinylation The
inability of the recombinant protein to constitute a
functional SDH complex was indicated by the
respira-tion-deficient phenotype of the WT-HA strain
(Fig 3A) and by the lack of enzymatic SDH activity
in the cellular extracts of engineered cells (data not
shown) Immunoblotting analysis with a-FAD
(Fig 3C) revealed only a faint band at 70 kDa in
mitochondria from WT-HA strains, which appeared to
migrate a little more slowly than the major band
rec-ognized by a-FAD in mitochondria from WT cells and
thus may represent a nonspecific reaction The 70 kDa
migrating protein in this position was identified as the
mitochondrial heat shock protein Ssc1p (theoretical
molecular mass 70.6 kDa) by ESI-MS⁄ MS As both
the band detected by a-HA (Fig 3B) and the one
rec-ognized by a-FAD in WT cells are four-fold enriched
in mitochondria as compared to cellular extracts, the
recombinant Sdh1-HAp is probably flavinylated poorly
or not at all Thus, Sdh1-HAp is a useful reagent for
the investigation of apoprotein synthesis and import
independently of flavin cofactor attachment or
avail-ability
Digitonin titration experiments, performed as in
Barile et al [27], proved that the 70 kDa HA-tagged
protein was released roughly like cytochrome c oxidase
activity, whereas the 66 kDa and 74 kDa proteins
followed kynurenine hydroxylase release (data not
shown) This suggests that the 70 kDa HA-tagged
pro-tein is localized in the inner mitochondrial membrane,
whereas the 66 and 74 kDa proteins are localized in
the outer membrane
The uncoupler carbonyl cyanide
p-(trifluorometh-oxy)-phenylhydrazone (FCCP) collapses the membrane
potential generated by the respiratory chain and there-fore inhibits import of proteins into the mitochondrion [28] WT-HA cells were incubated either in the absence
or presence of FCCP (25 lm) for 3, 5 or 24 h, and the HA-tagged proteins were monitored by SDS⁄ PAGE and immunoblotting As expected, three a-HA-reactive bands with molecular masses of about 74, 70 and
66 kDa were detected (Fig 4A) After 3 h of growth, each band represented about 30% of the total
Sdh1-HA proteins In the presence of FCCP (Fig 4A, lane 2), a 60% reduction of the total amount of
Sdh1-HA proteins was observed, the 70 kDa band, which presumably represents the mature Sdh1-HAp, being the most significantly reduced The relative amount of the 74 kDa band was unaffected by FCCP; it probably represents an extramitochondrial form of precursor Sdh1-HAp The intensity of the 66 kDa band was not changed by FCCP treatment, and it may be an N-ter-minally cleaved form generated in the outer mitochon-drial compartment After 5 h of growth, the intensity
of the 70 kDa band increased two-fold, and this increase was prevented by FCCP (Fig 4A, lane 4) After 24 h of growth, the abundance of the 70 kDa form was decreased even in the absence of FCCP (Fig 4A, lane 5), presumably because of degradation
of nonflavinylated protein The 74 kDa band also decreased, whereas the 66 kDa band remained con-stant Thus, the 66 kDa cleaved form seems to be more stable than the intramitochondrial mature pro-tein No a-HA-reactive bands were detectable in cells treated for 24 h with FCCP (Fig 4A, lane 6) No change in the amount of FAD-83p, used as an internal standard, was found under these experimental condi-tions (Fig 4A)
To determine how Flx1p controls the level of Sdh1p,
we used an flx1D-HA yeast strain, which carries both the FLX1 gene deletion and the SDH1-HA gene These cells were incubated in the absence or presence of FCCP (25 lm) for 3, 5 or 24 h, and the HA-tagged proteins were detected by SDS⁄ PAGE and immuno-blotting as above (Fig 4B) In the flx1D-HA mutant after 3 and 5 h of growth, both the 74 kDa precursor and the 70 kDa mature Sdh1-HA proteins were detect-able, but not the 66 kDa, putative cleaved form (Fig 4B, lanes 1 and 3) At 24 h, neither a-HA-reac-tive bands nor the internal standard, FAD-83p, were detected, presumably because generalized protein degradation correlated with the flx1D-HA growth defect (Fig 4B, lane 5) The total amount of Sdh1-HAp was reduced as compared to the WT-HA strain (by 86%, 90%, and 100%, respectively at 3, 5 and
24 h in the experiments reported in Fig 4) In four replicate experiments using different cellular extracts of
Trang 6glycerol-grown WT-HA and flx1D-HA cells, the total
amount of Sdh1-HAp was reduced in 73% and 81%
(means), respectively, at the 3 h and 5 h growth points
(P£ 0.01; Fig 4C) Extracts from ethanol-grown cells
exhibited a smaller but still significant reduction (45%
and 40% at 3 h and 5 h of growth, respectively;
P< 0.05; Fig 4C)
The 70 kDa mature Sdh1-HAp form was efficiently
generated and was more abundant than the full-length
precursor in the flx1D-HA cellular extracts at both 3 h
and 5 h Thus, its abundance seems to be solely limited
by the rate of precursor synthesis On treatment with
FCCP, the 74 kDa precursor band was almost the only
a-HA-crossreactive band detectable Its amount was
decreased by 78–80% in the flx1D-HA mutant strain
as compared to the total amount of protein found in
the WT-HA strain (Fig 4A,B, lanes 2 and 4) These
results are consistent with the proposal that Flx1p
con-trols Sdh1-HAp expression, rather than import and
processing of the precursor protein
Flx1p controls SDH1 expression
To substantiate the hypothesis that Flx1p controls
SDH1 expression, independently of cofactor
attachment, in a new yeast strain, namely WT-lacZ,
SDH1ORF was genomically replaced by the lacZ gene
coding for b-galactosidase (b-Gal) of Escherichia coli (gene reporter strategy), as described in Experimental procedures This transformed strain exhibits the same respiration-deficient phenotype as the WT-HA strain, as
it was able to grow as well as the WT cells on glucose, more slowly on ethanol, and not at all on glycerol (Fig 5A) In YEP liquid medium supplemented with these nonfermentable carbon sources, growth was reduced but not abolished (data not shown)
The b-Gal activity was 40 ± 7 lmolÆmin)1Æmg pro-tein)1 in cellular extracts of glucose-grown WT-lacZ cells up to 5 h The activity increased about six-fold and nine-fold at the 3 h time point when cells were grown on glycerol and ethanol, respectively, and reached a plateau after 5 h of growth on glycerol, whereas it still increased when ethanol was the carbon source (Fig 5Ba) As a control, to show that altered SDH1 expression was not a secondary effect of growth rate, the activity of the constitutive enzyme PGI was measured in the same extracts (Fig 5Bb) and showed
no difference between fermentable and nonfermentable carbon sources
We also constructed a double mutant, flx1D-lacZ, containing both the FLX1 gene deletion and the repor-ter gene This strain exhibited the same respiration-deficient phenotype as the flx1D and the WT-lacZ strains (Fig 5A) The b-Gal activity in extracts of
WT-HA
100
*
**
**
*
80
60
40
20
0
FAD-83p Sdh1-HAp
FCCP Growth time (h)
p
m
cl
Strain
A C
B Strain Lanes
FAD-83p Sdh1-HAp
FCCP Growth time (h)
Growth time (h) 3
Glycerol
Ethanol
p
m Lanes
1
– + – + – +
24
5
3
– + – + – +
24
5
3
2 3 4 5 6
1 2 3 4 5 6
Fig 4 Detection of Sdh1-HAp in cellular extracts from WT-HA and flx1D-HA cells incubated in the absence or presence of the uncoupler FCCP Glycerol-grown WT-HA (A) and flx1D-HA (B) cells were incubated in the presence (+) or absence ( )) of FCCP (25 l M ) for 3, 5 and
24 h Proteins from cellular extracts (10 lg) were separated by SDS ⁄ PAGE and transferred onto a poly(vinylidene difluoride) membrane, and the HA proteins were detected with a-HA The a-FAD-reactive band (FAD-83p) was used as an internal standard In (C), the total Sdh1-HAp amount of protein in flx1D-HA cellular extracts is reported as a percentage of that detected in WT-HA cellular extracts The values are the means (± SD) of four experiments performed with different cellular extract preparations Statistical evaluation was carried out according
to Student’s t-test (* P < 0.05; ** P £ 0.01).
Trang 7Growth time (h) 3 5
Glucose Glycerol Ethanol
ND ND
ND ND
Growth time (h) 3
0
20
40
60
80
100
120
140
160
0
70
140
210
280
350
0
100
200
300
400
500
800
900
b b
**
**
*
5 Glycerol
–1 )
–1 )
0
20
40
60
80
100
Ethanol
ND ND
A
B C
Fig 5 (A) Growth properties of WT-lacZ and flx1D-lacZ strains WT-lacZ, flx1D-lacZ, WT and flx1D strains were streaked on YEP solid
medium supplemented with different carbon sources The plates were incubated at 30 C for up to 5 days (B, C) b-Gal and PGI activities in
WT-lacZ and flx1D-lacZ strains b-Gal (a) and PGI (b) activities were measured in WT-lacZ (B) and flx1D-lacZ (C) cellular extracts obtained from
cells grown for up to 3, 5 or 24 h in YEP liquid medium supplemented with different carbon sources In (C), values are reported as
percent-age of the activities measured in the WT-lacZ cellular extracts The values are the means (± SD) of three experiments performed with
differ-ent cellular extract preparations ND, not determined Statistical evaluation was carried out according to Studdiffer-ent’s t-test (* P < 0.05;
** P £ 0.01).
Trang 8glucose-grown flx1D-lacZ and WT-lacZ cells was
simi-lar, indicating no significant differences in basal SDH1
expression (data not shown) However when
flx1D-lacZ cells were grown on glycerol for 3 or 5 h, SDH1
expression was reduced to 50% (P < 0.01) A 35%
reduction (P < 0.05) was observed when cells were
grown for up to 3 h in ethanol Extending the growth
time restored b-Gal activity No significant differences
in PGI activity were detected in the same extracts
(Fig 5Cb)
To exclude the possibility that the reduction in lacZ
expression levels was caused by the absence of
func-tionally active Sdh1p, we constructed diploid
heterozy-gous SDH1⁄ sdh1 strains (dWT-lacZ and dflx1D-lacZ)
The dWT-lacZ strain was able to grow on
nonferment-able carbon source, as expected for a recessive
disrup-tion mutation [29] The SDH1 expression level,
measured as b-Gal activity, was significantly reduced
in dflx1D-lacZ as compared to dWT-lacZ cells, the
reduction being more severe in glycerol than in ethanol
(Fig 6A) No significant change in PGI activity was
detected in these extracts (Fig 6B)
These results are consistent with the control of
SDH1 expression by Flx1p via a mechanism that
involves regulatory regions located upstream of the
SDH1ORF
To understand how this control is exerted, SDH1
mRNA level was measured by real-time RT-PCR
experiments, with ACT1 mRNA being used as an
internal control for gene expression As expected
[22,26], the relative amount of SDH1 mRNA was
5.5 times higher in glycerol-grown WT cells than in
glucose-grown WT cells (Fig 7) No change in the
rel-ative amount of SDH1 mRNA was found in the flx1D
mutant strain in comparison to the WT strain in both
the carbon sources used As changes in Sdh1p amounts
were not paralleled by changes in the SDH1 mRNA
level, we expected that the 5¢-UTR, defined as in de la
Cruz et al [25], rather than the promoter region is
involved in Flx1p–SDH1 crosstalk
Discussion
We have investigated the relationship between defects
in flavin cofactor homeostasis and the function of
mitochondrial FAD-binding enzymes Correlation of
these has been demonstrated in human pathologies,
including deficiencies of the flavoprotein subunit of
respiratory chain complex II [30] and in RR-MADD
[31,32], in which polypeptides involved in fatty
acyl-CoA and amino acid metabolism are impaired [3] The
molecular mechanism underlying these defects is
unknown, but one possibility is that low levels of
intramitochondrial FAD causes accelerated breakdown
of FAD-binding enzymes [31,33] Previously, we pro-posed that riboflavin cofactors may play a direct role
in transcriptional or translational regulation in RR-MADD [3] The hypothesis that riboflavin deficiency alters the affinity of transcription factors for DNA or modulates translational efficiency has also been pro-posed for HepG2 and in Jurkat lymphoid cells [34]
Saccharomyces cerevisiaeprovides a useful model for the alterations of flavoprotein biochemistry typical of
3 Growth time (h)
Glycerol Ethanol
ND ND
ND
**
100
80
60
40
20
0
0
40
80
120
160
**
*
A
B
ND
Fig 6 b-Gal and PGI activities in the diploid strains dWT-lacZ and dflx1D-lacZ b-Gal (A) and PGI (B) activities were measured in cellu-lar extracts obtained from dWT-lacZ and dflx1D-lacZ cells grown for
up to 3, 5 or 24 h in YEP liquid medium supplemented with differ-ent carbon sources The enzymatic activities, measured in dflx1D-lacZ, are reported as percentage of the activities measured in the dWT-lacZ cellular extracts The values are the means (± SD) of three experiments performed with different cellular extract prepara-tions ND, not determined Statistical evaluation was carried out according to Student’s t-test (* P < 0.05; ** P £ 0.01).
Trang 9RR-MADD, as the activity of the flavoenzymes
lipo-amide dehydrogenase and SDH can be reduced by
mutation or deletion of the genes encoding the
ribofla-vin membrane transporter (MCH5) [5], FAD
synthe-tase (FAD1) [14], and the mitochondrial FAD
transporter (FLX1) [13,15,35]
The reduced activity of SDH in FLX1 mutant⁄
deleted yeast strains was explained by an accelerated
breakdown of apoprotein in the absence of
mitochon-drial FAD, whose origin is still a matter of debate
[15,35] Previous studies reported that FAD synthetase,
Fad1p, was present only in the cytoplasm fraction and
not in mitochondria, so it was hypothesized that Flx1p
is responsible for FAD import into mitochondria in
exchange with FMN [14,15] We proposed an
alterna-tive hypothesis, in which FAD synthetase is present
inside mitochondria and Flx1p is involved in FAD
export from the organelle [13] Nevertheless, Flx1p seems not to be required for maintaining cytosolic FAD levels, at least under the experimental conditions used, as the activities of Gut2p and Dld1p (which reside on the outer face of the inner mitochondrial membrane) are unaffected by FLX1 gene deletion Direct measurements of flavin cofactor levels in sphe-roplasts confirm this conclusion (data not shown)
In the present study we have investigated how Flx1p enables mitochondrial succinate respiration and con-trols levels of Sdh1p, using epitope-tagged SDH1 Our data suggest that Sdh1-HAp is correctly imported and processed, but cannot be flavinylated either in the
WT-HA strain or in the flx1D-HA strain These experi-ments also showed that the availability and attachment
of flavin cofactors are not involved in the regulation of Sdh1p reduction Using their differential sensitivity to the uncoupler FCCP, we were able to distinguish pre-cursor and mature forms of Sdh1-HAp Accumulation
of the natural precursor of Sdh1p in the purified outer membrane has been previously reported in a proteomic study, using cells grown on nonfermentable carbon sources [36] We also postulated that an unexpected N-terminal cleavage product, presumably located in the outer mitochondrial compartments, is generated from a putative misfolded precursor by the mitochon-drial quality control system [37,38] In the flx1D-HA mutant strain, this cleaved form is not detectable, sug-gesting that import is favored over cleavage This is consistent with the reduced expression of precursor Sdh1-HAp, which prevents its accumulation in the outer membrane
Reporter gene experiments demonstrated that regu-lation of Sdh1p expression is exerted via the regulatory regions located upstream of the SDH1 ORF, and that regulatory sequences downstream of the SDH1 gene are not strictly required for the regulation of protein expression Thus, the reduced level of Sdh1p in an flx1D mutant strain is due to decreased precursor Sdh1p expression, rather than to its accelerated break-down
To rationalize the mechanism by which Flx1p modu-lates Sdh1p expression, we can speculate that, in a sort
of ‘retrograde’ crosstalk, Flx1p coordinates cofactor status inside mitochondria with apoprotein synthesis occurring outside, presumably on mitochondria-bound polysomes [36] In this pathway, Flx1p might function either as a ‘nutrient sensor’ [39,40] or as a flavin trans-porter (whatever the flavin transported is, FMN [15]
or FAD [13]), triggering a downstream cytosolic sig-naling pathway
The finding that apoprotein expression may be regu-lated by vitamins or vitamin-derived cofactors is not
Strain
WT
0
0.2
0.4
0.6
0.8
Glucose
WT Glycerol
Fig 7 Relative quantification of SDH1 mRNA level in WT and
flx1D cells by real-time RT-PCR Total RNA extracted from WT and
flx1D cells, grown for up to 5 h in YEP liquid medium
supple-mented with glucose or glycerol as carbon sources, were
reverse-transcribed and used in real-time RT-PCR assays, as described in
Experimental procedures SDH1 mRNA level was normalized to
ACT1 mRNA level, used as an internal standard, in order to correct
for differences in mRNA quantity between samples The SDH1
mRNA relative amount values reported are the means (± SD) of
four independent real-time RT-PCR reactions performed with two
different total RNA preparations Statistical evaluation was carried
out according to Student’s t-test.
Trang 10surprising This regulation might be exerted at a
tran-scriptional level by modulating the activity of specific
transcription factors as described for Vhr1p for biotin
[41,42], Pdc2p for thiamine diphosphate [43], and
Rip140 for pyridoxal 5¢-phosphate [44], or at a
post-transcriptional level by stabilizing or melting
RNA secondary structure (i.e via riboswitches or via
the internal ribosome entry site) with regulatory
conse-quences This control has been reported for biotin [45]
and more recently for vitamin B12, which binds specific
responsive elements in the 5¢-UTR of methionine
syn-thetase mRNA [46] Sequence analysis of the 5¢-UTR
of this mRNA also reveals the presence of two
upstream ORFs involved in regulating the translational
efficiency of the main ORF [47] Translational
effi-ciency may also be regulated by vitamin⁄ cofactors via
phosphorylation of translation initiation factors, as
suggested for riboflavin in riboflavin-deprived cells
[34]
Real-time RT-PCR experiments showed no change
in SDH1 mRNA level in the flx1D mutant strain as
compared to the WT strain This suggested that
regu-lation of SDH1 expression is exerted
post-transcrip-tionally, via a mechanism that involves the 5¢-UTR of
SDH1mRNA Searching for cis-acting elements in the
regulatory region located upstream of the SDH1 ORF
with bioinformatic tools [48,49], we found 12 highly
conserved motifs (six with an unknown function)
None of these were found in the 5¢-UTR, and no
upstream ORFs were found using the NCBI tool orf
finder Then, either allosteric rearrangements of the
5¢-UTR upon nutrient ⁄ protein binding or differential
phosphorylation of translation initiation factors might
be evoked to explain regulation of SDH1 mRNA
translation on the outer mitochondrial surface [36]
Owing to the high energy required to synthesize
apo-proteins, a translational response to flavin cofactor
level would be more ‘economic’ than the degradation
of translational products Such a control might also underlie the riboflavin-dependent restoration of com-plex II deficiencies in humans [30]
Experimental procedures
Materials All reagents and enzymes were obtained from Sigma-Aldrich Corp (St Louis, MO, USA), Fermentas Inc (Glen Burnie, MD, USA), Carl Roth GmbH+Co.KG (Kar-lsruhe, Germany) and Calbiochem (San Diego, CA, USA) Zymolyase was obtained from ICN Biomedicals (Aurora,
OH, USA) Bacto yeast extract and yeast nitrogen base were obtained from Difco (Lawrence, KS, USA), and anti-HA and anti-rat peroxidase conjugated IgG were obtained from Roche (Basel, Switzerland) and Jackson Immunoresearch (West Grove, PA, USA), respectively
Yeast strains The wild-type S cerevisiae strain (EBY157, WT), derived from the CEN.PK yeast series and the flx1D mutant strain (EBY167A, flx1D), constructed as described in Bafunno
et al [13], were used as recipient strains to obtain the new strains reported in Table 1
Genomic HA-tagging of SDH1 Three consecutive copies of the HA epitope were fused to the 3¢-end of the SDH1 ORF in the genome of both the
WT and EBY167-G418Sstrains, by using a modification of the PCR targeting technique [50] EBY167-G418Swas pre-viously obtained by transforming the flx1D mutant strain with the plasmid pSH47 to remove the kanMX marker in the FLX1 locus, according to Gu¨ldener et al [51] Plasmid pUG6-HA was used as a template to generate by PCR a
Table 1 Genotypes of S cerevisiae strains used in this study.
Haploid
Diploid
SUC2 ⁄ SUC2 + YCplac33URA3 FLX1 ⁄ FLX1 SDH1 ⁄ sdh1::lacZ-loxP-kanMX-loxP
YCplac33URA3 flx1::loxP-kanMX loxP ⁄ flx1::loxP-kanMX-loxP SDH1 ⁄ sdh1::lacZ-loxP-kanMX-loxP