Correspondence: Michele Saliola, Department of Cell and Developmental Biology, University of Rome ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Rome, Italy. Tel.: +39 06 4991 2544; fax: +39 06 4991 2351; e-mail: firstname.lastname@example.org
KlNDE1 and KlNDI1 code for two inner mitochondrial membrane transdehydrogenases involved in the maintenance of the intracellular NAD(P)H redox balance. The function of these genes during the utilization of fermentative and respiratory carbon sources was studied. During growth in glucose, deletion of KlNDE1 and KlNDI1 led to an altered kinetic of ethanol and glycerol accumulation compared with the wild type; in addition, KlndiΔ was unable to grow in respiratory substrates. Northern analysis and GFP-fusion experiments showed that KlNDE1 and KlNDI1 regulate the expression of KlGUT2, a component of the glycerol-3-phosphate shuttle. Moreover, both genes seem to be involved in the biogenesis of the mitochondrial tubular network.
Another important component of the redox-balancing system is the ethanol/acetaldehyde (Et–Ac) shuttle (Bakker et al., 2001; Overkamp et al., 2002). These metabolites, which can freely diffuse through mitochondrial membranes, are oxidized or reduced by cytosolic or mitochondrial alcohol dehydrogenases (ADH) transferring the NAD(P)H excess from one cellular compartment to the other. In S. cerevisiae, it has been suggested that the G3P and the Et–Ac shuttles operate as components of a supramolecular complex that also includes Nde1p and Nde2p (Grandier-Vazeille et al., 2001), two highly identical NADH: ubiquinone oxidoreductases located on the inner mitochondrial membrane, with the active site facing the outer membrane (Luttik et al., 1998; Small & McAlister-Henn, 1998). Likewise, Ndi1p, the inner mitochondrial membrane dehydrogenase facing the matrix side (Marres et al., 1991), these activities couple the oxidation of NADH to the respiratory chain. Kluyveromyces lactis has only one KlNde1p activity that is highly similar to those of S. cerevisiae (Tarrio et al., 2005), while KlNde2p, a second dehydrogenase, has greater similarity to the external calcium-dependent enzyme of Neurospora crassa (Tarrio et al., 2006). However, KlNde2p does not influence the growth on glucose, suggesting a distinct role from KlNde1p in the metabolism of K. lactis. Moreover, KlNde1p and KlNde2p, different from the corresponding dehydrogenases of S. cerevisiae, can accept both NADH and NADPH as substrates (Tarrio et al., 2005, 2006). Different from S. cerevisiae, K. lactis is a Crabtree-negative yeast (De Deken, 1966), and the characterization of activities involved in the maintenance of the redox balance will help to understand the biology of this yeast.
Recently, we reported the effects of the deletion of KlGUT2 on the metabolism of K. lactis (Saliola et al., 2008), showing that this gene influenced the transcription of KlNDE1 and KlNDI1 in a carbon source-dependent manner. Therefore, we deleted KlNDE1 and KlNDI1 to study their influence on the G3P and Et–Ac shuttles. We report that the deletion of either KlNDE1 or KlNDI1 highly reduced the transcription of KlGUT2, enabling these mutants to accumulate its activity in the mitochondria.
Cultures were grown under shaking conditions at 28 °C in YP (1% Difco yeast extract, 2% Difco Bacto peptone) or in minimal medium (6.7 g L−1 Difco yeast–nitrogen base) supplemented by different carbon sources at the concentration specified in the text. Geneticin (G418) concentration in selective plates was 100 μg mL−1. Minimal medium was supplemented with auxotrophic requirements at a final concentration of 10 μg mL−1. Escherichia coli strain DH5α was used for the propagation of plasmid DNA. Cultures were grown at 37 °C on Luria–Bertani medium (0.5% yeast extract, 1% Difco tryptone, 0.5% NaCl, supplemented with 100 μg mL−1 ampicillin).
Glucose, glycerol and ethanol concentrations in the culture supernatants were measured using commercial kits from R-Biopharma (Darmstadt, Germany) according to the manufacturer's instructions.
Amplification of KlNDE1 and KlNDI1 and construction of deletion cassettes
The KlNDE1 and KlNDI1 genes were amplified as XbaI fragments from the K. lactis genome with the following primers:
(Capital letters indicate inserted sequences to produce an XbaI site).
The primer sequences were located at −1294 bp upstream to the ATG and +624 bp downstream to the stop codon in the case of KlNDE1 and −1184 and +672 in the case of KlNDI1. The amplified KlNDE1was cloned as a 3.6-kbp XbaI fragment into the XbaI site of pTZ19 vector (Pharmacia) to harbour pTZ19/KlNDE1 and as a 3.4-kbp XbaI fragment to have pTZ19/KlNDI1. The deletion cassette was constructed in pTZ19/KlNDE1 digesting this plasmid with SpeI–BglII, and cloning in these sites the purified kanMX4 cassette digested with the same enzymes from the pFA6a vector (Wach et al., 1994). The plasmid containing the disrupting cassette, called pTZ19/Klnde1Δ, has a KlNDE1 deletion in the N-terminal part of the gene encompassing 40 bp of the promoter plus the first 500 bp of the ORF. The deletion cassette in the case of pTZ19/KlNDI1 was obtained in two sequential steps. In the first one, the plasmid was digested ClaI–BglII cloning in these sites a portion of the kanMX4 gene digested with the same enzymes (there are two ClaI sites in the gene). In the second step, the selected plasmid was digested using ClaI and the remaining kanMX4 ClaI fragment was cloned in the right orientation to harbour pTZ19/Klndi1Δ that contains 1.4-kbp deletion of the KlNDI1 gene. Both deletion cassettes were linearized by XbaI digestion and transformed into the MW179-1D strain for gene replacement. G418-resistant clones were first analysed by colony hybridization. To isolate putative disruptants, the filters were hybridized with 32P-DNA probes containing the deleted portions of the genes. Colonies that presented negative signals on the autoradiography were further analysed first by PCR and were finally verified by Southern analysis. The genomic DNA of three hypothetical Klnde1Δ deletants and of the wild-type strain were restricted by EcoRI. Restriction fragments, separated by gel electrophoresis, were transferred onto a nylon filter and probed with the EcoRI 3.6-kbp KlNDE1 fragment labelled with 32P. Autoradiography showed one signal of 3.6 kbp in the wild type and two signals of 3.4 and 1.05 kbp in the deletants, confirming the correct integration of the cassette into the KlNDE1 locus.
The genomic DNA of three putative Klndi1Δ disruptants and a wild-type strain were digested with EcoRI and NdeI, separated by gel electrophoresis and transferred onto a nylon filter. The filter was probed with the NdeI 2.85-kbp KlNDI1 fragment labelled with 32P. Autoradiography showed one signal of 2.85 kbp in the wild type and two signals of 1.85 and 1.15 kbp in the deletants, confirming the correct integration of the cassette into the KlNDI1 locus. We chose a deletant for each gene, named MW179-1D/Klnde1Δ and MW179-1D/Klndi1Δ, and these mutants were further used for the growth analysis.
The entire KlNDE1 and KlNDI1 genes were cloned in the XbaI site of the KCplac13 centromeric plasmid (provided by Prof. M. Wèsolowski-Louvel) for complementation analysis.
Construction of the chimeric KlGUT2/GFP and KlGUT2 centromeric vectors
The pTZ19/KlGUT2 vector (Saliola et al., 2008) was digested with BamHI–XbaI. The 1.27-kbp fragment obtained containing 1.1 kbp of promoter and the N-terminal part of the KlGUT2 gene (coding for the first 55 amino acids) was purified and cloned in the corresponding sites of the centromeric KCplac13 plasmid. This plasmid, once digested with XbaI, was ligated with the XbaI–SpeI-digested 3′ terminal part of the gene amplified by PCR with the following primers:
The PCR product contained 700 bp of the 3′ untranslated portion of the KlGUT2 gene starting from the STOP codon. The green fluorescent protein (GFP) gene was obtained from the pVT100U-mtGFP vector (Westermann & Neupert, 2000). Before we could clone the GFP gene in frame in the KlGUT2 construct, we inserted an extra XbaI site on its 5′ terminal part. This was achieved by digesting pVT100U-mtGFP with BglII–BamHI. The purified GFP fragment was cloned in the BamHI site of pTZ19 selecting for the GFP orientation that could be recut as an XbaI fragment. Finally, this fragment was cloned in the XbaI-digested KCplac13 vector containing the 5′ and 3′ DNA of KlGUT2 to harbour the chimeric KlGUT2/GFP vector. Because this plasmid can only be selected in leu2− strains, to express this plasmid in LEU2–HGT1 strains such as PM6-7A, we cloned in its unique SmaI site, the KanMX4 module, as a SmaI–EcoICR fragment. This vector, once transformed in yeast, confers resistance to G418. The shorter version of the KlGUT2 gene (KlGUT2M27) in which translation could only be achieved from the second methionine was amplified by PCR with the following primers:
The amplified fragment was digested with SpeI and cloned under the control of KlADH3 in the NheI site of pCG2/1, a derivative of the KCplac13 centromeric vector containing the KlADH3 promoter as a 1.3-kbp XbaI–HindIII fragment. In this plasmid (pCG2), HindIII was converted into Nhe1 following digestion, fill in with the Klenow fragment enzyme, religation, transformation and selection in E. coli to obtain pCG2/1. The shorter version of the chimeric KlGUT2/GFP gene, called KlGUT2M27/GFP, was amplified by PCR from the KlGUT2/GFP plasmid with the KlGUT2M27 and KlGUT2 reverse primers and cloned in the NheI site of pCG2/1. The chimeric KlGUT2/GFP gene devoid of its promoter, used as a control, was amplified by PCR from the KlGUT2/GFP plasmid with the KlGUT2M1 forward 5′-GGGACTAGTcgatgtttgcagctaaacgtgc-3′ and KlGUT2 reverse primers. The amplified fragment was digested with SpeI and cloned under the control of the KlADH3 promoter in the NheI site of pCG2/1 to harbour the plasmid KlGUT2M1/GFP.
DASPMI and GFP fluorescence microscopy
DASPMI [2-(4-dimethylaminostryl)-1-methylpyridinium iodide] mitochondria-specific staining of live cells was performed according to Yaffe (1995). Cells transformed with GFP plasmids were grown under selective conditions to the exponential phase and then shifted for 6 h in YP containing glucose or glycerol at the concentration specified in the text. To assess the fluorescent signal, cells were analysed using an Axio Observer inverted microscope (Zeiss, Oberkochen, Germany), scanned in a series of 0.3-μm sequential sections with the axiovision software (Zeiss) and a 2D reconstruction was obtained.
DNA manipulation, plasmid engineering and other techniques were performed according to standard procedures. Yeast transformation was performed by electroporation using a Biorad Gene-Pulser apparatus. Preparation of total RNA has been described previously (Saliola & Falcone, 1995).
Results and discussion
The KlNDE1 and KlNDI1 genes were amplified from the genomic DNA of K. lactis and cloned in commercial vectors. The resulting plasmids were used for the construction of the deletion cassettes in which the coding regions were replaced by the KanMX4 gene. The cassettes were introduced into the MW179-1D strain and the transformants selected for G418 resistance. Southern analysis confirmed the correct integration of the cassette in the transformants and we selected a mutant for each gene, namely Klnde1Δ and Klndi1Δ, which were used in the following analysis.
Carbon sources utilization in Klnde1Δ and Klndi1Δ
The Klnde1Δ and Klndi1Δ strains were analysed for their ability to grow in minimal medium supplemented with fermentative and respiratory carbon sources. As can be seen in Fig. 1, the strain devoid of KlNDE1 showed growth comparable to the wild type with all the substrates analysed (Tarrio et al., 2006). On the contrary, the strain devoid of KlNDI1 was unable to grow in all respiratory substrates, with the exception of ethanol, which sustained limited growth. The growth capabilities of the Klndi1Δ mutant were completely recovered in all substrates when we reintroduced the KlNDI1 gene into a centromeric vector (not shown). In S. cerevisiae, it has been reported that the deletion of NDI1 leads to reduced growth on all respiratory carbon sources (Dudley et al., 2005). However, although both K. lactis mutants were able to utilize ethanol, growth was severely impaired in Klnde1Δ and slightly in Klndi1Δ when hydrogen peroxide was added to an ethanol-containing medium (Fig. 1). These results indicated that the redox balance of these mutants, under oxidative stress conditions, could not be compensated by other activities/shuttles. Moreover, the inability of Klndi1Δ to grow in glycerol and lactate indicated that no other reoxidation route existed in K. lactis. In fact, the addition of small amounts of ethanol (Fig. 1), unable to sustain growth by itself, allowed a slight growth recovery in lactate medium. These data suggested that the Et–Ac shuttle transfers the redox excess from the mitochondria to the cytoplasm feeding the respiratory chain through KlNde1p/KlNde2p.
We also analysed glucose consumption and ethanol production during the growth of Klnde1Δ and Klndi1Δ. As reported in Fig. 2a, both mutants, although slightly delayed during the exponential phase, reached the same number of cells as those of the wild type in the stationary phase. While glucose consumption was similar for the three strains, the production of ethanol was identical (1.9 g L−1) only in the first 24 h of growth (Fig. 2b). In fact, compared with the wild type, this compound was reoxidized faster in Klnde1Δ between 24 and 33 h with a kinetic similar to the Klgut2Δ mutant (Saliola et al., 2008). In contrast, in Klndi1Δ cultures, ethanol continued to accumulate between 24 and 33 h, reaching a concentration of 2.3 g L−1. Therefore, the different production/consumption of ethanol in the two mutants might be due to the different ability to reoxidize the cytosolic NAD(P)H, a stress condition that cells counteract by the production of glycerol (Ansell et al., 1997; Bjoerkqvist et al., 1997). In order to test this hypothesis, we measured the kinetics of glycerol accumulation in the supernatant of the three strains. Indeed, as shown in Fig. 3, Klnde1Δ accumulated glycerol in the medium at a concentration (1.31 g L−1) that was twice the amount produced by the wild type and the Klndi1Δ mutant. This quantity was also considerably higher than the glycerol produced by the Klgut2Δ mutant (Saliola et al., 2008), which was included as a control unable to utilize glycerol. This data indicated that the G3P shuttle was unable to contribute to the neutralization of the NAD(P)H excess in Klnde1Δ, thus suggesting that the deletion of KlNDE1 could also affect KlGUT2 expression. In conclusion, Klnde1Δ and Klndi1Δ, to maintain the redox balance, accumulated glycerol and ethanol at different concentrations and with distinct kinetics compared with the wild type.
Transcriptional analysis of KlNDE1 and KlNDI1
We performed a transcriptional analysis of wild type and mutants to test whether KlNDE1 or KlNDI1 might regulate the expression of KlGUT2 and KlADH3 that, as components of the G3P and Et–Ac shuttles, intervene to neutralize the NAD(P)H redox excess. Compared with S. cerevisiae, K. lactis relies on a more respiratory metabolism with major differences in the regulatory circuits that govern their central metabolism (Breunig et al., 2000; González-Siso et al., 2000; 2009). Therefore, Northern analysis was performed from cultures grown in high glucose and after substrate shift to a low glucose concentration, as well as to glycerol and ethanol, substrates that regulate many respiratory genes differently (Saliola & Falcone, 1995; Lodi et al., 2001; Rodicio et al., 2008). The pattern of transcription of the four genes in the wild type and in the mutants is reported in Fig. 4. As can be seen, the deletions of KlNDE1 and KlNDI1 did not affect the expression of KlADH3 (lanes 1–12), indicating that the control of the Et–Ac shuttle is independent of the two transdehydrogenase genes. Interestingly, KlNDI1 showed two transcripts differently expressed in Klnde1Δ, the lower molecular weight mRNA being activated in 1% glucose (lane 6), while the longer transcript was highly induced under the ethanol condition (lane 8). KlNDE1 and KlGUT2 expressions are significantly reduced in the mutants, with the same reduction observed under all conditions (lanes 5–12). The highly reduced levels of the KlGUT2 mRNA (lanes 7–8 and 11–12) in Klnde1Δ and Klndi1Δ suggested an active role of both these genes in the control of the G3P shuttle. In addition, two KlGUT2 transcripts of different molecular weights appeared in both mutants (lanes 8, 11 and 12) compared with the wild type (lanes 3 and 4). Previously, we reported that the faster-migrating KlGUT2 mRNA was HGT1 dependent in that strains devoid of the high-affinity hexose transporter (HGT1) gene (Billard et al., 1996) only expressed the slower-migrating KlGUT2 transcript (Saliola et al., 2008). On the contrary, the isogenic respiratory-deficient Klcox14Δ strain, unable to assemble the cytochrome oxidase complex, did not express any KlGUT2 transcript (Saliola et al., 2008), suggesting that the regulation of KlGUT2 is dependent on the NAD(P)H redox balance interplay between the respiratory chain and the glycolytic pathway. Although mostly unexplained, these data altogether suggest the existence of a complex regulatory network for the expression of the genes involved in carbon metabolism.
KlNDE1 and KlNDI1 affect mitochondrial morphology and KlGut2p activity
KlNde1p and KlNdi1p are two transdehydrogenase activities that couple the oxidation of NAD(P)H to the respiratory chain, a crucial function for the growth and the maintenance of the redox balance. Therefore, we visualized the mitochondrial network in both mutants by means of the DASPMI vital assay (Fig. 5a). This analysis showed a mitochondrial network with a clear punctuated morphology, slightly more accentuated in Klndi1Δ, compared with the tubular network shown by the wild type and Klgut2Δ strains. These results suggested that the absence of either KlNde1p or KlNdi1p could alter the balance between fission and fusion events of the mitochondrial organelles. In addition, we introduced into these mutants a chimeric KlGUT2/GFP gene in a centromeric vector (see Materials and methods) in order to test the presence and the localization of KlGut2p. The wild type and Klgut2Δ strains showed comparable levels of fluorescence preferentially located in the mitochondria (Fig. 5b). On the contrary, in both Klnde1Δ and Klndi1Δ, this analysis showed, in agreement with the transcription data (Fig. 4), a pale and diffused staining in either 1% glucose (Fig. 5b) or glycerol (not shown) only after a long period of exposure. To better understand the role of the two differently regulated KlGUT2 transcripts, we introduced the chimeric KlGUT2/GFP gene into the HGT1-expressing strain PM6-7A (Saliola et al., 2008). The analysis of cells grown in the presence of 1% glucose (Fig. 6a) or glycerol (not shown) showed a preferential localization of the fluorescence within the mitochondria. On the contrary, cultures shifted to YPD5% a condition in which the slower-migrating KlGUT2 mRNA is not detected (Saliola et al., 2008) or treated with antimycin A, an inhibitor of the respiratory chain, showed the presence of residual mitochondrial fluorescence and an augmented GFP localization into the cytoplasm.
The analysis of the sequence of KlGut2p showed a second methionine (M27) at position 27. The presence of this methionine could suggest an alternative start for mRNA translation, which may lead to activities with distinct intracellular localization, a situation already observed in S. cerevisiae for Gpd1p and Gpd2p (Valadi et al., 2004). In order to test this hypothesis, we cloned under the control of the KlADH3 promoter a shortened version of the KlGUT2 gene (KlGUT2M27) in which translation could only occur from M27. Transformants of the Klgut2Δ mutant carrying this gene recovered the ability to grow in glycerol (Fig. 6b), indicating that the first 26 amino acids of KlGut2p are not required for its activity. To determine the localization of this protein, both KlGUT2/GFP and the shortened gene (KlGUT2M27/GFP) were put under the control of the KlADH3 promoter and introduced into the Klgut2Δ strain. Fluorescence analysis of transformants showed a diffused cytoplasmic localization of the KlGut2M27/GFP activity (Fig. 6c) and a mitochondrial localization of the control KlGut2M1/GFP protein (Fig. 6d). However, the diffused cytoplasmic fluorescence observed with the shorter protein did not exclude that a fraction of the KlGut2M27/GFP activity could localize within the mitochondria. To give direct evidence that M27 is an alternative start for mRNA translation that determines the appearance of the two KlGUT2 transcripts, we introduced both the wild type and the KlGUT2M27 genes into Klgut2Δ and analysed the size of the transcripts by Northern analysis. As shown in Fig. 6e, compared with the PM6-7A strain expressing both transcripts (lane 1), the Klgut2Δ mutant transformed with the wild-type KlGUT2 gene showed only the slow-migrating mRNA (lane 2), while the transformants harbouring KlGUT2M27 showed the presence of the sole faster-migrating mRNA (lanes 3 and 4). Therefore, although both protein forms could be obtained from alternative translations of the same mRNA, the expression of two transcripts encoding activities with distinct localization, depending on the physiological growth conditions, has been preferred. In summary, the slower-migrating mRNA is induced under respiratory conditions and codes for an activity probably localized on the outer side of the inner mitochondrial membrane feeding reduced equivalents to KlNde1p. The faster-migrating mRNA form, constitutively expressed in HGT1 strains, might code for a cytoplasmic-localized activity involved in glycerol metabolism. Therefore, under fermentative conditions, a putative cytoplasmic activity may be required to fulfil the reoxidation of the reduced equivalents in this cellular compartment.
The presence of an M26 in Gut2p of S. cerevisiae might suggest an evolutionarily conserved cytoplasmic role of this activity under different physiological conditions.
This work was funded by the grant ‘Ateneo’ 2008 from the University of Rome – La Sapienza.