The Suppressor of AAC2 Lethality SAL1 Modulates Sensitivity of Heterologously Expressed Artemia ADP/ATP Carrier to Bongkrekate in Yeast

The ADP/ATP carrier protein (AAC) expressed in Artemia franciscana is refractory to bongkrekate. We generated two strains of Saccharomyces cerevisiae where AAC1 and AAC3 were inactivated and the AAC2 isoform was replaced with Artemia AAC containing a hemagglutinin tag (ArAAC-HA). In one of the strains the suppressor of ΔAAC2 lethality, SAL1, was also inactivated but a plasmid coding for yeast AAC2 was included, because the ArAACΔsal1Δ strain was lethal. In both strains ArAAC-HA was expressed and correctly localized to the mitochondria. Peptide sequencing of ArAAC expressed in Artemia and that expressed in the modified yeasts revealed identical amino acid sequences. The isolated mitochondria from both modified strains developed 85% of the membrane potential attained by mitochondria of control strains, and addition of ADP yielded bongkrekate-sensitive depolarizations implying acquired sensitivity of ArAAC-mediated adenine nucleotide exchange to this poison, independent from SAL1. However, growth of ArAAC-expressing yeasts in glycerol-containing media was arrested by bongkrekate only in the presence of SAL1. We conclude that the mitochondrial environment of yeasts relying on respiratory growth conferred sensitivity of ArAAC to bongkrekate in a SAL1-dependent manner.


Introduction
Embryos of the brine shrimp Artemia franciscana exhibit a type of extremophilia characterized by tolerance of anoxia at room temperature for several years [1] [2] with no evidence of apoptotic or necrotic cell death [3]. Protracted anoxia in mammalian species opens the so-called mitochondrial permeability transition pore (PTP) [4], [5], [6]. This pore is of a sufficient size (cut-off ,1,5 kDa) to allow passage of solutes and water, causing swelling and ultimately rupture of the organelle. Activation of the PTP by Ca 2+ overload with the ensuing loss of mitochondrial function results in a severely diminished capacity for energy production and is a final common pathway of cell death [7].
Mitochondria isolated from the embryos of Artemia franciscana lack a Ca 2+ -induced PTP [8], despite the capacity for a profound storage for calcium. While the salient components of the PTP have just been unravelled [9], [10] but see [11] the adenine nucleotide carrier (AAC) remains a well-established modulatory component of this pore [12], [13]. All known ligands of the AAC modulate the probability for PTP opening [12]. Relevant to this, Artemia franciscana is the only species known in which adenine nucleotide exchange operated by AAC is refractory to the naturally occurring inhibitor, bongkrekic acid (BKA) [14]. In species where Ca 2+induced mitochondrial PTP can be demonstrated, BKA inhibits adenine nucleotide exchange mediated by the AAC isoforms they express, and decreases the probability of PTP opening [15], [16], [17], [12]. Recently however, we have reported that mitochondria obtained from brown shrimp (Crangon crangon) and common prawn (Palaemon serratus) exhibit BKA-sensitive mitochondrial adenine nucleotide transport while lacking a Ca 2+ -induced permeability transition [18].
A plausible hypothesis is that AAC interacts with a yet to be identified protein that initiates PTP opening, and that this protein interaction is affected by BKA. An obvious impediment in the elucidation of the role of AAC in PTP and its inhibition by BKA is the lack of knowledge regarding the binding site of BKA on AAC [12], [19]. In the AAC protein sequence found in Artemia, the region between residues 198 and 225 exhibits a low degree of similarity with AAC sequences from other species [14]. We therefore postulated that the BKA binding site may reside somewhere within this Artemia AAC region. To verify this hypothesis, a base-by-base alteration is warranted to pinpoint residues that are critical for the binding of BKA. Insufficient information regarding the genetic background of Artemia franciscana prompted us to express Artemia AAC (ArAAC) in a heterologous environment amenable to genetic manipulations. Yeast is an excellent platform for such experiments. However, as it will become evident from the 'Results' section below, adenine nucleotide exchange mediated by heterologously expressed ArAAC expressed in Saccharomyces cerevisiae was sensitive to BKA. In addition, due to substitution of endogenous yeast AAC2 carriers which are also important for cell respiration and viability with ArAAC, it was necessary to manipulate the presence of the suppressor of AAC2 lethality, SAL1. SAL1 is required for growth of yeasts when AAC2 is absent or inhibited by BKA [20]. Contrary to our expectation, the viability of yeasts expressing ArAAC under non-fermenting conditions was arrested by BKA only when SAL1 was coexpressed while in the absence of Sal1p, growth of yeasts expressing ArAAC was BKA resistant. In direct contrast, under fermenting conditions the viability of yeasts expressing ArAAC was arrested by BKA only when SAL1 was absent, indicating the lethality of ArAAC sal1D.

Expression of ArAAC in Yeast Cells
We integrated the ArAAC gene into the locus of the main yeast AAC gene -AAC2 ( Figure 1). As the double aac2 sal1 deletion strain is lethal and a functional Sal1p is required for growth of yeast in the presence of BKA which blocks the operation of AAC2 protein [20], the ArAAC was expressed in SAL1 and sal1::NatMX4 deletion background. We first amplified ArAAC using cDNA from reverse transcribed total Artemia franciscana RNA as template and cloned it into a TOPO-TA Cloning Vector (TOPO TA CloningH Kits for Sequencing, Invitrogen). The ArAAC integration cassette containing the ArAAC-HA tagged gene and the hygromycin resistance gene HphNTI was constructed as described in 'Materials and Methods'. The cassette DNA was transformed into a strain bearing deletions of two other AAC genes present in yeast, AAC1 and AAC3 (RKY67-1C), resulting in MWY79/15 and MWY79/ 17 clones bearing ArAAC-HA gene in the locus of AAC2 (Table 1). We then deleted the SAL1 gene in control and ArAAC expressing strains MR6, RKY67-1C and MWY79/15 by transforming them with sal1::NatMX4 cassette (see under 'Materials and Methods'), resulting in strains: MWY85/9, MWY84/3, MWY83/1 and 5 (Table 1). ArAAC could not rescue yeasts in sal1 background, thus the deletion of SAL1 gene in ArAAC background was done in the presence of a wild type copy of yeast AAC2 on a Yep352 plasmid (MWY83 strains). A most appropriate isogenic control for our ArAAC-expressing constructs would be to reintroduce the AAC2 gene in the same manner, with the HA tag and the resistance gene cassette in the same positions, as performed in [21] and [22]. However, proteomic analysis of yeast mitochondria expressing ArAAC verified that the sequence is identical to that expressed in Artemia franciscana (see below). Moreover, mitochondria from the MWY79/15 strain, where no endogenous AAC was present, achieved 85% of total attainable membrane potential, and ADP-ATP exchange was readily demonstrated (see below). These results afford a reasonable degree of assurance that the genetic manipulations in our strains did not confer confounding variables to the results.
The proper localization of ArAAC protein in mitochondria was verified by yeast cell extract fractionation, Western blotting and estimations of citrate synthase activity in the various fractions ( Figure 2). Although there is an extramitochondrial citrate synthase in yeasts [23], the overall activity is mainly due to that residing in the mitochondrial matrix. During preparation of the mitochondria, fractions from the two strains expressing ArAAC with (MWY79/15) and without (MWY83/1) Sal1p were probed for the presence of ArAAC (through its HA tag) and porin by Western blotting and citrate synthase activity. Furthermore, mitoplasts were also generated from mitochondria and these were also probed for the presence of ArAAC (through its HA tag) and porin by Western blotting and citrate synthase activity. The fractions were: T (total yeast homogenates), S (supernatant), M (mitochondria), and MP (mitoplasts). The probing for ArAAC (through its HA tag) and porin was performed in the same blot, thus the bands appear very close to each other due to the similar molecular weights of ArAAC and porin. As shown in Figure 2, the band representing ArAAC in the Western blot and the corresponding citrate synthase activity in the MP fraction is greater than the M fraction, but at the same time the band corresponding to porin in the M fraction is greater than the signal from the MP fraction. This is consistent with the notion that porin resides in the outer mitochondrial membrane, which is partially removed during the preparation of the MP fraction, while ArAAC is in the inner membrane/matrix side. Thus, ArAAC protein was expressed and targeted into the yeast inner membrane/matrix side. The experiments shown below (especially for the strain MWY79/15 where no endogenous AAC is present) argue that there was indeed ADP-ATP exchange, and validate the claim that ArAAC was correctly targeted to the inner mitochondrial membrane.

Effect of BKA on Viability of the Modified Strains
AAC2 is responsible not only for yeast growth in nonfermentable sources, termed the respiratory function (R function), but also possesses an essential role in maintaining cell mitotic viability on fermentable carbon sources, termed the V function [24]. Mindful that a functional Sal1p is required for growth of yeasts when Aac2p is absent or inhibited by BKA [20], we addressed the function of ArAAC and its response to BKA by performing growth analysis on fermentable (glucose) and nonfermentable (glycerol) carbon sources in the presence or absence of BKA. BKA enters intact yeast cells at pH 4, thus the growth was tested at this pH [20]. Two different clones harboring the same genetic manipulations were used in order to gain additional reinforcement that the observed results were genuine, and not confounded by the genetic background: therefore, experiments shown in lanes 4/5, 6/7, and 9/10 harbor the same genetic manipulations but were performed in different yeast strains (top table within Figure 3). As shown in Figure 3, yeast cells expressing AAC2 required a functional SAL1 gene for surviving on glucosecontaining media (V function) in the presence of BKA (compare lanes 4 and 5 with lane 8 of panels B and C). The strains expressing ArAAC in lieu of AAC2 also required a functional SAL1 gene for surviving on glucose-containing media in the presence of BKA, indicating that ArAAC together with the sal1::NatMX4 deletion is lethal (compare lanes 6 and 7 with lanes 9 and 10 in panels B and C). We confirmed this phenotype by testing the ability of MWY83 strains to grow on 5-Fluoroorotic Acid (5-FOA) plates. 5-FOA is converted to a toxic form (i.e., 5-flurouracil) in strains expressing the functional URA3 gene coding for orotine-5monophosphate decarboxylase that is involved in the synthesis of uracil. Yeast strains that are phenotypically Ura+ become Uraand 5-FOA(R) (resistant) after selection. We considered that the URA3 plasmid-born copy of AAC2 gene in the presence of 5-FOA in the medium could be lost. However, the MWY83 strains were not able to grow on this medium, confirming the ArAAC sal1::NatMX4 lethality (data not shown). Regarding yeast viability in glycerol media where cells rely on oxidative phosphorylation (R function) in relation to SAL1, ArAAC complemented the R function of yeast AAC2. This can be seen in panels D and E when comparing the triple deletion lane 1 (where no AAC isoform was expressed and despite the presence of Sal1p no yeast growth was observed), and lanes 4 and 5 (where both AAC2 and SAL1 were expressed), to lanes 6 and 7 (where ArAAC was expressed in lieu of AAC2). It is noted though, that yeast expressing ArAAC exhibited slower respiratory growth than those expressing AAC2 (compare the growth of strains in lanes 1, 4 and 5 with lanes 6 and 7 on glycerol medium of panels D and E in Figure 3). In the presence of Sal1p, ArAAC expressed instead of AAC2 conferred sensitivity to BKA; this is reflected in lanes 6 and 7 of panel F, whereas in the absence of Sal1p, it conferred resistance to BKA, inferred from the growth of cells shown in lanes 9 and 10 of panel F where ArAAC and AAC2 (from the plasmid) were expressed. Because sal1 yeasts expressing ArAAC can grow in the presence of BKA in glycerol but not glucose media, we concluded that ArAAC could not complement the viability (V) function of yeast Aac2p.

Adenine Nucleotide Exchange Mediated by Yeastexpressed ArAAC is Sensitive to BKA
To test the effect of BKA on ADP/ATP exchange by ArAAC expressed in yeasts, we relied on the electrogenic property of the carrier during adenine nucleotide exchange. Mitochondria isolat-  ed from the yeast strains detailed below were energized by ethanol, and the effect of BKA during ADP-induced depolarization was evaluated by recording membrane potential (m) with rhodamine 123. To calibrate the rhodamine fluorescence signal, at the end of all experiments the uncoupler SF 6847 (SF, 1 M) was added to induce complete depolarization. The yeast strains used for these experiments are shown in the table on the top of Figure 4. As shown, they either express AAC2 plus SAL1 (RKY67-1C), or ArAAC plus SAL1 (MWY79/15) or only AAC2 (MWY84/4) or ArAAC plus AAC2 from a plasmid (MWY83/5). Because BKA needs to be protonated in order to exert its action on the carrier [25], [26], it becomes less effective at increasing pH. We therefore tested the effect of BKA versus its vehicle (NH 4 OH) in the pH range of 7.1-7.4. Isolated mitochondria (1 mg) were added in a buffer the composition detailed in 'Materials and Methods' with a pH indicated in the panels of Figure 4, in the presence of rhodamine 123 (0.5 M). Ethanol (20 l of 96%) was immediately added in order to energize mitochondria, and this was reflected by a decrease in rhodamine fluorescence. In ArAAC-expressing cells (green or magenta traces) there was a 15% decrease in mitochondrial polarization as compared to other strains (black or grey traces). After 50 sec, 2 mM ADP was added as indicated in Figure 4, resulting in an immediate depolarization due to the electrogenic exchange of ATP 42 for ADP 32 through the adenine nucleotide carrier. As soon as a new plateau of m was established, BKA (1 M) or NH 4 OH (1 mM) was added as indicated in Figure 4. In accordance with the results reported previously [14], NH 4 OH exerted a minor depolarizing effect. It is also apparent from all panels that mitochondria obtained from all strains exhibited BKA-induced repolarizations. As expected, the extent of BKA-induced repolarization was dependent on pH, but even at a condition at pH 7.4 where BKA had only a minimal effect on Aac2p-mediated adenine nucleotide exchange (black trace bottom right), it potently abolished ADP-induced depolarization in ArAAC-expressing mitochondria (same panel, green trace). Since BKA induced repolarization in mitochondria obtained from all strains, we concluded that the ArAAC expressed in yeasts became sensitive to BKA. Moreover, in contrast to the viability results shown above, ArAAC was sensitive to BKA in a manner independent of SAL1.

Mass-spectrometry of Mitochondria Isolated from Artemia Franciscana Embryos and from Yeasts Expressing ArAAC
The apparent discrepancy between the above observation showing sensitivity of ArAAC expressed in yeasts to inhibition by BKA, while Artemia mitochondria are refractory to this poison [14], prompted us to investigate if ArAAC expressed in yeast is indeed identical to that expressed in Artemia. To answer this question, we performed mass spectrometric analysis of isolated mitochondria from strain MWY79/15 expressing the ArAAC, and from isolated mitochondria from embryos of Artemia franciscana, see Table S1 in the Supplemental Material. Samples were treated as detailed in 'Materials and Methods'. As shown in Figure 5, we detected 62% sequence coverage of the ArAAC protein expressed in yeast mitochondria, and 70% of the ArAAC found in Artemia. Despite the fact that sequence coverage was not 100%, peptide fragments of ArAAC expressed in either yeasts or Artemia were detected near the N-and C-termini, arguing against a truncated form of ArAAC expressed in yeasts. By comparing the ArAAC expressed in yeasts to that expressed in Artemia, it could be concluded that the proteins are identical. The homologues of Sal1p, SCaMC-2 (isoform 1) and SCaMC-3 were identified in the mass spectrometric analysis of the Artemia mitochondria (see Excel Sheet S1 in the Supplemental Material). However, the searches of the spectra obtained by mass spectrometry were made using the mouse MitoCarta database [27]. Ongoing efforts are under way to sequence the transcriptome of Artemia franciscana which would be optimal for correctly identifying all proteins from the Artemia mitochondria peptide fragments detected by mass spectrometry.

Discussion
We investigated the effect of heterologously expressing AAC of Artemia franciscana in Saccharomyces cerevisiae. The most important findings of this study are as follows: i) respiratory growth of yeasts expressing ArAAC was arrested by BKA only when SAL1 was coexpressed; ii) fermentative growth of yeasts expressing ArAAC was arrested by BKA only when SAL1 was absent, and iii) adenine nucleotide exchange mediated by ArAAC expressed in yeasts became sensitive to BKA, in a manner independent of SAL1.
In order to explain the sensitivity of adenine nucleotide exchange mediated by ArAAC expressed in yeasts to BKA, the challenge of reconciling two findings emerges: the first is that ArAAC expressed in Artemia franciscana is refractory to BKA [14], [18] and the second is that BKA binds directly on the carrier [28], [29], [reviewed in [12]], making it unlikely that an as yet to be identified protein conferring AAC sensitivity to BKA is absent in mitochondria from Artemia franciscana. However, the opposite scenario seems possible: that an as yet to be identified protein conferring ArAAC resistance to BKA is present in mitochondria from Artemia franciscana. As mentioned in the 'Results' section, ongoing efforts are under way to sequence the transcriptome of Artemia franciscana which would allow the identification of more proteins from the Artemia mitochondria peptide fragments. After this database is assembled, proteins interacting with ArAAC expressed in Artemia mitochondria can be sought.
The acquired sensitivity of ArAAC expressed in yeast to BKA outlines a limitation in a common strategy of heterologously expressing proteins in this organism for the purpose of genetic manipulation; it is neither necessary nor expected for a heterologously expressed protein to behave exactly like in its native environment. Relevant to this, the lipid environment in which AAC is embedded is a critical component for exchange activity in both yeast and mammals [30], [31], [32]. Indeed, a high sensitivity of yeast AAC2 to the cardiolipin content has been previously demonstrated [33], [34]. The lipid composition of the inner mitochondrial membrane of Artemia in which AAC is embedded may be very different from that in yeasts or any other organism to the extent that affords BKA resistance.
The ArAAC-SAL1 interaction suggests a departure from what has been established for AAC2-SAL1 interaction: in yeast cells with no AAC2 expression the presence of SAL1 is required for survival in glucose-containing media and vice versa [24]. This is because besides the respiratory function, AAC2 exhibits an essential role for maintaining cell mitotic viability on fermentable carbon sources; co-inactivation of SAL1 and AAC2 leads to defects in mitochondrial translation and mitochondrial DNA (mtDNA) maintenance [35]. In the results presented above, ArAAC-expressing yeasts in lieu of AAC2 and also coexpressing SAL1 failed to exhibit respiratory growth in the presence of BKA; on the other hand, coexpression of SAL1 in ArAAC-expressing yeasts in lieu of AAC2 in glucosecontaining media rescued cell growth from BKA. In the absence of BKA, expression of ArAAC in lieu of AAC2 did not result in growth arrest irrespective of SAL1 and independent from growth conditions. These results suggest that ArAAC restored the R function but not the V function, originally attributed to AAC2.
Based on the observations that Sal1p allowed yeast growth in the presence of bongkrekic acid on respiratory media, while BKA blocked ADP-ATP exchange in isolated mitochondria from the same strains, one might propose that Sal1 prevents BKA from hindering AAC in maintaining growth during respiratory conditions in a manner unrelated to adenine nucleotide exchange. This conclusion is at odds with earlier reports with yeast mutants exhibiting BKA resistance during respiration [36], in view of the presumption that these strains possessed Sal1p in their mitochondria. The reason(s) for this discrepancy is unknown; however, the exact role of Sal1p regarding yeast viability has not been unequivocally determined and is discussed below.
Sal1p mediates a Ca 2+ -dependent import of ATP-Mg from the cytosol to the mitochondria under conditions where these organelles are ATP consumers [37], [38], thereby maintaining or promoting cell survival. The mammalian homologue SCaMC (small calcium-binding mitochondrial carrier protein) also operates as a Ca 2+ -dependent ATP-Mg/P i carrier. It has been shown that SCaMC-1 promotes cancer cell survival by desensitizing mitochondrial permeability transition via ATP/ADP-mediated matrix Ca 2+ buffering [39]. SCaMC-3 is critical for the increase in oxidative phosphorylation in rodent liver mitochondria in response to glucagon and Ca 2+ -mobilizing agents, an effect that that was likely due to allowing Ca 2+ -dependent accumulation of mitochondrial adenine nucleotides [40]. Furthermore, although it has not been explicitly sought, the presence of SCaMC isoforms in the livers of transgenic mice engineered to lack their own adenine nucleotide carriers is probably the reason why these mice did not exhibit embryonic lethality [13]. However, the requisite nature of SCaMC in mammalian systems has not been attributed to maintenance of matrix adenine nucleotide levels. Likewise, in yeast it has not been established if Sal1p complements for Aac2p by means of transporting adenine nucleotides, hence the opposing conclusions by the groups of Chen and Kolarov [35], [20]. Nonetheless, yeast as a model system enjoys greater versatility because its growth can be examined during fermenting (V function) versus respiratory (R function) conditions. With such a comparison, we addressed the contribution of SAL1 and AAC2. The availability of our yeast strains expressing ArAAC in lieu of AAC2 and the effects it exhibits in relation to SAL1 may help to elucidate the role of SAL1 to cell survival.

Strains and Genotypes
The S. cerevisiae strains and their genotypes are listed in Table 1.

Construction of Yeast Strains
To construct AAC1 and AAC3 deletions in MR6 background the aac1::KanMX4 and aac3::KanMX4 cassettes were amplified by PCR using as templates total DNA isolated from strains bearing deletions of AAC1 and AAC3 genes from the EUROSCARF collection. MR6 strain was transformed with the deletion cassettes as described previously [45] and transformants were selected on YPGA medium supplemented with 200 mg/ml geneticin G418. The correct integration of the cassettes was verified by PCR using AAC1Ver, AAC3Ver and KanMX-Up primers, and by phenotypic analysis. The RKY67-1C double aac1::KanMX4 aac3::K-anMX4 strain was constructed by crossing the single deletion strains aac1::KanMX4 and aac3::KanMX4, sporulation of the diploid AAC1/aac1::KanMX4 AAC3/aac3::KanMX4 strain followed by tetrad dissection. To integrate ArAAC into yeast genome the ArACC gene was first amplified with primers ArAnt F and ArAnt R (all primers are listed in Table 2) using cDNA from reverse transcribed total Artemia franciscana RNA as template and cloned into TOPO-TA Cloning Vector. The ArACC cassette for integration into yeast DNA was constructed by multiple PCR reactions; all cloning steps were verified by sequencing [46]. The cassette included ArACC-HA and HphNTI genes flanked by yeast ACC2 gene promoter and terminator sequences, see Figure 1. The HA, HphNTI, AAC2 gene promoter and terminator sequences were amplified separately using pYM16 or total DNA from yeast as templates, respectively. Before the HphNTI gene the sequences of ADH terminator and TEF promoter are present [46]. The AAC2 gene promoter was then fused to HA by one PCR reaction (product A), HA was fused to ADH terminator and HphNTI gene by a second PCR reaction (product B) and HA and HphNTI gene were fused to AAC2 gene terminator in a third PCR reaction (product C). The final PCR product was assembled using products A, B and C as templates and cloned into pGEM T-easy vector resulting in plasmid pMW77/11. The EcoRV restriction site was engineered before the HA tag sequence. The ArACC gene was amplified from the pCR 2.1-TOPO vector (Invitrogen) using primers 9upANT and 10upANT, introducing EcoRV restriction sites. The 915-bp product digested by EcoRV was cloned into EcoRV site of pMW77/11 resulting in pMW83 plasmid. The integration cassette was cut off from pMW83 by XbaI enzyme and used to transform RKY67-1C strain. Transformants were selected on YPGA medium supplemented with 300 mg/ml hygromycin B. Homologous integration of ArACC in MWY79/15 strain was verified by PCR using primers AAC2veri and AAC2lov and by phenotypic analysis. To construct sal1::NatMX4 deletion, the SAL1 coding sequence and its promoter and terminator region (2193 bp upstream and 226 bp downstream) was PCR-amplified using primers Sal1-Up and Sal1-Low. The PCR product was cloned into pJet vector resulting in pMW100/6 plasmid. The construct sal1D::NatMX4 gene and the PvuII/EcoRV fragment from plasmid pAG25 [47] including NatMX4 drug resistance gene was cloned into KpnI blunt-ended with Klenow and EcoRV sites of SAL1 gene on pMW100/6 plasmid resulting in pMW105/5 plasmid. A linear DNA fragment including sal1D::NatMX4 cassette was cut off from this plasmid and introduced by transformation into RKY67-1C and MWY79/15 strain bearing the wild type yeast AAC2 gene on a Yep352 URA3 plasmid. Transformants were selected on YPGA medium supplemented with nourseothricin (nat), 100ug/ml. Homologous integration of sal1::NatMX4 cassette in MWY84/3 and MWY83/5 strains was verified by PCR using the primers SalVerif and SalR and by phenotypic analysis.

Yeast Mitochondrial Isolation
Mitochondria from the yeast strains were prepared as described previously [48]. Typical yield was ,20 mg of mitochondria from 2 liters of yeast cultures (OD = 4). Protein was determined by the method of Lowry.

Isolation of Mitochondria from Artemia Franciscana
No permits were required for the described study, which complied with all relevant regulations. Mitochondria from embryos of Artemia franciscana were prepared as described elsewhere, with minor modifications [2]. Dehydrated, encysted gastrulae of Artemia franciscana were obtained from Salt Lake, Utah through Artemia International LLC (Fairview, Texas 75069, USA) and stored at 4uC until used. Embryos (15 g) were hydrated in 0.25 M NaCl at room temperature for at least 24 h. After this developmental incubation, the embryos were dechorionated in modified antiformin solution (1% hypochlorite from bleach, 60 mM NaCO3, and 0.4 M NaOH) for 30 min, followed by a rinse in 1% Na + -thiosulfate (5 min) and multiple washings in icecold 0.25 M NaCl as previously described [49]. After the embryos were filtered through filter paper, ,10 g were homogenized in icecold isolation buffer consisting of 0.5 M sucrose, 150 mM KCl, 1 mM EGTA, 0.5% (wt/vol) fatty acid-free BSA, and 20 mM K + -HEPES, pH 7.5, using a glass-Teflon homogenizer at 850 rpm for ten passages. The homogenate was centrifuged for 10 min at 300 g at 4uC, the upper fatty layer of the supernatant was aspirated and the remaining supernatant was centrifuged at 11,300 g for 10 min. The resulting pellet was gently resuspended in the same buffer, avoiding the green core. The green core was discarded, and the resuspended pellet was centrifuged again at 11,300 g for 10 min. The final pellet was resuspended in 0.4 ml of ice-cold isolation buffer consisting of 0.5 M sucrose, 150 mM KCl, 0.025 mM EGTA, and 20 mM K + -HEPES, pH 7.5 and contained ,80 mg protein/ml.

Determination of Mitochondrial Membrane Potential (m)
m of isolated yeast mitochondria was estimated as described previously [50], by monitoring the quenching of rhodamine 123 fluorescence (0.5 mM) using a l ex of 485 nm and a l em of 533 nm

Determination of Citrate Synthase Activity
Citrate synthase activity for the fractions obtained during mitochondrial and mitoplast isolations was performed as detailed in [51]. Citrate synthase specific activity is expressed as nmol of DTNB (5,59-dithiobis-(2-nitrobenzoic acid) reduced to NTB (2nitro-5-thiobenzoate) per min per mg of protein. Mitoplasts were generated by subjecting 10 mg of mitochondria in a 2 ml buffer composed of 20 mM Tris pH 6.8 for 30 min, in the presence of 0.05% digitonin. After spinning at 10,000 g for 10 min, the supernatant was removed, and the pellet consisting of mitoplasts was resuspended in 0.1 ml of the same buffer.

Mass Spectrometry: Yeast Mitochondria
2 ml of yeast mitochondria (6 mg/ml) containing 5 pmol betalactoglobulin as an internal standard was incubated for 30 min at 60uC with 1 ml reagent mixture containing 0.33 w/w% RapiGest SF and 33 mM dithiothreitol in a total volume of 20 ml. This was followed by alkylation for 30 min in the dark at room temperature in the presence of 1 ml 200 mM iodoacetamide and 5 ml 200 mM NH 4 HCO 3. Digestion was performed either by trypsin (1 ml, 20 mM) at 37uC for 90 min or 180 min or by chymotrypsin (1 mL, 39 mM) at 37uC for 90 min or overnight. Digestion was quenched by 1 ml formic acid (30 min at 37uC) and the reaction product was centrifuged at 17,000 g for 10 min. The sample was diluted two times prior to analysis and a 2 mL sample was injected on to the capillary LC column. LC-MS/(MS) experiments were carried out using a nanoflow UPLC system (nanoAcquity UPLC, Waters, Milford, MA, USA) coupled to a Q-TOF Premier mass spectrometer (Waters, Milford, MA, USA). Before separating the peptides on a reverse phase analytical column (C18, 75 mm i.d. 6 150 mm, 1.7 mm BEH300 particles, Waters, Milford, MA, USA), samples were desalted online on a Symmetry C18 trap column (180 mm i.d. 6 20 mm, Waters, Milford, MA, USA). A gradient was applied using a ow rate of 450 nanol/min and column temperature 55uC for 100 or 200 min, as described previously [54] using aqueous and acetonitrile-containing solvents, both in the presence of 0.1% formic acid. Peptides were identified by tandem mass spectrometry in two separate data dependent acquisition modes (DDA). In the first case 2 sec cycles were used, consisting of a full scan spectrum (m/z: 500-1999) and MS/MS spectra of the three most abundant ions. In the second case 3 sec cycles were applied, consisting of a full scan spectrum (m/z: 400-1500) and MS/MS spectra of the three most abundant ions included in the predefined inclusion list. The inclusion list contained the masses of the expected tryptic or chymotryptic peptides of ArAAC. Ar collision gas was used in the tandem mass spectrometry measurements. ProteinLynx Global Server v.2.3 (Waters, Milford, MA, USA) was used to process data of DDA experiments. Mascot Server version 2.2 (Matrix Science, London, UK) was used to analyze samples searching against Artemia franciscana taxonomy assuming digestion enzymes trypsin or chymotrypsin and allowing for two missed cleavages. Iodoacetamide derivatives of cysteines and oxidation of methionines were specified as fixed and variable modifications, respectively. Fragment ion mass tolerance of 0.15 Da and a parent ion tolerance of 50 ppm were used and peptides were considered if they could be established at greater than 95.0% probability. To identify further peptide fragments, additional error-tolerant Mascot searches were performed against Artemia franciscana taxonomy.

Sample Preparation for Artemia Franciscana Mitochondria Lysed in RIPA Buffer
Artemia franciscana mitochondria in RIPA (radioimmunoassay) buffer were thawed and centrifuged at 16,000 g for 10 min. The pellet was solubilized in 5 M urea, 2 M thiourea, 40 mM DTT and 0.1% SDS, and both the pellet and the supernatant were cleaned up using the GE Healthcare 2-D Clean-Up Kit per the manufacturer's instructions (GE Healthcare Bio-Sciences Corp., Piscataway NJ). After the clean up, the pellets were extracted with 100 mM ammonium bicarbonate, pH 7.8 and sonicated for 10 min in a bath sonicator. After centrifugation, the extraction and sonication steps were repeated on pellets still remaining. Supernatants were combined and the protein concentration was determined using the Pierce 660 nm assay as per the manufacturer's instructions (Pierce, Rockford IL). Three 60 g protein portions were electrophoresed by SDS-PAGE using a 7 cm 10% Mini-Protean TGX gel from BioRad (BioRad, Hercules CA) and stained with BioSafe Coomassie Blue (BioRad, Hercules CA). The gel lanes were digested with trypsin following reduction with DTT and alkylation with iodoacetamide, followed by peptide extraction from the gel lane [55]. The peptide digest was cleaned up using Spec-C18 cartridges (Varian, Lake Forest, CA).

LC-LC-MS/MS Analysis of the RIPA-lysed Mitochondria
For LC-LC-MS/MS analysis, a microbore HPLC system (Surveyor, Thermo Fisher Scientific, San Jose, CA) was used with separate strong cation exchange (SCX) and reversed phase (RP) columns: a 100 mm I.D. capillary packed with 3.5 cm of 5 mm PolySulfoethyl-Asp strong cation exchanger (PolyLC Inc., Columbia, MD) and a separate 100 mm I.D. capillary packed with 7.5 cm of 5 mm Zorbax Eclipse XDB-C18 material (Agilent, Santa Clara, CA). Peptides were acidified using trifluoroacetic acid and 30 g was manually loaded by pressure packing onto the SCX column. Peptides were eluted at 400 nl/min by a reverse phase gradient using Buffer A (water/0.1% formic acid), Buffer B (acetonitrile/0.1% formic acid), preceded by a salt bump using Buffer C (250 mM ammonium acetate), and Buffer D (1.5 M ammonium acetate). Twelve steps were then performed as follows: (step 1) 0% C with an RP gradient of 5-50% B over 90 minutes followed by a column clean-up of 5 minutes using 50-98% B and an equilibration of 20 min with 5% B. (steps 2-11) X% C (where X = 10-100% C increased in increments of 10%; the remaining % was Buffer A) loaded over 4 min and then washed with 5% B for 7 min followed by a RP gradient of 5-50% B over 60 min. Each RP gradient was followed by a column clean-up of 5 min using 50-98% B and an equilibration of 20 min with 5% B.
Step 12:50% D loaded over 4 min and then washed with 5% B for 7 min followed by a gradient of 5-50% B over 60 min. The flow rate was 1000 nl/min for the 7-minute washes following each salt bump and for each final 5% B equilibration step. Peptides were directly sprayed into a ThermoFinnigan LTQ linear ion trap mass spectrometer (Thermo Fisher Scientific, San Jose, Ca) using a custom-built nanoelectrospray ionization source. Electrospray voltage of 2.0 kV was applied using a gold electrode via a liquid junction upstream of the column. Spectra were scanned over the range 400-1500 atomic mass units (amu). Automated peak recognition, dynamic exclusion (45 seconds), and daughter ion scanning of the top seven most intense ions were performed using the Xcalibur v 1.4 SR1 data system (Thermo Fisher Scientific, San Jose CA) [56]. The LC-LC-MS/MS analysis was repeated twice for a total of three replicates and the data were combined. Tandem MS spectra of peptides were analyzed with TurboSE-QUEST TM v 3.1, a program that allows the correlation of experimental tandem MS data with theoretical spectra generated from known protein sequences [57]. The peak list for the search was generated by Bioworks 3.1 (Thermo Fisher Scientific, San Jose, CA) considering fully tryptic peptides with up to two missed cleavages. Iodoacetamide derivatives of cysteines and oxidation of methionines were specified as variable modifications. Parent peptide mass error tolerances were set at 1.5 amu and fragment ion mass tolerance set at 0.5 amu during the search. Preliminary peptide identifications were made using the following Xcorr filters: peptide precursor ions with a +1 charge having a Xcorr .1.8, +2 Xcorr .2.5 and +3 Xcorr .3.5. A deltaCn score .0.08 was also used as filtering criteria for reliable matched peptide identification [58], [59]. Tandem mass spectra were searched against mouse genes encoding proteins with strong support of mitochondrial localization downloaded from the Broad Institute (http://www. broadinstitute.org/pubs/MitoCarta/index.html), on July 27, 2012. At the time of the search this Mouse MitoCarta protein database contained 1,173 entries. Additionally, the tandem mass spectra were searched against Crustacea proteins downloaded from NCBI on July 25, 2012 and appended with common contaminant proteins (eg., human keratins, trypsin, etc). At the time of the search this custom protein database contained 108,788 entries. The results were also validated using the search engine X!Tandem [60], and displayed with Scaffold v 3.6.1 (Proteome Software Inc., Portland OR), a program that relies on various search engine results (i.e: Sequest, X!Tandem, MASCOT) and which uses Bayesian statistics to reliably identify more spectra [61].

Sample Preparation for Artemia Franciscana Mitochondria Lysed by Freeze/Thaw in Water
SDS was added to the lysate at a final concentration of 0.1% and the lysate was sonicated for 5 sec using a probe sonicator. Proteins were precipitated with acetone. After centrifugation at 16,000 g for 10 min the topmost lipid layer and the supernatant were removed and discarded. The pellet was extracted with 8 M urea/1 M guanadine HCl, sonicated as described above, and subjected to a round of freeze/thawing. After centrifugation the supernatant was saved and the pellet was extracted with 8 M urea then centrifuged and the supernatant was removed and saved. The combined supernatants were assayed for protein concentration using the Pierce 660 nm reagent as per the manufacturer's instructions (Pierce, Rockford IL). The pellets from the processing steps described above were dissolved in Laemmli sample buffer, combined, and assayed for protein concentration as described previously. One mg portions of the supernatant fraction were either delipidated ( [62] or cleaned up using the GE Healthcare 2-D Clean-Up Kit as per the manufacturer's instructions (GE Healthcare, Piscataway NJ). The samples were fractionated by SDS-PAGE; 10 g each of the delipidated sample, the kit-processed sample, and the pellets was loaded on a 15% acrylamide Criterion gel (BioRad, Hercules, CA), and after electrophoresis the gel was stained with BioSafe Coomassie Blue (BioRad, Hercules CA). Each gel lane was cut into 8 sections and the sections were digested with trypsin following reduction with DTT and alkylation with iodoacetamide, followed by peptide extraction from the gel sections [55].

LC MS/MS Analysis of the Freeze/Thaw Lysed Artemia Franciscana Mitochondria
LC-MS/MS analysis of trypsin-digested gel sections was carried out using a LTQ Orbitrap Velos mass spectrometer (Thermo Fisher Scientific, San Jose, CA) equipped with an Advion nanomate ESI source (Advion, Ithaca, NY), following ZipTip C18 sample clean up according to the manufacturer's instructions (Millipore, Billerica, MA). Peptides were eluted from a C18 precolumn (100-mm id62 cm, Thermo Fisher Scientific) onto an analytical column (75-mm ID 6 10 cm, C18, Thermo Fisher Scientific) using a 5% hold of solvent B (acetonitrile, 0.1% formic acid) for 5 min, followed by a 5-7% gradient of solvent B over 5 min, 7-15% gradient of solvent B over 45 min, 15-35% gradient of solvent B over 60 min, 35-40% gradient of solvent B over 28 min, 40-85% gradient of solvent B over 5 min, 85% hold of solvent B for 10 min and finally a return to 5% in 1 minute and another 10 minute hold of 5% solvent B. All flow rates were 400 nl/min. Solvent A consisted of water and 0.1% formic acid. Replicate runs were also performed using a shorter 125 minute RP gradient (5% solvent B hold for 10 min, 5-20% gradient of solvent B over 65 min, followed by a 20-35% gradient of solvent B over 25 min, 35% solvent B hold for 9 min, 35-95% gradient of solvent B over 5 min, and finally by a 95% solvent B hold for another 5 minutes). Data dependent scanning was performed by the Xcalibur v 2.1.0 software [56] using a survey mass scan at 60,000 resolution in the Orbitrap analyzer scanning mass/charge (m/z) 400-1600, followed by collision-induced dissociation (CID) tandem mass spectrometry (MS/MS) of the 14 most intense ions in the linear ion trap analyzer. Precursor ions were selected by the monoisotopic precursor selection (MIPS) setting with selection or rejection of ions held to a +/210 ppm window. Dynamic exclusion was set to place any selected m/z on an exclusion list for 45 seconds after a single MS/MS. All MS/MS spectra were searched against the protein databases as described above using Thermo Proteome Discoverer 1.3 (Thermo Fisher Scientific) considering fully tryptic peptides with up to two missed cleavages. Iodoacetamide derivatives of cysteines and oxidation of methionines were specified as variable modifications. Proteins were identified at 99% confidence with XCorr score cut-offs [59] as determined by a reversed database search. The protein and peptide identification results were also visualized with Scaffold v 3.6.1 (Proteome Software Inc., Portland OR), a program that relies on various search engine results (i.e.: Sequest, X!Tandem, MASCOT) and which uses Bayesian statistics to reliably identify more spectra [61]. Proteins were accepted that passed a minimum of two peptides identified at 95% peptide confidence and 99.9% protein confidence by the Peptide and Protein Profit algorithms, respectively, within Scaffold.

Supporting Information
Table S1 ArAAC peptides identified by various mass-spectrometric methods. Samples were purified Artemia franciscana mitochondria treated by various approaches as described in the headers of the tables and detailed in Materials and Methods. (PDF) Excel Sheet S1 Mass spectrometric analysis of Artemia mitochondria. Homologues of identified Sal1p, SCaMC-2 (isoform 1) and SCaMC-3 are listed in six sheets, each representing the results of an independent experiment. (PDF)