Dual-localized enzymatic components that constitute the mitochondrial and plastidial fatty acid synthase systems

We report the identification and characterization of genes encoding three enzymes that are shared between the mitochondrial and plastidial-localized Type II fatty acid synthase systems (mtFAS and ptFAS, respectively). Two of these enzymes, β-ketoacyl-ACP reductase (pt/mtKR) and enoyl-ACP reductase (pt/mtER) catalyze two of the reactions that constitute the core, 4-reaction cycle of the FAS system, which iteratively elongate the acyl-chain by 2-carbon atoms per cycle. The third enzyme, malonyl-CoA:ACP transacylase (pt/mtMCAT) catalyzes the reaction that loads the mtFAS system with substrate, by malonylating the phosphopantetheinyl cofactor of acyl carrier protein (ACP). GFP-transgenic experiments determined the dual localization of these enzymes, which were validated by the characterization of mutant alleles, which were transgenically rescued by transgenes that were singularly retargeted to either plastids or mitochondria. The singular retargeting of these proteins to plastids rescued the embryo-lethality associated with disruption of the essential ptFAS system, but these rescued plants display phenotypes typical of the lack of mtFAS function. Specifically, these phenotypes include reduced lipoylation of the H subunit of the glycine decarboxylase complex, the hyperaccumulation of glycine, and reduced growth; all these traits are reversible by growing these plants in an elevated CO2 atmosphere, which suppresses mtFAS-associated, photorespiration-dependent chemotypes.


39
In plants, de novo fatty acid biosynthesis occurs in two distinct subcellular 40 compartments, the plastids and mitochondria (1,2). The two FAS systems utilize an acyl 41 carrier protein (ACP)-dependent, multi-component Type II

111
In these experiments, organelles were identified by a combination of two fluorescence transgene that is plastid targeted (23) ( Figure 1A). These control markers show distinct 117 patterns that are consistent with mitochondrial or plastid localization, and these are 118 distinct from the GFP-signal observed with the non-targeted GFP control, which is 119 located in the cytosol and nucleus, as previously characterized (24) ( Figure 1A).

120
The fluorescence observed in transgenic plants carrying C-terminal GFP-fusions with 121 each full-length candidate mtFAS component-protein show an assortment of organelle 122 localizations. For example, the full-length AT3G45770-encoded protein (rows 1 of Figure   123 1B) and the N-terminal pre-sequence of this protein (rows 2 of Figure 1B) direct GFP 124 into mitochondria. In contrast, when the N-terminal pre-sequence is removed from the 125 AT3G45770-encoded protein, the GFP is guided to the cytosol (rows 3 of Figure 1B).

126
The other 3 candidate mtFAS component-proteins guide GFP location to both 127 mitochondria and plastids, with AT2G05990 guiding expression predominantly to 128 mitochondria, whereas AT2G30200 and AT1G24360 guide GFP predominantly to 129 plastids ( Figure 1C to 1E). These phenomena were further dissected by fusing the GFP-130 8 marker to the N-terminal, potential organelle-targeting pre-sequence of each protein, or 131 fusing the GFP-marker to the C-terminus of each protein that lacks this potential 132 organelle-targeting pre-sequence. These experiments revealed that both the N-terminal, 133 potential organelle-targeting segments, and the mature segments of these proteins 134 contribute to differential targeting of GFP to these two organelles. Specifically, the N-135 terminal pre-sequences of AT2G30200 (residues 1 to 72) and AT1G24360 (residues 1 to 136 78) direct the GFP-fusions to accumulate in mitochondria (row 2 of Figures 1C and 1D), 137 but the mature proteins that lack these N-terminal pre-sequences direct the expression 138 of GFP fusion to plastids. However, when these N-terminal pre-sequences are removed 139 from each protein, the remaining segments direct the accumulation of the fused GFP 140 protein to plastids (row 3 of Figures 1C and 1D). These results demonstrate therefore, 141 that the N-terminal pre-sequences of AT2G30200 and AT1G24360 carry mitochondrial-142 targeting signals. In the case of the AT2G05990, the N-terminal pre-sequence (residues 143 1 to 87) directs GFP to plastids (row 2 of Figure 1E), but the segment, which lacks this 144 N-terminal pre-sequence, guides GFP to mitochondria (row 3 of Figure 1E).

145
In summary therefore, proteins encoded by AT2G30200, AT1G24360 and AT2G05990 146 appear to encode dual localization signals, to both mitochondria and plastids; one of 147 these signals resides in the N-terminal pre-sequence, and the other in the mature portion 148 of these proteins. Furthermore, the signal that resides in the N-terminal pre-sequence 149 appears to predominate in the case of AT2G30200 and AT2G05990, whereas with 150 AT1G24360 the organelle targeting signal in the mature portion of the protein 151 predominates. In contrast, the AT3G45770-protein is singly targeted to mitochondria via 152 a N-terminal pre-sequence signal peptide.

164
The yeast mutant strains that lack mtFAS functions cannot utilize glycerol as a sole

174
In the case of the enoyl-ACP reductase components of the mtFAS system, recombinant 175 purified proteins encoded by AT2G05990 or AT3G45770 (expressed in E. coli) were also 176 evaluated in vitro for their ability to catalyze the expected chemical reaction. Because 177 enoyl-ACP reductases are known to be active with both enoyl-ACP (their native 10 substrate) and enoyl-CoA (32), in these experiments each protein was tested for the 179 ability to reduce enoyl-CoA substrates. These assays were conducted with Δ2 trans -10:1-

180
CoA and Δ2 trans -16:1-CoA, and activity was monitored by the decrease in A340 due to 181 the coupled oxidation of the pyrimidine nucleotides (NADH or NADPH). Both 182 AT2G05990 and AT3G45770 proteins were capable of reducing the enoyl-CoA 183 substrates, and they exhibit comparable K m , V max and catalytic efficiency (k cat /K m ) with 184 both tested substrates ( Figure 2B). Moreover, these assays establish that AT2G05990 is 185 an NADH-dependent reductase, and its activity with NADPH is undetectable. In contrast,

186
AT3G45770 is an NADPH-dependent reductase, and its activity with NADH is 187 undetectable.

188
In combination therefore, the GFP-transgenic fluorescence data, the yeast genetic

229
In contrast, the transgenic plants expressing the N-terminally truncated pt/mtMCAT or 230 pt/mtKR proteins (i.e., these proteins are predicted to be plastid localized, but not 231 targeted to mitochondria) exhibit reduced size ( Figure 3A). Most significantly, when 232 these plants were grown in a 1% CO 2 atmosphere, where photorespiration deficiency is 233 suppressed, the stunted growth morphology is reversed ( Figure 3A).    Table 1). Collectively therefore, we 287 conclude that the mtFAS system appears to be redundantly enabled by two enoyl-ACP

359
This functional conservation may therefore lead to the retention of dual-targeting events 360 by ensuring that proteins from these two organelles are functionally interchangeable.

361
Mechanistically, for dual-targeting between mitochondria and plastids, the targeted 362 protein has to be recognized by the import machinery of both organelles (34).

363
Evolutionarily, the protein import machineries in these organelles arose independently, 364 are non-homologous, and therefore would normally be expected to recognize different 365 18 targeting signals (35). Indeed, in the case of pt/mtMCAT, pt/mtKR and pt/mtER, these 366 proteins appear to utilize bi-partite targeting signals, an N-terminal signal unique for one 367 organelle and an internal signal that specifies import to the other organelle. Deletion of 368 the N-terminal signal of pt/mtMCAT and pt/mtKR abolishes import into mitochondria and 369 enhances import into plastids, whereas the contrary was observed for pt/mtER. Similar

425
In complementation transgene experiment, pt/mtMCAT          -GFP <-Pt <-Pt <-Pt (A) Genetic complementation of yeast mtFAS mutants (mct1, oar1 and etr1) by expression of Arabidopsis mtFAS candidate genes (AT2G30200, AT1G24360, AT3G45770, and AT2G05990); expression of the WT yeast homolog (MCT1, OAR1, and ETR1) served as a positive control. Gene expression was transcriptionally controlled by the phosphoglycerate kinase promoter (pPGK) and terminator (tPGK). Mitochondrial pre-sequence of yeast COQ3 protein was fused to N-terminus of each protein to ensure the mitochondrial localization. All yeast strains, carrying the indicated expression cassettes were grown on media containing either glycerol or glucose as the sole carbon source, and a dilution series served as the inoculum for each strain.

(B)
In vitro characterization of the catalytic capability of purified recombinant Arabidopsis mtER and pt/ mtER enzymes. Substrate concentration dependence of the enoyl reductase activity was assayed with increasing concentrations of either trans-Δ 2 -10:1-CoA or trans-Δ 2 -16:1-CoA as substrates. The tabulated Michaelis-Menten kinetic parameters were calculated from 3 to 6 replicates for each substrate concentration.      (C) Glycine accumulation. Plants of the indicated genotype were grown in either ambient air or in a 1% CO 2 atmosphere.