A structure and function relationship study to identify the impact of the R721G mutation in the human mitochondrial lon protease

Abstract

Lon is an ATP-dependent protease belonging to the “ATPase associated with diverse cellular activities” (AAA+) protein family. In humans, Lon is translated as a precursor and imported into the mitochondria matrix through deletion of the first 114 amino acid residues. In mice, embryonic knockout of lon is lethal. In humans, some dysfunctional lon mutations are tolerated but they cause a developmental disorder known as the CODAS syndrome. To gain a better understanding on the enzymology of human mitochondrial Lon, this study compares the structure-function relationship of the WT versus one of the CODAS mutants R721G to identify the mechanistic features in Lon catalysis that are affected. To this end, steady-state kinetics were used to quantify the difference in ATPase and ATP-dependent peptidase activities between WT and R721G. The K_m values for the intrinsic as well as protein-stimulated ATPase were increased whereas the k_cat value for ATP-dependent peptidase activity was decreased in the R721G mutant. The mutant protease also displayed substrate inhibition kinetics. In vitro studies revealed that R721G did not degrade the endogenous mitochondrial Lon substrate pyruvate dehydrogenase kinase isoform 4 (PDK4) effectively like WT hLon. Furthermore, the pyruvate dehydrogenase complex (PDH) protected PDK4 from hLon degradation. Using hydrogen deuterium exchange/mass spectrometry and negative stain electron microscopy, structural perturbations associated with the R721G mutation were identified. To validate the in vitro findings under a physiologically relevant condition, the intrinsic stability as well as proteolytic activity of WT versus R721G mutant towards PDK 4 were compared in cell lysates prepared from immortalized B lymphocytes expressing the respective protease. The lifetime of PDK4 is longer in the mutant cells, but the lifetime of Lon protein is longer in the WT cells, which corroborate the in vitro structure-functional relationship findings.

1. Introduction

Human Lon is synthesized in the cytoplasm as precursors and then imported into the mitochondrial matrix where it is processed to become the mature enzyme by cleaving off the mitochondria targeting sequence (MTS) consisting of the first 114 amino acid residues at the N-terminal [1]. In this manuscript, the full-length human Lon is referred to as LonP1 whereas the matured Lon in the mitochondrial matrix, which lacks residues 1 to 114 of LonP1, is referred to as hLon. Human Lon (hLon) protects the mitochondria by degrading oxidatively damaged proteins, certain transiently expressed regulatory proteins and uncomplexed enzyme subunits [2]. The disruption of mitochondria integrity has been implicated in diseases such as Parkinson’s [3], Friedreich ataxia (FRDA) [4] and cerebral, ocular, dental, auricular, skeletal anomalies (CODAS) syndrome [5]. CODAS is a rare, multi-system developmental disorder related to the mutations of hLon [5]. The phenotype consists of developmental delay, craniofacial abnormalities, dental malformation, short stature, epiphysis bone formation noticeable delay, coronal plane cracked, and hip dislocated [6]. Because CODAS syndrome is homozygous recessive, the human has to inherit the mutated lonP1 gene from both parents, who carry one copy of wild-type and one copy of mutated lonP1 gene. Interestingly, the parental carriers do not exhibit any CODAS phenotypes. There are 12 mutants of hLon found related to CODAS disease [5,7] Among these mutants, R721G has a high allele frequency among the Amish [5].

Immortalized B lymphocytes that are homozygous recessive for mutant R721G displayed some mitochondrial abnormalities but cells can grow under standard culture conditions [5]. Therefore this cell line could be used to validate in vitro mechanistic findings. Since one of the functions of Lon is to maintain mitochondria homeostasis by degrading uncomplexed protein subunits, abnormalities in the mitochondria could be attributed in part to the dysfunctional proteolytic activity of R721G hLon. In choosing an endogenous substrate in the B lymphocytes to monitor Lon activity, a mitochondrial protein such as the pyruvate dehydrogenase kinase isoform 4 (PDK4) [8] is a desirable candidate. The PDK4 participates in metabolic flexibility by inhibiting the pyruvate dehydrogenase complex (PDH) through phosphorylation. In vivo, PDK4 is rapidly degraded by hLon with a half-life of 1 h. The immortalized B lymphocytes utilized in this study was used as a cell model to develop the detection threshold for monitoring the Lon-mediated metabolic flexibility in future cell-based studies.

Strauss et al. demonstrated that R721G hLon exhibited reduced enzymatic activity at a single substrate and ATP concentration, but that study did not reveal how substrate binding and enzyme turnover were affected [5]. More importantly, it was not clear how R721G retained the observed level of enzymatic activity. Using an immunoprecipitation study, the Strauss study also detected impaired homo-oligomerization of R721G hLon in cell lysates transiently expressing this mutant. However, the oligomeric state of R721G hLon was not determined. A computational structural model to predict the function of R721 in hLon was constructed based on sequence homology modeling with bacterial Lon [5]. The structural model predicted that R721 was important for binding to MgATP but detailed activity data to evaluate the impact of R721G mutation on hLon, have not been reported. Furthermore, this model could not identify any structural perturbations within the entire protein caused by the mutation. Given these considerations, this study was performed to obtain mechanistic insights at a molecular level. Steady-state kinetics techniques and hydrogen deuterium exchange/mass spectrometry (HXMS) were used to probe the impact of R721G on the structural function and structural dynamic relationship of hLon in solution. For quantitative characterization of functions, a radioactive ATPase assay and a fluorogenic ATP-dependent peptidase assay were used for kinetic characterization of Lon proteases and compared with the k_cat and K_m values published for WT hLon [9]. Negative stain electron microscopy was used to image the oligomeric states of purified WT versus R721G hLon to aid the interpretation of the steady-state kinetic profiles of R721G hLon. In vitro stability of WT versus R721G hLon was compared by limited tryptic studies. Western blots were used to monitor the stability of WT versus R721G hLon as well as the endogenous protein substrate PDK4 in cell lysates to provide physiological meaning to the in vitro studies using purified Lon proteins.

2. Materials and methods

2.1. Materials

Epstein-Barr virus (EBV)-transformed B lymphocyte 30936 was generated from an Amish CODAS-syndrome affected patient homozygous for LONP1 c.2161C > G (p.Arg721-Gly) by the Lineberger Comprehensive Cancer Center, University of North Carolina [5]. This cell line was authenticated in this study showing the mutant hLon R721G indeed exhibits reduced proteolytic activity in cells. Epstein-Barr virus (EBV)-transformed B lymphocyte CCL104 expressing WT hLon was purchased from ATCC, who authenticated this cell line. Synthetic peptides used in this study were custom synthesized by Genscript. Radioactive ³²P-gamma ATP was purchased from PerkinElmer. Chemicals used to prepare buffers, Western blot transfer, protein purification were purchased from Sigma-Aldrich and Thermo Fisher Scientific. Porcine heart pyruvate dehydrogenase was purchased from Sigma. Unless specified, secondary antibodies used in western blots were purchased from Thermo Fisher Scientific. The GenBank accession number AAD24414 was used for the full length LonP sequence.

2.2. Human Lon wildtype and mutant R721G purification

Wildtype hLon was expressed and purified as previously described [9] with some improvements described below: WT hLon expressed in BL21*(DE3) cells were grown at 37 °C, 200 rpm shaking in super broth (SB) containing 50 μg/mL kanamycin until OD₆₀₀ reached 0.7. Then the cells were induced with 1 mM IPTG for 1 h at 37 °C and 200 rpm shaking. After induction, cells were harvested at 4000×g spinning at 4 °C for 10 min. Pelleted cells were combined and flash-frozen in liquid nitrogen. After that, cells were thawed and resuspended in lysis buffer (containing 25 mM Tris pH 7.5 at 4 °C, 0.2 M NaCl, 5 mM EDTA, 2 mM BME, 20% glycerol, and 0.01% Tween 20). Cells were sonicated in a cooling bath prepared with ice and NaCl for 5min (15sec pause every 10sec sonication). Sonicated cell lysate was centrifuged at 23,000×g for 2hrs at 4 °C. Cleared lysate was immediately loaded onto a P11 cation exchange column (Whatman) equilibrated in lysis buffer. The column was then washed with low salt buffer (25 mM Tris pH 7.5 at 4 °C, 0.2 M NaCl, 2 mM BME, 20% glycerol, and 0.01% Tween 20) until protein was no longer coming off the column. Next step, WT hLon was eluted with a linear gradient of 0.2 M NaCl to 1 M NaCl buffer, collected in 10 mL fractions. Fractions were analyzed by Bradford dye, and SDS-PAGE further analyzed the Bradford test positive fractions. Fractions containing WT hLon were combined in dialysis bags (MWCO 12,000–14, 000 Da) and precipitated in saturated ammonium sulfate (pH 7). Precipitated WT Lon was pelleted by spinning at 25,000×g, 4 °C for 2 h. Pelleted WT hLon was resuspended in 500 μL of buffer containing 50 mM Tris (pH 7.5 at 4 °C) and 150 mM NaCl, then immediately loaded onto a superpose 6 gel filtration column, which pre-equilibrated in Lon storage buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 75 mM KPi, 1 mM Mg(OAc)2, 5 mM DTT, 20% glycerol, and 0.05% Tween 20). Elutions were collected in fractions and tested with Bradford dye. SDS-PAGE analyzed Bradford test positive fractions. Fractions containing pure WT hLon were combined, concentrated using Amicon centrifugal filter and quantified by Bradford assay, aliquoted, and stored at −80 °C.

Human Lon R721G lost activity readily at 4 °C during the gel purification step, which we attributed to the time-dependency in purifying the protein. To alleviate this problem, we used Ni(II)NTA chromatography to more rapidly purify his-tagged R721G hLon. We have compared the activity of his-tagged (at the N-terminal) versus non-his tagged WT hLon and observed no difference in enzymatic activity. More importantly, we compared the activity of R721G hLon (with and without an N-term his-tag) after the P11 chromatography step and detected no difference in the enzymatic activity between the two proteins. As such, human Lon mutant R721G was expressed and purified using a similar procedure as WT hLon with the following modifications: hLon mutant R721G expressed in Rosetta DE3 cells were grown at 37 °C, 200 rpm shaking in super broth (SB) containing 50 μg/mL kanamycin and 34 μg/mL Chloramphenicol until OD₆₀₀ reached 0.7. The cell lysate was eluted from the P11 cation exchange column (Whatman) with 1 M NaCl high salt buffer (25 mM Tris pH 7.5 at 4 °C, 2 mM BME, 20% glycerol, and 0.01% Tween 20) directly onto a Ni-NTA column equilibrated in a high salt buffer and the flow-through was collected. The column was washed with 50 mM Imidazole buffer (all Ni-NTA buffers contain 50 mM HEPES pH 8, 300 mM NaCl, 0.1 Mg(OAc)₂, 1 mM TCEP, and 20% glycerol). Finally, hLon mutant R721G was eluted in a stepwise gradient of 0.1, 0.2, and 0.4 M imidazole, and collected in 3 mL per fractions. SDS-PAGE analyzed fractions, and hLon mutant R721G pure ones were combined. Tween 20 was added to the combined fractions to a final concentration of 0.01%. Then hLon mutant R721G was concentrated using an Amicon centrifugal filter, quantified by Bradford assay, aliquoted, and stored at −80 °C.

2.3. Human Lon R721G intrinsic ATPase activity

Reaction mixture containing 50 mM Tris pH 8, 150 mM NaCl, 50 mM Mg(OAc)₂, 2 mM DTT. 1.2 μM hLon R721G, varying concentration of [gamma-³²P] ATP (0.8mM–40mM) was incubated at 37 °C. Aliquots were quenched with 0.5 N formic acid at time 0, 10min, 20min, 30min, 40min, and separated on a PEI-cellulose TLC plate and developed in 0.3 M KPi (pH 7). The amount of Pi produced was determined from using equation 1

$$[Pi]=\frac{DLU(Pi)}{DLU(Pi)+DLU(ATP)}\times [ATP{]}_i$$ (1)

Where [Pi] is the amount of Pi produced, DLU is density light units quantified, and [ATP]_i is the initial concentration of ATP. Rates of ATP hydrolysis were converted to k_obs values by dividing by the enzyme concentration. Values for n, k_cat, K_m and K_i were determined by fitting k_obs values determined for varying amounts of ATP to equation (2) [10].

$$k_{obs}=\frac{k_{cat}\times [ATP{]}^n}{K_m^n+[ATP{]}^n\times \left(1+\frac{[ATP{]}^n}{K_i^n}\right)}$$ (2)

where k_obs is the steady-state rate of ATP hydrolysis (sec⁻¹), k_cat is the rate constant for hydrolysis (sec⁻¹), [ATP] is the concentration of ATP in the reaction, n is the Hill coefficient, K_m is the Michaelis constant (μM), and K_i is the dissociation constant for the inhibitory SES ternary complex (μM).

2.4. Human Lon R721G casein stimulated ATPase activity

Reaction mixture containing 50 mM Tris pH 8, 150 mM NaCl, 50 mM Mg(OAc)₂, 2 mM DTT. 1.2 μM hLon R721G, 20 μM alpha-casein, varying concentration of [gamma-³²P] ATP (0.2mM–35mM) was incubated at 37 °C. Aliquots were quenched with 0.5 N formic acid at time 0, 10min, 20min, 30min, 40min, and separated on a PEI-cellulose TLC plate and developed in 0.3 M KPi (pH 7). The amount of Pi produced was determined by using equation (1). Rates of ATP hydrolysis were converted to k_obs values by dividing by the enzyme concentration. Values for k_cat, K_m and K_i were determined by fitting k_obs values determined for varying amounts of ATP to equation (2).

2.5. Human Lon preparation with AMP-PNP, negative stain EM and image processing

The reaction mixture containing 50 mM Tris pH8, 50 mM Mg(OAc)₂, 150 mM NaCl, 2 mM DTT, 2 μM WT hLon or R721G hLon or a mixture of 1 μM WT hLon and 1 μM R721G hLon, and 2 mM or 30 mM AMP-PNP was incubated at room temperature for 10min. For all negative stain electron microscopy, prepared samples were incubated on carbon-coated EM grids and stained with 2% uranyl acetate. Imaging was performed on a TF20 (FEI Co.) electron microscope at 200 keV, and images were acquired using a Tietz 4k x 4k CCD camera at 80,000x magnification. Initially, hLon particles were manually selected in RELION to generate automatic picking templates [11]. Particles were then imported into cryoSparc [12] for further 2D and 3D class processing. In total, 90,000 WT particles from 157 micrographs and 140,000 mutant particles from 239 micrographs were selected and subjected to 2D classification to identify homogeneous views of the hLon assembly. Ab-initio models were generated with a C1 symmetry. A C6 symmetry was used in 3D refinement of WT particles.

2.6. Cleavage efficiency comparison between 89 and 98Bz and 89-98

Reaction mixture containing 50 mM Tris pH 8, 150 mM NaCl, 10 mM Mg(OAc)₂, 2 mM DTT, 300 nM WT hLon was incubated at 37 °C for 2min. Then the reaction was initiated by 3 mM ATP and 100 μM substrate cocktail. The cocktail used was 10% FR/90% 89-98Bz or 10% FR/90% 89–98. Fluorescent emission upon cleavage of the peptide was monitored at 420 nm (λ_ex = 320 nm) for 600 s. Trypsin was used for calibration.

2.7. Human Lon R721G peptidase activity

Reaction mixture containing 50 mM Tris pH 8, 150 mM NaCl, 10 mM Mg(OAc)₂, 2 mM DTT, 1 μM hLon R721G was incubated at 37 °C for 2min. Then the reaction was initiated by 5 mM ATP and varying [89–98] (89–98 sequence: YRGITCSGRQK, concentration 250μM-15mM). For peptide concentrations less than 2 mM, a 5% FR-89-98 stock solution was used. For peptide concentrations greater than 2 mM and less than 15 mM, a 1%FR/99%NF89-98 stock solution was used. For peptide concentration greater than 15 mM, a 0.5%FR/99.5%NF89-98 stock solution was used to avoid complications due to the inner-filter effect. Fluorescent emission upon cleavage of the peptide was monitored at 420 nm (λ_ex = 320 nm) for 1200 s and converted to the amount of product based on the maximum fluorescence generated per micromolar peptide by complete digestion with trypsin. The steady-state rate of peptide cleavage was determined by the slope of a line tangent to the linear phase of the time course and converted to k_obs by dividing by the enzyme concentration used in the assay. Values for k_cat, K_m and K_i were determined by fitting k_obs values determined for varying amounts of 89–98 to equation (2).

2.8. Digestion of PDK4 by human Lon proteins in the absence and presence of PDH

Reactions contained 50 mM HEPES pH8.1, 50 mM Mg(OAc)2, 200 mM NaCl, 2 mM DTT, 15 μM PDK4, 2 μM hLon, 10 mM ATP with or without the presence of 15 μM PDH. The reaction mixture was quenched at 0, 30, 60, 90min for PDK4. Quenched samples were separated on 4–20% SDS-PAGE and transferred onto a nitrocellulose membrane. The membrane was blocked with 1% BSA/TBS for 1h at room temperature, then incubated with 1/1000 anti-Lon/anti-PDK4 overnight at 4 °C. The secondary antibody used was 1/3000 anti-rabbit-AP in 1% BSA/TBS.

The Western blot pictures were analyzed using Image J. The band intensities of PDK4 were measured and divided by band intensity of human Lon on each lane to get [PDK]/[Lon]. Afterward, [PDK]/[Lon] at each time point was divided by [PDK]/[Lon] at time zero to get the ratio of PDK left at each time point. The ratio of PDK left was plotted versus time (min). The experiment was performed in triplicates.

2.9. Limited tryptic digestion of WT versus R721G hLon

The limited tryptic digestion data presented in this manuscript monitored the protection of a ~72 kDa hLon fragment from tryptic digestion in WT but not R721G. Protein sequencing indicated this fragment corresponded to mature hLon missing the first 139 amino acids at the N-terminal, with trypsin cleaving at Arg 139 of hLon [13]. As such, replacement of R721 by G should not affect the limited tryptic cleavage profile. Reaction mixture containing 50 mM Tris pH8, 50 mM Mg(OAc)₂, 150 mM NaCl, 2 mM DTT, 3 μM WT or R721G hLon, 2 mM or 30 mM ATP, along with 0.27 μg/mL trypsin (trypsin/protein ratio 1:1000) was incubated at 37 °C. The reaction mixture (7.5 μL) was quenched with 20 μg SBTI (soybean trypsin inhibitor) at 0, 5min, 15min, 30min, 60min on ice. Laemmli dye was added to each quenched sample followed by heating at 100 °C for 5min. Then the digested products were separated on 12.5% SDS-PAGE and the gel was stained by Coomassie blue.

2.10. Western Blots/CHX chase of B lymphocytes cell lysates

Cells expressing WT (ATCC-CCL104) or R721G (30936) were maintained in RPMI 1640 with 10% fetal bovine serum and 1% penicillin and streptomycin. To assess protein stability cells were treated with 100 μM of the protein synthesis inhibitor cycloheximide (CHX) for 1, 2 and 8 h, harvested, and total protein was extracted using M-PER Mammalian Protein Extraction reagent (Thermo Fisher Scientific). Western blots were used to detect Lon protein and the endogenous substrate pyruvate dehydrogenase kinase isoform 4 (PDK4), with GAPDH as the internal loading control. Rabbit polyclonal antibodies against human Lon was custom produced by Lampire Biological Lab. Polyclonal antibodies against human/mouse PDK4 were generated as described by Crewe et al. [8] as well as purchased from Invitrogen; comparable results were obtained. Proteins were detected using their respective primary antibodies and IR dye-conjugated secondary antibody (IRDye 800CW and IRDye 680RD from LI-COR Biosciences). GAPDH was used as a loading control (GAPDH antibody was from Millipore, cat. no. MAB374). Western blots were imaged under both 700 nm and 800 nm channels using Li-COR imaging system. Anti-mouse 680RD is the secondary antibody for detecting anti-GAPDH, while anti-rabbit 800CW is the secondary antibody for detecting anti-Lon or anti-PDK4. The proteins were quantified using an Odyssey infrared imaging system (LI-COR Biosciences). The dye intensity of GAPDH was used as an internal reference for comparing the amount of endogenous Lon protein or PDK4 at a given time point.

2.11. HXMS studies of WT versus R721G hLon

WT and R721G hLon were overexpressed in Rosetta (DE3) (Novagen). Cells were grown in super broth (SB, per L: 30 g tryptone, 20 g yeast, 10 g mops, pH 7.5, 34 μg/mL cam, and 100 μg/mL amp) at 37 °C and induced with 1 mM IPTG at OD₆₀₀ = 0.6 for 1 h. Cells were then harvested with Lon lysis buffer (50 mM Tris, 5 mM EDTA, 0.3 M NaCl, 20% glycerol, 0.005% Tween 20, 1 mM BME, pH 7.5), and the resulting lysate was loaded on a preequilibrated phosphocellulose column. The column was then washed with wash buffer (50 mM Tris, 0.5 M NaCl, 10% glycerol, 0.005% Tween 20, 1 mM BME, pH 7.5), and then eluted onto a nickel column with elution buffer (50 mM Tris, 1 M NaCl, 10% glycerol, 0.005% Tween 20, 1 mM BME, pH 7.5). The nickel column was washed with 0.05 M imidazole buffer (50 mM Hepes-KOH, 0.3 M NaCl, 20% glycerol, 0.1 M MgCl₂, 1 mM TCEP, 0.05 M imidazole, pH 8). Lon was then eluted off with a step-wise gradient of 0.1 M, 0.2 M and 0.4 M imidazole buffers. Lon positive fractions were pooled, concentrated, and initially quantified with Bradford assays using BSA as a standard, and subsequently verified by UV absorbance at 280 nm using equation 3

$${\epsilon }_{280}=\mathrm{W}(5500)+\mathrm{Y}(1490)+\mathrm{C}(125)$$ (3)

where ε₂₈₀ is the molar absorptivity at 280 nm, W is the number of tryptophan, Y is the number of tyrosine, and C is the number of cysteine [14,15]. An SDS-PAGE gel of each mutant showed differences in band intensity after Coomassie staining. In our hands, we discovered that R721G but not WT hLon precipitated out of solution at high concentration in the absence of Tween 20. Since Tween 20 was incompatible with the liquid chromatography/mass spectrometer used for this study, Tween 20 was omitted in the buffer for the HXMS experiments. As such, it was possible that the concentration of R721G was underestimated at A280. To account for these differences, the band intensities of each variant were determined using the program ImageJ and compared to WT Lon [16]. Mutant concentrations were then corrected based on the comparison.

2.12. Peptide mapping

A total of 100 pmol (10 μM) of WT R721G hLon in 10 μL of Lon buffer [50 mM Tris (pH 8), 60 mM Mg(OAc)₂, 2.5 mM tris(2-carboxyethyl) phosphine] was mixed with quench buffer [300 mM sodium phosphate (pH 2.3), 10 mM tris(2-carboxyethyl)phosphine, 10% 1 M HCl] and injected into a 20 °C pepsin column for digestion of protein into peptic fragments. Mass spectra of desalted peptic fragments purified by a gradient of 5%–35% acetonitrile in 7 min was acquired by Waters HDX with nanoACQUITY reverse-phase ultraperformance liquid chromatography (RP-UPLC) with HDX technology (Waters) in resolution mode (m/z 100–2000) on a SynaptG2 mass spectrometer with a standard electrospray ionization source. Peptide identities were confirmed by MS^E (mass spectrometry elevated energy) analysis. Data was processed using Protein Lynx Global Server 2.5 and manually confirmed.

2.13. Deuterium labeling

D₂O was purchased from Cambridge Isotope Laboratories, Inc (Andover, MA). Deuterium labeling was performed at room temperature by first incubating Lon for 5 min in Lon buffer, and then dilution in deuterated Lon buffer with a pD of 8. For the samples with nucleotide, Lon was preincubated with 20 mM nucleotide before dilution into deuterated buffer containing nucleotide. The samples were quenched at 5, 50, 500, 1000 and 3180 s with quench buffer at 4 °C and immediately injected. Each time point was run at least in duplicate.

2.14. Data analysis

The deuterium uptake by the identified peptides over time was analyzed using Water’s DynamX 3.0 software. The relative amount of deuterium uptake for each peptic digested Lon peptide was calculated from the difference in deuterium uptake between WT and R271G mutant at each quenched time point using the duplicate runs at 5s, 50s, 500s, 1000s, and 3180 s. To identify regions showing noticeable deviation between WT and R721G, the percentage difference between the time constant (Tau) in the HD exchange time courses of WT versus R721G for each hLon peptide was determined. A more than 70% difference in Tau between R721G and WT was considered significant. To determine Tau, the HD exchange time courses were plotted as deuterium uptake against deuterium exposure time. The plots were then fitted with equation (4) to yield the rate constant for HD exchange. In equation (4), Y_max is the maximum mass difference, x is the deuterium exposure time, k is the rate constant, Y₀ was set to zero. The reciprocal of k defines Tau. Equation (4) is included in GraphPad Prism version 5.0.0 for Windows, GraphPad Software, San Diego, California USA.

$$Y=Y_{max}\times (1-\text{exp}(-k\times x))$$ (4)

3. Results and discussion

3.1. ATP hydrolysis activity of hLon mutant R721G

Like WT hLon, the intrinsic ATPase activity of R721G was stimulated by protein substrates such as casein or a peptide sequence constituting residues 89–98 of the bacteriophage λN protein, an endogenous substrate of E. coli Lon [17,18]. Since the K_m and k_cat parameters of WT hLon have already been determined [9], the impact of R721G mutation on the ATPase activity of hLon can be quantitatively assessed by comparing the K_m and k_cat of the purified mutant. To this end, radioactive [gamma-³²P] ATP was used to monitor the ATPase activity. Human Lon hydrolyzed [gamma-³²P] ATP to yield ADP and [gamma-³²P] Pi, which was resolved and quantified from one another using thin-layer chromatography (TLC) as described in methods and materials [19]. As shown in Fig. 1, the intrinsic ATPase activity of R721G was stimulated by the protein substrate, alpha-casein, which contains multiple Lon cleavage sites. This substrate was used to ensure the processive translocation step associated with the ATPase activity was being monitored and could be compared with the values obtained for WT hLon published in an earlier study [9]. According to Fig. 1, the k_obs values for the intrinsic and casein-stimulated ATPase increase with [ATP] until 10 mM and 5 mM, respectively. After which, the k_obs, value decreases with further increase in [ATP] and approaches zero at 40 mM ATP. This kinetic behavior is indicative of a complete substrate inhibition pattern with the rate of enzyme catalysis (k_obs) approaching zero at high [ATP] [10]. Fitting the data shown in Fig. 1 with equation (2) yields the kinetic parameters summarized in Table 1, which when compared with the kinetic parameters determined for WT hLon (Supplementary Table 1) [9], reveals that K_m but not k_cat of the ATPase are affected by the R721G mutation. The K_m for casein-stimulated and intrinsic ATPase activity of R721G are 9.3-fold and 82-fold higher, respectively, than the WT enzyme. In an AAA + protein, the ATP binding site is formed by the interaction of two enzyme subunits [20]. According to the hLon model constructed from the crystal structure of B. subtilis Lon, R721 forms a salt bridge with Glu 654 in the adjacent enzyme subunit [21]. Replacement of R721 with G abolishes this salt bridge interaction, which leads to a “loosened ATP binding pocket”, hence a reduced K_m. Previous study indicated that oligomeric Escherichia coli Lon displayed a half-site reactivity [22]. This means each hexameric Lon has six ATP binding sites but only three of these sites catalyze MgATP hydrolysis whereas the remaining sites bind MgATP. Furthermore, a dimer interface at the ATPase domain can catalyze ATP hydrolysis. Assuming hLon adopts the same ATPase mechanism, a half-site reactivity is anticipated. Given the k_cat of R721G and WT hLon are comparable, it is plausible that WT and R721G use the same number of ATP binding sites to catalyze MgATP hydrolysis regardless of their oligomeric states. This hypothesis should be testable by monitoring the burst amplitude in the ATPsae activity of R721G by rapid quench transient kinetics, which is currently being conducted in our lab.

Fig. 1.

The intrinsic (red) and casein stimulated ATPase activity (blue) of R721G hLon. The initial rates of ATP hydrolysis were determined from the radioactive ATPase assay as described in Methods and Materials. The k_obs values, defined by rate/[Lon monomer], at 0.8 mM, 1.5 mM, 3 mM, 5 mM, 10 mM, 15 mM, 20 mM, 30 mM and 40 mM ATP are showed in the red plot. The k_obs values for casein stimulated ATPase activity were determined at 0.2 mM, 0.5 mM, 0.8 mM, 1 mM, 1.2 mM, 1.5 mM, 3 mM, 5 mM, 7 mM, 9 mM, 10 mM, 12 mM, 16 mM, 18 mM, 20 mM, 25 mM, 35 mM ATP are shown in the blue plot. The data were fitted with equation (2) to yield the kinetic parameters summarized in Table 1.

Table 1.

Steady-state Kinetic Parameters for the ATPase and Peptidase Activity of R721G human Lon.

	kcat (sec−1)	n	Km (μM)	Ki (μM)
ATPase-casein stimulated	1.1 ± 0.1	3.1 ± 0.5	1500 ± 100	17,000 ± 1000
ATPase-intrinsic	0.4 ± 0.1	3.1 ± 0.5	4200 ± 1400	23,000 ± 7000
Peptidase	1.6 ± 0.2	3.1 ± 0.5	1000 ± 200	5900 ± 800

The K_i for ATP towards the casein-stimulated and intrinsic ATPase activity in R721G are 16.4 mM and 23.4 mM, respectively. The cellular concentration of ATP ranges between 3 and 5 mM. As such, the K_i values are likely too high to have any physiological relevance; but their presences support the existence of at least one non-catalytic or less active enzyme form in the mutant whose formation is favored at high [ATP]. A Hill coefficient n greater than 1 indicates positive cooperativity; binding of ATP to one active site of the hLon mutant R721G complex enhances the binding and hydrolysis of ATP of the other active sites. Since the functional oligomeric state of hLon is hexameric, the detection of active ATPase in R721G hLon indicates that this mutant can exist as functional hexamers, at least prior to substrate inhibition.

To determine if there was more than one enzyme form present in R721G hLon, we compared the negative stain EM images of R721G with WT hLon. Fig. 2 shows the negative stain micrographs of purified recombinant forms of WT and R721G incubated with MgAMPPNP, a non-hydrolyzable analog of MgATP (Fig. 2A and D). The particle density, or abundance, of the assembled oligomer for WT hLon was observed to be greater than the R721G. From these images, 2D class averages were generated and the end-on view of the WT hLon hexamers were primarily detected (Fig. 2B). A low-resolution volume was obtained applying a C6 symmetry (Fig. 2C). Conversely, a mixture of hexamers (Fig. 2E) and pentamers (Fig. 2F) were observed with R721G hLon, and fewer hexamers were present. Additional images of the mutant R721G were taken to increase the number of particles, since fewer assembled complexes were identified. And even with a comparable number of particles, the 2D class average was not as homogeneous, which is consistent with an impaired ability of the mutant to form homogeneous assemblies. It was not possible to generate reliable 3D volumes of either suspected R721G oligomer due to the heterogeneity. Since Lon generally exists as a hexamer, the pentameric R721G could be misassembled complexes. This finding lends support to the proposal that the substration inhibition kinetic could be attributed to the presence of non-catalytic enzyme form in R721G.

Fig. 2.

(A) Negative stain micrograph of hLon WT. Circles denote examples of particles. (B) 2D class averages of hLon WT particles. (C) 3D volume of hLon WT with a C6 symmetry. (D) Negative stain micrograph of hLon R721G. (E) 2D class averages of suspected hexamer particles of hLon R721G. (F) 2D class averages of suspected pentamer particles of hLon R721G.

3.2. Peptide cleavage activity of hLon mutant R721G

Previously we developed a fluorogenic peptidase assay to quantify the ATP-dependent peptidase activity of Lon [9,18]. This assay contains a mixture of two peptides, FR89-98 and Bz89-98, both bear residues 89–98 of the lambda N protein, an endogenous substrate of E. coli Lon [23]. The FR89-98 peptide contains a nitrotyrosine at the N-terminal and a 2-aminobenzylamide (Abz) functionality at the C-terminal. When intact, the fluorescence emitted by Abz is internally quenched by nitrotyrosine. Once cleaved by Lon at the single cleavage site, Abz is separated from nitrotyrosine such that the extent of peptide cleavage could be continuously monitored and quantified by fluorimetry. The mixture of FR89-98 and Bz89-98 peptide allows for fluorescence quantification at substrate concentrations that would have suffered from inner filter effect if a 100% FR89-98 is used [18]. To simplify the preparation of this fluorogenic peptidase assay, we substituted Bz89-98 with the unmodified peptide containing only residues 89–98 of lambda N (N89-98). The peptide N89-98 could be readily generated in high yield by standard solid phase peptide synthesis techniques. Supplementary Figure 1 shows that N89-98 could substitute 89-98Bz in the fluorogenic peptidase assay to quantify the activity of Lon protease, as trypsin, WT as well as R721G generated comparable fluorescent signals when cleaving the peptide cocktail containing FR89-98/Bz89-98 or FR89-98/N89-98. In this study, a cocktail of FR89–98 and N89-98 was used to determine the steady-state kinetic parameters of R721G hLon to avoid the inner filter effect. Based on the ATPase finding discussed above, 5 mM ATP was used in this set of reactions to attain maximal reaction rates without encountering ATP inhibition.

As shown in Fig. 3, the peptidase activity of R721G hLon increases with the concentration of peptide substrate until 2.5 mM. After which the peptidase k_obs decreases and approaches zero at 20 mM. Fitting the data shown in Fig. 3 with equation (2) yield the kinetic data summarized in Table 1. The ATP-dependent peptidase data for WT hLon, which was determined in an earlier study, is included in Supplementary Table 1 for comparison. As in the case of the ATPase, the peptidase activity of R721G exhibits substrate inhibition kinetics, with a K_i value of 5.9 mM, 5.8-fold higher than the K_m of the peptide substrate. Comparing the peptidase data for R721G (Table 1) with WT hLon (Supplementary Table 1) reveals that the K_m and the hill coefficient of the peptidase activity are comparable but the k_cat in R721G is 3.8-fold lower that WT hLon. The reduced k_cat in the peptidase activity of R721G is likely attributed to the compromised hexamer integrity and/or a reduction in the population of functional hexamer as revealed in the negative EM imaging study. As in the case of the ATPase activity, the substrate inhibition pattern detected in Fig. 3 is attributed to high substrate concentrations sequestering the non-productive Lon complexes.

Fig. 3.

The ATP-dependent peptidase activity of R721G hLon. The initial rates of Lon peptide cleavage were monitored by fluorimetry with excitation wavelength at 320 nm and emission wavelength at 420 nm. The initial rates were calibrated as described in Methods and Materials. The k_obs values, defined by rate/[Lon monomer], were obtained at 0.25 mM, 0.75 mM, 1.5 mM, 2 mM, 2.5 mM, 3.5 mM, 6 mM, 10 mM and 15 mM peptide. The k_obs versus [peptide] data were fitted with equation (2) to yield the kinetic parameters summarized in Table 1.

3.3. Compare the proteolytic activity of R721G versus WT hLon in vitro and in B-lymphocytes

In the mouse cardiac muscle cells HL-1, Lon has been shown to digest PDK4 in mitochondria [8]. The function of PDK4 is to inhibit the activity of the pyruvate dehydrogenase complex (PDH) through phosphorylation, thereby reducing glucose metabolism [8,24] According to blast sequence alignment analysis [25], human and murine PDK4 share 93% sequence identity and 97% sequence homology. As murine PDK4 was used in the original study [8], we decided to use it here to bridge the comparison of in vitro (this section) with the cell-lysate (see below) results. The degradation of PDK4 by WT versus R721G hLon was monitored by SDS-PAGE. Fig. 4A shows that purified WT but not R721G hLon digests murine PDK4 in an ATP-dependent manner. In vivo, PDK4 interacts with PDH. To determine if PDH interferes with PDK4 degradation by hLon, we compared the time courses of purified recombinant murine PDK4 degradation by purified recombinant WT hLon in the absence and presence of purified porcine heart PDH, which has 98% sequence identity with the murine enzyme [25]. The protein band intensities of PDK4 and hLon were quantified by Image J to generate the kinetic time courses shown in Fig. 4B. The band intensity of hLon was used as the internal loading reference in each lane. It is discerned in Fig. 4B that in the presence of PDH, WT hLon degrades PDK 4 at a slower rate, indicating PDH protects PDK4 from proteolysis.

Fig. 4.

(A) In vitro, purified R721G hLon fails to degrade PDK4 with the presence of ATP. 10 μM PDK4 was incubated with 2 μM WT hLon or R721G hLon at 37 °C with or without the presence of ATP. Aliquots were quenched from 0 to 90min and the reaction products were separated on 12.5% SDS-PAGE followed by Coomassie stain. (B) Fifteen micromolar of PDK4 was incubated with 2 μM Lon with or without the presence of 15 μM PDH at 37 °C. Aliquots were quenched from 0 to 90min. The reaction products were separated on 4–20% SDS-PAGE and analyzed on Western blot. (C) The band intensity of PDK4 was measured and divided by band intensity of human Lon on each lane to get [PDK4]/[Lon]. Afterward, [PDK]/[Lon] at each time point was divided by [PDK4]/[Lon] at time zero to get the relative PDK4 left at each time point. The relative PDK4 left was plotted versus time (min). With the presence of PDH, PDK4 degradation rate reduced slightly.

Since the regulation of PDH by PDK4 is a central metabolic pathway in mammalian cells, we anticipate this pathway to be present in all cell types, including resting B lymphocytes in humans, despite at a relatively low level of PDK4 expression [26]. Our goal in this study is to monitor the intrinsic stability of PDK4 in WT versus Lon R721G expressing B lymphocytes, thereby obtaining insight into the impact of R721G mutation on the ability of hLon to degrade PDK4. B lymphocytes are readily obtained from human blood samples, and the ability to correlate PDK4 stability with the expression of defective hLon in blood cells should contribute to the development of a diagnostic assay to profile changes in mitochondrial Lon activity.

We analyzed the relative protein stabilities of hLon (WT and R721G) and human PDK4 using a cycloheximide chase assays wherein GAPDH was used as an internal reference. The intensity of each target protein band was quantified by the infrared dye attached to the secondary antibodies used in the western blots, which when divided by the signal from GAPDH, yielded the ratio of hLon protein/GAPDH or PDK4/GAPDH. Fig. 5 shows the plots relating the signal intensities of Lon protein/GAPDH or PDK4/GAPDH before and after CHX treatment. Fig. 5A compares the relative stability of WT versus R721G hLon as well as endogenous PDK4 in the mitochondria expressing the respective Lon protein prior to the addition of CHX (t = 0). The WT hLon/GAPDH ratio is comparable to that from R721G hLon/GAPDH. Comparing the signal intensities of PDK4/GAPDH in cell lysates containing WT versus R721G hLon before the addition of CHX indicates that there is 2.2-fold more PDK4 in cells expressing R721G than WT hLon (Fig. 5A).

Fig. 5.

(A) The cellular level of PDK4 was not significantly affected by the presence of the proteolytically defective Lon (R721G) subunits in the mitochondria of immortalized B lymphocytesATCC-CCL104 cells express WT LonP1 whereas mutant 30936 cells express only R721G LonP1. Western blots were used to detect PDK 4 in immortalized B lymphocytes. GAPDH served as the internal reference for quantification. (B) Cells expressing WT (ATCC-CCL104) or R721G (30936) were treated with 100 μM of the protein synthesis inhibitor cycloheximide (CHX) for 1, 2 and 8 h. Western blots of the respective cell lysates were obtained and imaged as described in methods and materials. GAPDH was used as the internal loading control.

According to Fig. 5B, the WT Lon/GAPDH ratio increased whereas the R721G Lon/GAPDH ratio decreased with increased CHX treatment time (Fig. 5B). An increase in WT hLon/GAPDH over time indicates WT hLon is more stable than GAPDH whereas a decrease in R721G hLon/GAPDH over time indicates mutant Lon is less stable than GAPDH. At the different CHX treatment times, the PDK4/GAPDH ratio is relatively constant in the WT cell lysates but decreases in the R721G cell lysate. The half-life of PDK4 in HL-1 cells is 1 h [8]. The PDK4/GAPDH ratio in R721G compared to WT hLon is 1.7-fold higher at the 1 h CHX post-treatment time point, 1.8-fold higher at 2 h and become comparable at 8 h. The relatively higher level of PDK4 detected in the 30936 cells at 1 and 2 h post CHX treatment could be attributed to R721G hLon failing to effectively degrade PDK4 in the mitochondria. Upon 8 h of treatment with CHX, which likely induces significant amount of cell stress due to extensive inhibition of protein synthesis, non-specific proteolysis degrades PDK4, which gives rise to the comparable level of PDK4 found in WT versus 30963 cells reported in the bottom panel of Fig. 5B. As discerned in the in vitro kinetic studies shown above, R721G possesses ATP-dependent peptidase activity with a reduced k_cat. It is plausible that it takes up to 8 h for R721G hLon to reduce PDK4 to a basal level in the 30936 cells. Since PDK4 is one of the enzymes essential for maintaining metabolic flexibility, its expression in the mitochondria is closely regulated. In WT Lon cells, the relatively constant levels of PDK4 is likely due to the presence of stable WT Lon activity. In the homozygous R721G cells, the ATP-dependent proteolytic activity of hLon is compromised by the mutation. As such, the level of PDK4 is higher at the 1 and 2-hr post CHX treatment time points, which is consistent with the half-life of PDK4 of 1 h in cells. The observation that PDH protected PDK4 from hLon degradation could account for the residual level of PDK4 detected in the CHX study presented above (Fig. 5B). Even after 8 h of CHX treatment, PDK4 was still detected in the cell lysate of WT as well as R721G because the kinase was protected from proteolysis through binding to PDH. Given the degradation of PDK4 could be monitored as a Lon activity reporter in human B lymphocytes, we propose that these cells can be used for studying the physiological molecular mechanism of hLon actions and serve as a platform for screening small molecules to target metabolic diseases associated with the deregulation of the PDH/PDK4/hLon pathway. Effects of mitochondrial protein-protein interactions such as the protective effect of PDH on PDK4, that are not reflected in purified protein digestion assays would be readily detected in cell lysates as a complementary assay.

3.4. Comparing the structural stability of WT versus R721G hLon in vitro

To further compare the structural stability of WT versus R721G hLon, limited tryptic digestion experiment was done on purified recombinant Lon proteins in the absence, and independently in the presence of MgATP. Fig. 6 shows that MgATP protects WT but not R721G hLon from tryptic digestion, as evidenced by the relative persistence of the intact hLon (>96 kDa) and the 72 KDa hLon fragment (see arrow) detected in the WT (top panel) but not R721G tryptic digestion gel data (bottom panel). In the absence of MgATP, WT and R721G hLon show a comparable tryptic digestion pattern. In the presence of MgATP, WT hLon is digested to a ~72 KDa protein fragment, which is detected along with intact hLon in the SDS-gel from 5 to 60 min into the tryptic digestion time course (top panel). In contrast, the intact R721G protein as well as the 72 KDa protein fragment are barely detectable under the same digestion times (bottom panel). This result indicates that the WT hLon: MgATP complex is structurally more stable and R721G hLon either does not bind MgATP properly, or despite binding to MgATP, R721G fails to undergo the same conformational changes as WT hLon.

Fig. 6.

Limited tryptic digestion indicates that R721G is structurally less stable than WT hLon. (A) Wildtype hLon was digested by trypsin with the presence of 0, 2 mM, or 30 mM ATP. Red arrow indicates intact Lon band and yellow arrow indicates trypsin inhibitor SBTI band. (B) Human Lon mutant R721G was digested by trypsin with the presence of 0, 2 mM, or 30 mM ATP. Red arrow indicates the 72 KDa Lon fragment band and yellow arrow indicates SBTI band.

3.5. Comparison of the structural dynamics R721G hLon to WT hLon by HXMS

While the limited tryptic digestion study showed WT hLon underwent a more compact conformational change than R721G when bound to MgATP, it did not reveal the regions in Lon protein that participated in the conformational change. To identify the regions in WT versus R721G that differ upon binding to MgATP, HXMS was used. To this end, WT or R721G hLon was incubated in deuterated buffer in the presence of MgATP at 4 °C at different times followed by quenching with acid. Supplementary Figure 2 shows that WT and R721G hLon both display ATP-dependent peptidase activity under the time frame used to monitor HXMS. Therefore, the HXMS experiments monitored the functional enzyme form(s). As hLon proteins were incubated with MgATP at 4 °C, where ATP hydrolysis was slow, the predominant enzyme form detected in the HXMS experiment was hLon:MgATP. Mass spectrometry identified 245 pepsin-digested peptides that covered 93% of the hLon sequence. Table 2 summarizes the amino acid sequences within the two Lon proteins that exhibit >70% difference in the Tau value of the HD exchange kinetic time courses; the peptide sequences in the ATPase domain are highlighted in a box. The Tau values were determined from the HD exchange kinetic time courses as described in Methods and Materials. Fig. 7 shows the HD exchange time courses of the peptides encompassing the R721G regions listed in Table 2. Many of these peptides are located at the N-terminal end, upstream from the R721G mutation. Fig. 8 shows the peptide sequences exhibiting different HD exchange kinetics in the cryoEM hLon structure (PDB 7KSL) lacking the first 420 residues in LonP, bound to MgADP [27]. The HXMS experiments monitored hLon proteins bound to MgATP. A high-resolution structure of hLon or bacterial Lon bound to MgATP is not yet available. The mutant R721 is located in a helix containing the sequence ⁷⁰³YCRESGVRNLQKQVEKVL⁷²¹RKSAYKIVSGEAES⁷³⁴. The helical structure in R721G is likely distorted by the Gly mutation, a helix breaker, which weakens its interaction with the juxtapose helix containing ⁶⁷¹ERYLVPQARALCGL⁶⁸⁴, one of the peptides that shows difference in the HD exchange kinetics between WT and R721G hLon. The weakened subunit interaction is propagated to residues 455–462, 496–510, 671–684, 703–734, and 755–767 (Table 2), which line the subunit interface. As ⁶⁷¹ERYLVPQARALCGL⁶⁸⁴ is located in proximity to the helices containing ⁵²⁰CFYGPPGVGKTSIARSIA⁵³⁷ and ⁴⁹⁶DVKKRILEFIAVSQL⁵¹⁰, respectively, it is possible that the interactions of these three regions are indirectly disrupted by the broken helix in R721G. Since Lon undergoes conformational changes upon binding to MgATP to unfold and translocate substrate to the proteolytic chamber, it is possible that the difference in HD exchange profiles observed in the peptide sequence distal from the ATPase site reflect the consequence of a weakened subunit interaction in the mutant, which affect protein or peptide substrate binding and translocation. As hLon degrades proteins by unfolding and translocating the unfolded protein substrate through the tunnel formed by the hexamer, the structural failure in mutant R721G will compromise function, which accounts for the detection of damaged protein aggregates in the mitochondria, leading to mitochondria dysfunction [5].

Table 2.

Peptide sequences exhibiting difference in HD exchange in WT versus R721G hLon.

Fig. 7.

Time courses of HD exchange of WT (blue) versus R721G (red) hLon peptide sequences exhibiting > 70% difference in the Tau values. Deuterium update in atomic mass unit (ΔD) were plotted against log₁₀ (deuterium exposure times).

Fig. 8.

A dimer of the cryoEM structure of LonP1 lacking residues 1 to 420 (PDB 7ksl). Peptide sequences exhibiting different HD exchange kinetics between WT and R721G hLon are colored in blue. R721 was marked with a red arrow. Because 7ksl does not contact the N-terminal of the intact WT hLon, the sequence ³⁹⁹QLKIIKKELGL⁴⁰⁹ listed in Table 2 is not shown in this figure. It is discerned that the peptide sequences flanking residues 455–462, 496–510, 671–684, 703–734, and 755–767 are located at the enzyme subunit interface.

4. Conclusions

This study investigated the structure-functional relationship of the R721G human Lon protease. Using steady-state kinetic techniques, the impact of the mutation on Lon’s ATPase and peptidase activities were quantified. When interpreted in conjunction with the EM and HXMS data, it is concluded that the R721G mutation causes deformation in the enzyme structure at multiple regions, which hinders productive hexamer formation, ATP-binding and peptide translocation to the proteolytic site. To assess the physiological significance of the in vitro findings, we performed complementary analyses using immortalized B lymphocytes expressing only WT (ATCC-CCL104) or R721G (30936) hLon. In agreement with the in vitro findings, R721G is less stable than WT hLon in cells. Furthermore, the endogenous hLon substrate PDK4 is more table in 30936 cells than in ATCC-CCL104 cells during the first 2 h of protein synthesis inhibition, indicating R721G hLon has compromised proteolytic activity. The relatively constant level of PDK4 found in CCL104 and 30936 cells after 8 h CHX treatment can be explained by PDH protects PDK4 from proteolytic degradation. Therefore, Lon, PDK4 and PDH together will serve as an excellent system for studying the dynamic mechanism of Lon-mediated degradation of uncomplexed mitochondrial protein in vitro and in cells.

The integrative approaches applied in this study can be applied to investigate the structural-function relationship of other naturally occurring hLon mutants, which advance the physiological enzymology of Lon proteases in general. The HXMS and negative stain EM studies can be conducted under the conditions where the mutant proteins are active, as judged by the ATP-dependent peptidase and ATPase assays. The degradation of PDK4 by mutant hLons and relative stability of the mutant proteins in cells can be monitored in B lymphocytes immortalized from the blood of the donors harboring the mutant genes. Since immortalized B lymphocytes can be readily prepared from human blood, the strategy of combining cell-based experiments with in vitro characterization as described in this study could be used to investigate the structure-function relationship of other mutations in human mitochondrial Lon, thereby advancing the physiological enzymology of this protease.

Abbreviations

FRDA

Friedreich ataxia

CODAS

cerebral, ocular, dental, auricular, skeletal anomalies syndrome

Lon

protease La

WT

wildtype

hLon

human Lon protease

eLon

E. coli Lon protease

MTS

mitochondria targeting sequence

PDK1

pyruvate dehydrogenase kinase isoform 1

PDK4

pyruvate dehydrogenase kinase isoform 4

PDH

pyruvate dehydrogenase complex

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

ATP

adenosine triphosphate

CHX

protein synthesis inhibitor cycloheximide

SBTI

soybean trypsin inhibitor

SDS-PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

BSA

bovine serum albumin

TBS

tris-buffered saline

BL21*(DE3)

Chemically Competent E. coli BL21 Star (DE3)

OD₆₀₀

optical density of a sample measured at a wavelength of 600 nm

SB

super broth

IPTG

Isopropyl β-d-1-thiogalactopyranoside

Tris

tris(hydroxymethyl) aminomethane

NaCl

sodium chloride

EDTA

ethylenediaminetetraacetic acid

BME

beta-mercaptoethanol

KPi

potassium phosphate

HEPES

(4-(2-hydroxyethyl)-1 piperazineethanesulfonic acid)

MOPS

3-(N-morpholino) propanesulfonic acid

[gamma-³²P] ATP

adenosine triphosphate, labeled on the gamma phosphate group with ³²P

TLC

thin-layer chromatography

FR

fluorescence resonance energy transfer

FR89-98

Y(NO₂)RGITCSGRQK(Abz)

Abz

anthranilamide

89-98Bz

YRGITCSGRQK(Bz)

Bz

benzoic acid amide

89–98

YRGITCSGRQK

k _obs

rate/[enzyme]

k _cat

v_max/[enzyme]

K _m

Michaelis constant

K_i

Inhibition constant

EM

electron microscopy

HXMS

hydrogen/deuterium exchange-mass spectrometry

RP-UPLC

reverse-phase ultraperformance liquid chromatography

MS^E

mass spectrometry elevated energy

RPMI 1640

Roswell Park Memorial Institute 1640 Medium

PBS

phosphate-buffered saline