Molecular Basis Behind Inability of Mitochondrial Holocytochrome c Synthase to Mature Bacterial Cytochromes

Mitochondrial holocytochrome c synthase (HCCS) is required for cytochrome c (cyt c) maturation and therefore respiration. HCCS efficiently attaches heme via two thioethers to CXXCH of mitochondrial but not bacterial cyt c even though they are functionally conserved. This inability is due to residues in the bacterial cyt c N terminus, but the molecular basis is unknown. Human cyts c with deletions of single residues in α helix-1, which mimic bacterial cyt c, are poorly matured by human HCCS. Focusing on ΔM13 cyt c, we co-purified this variant with HCCS, demonstrating that HCCS recognizes the bacterial-like cytochrome. Although an HCCS-WT cyt c complex contains two covalent links, HCCS-ΔM13 cyt c contains only one thioether attachment. Using multiple approaches, we show that the single attachment is to the second thiol of C15SQC18H, indicating that α helix-1 is required for positioning the first cysteine for covalent attachment, whereas the histidine of CXXCH positions the second cysteine. Modeling of the N-terminal structure suggested that the serine residue (of CSQCH) would be anchored where the first cysteine should be in ΔM13 cyt c. An engineered cyt c with a CQCH motif in the ΔM13 background is matured at higher levels (2–3-fold), providing further evidence for α helix-1 positioning the first cysteine. Bacterial cyt c biogenesis pathways (Systems I and II) appear to recognize simply the CXXCH motif, not requiring α helix-1. Results here explain mechanistically how HCCS (System III) requires an extended region adjacent to CXXCH for maturation.

Previous genetic studies have shown that bacterial cytochromes are incompatible substrates for HCCS-mediated maturation (25,27,28,30). However, it has also been shown by two independent groups that making three mutations in the N-terminal ␣ helix-1 region of bacterial cytochromes can convert them into a substrate that is recognized and matured by HCCS (27,28). c-type cytochromes from the alphaproteobacteria Rhodobacter and Paracoccus have glutamate residues at the sequence positions equivalent to Lys 8 and Ile 10 found in human cyt c (see Fig. 2A, blue asterisks). Furthermore, following an important Phe residue found in human (Phe 11 ) and some bacterial cytochromes (21,27), a residue at the sequence position corresponding to Met 13 in human cyt c (see Fig. 2A) is missing in certain bacterial cytochromes. Substitution of the glutamate residues and insertion of an alanine residue at the corresponding positions in Rhodobacter capsulatus cyt c 2 resulted in its maturation by HCCS (27), thereby suggesting a critical role of the ␣ helix-1 adjacent to the CXXCH motif in mediating interactions between the substrate and enzyme.
The N-terminal ␣ helix-1 of human cyt c consists of an 11-residue sequence (Val 4 -Lys 14 ) immediately followed by the conserved C 15 XXC 18 H 19 heme attachment motif (Fig. 2, A and  B). ␣ helix-1 includes the Lys 8 , Ile 10 , and Phe 11 residues that appear to be important for HCCS-mediated maturation due to their putative roles in substrate recognition by HCCS (21,25,27,28). Cys 15 and Cys 18 supply the reactive thiol side chains that, respectively, form thioether attachments to the 2-and 4-vinyl groups of heme (2,3). Recently, it was shown that His 19 is recognized as one of two axial heme ligands in the HCCS-cyt c complex (31); the other is His 154 of HCCS (27). Furthermore, cyt c His 19 plays crucial roles in facilitating thioether formation at the active site of HCCS (31).
First, HCCS non-covalently binds to reduced heme with His 154 serving as an axial ligand. Second, heme-bound HCCS recognizes, binds to, and positions apocyt c adjacent to the heme with His 19 from cyt c acting as the second axial ligand. Third, heme spontaneously forms two thioether bonds to the apocytochrome cysteines at the CXXCH motif. Fourth, the heme-attached cyt c is released from HCCS.
The exact roles of the N-terminal ␣ helix-1 residues (e.g. Phe 11 ) in recognition (step 2) and/or thioether attachment (step 3) are unknown. In this study, we engineered human cyt c variants (e.g. deletion of the Met 13 residue (⌬M13)) to reflect the sequence properties of bacterial cyt c with resultant poor maturation by HCCS. We examined the biochemical basis behind the inability of HCCS to mature the ⌬M13 cyt c variant. We discovered that substrate recognition by HCCS is not defective with the ⌬M13 cyt c variant. Rather, the N-terminal ␣ helix-1 of cyt c is responsible for aligning the Cys 15 thiol for covalent attachment to the heme 2-vinyl. The ⌬M13 cyt c variant does not possess a thioether at the Cys 15 position when bound at the HCCS active site. However, in the ⌬M13 cyt c variant, His 19 forms the axial heme ligand and positions the Cys 18 thiol for heme 4-vinyl attachment. These results provide mechanistic insight into step 2 (and consequently step 3) of HCCS-mediated cytochrome c maturation and are depicted in an updated model in Fig. 1. We confirmed these conclusions by engineering a cyt c substrate with a repositioned "Cys 15 " thiol in the ⌬M13 background, creating a unique CXCH motif, which now was matured by HCCS.

Results
Residue Spacing in ␣ Helix-1 of Cyt c Is Important for HCCSmediated Maturation-To determine how sequence variations in the N-terminal region of bacterial c-type cytochromes affect the HCCS-mediated biogenesis of eukaryotic cyt c, we co-expressed human HCCS with human cyt c variants that are analogous to bacterial cytochromes (Fig. 2, A-C). Ser 16 and Gln 17 were substituted with alanine residues (S16A/Q17A) because the equivalent positions in bacterial cytochromes are variable. These substitutions resulted in near-wild-type (WT) levels of cyt c (Fig. 2C) and thus do not appear to be involved in maturation by HCCS. Changing Lys 8 and Ile 10 in human cyt c to glutamate (K8E/I10E) resulted in a marked increase in cyt c yield (Fig. 2C). We have not investigated this effect further because ⌬M13, which mimics the shifted residue spacing found in bacterial cytochromes ( Fig. 2A), resulted in at least a 10-fold decrease in maturation by HCCS (Fig. 2C). The decrease in cyt c was confirmed by UV-visible spectra of purified soluble ⌬M13 cyt c when compared with WT cyt c from 1-liter cultures (Fig.  2D). This purified cyt c is termed "released cyt c" because it is released from membrane-bound HCCS into the soluble fraction as described previously (27). Levels of released cyt c for the S16A/Q17A variant were the same as WT, whereas the levels of the K8E/I10E and K8E/I10E/⌬M13 variants were 1.8-fold more and 10-fold less than WT, respectively. Other N-terminal ␣ helix-1 deletion variants of cyt c (i.e. ⌬I12, ⌬K14, ⌬M13/⌬S16, and ⌬M13/CSAQC) also were poorly matured as determined by UV-visible spectroscopy (Fig. 2, A and E). Combining the Met 13 deletion with the Lys 8 /Ile 10 substitutions (K8E/I10E/ ⌬M13) resulted in a dramatic reduction of holocyt c product (Fig. 2C), effectively negating the increase in cyt c yield gained by the K8E/I10E double mutation. These data indicate that the spatial organization of residues in the N-terminal ␣ helix-1 is of major importance for maturation by HCCS and the major basis behind the inability of HCCS to mature bacterial c-type cytochromes.
Cyt c Variants Containing N-terminal Residue Deletions Form Single Thioether Complexes with HCCS-Following coexpression, a fraction of cyt c can be co-purified in complex with membrane-bound HCCS (27,31,32), effectively trapping a transitional intermediate during the maturation process. For example, single cysteine or histidine (from the CXXCH motif) FIGURE 2. Spatial context of cyt c sequence required for efficient HCCSmediated maturation. A, ClustalW-generated amino acid sequence alignment of N-terminal region from bacterial (Rhodobacter capsulatus (UniProt ID P00096) and Paracoccus denitrificans (UnitProt ID P00094)) and eukaryotic (Homo sapiens (UniProt ID P99999)) cyts c along with the cyt c variants constructed for this study. Identical residues are shaded in light blue. Blue asterisks (*) denote residue positions in human cyt c that were substituted with glutamate. Red asterisks (*) denote residue positions in human cyt c that were deleted or substituted for this study. Residue numbering represents that of full-length human cyt c sequence with the initiating methionine as position 1. B, PyMOL-generated structure of human cyt c N terminus (modified from Protein Data Bank code 3ZCF chain D) with residues from Gly 2 to His 19 shown. Residues in ␣ helix-1 and the heme attachment site are color-coded as follows: Phe 11 , pink; Ile 12 , gray; Met 13 , magenta; Lys 14 , cyan; Cys 15 and Cys 18 , yellow; His 19 , blue. Heme is depicted in red. C, heme stain of released cyt c. The indicated cyts c were co-expressed with HCCS in ⌬ccm E. coli. Cells were lysed with Bacterial Protein Extraction Reagent (BPER), and 100 g of total protein was resolved by SDS-PAGE and transferred to nitrocellulose. Values below represent densitometric quantification of detected heme stain signal relative to WT, which is set to 1. D, UV-visible absorption spectrum showing the levels of HCCS-matured WT (black) and ⌬M13 (red) cyt c purified from 1-liter cultures. The wavelength (nm) at the ␣ peak absorption maxima is indicated. E, UV-visible absorption spectrum showing the ␣/␤ region of HCCS-matured WT 10ϫ diluted (black), ⌬I12 (red), ⌬K14 (blue), ⌬M13/⌬S16 (green), and ⌬M13/CSAQCH (orange) cyt c purified from 1-liter cultures. Spectra are offset for clarity. Abs, absorbance.
substitutions are trapped in complex with HCCS, whereas the triple mutant (in which both cysteines and the histidine residue are changed) is not (31). We used this co-purification approach to determine whether ⌬M13 cyt c was defective in recognition by HCCS (step 2) or perhaps another step in maturation. HCCS complexes containing cyt c exhibit spectral characteristics that provide insight into the heme environment within the complex (31). To determine whether cyt c ␣ helix-1 deletion variants interact with HCCS, we purified recombinantly expressed HCCS from Escherichia coli membranes and probed for co-purified cyt c (Fig. 3, A-C). Similar to WT cyt c (Fig. 3A, lane 1), each deletion variant co-purified with HCCS in a heme-attached state (Fig. 3A). The heme stain shows significant holocyt c in all cases co-purifying with HCCS. This was confirmed by total protein stain (Fig. 3B) and by immunoblotting with antibodies that specifically recognize cyt c (Fig. 3C). However, when we evaluated the UV-visible spectra of the WT, ⌬M13, and ⌬K14 complexes (Fig. 3D, panels i-iii), complexes with deletion variants exhibited spectral features similar to a C15S cyt c complex (Refs. 27 and 31 and Fig. 3D, panel iv). When co-purified with HCCS, C15S cyt c is only able to form a single covalent attachment (at Cys 18 ) with heme, yielding a split ␣ absorption peak at wavelength 555-560 nm upon chemical reduction (Refs. 27 and 31 and Fig. 3D, panel iv, red line). Complexes with C18A cyt c form a single thioether with heme at Cys 15 and yield an ␣ peak at ϳ560 nm (27,31). The above results suggest that ⌬M13 cyt c may only have Cys 18 covalently attached when in complex with HCCS. These results were further supported by dithionite-reduced pyridine hemochrome spectra for each of the complexes, which showed ␣ peak maxima at 552.8, 552.4, and 552.5 nm for ⌬M13, ⌬K14, and the C15S cyt c complexes, respectively (Fig. 3E), values indicative of a single thioether (31,34,35). Likewise, other deletion variants co-purified with HCCS ( Fig. 3, A-C) and showed similar ␣ peak maxima values in the reduced pyridine (Fig. 3F), all diagnostic for the presence of a single covalent attachment (31,34,35). As expected, the complex with WT cyt c exhibited an ␣ peak maximum at 550.8 nm in the pyridine hemochrome spectrum (Fig. 3E), a value that is consistent with heme containing two covalent attachments (31,32,36). Collectively, these data suggest that, once bound to HCCS, cyt c variants containing deletions in ␣ helix-1 readily form only a single thioether with heme, although both reactive cysteines in the C 15 XXC 18 H 19 heme attachment motif are intact. Furthermore, based on the spectral similarities of these variants with the C15S cyt c mutant complex, it is possible that formation of the single thioether takes place at the Cys 18 posi- The ␣/␤ peaks in the reduced spectra are enlarged 3-fold for clarity. Absorption maxima values are indicated. E, sodium dithionite-reduced pyridine hemochrome spectra of HCCS co-purified with WT (black), ⌬M13 (red), ⌬K14 (blue), or C15S (green) cyt c. Arrows depict the wavelength (nm) of ␣ peak absorption maxima. F, sodium dithionite-reduced pyridine hemochrome spectra of HCCS co-purified with ⌬I12 (red), ⌬M13/⌬S16 (green), or ⌬M13/CSAQSH (orange) cyt c. Arrows depict the wavelength (nm) of ␣ peak absorption maxima. Abs, absorbance; WB, Western blotting.
tion. All complexes exhibit spectra demonstrating that the His 19 axial ligand is formed (with the HCCS His 154 as the other axial ligand) (data not shown). This includes the shift in the Soret peak (i.e. of reduced versus oxidized) and increased absorption of the reduced ␣/␤ peaks as discussed previously (31).

Deletion of Cyt c Met 13 Abrogates Formation of Cys 15 Thioether Linkage in the HCCS Complex-We focused on
⌬M13 cyt c to test the hypothesis that HCCS complexes with cyt c deletion variants in ␣ helix-1 form only a single covalent attachment to heme at Cys 18 . We generated single cysteine substitutions (of CXXCH) in the ⌬M13 background ( Fig. 2A). Substituted residues in this background are denoted with wild-type numbering of the corresponding residue to maintain clarity for the residue order, and tick marks (Ј) are used to imply altered spacing from wild type. When we co-expressed HCCS with either ⌬M13/C15ЈA or ⌬M13/C18ЈA cyt c variants, purified complexes were obtained (Fig. 4, A-C), indicating that a single cysteine substitution combined with the deletion of Met 13 did not abolish recognition/interaction of the cytochrome with HCCS. However, the ⌬M13/C18ЈA cyt c variant no longer possessed covalently attached heme (Fig. 4B, last lane). These data were further validated by UV-visible and reduced pyridine hemochrome spectra of the purified complexes (Fig. 4, D and E). Spectral properties of ⌬M13/C15ЈA cyt c were nearly identical to those of the HCCS complex with ⌬M13 cyt c (compare Fig. 4D, panel i, with Fig. 3D, panel ii), whereas the ⌬M13/ C18ЈA complex appeared similar to recombinant HCCS alone (Fig. 4D, panel ii, and Refs. 27, 31, and 32). Recombinant HCCS is purified with endogenous heme that is non-covalently bound (27,31,32), and the absence of any covalent linkages to the heme present in the ⌬M13/C18ЈA complex was confirmed by the reduced pyridine hemochrome ␣ absorption maximum value of 555.5 nm (Fig. 4E, red line). The presence of a single thioether in the ⌬M13/C15ЈA complex was confirmed by its pyridine hemochrome spectrum (Fig. 4E, black line), which again mimicked the complex with ⌬M13 (Fig. 3E, red line), both exhibiting ␣ absorption maxima around 552 nm. These data demonstrate that the ⌬M13 cyt c variant forms a single thioether with heme at the Cys 18 position in the C 15 XXC 18 H 19 heme attachment site of cyt c; thus, the Cys 15 side chain is free in the complex with HCCS and not covalently attached.
Deletion of Met 13 Places Cys 15 Out of Position Relative to Phe 11 in N-terminal ␣ Helix-1-Based on our data, we hypothesized that the Cys 15 residue in ⌬M13 cyt c was spatially repositioned from the heme 2-vinyl when bound at the active site of HCCS, thereby preventing its covalent attachment. To gain more insight into the potential structural abnormalities elicited by this deletion, we used PEP-FOLD (37,38) to predict the N-terminal structure of human cyt c (from Asp 3 to His 19 ), including ␣ helix-1 and the heme attachment motif (Fig. 5). The ␣ helical character in the structural prediction of the WT cyt c N-terminal region appeared similar to that found in the cyt c solved structure (compare Fig. 5A with Fig. 2B). PEP-FOLD structures displayed in Fig. 5 anchor Phe 11 in the same position; thus, the remaining residues are displayed relative to anchored Phe 11 . In the predicted structure for the N-terminal region of ⌬M13 cyt c (Fig. 5B), it is clear that Cys 15 Ј is displaced one residue away from its WT location relative to Phe 11 . Rather than a backbone hydrogen bond between Phe 11 and Cys 15 (a bonding pattern that occurs in ␣ helices (39)), in the ⌬M13 cyt c ␣ helix-1, the backbone hydrogen bond would link Phe 11 to Ser 16 Ј. A PEP-FOLD model in which Ser 16 is substituted with cysteine (in a ⌬M13/C15ЈA background) (Fig. 5C) shows that the S16ЈC residue is now positioned where a covalent attachment might occur more frequently, that is if the ␣ helix-1 (including Phe 11 ) anchors the first cysteine residue next to 2-vinyl of heme at the HCCS active site.
Engineering an Unconventional CXCH Heme Attachment Site Partially Restores Thioether Formation at the Proximal Cysteine Residue in a ⌬M13/C15ЈA Cyt c Variant-The theoretical PEP-FOLD models and crystal structure of human cyt c both indicate some flexibility of the loop between the two cysteines (Figs. 2B and 5, A-C). Thus, if cyt c ␣ helix-1 is anchoring Cys 15 and the His 19 axial ligand positions the Cys 18 thiol in a ⌬M13 background, substituting Ser 16 Ј with cysteine may facilitate thioether attachment. To make sure the native Cys 15 is unable to form a covalent bond, we used the ⌬M13/C15ЈA variant as a starting point. We engineered a S16ЈC substitution in the ⌬M13/C15ЈA cyt c background ( Fig. 2A) (thus creating a CXCH motif) to determine whether the predicted PEP-FOLD model (Fig. 5C) and the binding data (Figs. 3 and 4) described above would again "anchor" the new cysteine residue. To compare an alternate residue spacing arrangement, we also engineered a cysteine substitution at Lys 14 Ј in the ⌬M13/C15ЈA cyt c variant background ( Fig. 2A), thus creating a unique CXXXCH heme attachment motif. To determine whether these variants could stably interact with HCCS, we purified complexes and evaluated whether cyt c co-purified (Fig. 6, A-C). The ⌬M13/C15ЈA/K14ЈC variant was undetectable in the com- , and ⌬M13/C15ЈA/S16ЈC (C) cyts c are depicted. PyMOL was used to display each predicted model. Select residues are shown for each structure, including the cysteine residues in yellow and His 19 in blue. Phe 11 is used to anchor each structure. plex (Fig. 6, A-C), suggesting that the CXXXCH cyt c substrate variant is not recognized by HCCS. However, ⌬M13/C15ЈA/ S16ЈC cyt c could be detected in complex with HCCS with heme attached (Fig. 6, A-C). The UV-visible absorption spectrum of the HCCS complex with ⌬M13/C15ЈA/K14ЈC cyt c appeared similar to that expected for HCCS alone (Fig. 6D, panel i, and Refs. 27, 31, and 32), and the reduced pyridine spectrum of HCCS co-expressed with this variant confirmed that only b-type heme was present (data not shown), thus confirming that this variant does not interact with HCCS (Fig. 6, A-C). Absorption spectra for the HCCS complex with ⌬M13/C15ЈA/ S16ЈC cyt c exhibited broad ␣/␤ peaks upon reduction (Fig. 4D, panel ii) with ␣ absorption maxima values similar to those of the complex with ⌬M13/C15ЈA cyt c (Fig. 4D, panel i) (see "Discussion"). The UV-visible spectrum of released S16ЈC variant appeared strikingly similar to that of WT cyt c, exhibiting the c-type cytochrome signature ␣ absorption peak at 549 nm, whereas the low levels of released ⌬M13/C15ЈA showed a 552-nm maximum (Fig. 6E). Moreover, the S16ЈC cyt c (CXCH variant) was released at a level 2.5-fold higher than the ⌬M13/ C15ЈA parent (Fig. 6E), suggesting that two thioethers are formed, thus facilitating its release (i.e. step 4 of cyt c maturation). Accordingly, the reduced pyridine hemochrome of the released ⌬M13/C15ЈA/S16ЈC cyt c gave an ␣ peak maximum at 550 nm, indicating that both thioethers formed at the CXCH site (Fig. 6F).
N-terminal Region of Cyt c Protected from Protease Cleavage When in Complex with HCCS-Results presented above indicate that the cyt c ␣ helix-1 and CXXCH motif are directly bound to the HCCS active site. As a complementary approach, we initiated a biochemical analysis of the HCCS-WT cyt c complex to establish which regions of cyt c are bound. Upon resolving HCCS-WT cyt c complexes by SDS-PAGE, we observed heme-stainable polypeptides migrating below full-length cyt c (Fig. 7A, arrows). These polypeptides could be detected by protein stain when increased amounts of co-complex were loaded (Fig. 7B, arrow). We hypothesized that these small peptides are degradation products of complexed cyt c with natural protection from proteolysis. Notably, matured cyt c (released from HCCS) that is purified from the soluble fraction does not contain this truncated fragment when resolved by SDS-PAGE (data not shown). To determine what region of cyt c was "protected" while in complex with HCCS, we isolated the polypeptides (Fig.  7B, arrow) for mass spectrometry analysis. Mass spectrometry of tryptic digests identified the N-terminal portion (Ile 10 -Lys 54 ) of cyt c containing the heme attachment site (Fig. 7C, tryptic peptides detected in red) (Note that peptides containing heme (i.e. as expected with the peptide containing the CSQCH motif) are typically poorly detected by mass spectrometry (40,41); however, the presence of heme within the truncated fragment was confirmed by heme stain (Fig. 7A, top arrow)). Importantly, the ␣ helix-1 tryptic peptide was detected in the protected fragment (Fig. 7C, underlined sequence). We conclude that the cyt c N-terminal ␣ helix-1 is likely bound at the active site of HCCS, consistent with the ⌬M13 cyt c results.

Discussion
We established that the N-terminal ␣ helix-1 of cyt c plays a major role in steps 2 and 3 ( Fig. 1) of HCCS-mediated maturation by positioning Cys 15 of the C 15 XXC 18 H 19 motif for heme attachment. Specifically, the spatial context of the cyt c N-terminal ␣ helix-1 sequence relative to the CXXCH motif is crucial for optimal heme attachment to the polypeptide, which, in large part, explains why bacterial cytochromes are not matured by HCCS. Although the bacterial CCM system (i.e. System I) can successfully mature a wide range of mono-and multiheme cytochromes, containing simply a CX n CH motif (where X n ϭ 1-4 intervening residues) (21,24,42,43), additional structural features of eukaryotic cytochromes contribute to the strict substrate specificity of HCCS. For example, previous genetic studies have suggested that the conserved phenylalanine residue in cyt c ␣ helix-1 is critical for HCCS maturation by serving as a determinant for substrate recognition (step 2) because substitution of this residue can result in poor cyt c yields (21, 25) (Phe 11 in human cyt c; the equivalent of this position in bacterial cytochromes is not fully conserved (21,44)). Here we show deficiencies in cyt c production with the human ⌬M13 cyt c variant containing residue spacing similar to some bacterial cytochromes. However, our data indicate that poor substrate recognition of this variant is not the basis for defective maturation because complex formation with HCCS is unaffected (Fig.  3). Conclusive evidence was obtained that thioether formation (step 3) at the Cys 15 position is specifically impacted by the altered residue spacing caused by the deletion of Met 13 (and other ␣ helix-1 residue deletions) (Fig. 4). Accordingly, we discovered that changing the position of the proximal cysteine in the context of the Met 13 deletion (i.e. ⌬M13/C15ЈA/S16ЈC) partially restored thioether formation (Fig. 6) by correcting the spatial orientation of the cysteine with the residues in the N-terminal ␣ helix (Fig. 5).
The human HCCS recognizes both cyt c and cyt c 1 of the membrane cytochrome bc 1 complex III (6,45). Some lower eukaryotes such as Saccharomyces cerevisiae have two related synthases: HCCS that recognizes cyt c and HCC 1 S that recognizes cyt c 1 (8,46). Hamel and co-workers (45) have shown in S. cerevisiae that human HCCS (termed CCHL for cytochrome c heme lyase) matures yeast cyt c and cyt c 1 , so it has broader specificity. They also showed that, when overexpressed, the yeast HCCS matures cyt c 1 and that suppressor substitutions of the cyt c 1 heme attachment motif, CAACH (second alanine), allow recognition by yeast HCCS (45,47). Our study shows there is little impact on substrate recognition by human HCCS with serine or glutamine substitutions in the human cyt c CSQCH motif (Fig. 2) in keeping with broader substrate specificity for the human HCCS. Our results are consistent with the genetic analyses in yeast by Hamel and co-workers (45) that suggest that the CXXCH motif, as well as the region N-terminal to this site, are important determinants for HCCS maturation of cyt c. Here we define that the cyt c ␣ helix-1 is a component of the recognition elements but is specifically needed for positioning the first cysteine residue at the HCCS active site. A PEP-FOLD analysis of the human cyt c 1 region (IRRGFQVYKQVC-ASCH) where Tyr replaces the Phe 11 of cyt c yields a similar ␣ helix region before the CXXCH motif (data not shown). We have previously shown that a Tyr can replace the cyt c Phe 11 with normal HCCS-mediated maturation unlike an alanine substitution (27). It is likely that the mechanisms for heme attachment by HCCS described here for cyt c are applicable to cyt c 1 .
⌬M13 cyt c, as well as all other ␣ helix-1 deletion variants reported here, are poorly released from HCCS with yields between 5 and 10% of WT cytochrome c (Fig. 2, C and D), reminiscent of other HCCS-matured, single thioether cyt c variants (e.g. C15S and C18A (27,48)). Accordingly, poorly released cyt c variants are co-purified in complex with HCCS at higher levels than WT cyt c (e.g. Fig. 4, A-C). These data further support the contention that formation of both thioethers is required for efficient release (i.e. step 4) of the cytochrome product from the HCCS active site (27). Because the HCCS complex with the engineered CXCH heme attachment site (i.e. ⌬M13/C15ЈA/S16ЈC cyt c) exhibited some spectral characteristics similar to the complex containing the single thioether M13/C15ЈA cyt c parent, we cannot conclude whether both cysteines are thioether-attached. However, we also do not know what spectral features to expect for a HCCS complex with a CXCH cyt c covalent link. We base our conclusion of a repositioned S16ЈC on the fact that HCCS releases 2.5-3-fold more ⌬M13/C15ЈA/S16ЈC cyt c relative to ⌬M13/C15ЈA (Fig. 6E) and other deletion variants. We propose that the positioning of the cysteine substitution at Ser 16 Ј renders the variant more competent to form both thioethers at the HCCS active site, causing further distortion of the heme (31), which facilitates increased release of the CXCH cyt c. We note that the low levels of released and purified cyt c from each deletion variant exhibited ␣ peak absorption maxima at 550 nm (Fig. 2, D and E), similar to that expected for a cytochrome with two thioethers like WT (27,31,49). Although ⌬M13 cyt c, for example, contains an intact heme attachment site (CSQCH), only a single thioether (at Cys 18 Ј) is present in complex with HCCS (Figs. 3  and 4). It is possible that only rare doubly attached ⌬M13 cyt c variants are ultimately released from HCCS. Alternatively, it is also possible that the low level of ⌬M13 cyt c that is released forms the second thioether upon folding and/or while free in the cell. In either scenario, the conclusions here remain the same.
It is noteworthy that the ⌬M13/C15ЈA/S16ЈC cyt c variant containing the unconventional CXCH heme attachment site was biosynthesized by HCCS at significant levels (ϳ22% of WT levels). A previous study using recombinant yeast HCCS (termed CCHL) attempted to mature an engineered cyt c variant from horse heart containing a CXCH heme attachment site but could not detect product (21). Our results are consistent with that study because efficient maturation and release of the CXCH variant required "repositioning" of the first cysteine and shortening of ␣ helix-1 mediated by the deletion of Met 13 . Thus, two regions in mitochondrial cyt c form the basis for recognition and thioether attachment by HCCS: the ␣ helix-1 and the CXXCH motif where the histidine plays multiple critical roles as described previously (31). We demonstrate in this work that both of these substrate maturation determinants can be modulated and potentially exploited to produce other unconventional c-type cytochromes.

Experimental Procedures
Construction of Strains and Plasmids-Plasmids used in this study, pRGK403 (pGEX-HCCS: N-terminal GST-tagged human WT HCCS), pRGK405 (pBAD CYCS: WT human cytochrome c (CYCS)), and pRGK417 (pBAD CYCS (C15S cytochrome c)) have been described previously (27). Oligonucleotide primer sequences and plasmids are reported in Table 1. Nucleotide substitutions were engineered using the Quik-Change II site-directed mutagenesis kit (Agilent Technologies) according to the manufacturer's specifications. All cloning steps were confirmed by sequencing. Verified clones were transformed into the E. coli ⌬ccm strain RK103 (50) or C43 (51).
Bacterial Protein Extraction Reagent Functional Assay-E. coli strains were grown overnight and used to inoculate 5 ml of LB supplemented with appropriate antibiotics. These cultures were grown at 37°C with shaking at 200 rpm for 3 h followed by induction with 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside and 0.8% arabinose (w/v) for an additional 3 h. Cells were harvested by centrifugation at 4,500 ϫ g, and the cell pellet was lysed in 200 l of Bacterial Protein Extraction Reagent (Thermo Scientific). 100 g of extracted protein was resolved by SDS-PAGE, transferred to nitrocellulose, and analyzed by heme stain. Heme stained cytochrome c signals were quantified using Image Studio Lite Version 5.2 software (LI-COR Biosciences).
HCCS Protein Expression and Purification-GST-HCCS proteins were co-expressed with cytochrome c (WT or variant) and purified from E. coli ⌬ccm (RK103 or C43) as described previously (27). Briefly, 100 ml of starter cultures were grown overnight at 37°C with shaking and used to inoculate 1 liter of LB supplemented with the appropriate antibiotics. Following 1 h of growth of the 1-liter cultures at 37°C with shaking at 120 rpm, the cultures were induced with 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside for expression of pGEX-HCCS. For coexpression of the pBAD CYCS (WT or variants), the cultures were induced with 0.2% arabinose (w/v) 2 h after the induction of HCCS expression. Cells were harvested by centrifugation at 4,500 ϫ g, resuspended in PBS with 1 mM PMSF, and sonicated. The crude sonicate was cleared by centrifugation at 24,000 ϫ g for 20 min, and the membrane fraction was isolated by ultracentrifugation at 100,000 ϫ g for 45 min. Membrane pellets were solubilized in 50 mM Tris (pH 8), 150 mM NaCl, 1% Triton X-100 on ice for 1 h. Solubilized membranes were loaded onto glutathione-agarose (Pierce) for an overnight batch pulldown of GST-HCCS protein (with any co-purified cytochrome c). Bound GST-HCCS complexes were eluted with 20 mM reduced glutathione in 50 mM Tris (pH 8), 150 mM NaCl, 0.02% Triton X-100 and concentrated in an Amicon Ultra Centrifugal Filter (Millipore) with a 100,000 molecular weight cutoff, and the total protein concentration was determined using the Bradford reagent (Sigma).
Cytochrome c Purification-The purification of HCCS-released cytochrome c was performed as described previously (31). Briefly, ⌬ccm E. coli carrying plasmids for GST-HCCS and cytochrome c were inoculated into 100 ml of LB supplemented with the appropriate antibiotics, grown overnight at 37°C with shaking, and used to inoculate 1 liter of LB. Following 1 h of growth of the 1-liter cultures at 37°C with shaking at 120 rpm, the cultures were induced with 0.1 mM isopropyl 1-thio-␤-Dgalactopyranoside for expression of pGEX-HCCS. 2 h after the induction of HCCS expression, arabinose was added to 0.2% (w/v) to induce the expression of pBAD CYCS overnight. Cells were harvested by centrifugation at 4,500 ϫ g, resuspended in 50 mM Tris (pH 8) with 1 mM PMSF, and sonicated. The crude sonicate was cleared by centrifugation at 24,000 ϫ g for 20 min, and the soluble fraction was isolated by ultracentrifugation at 100,000 ϫ g for 45 min. The supernatant was loaded onto CM Sepharose Fast Flow resin (GE Healthcare) for an overnight batch pulldown of positively charged proteins (including cytochrome c). Bound proteins were eluted with 50 mM Tris (pH 8), 500 mM NaCl and concentrated in an Amicon Ultra Centrifugal Filter with a 3,000 molecular weight cutoff, and total protein concentration was determined using the Bradford reagent.
Heme Stains, SYPRO Ruby and Coomassie Protein Staining, and Immunoblotting-Heme stains and SYPRO Ruby blot staining were performed as described previously (31,52). Briefly, to preserve the heme signal, protein samples were prepared for SDS-PAGE with loading dye at 1:1 (v/v) that did not contain reducing agents, and the samples were left unboiled. Following electrophoresis, proteins were transferred to nitrocellulose membranes, and the chemiluminescence signal for the heme stain was developed using the SuperSignal Femto kit (Thermo Scientific) and detected with either the ImageQuant LAS4000 Mini detection system (Fujifilm-GE Healthcare) or the Odyssey Fc Imaging System (LI-COR Biosciences). Following heme staining, membranes were washed in PBS and treated with fixing solution (7% acetic acid, 10% methanol (v/v)) for 15 min. The membranes were washed in deionized water, stained with SYPRO Ruby protein blot reagent (Molecular Probes) for 15 min, and washed again in deionized water. SYPRO-stained proteins were visualized with the ImageQuant LAS4000 Mini detection system using the Y515-Di filter or with the Odyssey Fc Imaging System. The membranes were then blocked in PBS containing 5% milk protein (w/v) for 1 h. Blots were probed for 1 h with a 1:2,000 dilution of antiserum against cytochrome and then washed with PBS. Cyt c antiserum was raised (Cocalico Biologicals) against horse heart apocyt c. This cyt c antiserum reacts well with human cyt c. Protein A-peroxidase (Sigma) was used as a secondary label for detection. Following several washes with PBS, the chemiluminescence signal for anti-cytochrome c was developed using the Immobilon Western kit (Millipore) and detected by either the ImageQuant LAS4000 Mini detection system or the Odyssey Fc Imaging System. For Coomassie stains, proteins resolved by SDS-PAGE were stained in Coomassie solution (50% methanol (v/v), 10% acetic acid (v/v), 0.25% Coomassie Brilliant Blue (Sigma) (w/v)) for 1 h. Background staining was removed with several applications of a destaining solution (50% methanol (v/v), 10% acetic acid (v/v)). Stained proteins were imaged with the LAS-1000 Plus detection system (Fujifilm-GE Healthcare).
UV-Visible Absorption Spectroscopy-UV-visible absorption spectra were recorded with a Shimadzu UV-1800 spectrophotometer at room temperature. All spectra were obtained in the same buffer in which the proteins were purified (for membrane protein complexes, 50 mM Tris (pH 8), 150 mM NaCl, 0.02% Triton X-100; for cytochrome c, 50 mM Tris (pH 8), 500 mM NaCl). Spectra of chemically reduced proteins were generated by the addition of solid sodium dithionite (sodium hydrosulfite) to the purified sample. Pyridine hemochrome extraction of purified proteins was performed as described previously (36). Briefly, 0.5 M NaOH and pyridine were added to 100 g of purified protein to yield final concentrations of 100 mM NaOH and 20% pyridine (v/v). Samples were chemically reduced with the addition of solid sodium dithionite (sodium hydrosulfite), and UV-visible spectra were recorded from 500 to 600 nm.
Three-dimensional Models-Three-dimensional models of the human cytochrome c N terminus (WT and variants) were generated by PEP-FOLD (37,38) in Protein Data Bank format and subsequently illustrated in PyMOL (PyMOL Molecular Graphics System, Version 1.8, Schrödinger, LLC). Of the five PEP-FOLD-generated peptide predictions for each cytochrome variant, the selected peptide models represent the lowest energy conformation and thus the most probable structural arrangement. The x-ray crystal structure for recombinant human cytochrome c (Protein Data Bank code 3ZCF) (53) was obtained from the Research Collaboratory for Structural Bioinformatics Protein Data Bank and modified in PyMOL.
In-gel Trypsin Digestion and Peptide Extraction for Mass Spectrometry-100 g of purified GST-HCCS-WT cyt c complexes was resolved by SDS-PAGE on a 15% polyacrylamide gel. Resolved proteins were stained with Coomassie solution for 1 h and destained in destaining solution overnight. The Coomassiestained band containing the protein of interest was excised from the gel and cut into 1-mm 2 pieces. The gel pieces were washed with vigorous agitation at room temperature with 200 l of Milli-Q (EMD Millipore) water for 5 min, 200 l of acetonitrile for 10 min, and 200 l of 50 mM NH 4 HCO 3 , 50% acetonitrile for 15 min. Washes with the NH 4 HCO 3 , 50% acetonitrile solution were repeated until the Coomassie stain was completely removed. The gel pieces were washed once more with 200 l of acetonitrile for 5 min with vigorous agitation at room temperature, then incubated in 30 l of trypsin digestion buffer (6% 100 ng/l trypsin, 47% Milli-Q water, 47% 100 mM NH 4 HCO 3 ) for 10 min, and covered with no more than 20 l of 50 mM NH 4 HCO 3 . The sample was incubated at 37°C overnight after which the sample was incubated with 25 l of 1% formic acid, 2% acetonitrile at room temperature with vigorous agitation for 20 min. The supernatant containing the extracted peptides was transferred to a clean tube and dried down in a refrigerated SpeedVac centrifuge for 30 min. The dried peptide sample was resuspended in 10 l of 0.1% formic acid, vortexed, cleaned with micro-C 18 reversed phase ZipTips (EMD Millipore), and eluted in 12 l of 70%acetonitrile, 0.1% formic acid solution, which was loaded into a 96-well plate reader in an LTQ-Orbitrap Velos Pro (Thermo) mass spectrometer. The collected data were analyzed using Mascot software (Matrix Science) for protein identification and Scaffold (Proteome Software) to validate identified proteins.
Author Contributions-S. E. B. designed and performed experiments, analyzed data, and wrote the manuscript. J. H. performed experiments. R. G. K. designed experiments, analyzed data, and cowrote the manuscript with S. E. B.