A Single Amino Acid Change (Substitution of Glutamate 3 with Alanine) in the N-terminal Region of Rat Liver Carnitine Palmitoyltransferase I Abolishes Malonyl-CoA Inhibition and High Affinity Binding*

We have recently shown by deletion mutation analysis that the conserved first 18 N-terminal amino acid residues of rat liver carnitine palmitoyltransferase I (L-CPTI) are essential for malonyl-CoA inhibition and binding (Shi, J., Zhu, H., Arvidson, D. N., Cregg, J. M., and Woldegiorgis, G. (1998)Biochemistry 37, 11033–11038). To identify specific residue(s) involved in malonyl-CoA binding and inhibition of L-CPTI, we constructed two more deletion mutants, Δ12 and Δ6, and three substitution mutations within the conserved first six amino acid residues. Mutant L-CPTI, lacking either the first six N-terminal amino acid residues or with a change of glutamic acid 3 to alanine, was expressed at steady-state levels similar to wild type and had near wild type catalytic activity. However, malonyl-CoA inhibition of these mutant enzymes was reduced 100-fold, and high affinity malonyl-CoA binding was lost. A mutant L-CPTI with a change of histidine 5 to alanine caused only partial loss of malonyl-CoA inhibition, whereas a mutant L-CPTI with a change of glutamine 6 to alanine had wild type properties. These results demonstrate that glutamic acid 3 and histidine 5 are necessary for malonyl-CoA binding and inhibition of L-CPTI by malonyl-CoA but are not required for catalysis.

lar mechanism of the regulation of CPTI by malonyl-CoA is important in the design of drugs for control of excessive fatty acid oxidation in diabetes mellitus (10) and in myocardial ischemia where accumulation of acylcarnitines has been associated with arrhythmias (11).
We developed a novel high level expression system for rat L-CPTI and human heart M-CPTI in the yeast Pichia pastoris, an organism devoid of endogenous CPT activity (6,12,13).
Using this system, we demonstrated conclusively that L-CPTI and M-CPTI are active, distinct, malonyl-CoA-sensitive CPTIs that are reversibly inactivated by detergents. We recently showed that deletion of the conserved first 18 N-terminal amino acid residues of rat L-CPTI abolishes malonyl-CoA inhibition and high affinity malonyl-CoA binding (14). In this study, we have constructed and characterized rat L-CPTI deletion mutants of the first 12 and 6 N-terminal amino acid residues. To identify specific residue(s) involved in malonyl-CoA binding and inhibition of L-CPTI, we also constructed three substitution mutations within the conserved first 6 Nterminal amino acid residues (Glu 3 3 Ala, His 5 3 Ala, and Gln 6 3 Ala).

Construction of Plasmids for the N-terminal Point and Deletion Mutants of Rat L-CPTI-
The cDNA used and the construction of the plasmids for the wild type, deletion, and point mutants for rat L-CPTI expression in P. pastoris was as described in our previous publications (12,14). A HindIII-KpnI fragment (1-566 base pair; cDNA sequence; Refs. 12 and 15) was excised from pYGW9, a plasmid containing the full-length rat L-CPTI in pUC119 to generate the plasmid pYGW12. The ⌬12 mutant was constructed by polymerase chain reaction amplification of a 662-base pair HindIII-EcoRI fragment using the plasmid pYGW9 as a template with the primers RL665 (5Ј-CCACCAGGATTT-TAGCT-3Ј) and RLD12 (5Ј-CAAGCTAAGCTTGAATTCATGACTGT-CACCCCCGATGGCAT-3Ј). HindIII and EcoRI enzyme restriction sites were introduced in primer RLD12. An ATG start codon (shown in bold type) was added immediately after the EcoRI site and before the fourth amino acid alanine. The polymerase chain reaction product was digested with HindIII and KpnI and then ligated into pYGW12 to generate plasmid pYGWD12. An EcoRI fragment of pYGWD12 containing the mutant ⌬12 was then ligated into the EcoRI-cut P. pastoris expression vector pHW010 (12,16). The DNA sequences of the deletion and point mutants were confirmed by sequencing.
The expression plasmids were linearized and integrated into the HIS4 locus of P. pastoris strain GS115 by electrotransformation (17). Histidine prototrophic transformants were selected on YND plates and grown on YND medium. Mitochondria were isolated from the wild type and mutant L-CPTIs as described previously (12).
CPT Assay-CPT activity was assayed by the forward exchange method using L-[ 3 H]carnitine as described previously (12,18). The K m for palmitoyl-CoA was determined by varying the palmitoyl-CoA concentration in the presence of a fixed albumin concentration (1%) or a fixed molar ratio (6.1:1) of palmitoyl-CoA/albumin (19,20). The K i for malonyl-CoA inhibition of the yeast-expressed wild type and mutant (E3A) L-CPTIs was determined by assaying CPT activity with varying concentrations of palmitoyl-CoA (12.5, 25, 50, 75, and 100 M) in the presence of 0, 1, 2, and 5 M malonyl-CoA (wild type), or 0, 2, 25, and 200 M malonyl-CoA (E3A), respectively. The concentration of the second substrate, carnitine, was fixed at 200 M in all the assays. 106 g (wild type) and 118 g (E3A) of mitochondrial protein were used, and all incubations were performed at 30°C for 3 min.
[ 14 C]Malonyl-CoA Binding Assay-[ 14 C]Malonyl-CoA binding was determined by a modified centrifugation assay as described previously (14,21). Isolated mitochondria from wild type and mutants were suspended in 0.5 ml of ice-cold medium composed of 72 mM sorbitol, 60 mM KCl, 25 mM Tris/HCl (pH 6.8), 1.0 mM EDTA, 1.0 mM dithiothreitol, and 1.3 mg/ml fatty acid-free bovine serum albumin (22). This was followed by addition of 0.1-1000 nM [2-14 C]malonyl-CoA, and the suspension was incubated at 4°C for 30 min with periodic vortexing. All subsequent procedures were as described previously (14). The CPT activity and C 50 values are given as the means Ϯ S.D. for at least three independent assays with different preparations of mitochondria. The K D values are averages of at least two independent experiments.
Western Blot Analysis-Proteins were separated by SDS-polyacrylamide gel electrophoresis in a 7.5% gel and transferred onto nitrocellulose membranes. Immunoblots were developed by incubation with the CPTI-specific polyclonal antibody (1:4000 dilution) followed by an antirabbit IgG conjugated to horseradish peroxidase (1:10,000 dilution) as described previously (12). The antigen-antibody complex was detected using an ECL-enhanced chemiluminescence detection kit (Amersham Pharmacia Biotech).
Other Materials and Procedures-DNA sequencing was performed at the Oregon Regional Primate Research Center core facility using an automatic DNA sequencer (23). Protein was determined by the Lowry procedure (24). All restriction enzymes were from New England Biolabs (Beverly, MA). L-[ 3 H]Carnitine and [2-14 C]malonyl-coenzyme A were from Amersham Pharmacia Biotech. Nucleotides were from Amersham Pharmacia Biotech, palmitoyl-CoA was from Boehringer Mannheim, and malonyl-CoA was from Sigma.

Generation of Deletion and Point Mutants and Expression in
P. pastoris-Construction of plasmids carrying the N-terminal deletions ⌬6 and ⌬12 and point mutants Glu 3 3 Ala, His 5 3 Ala, and Gln 6 3 Ala of rat L-CPTI ( Fig. 1) was performed as described under "Experimental Procedures." Mutations were confirmed by DNA sequencing. P. pastoris was chosen as an expression system for L-CPTI and the mutants, because it does not have endogenous CPT activity (6,(12)(13)(14). The P. pastoris expression plasmids expressed L-CPTI under control of the P. pastoris glyceraldehyde-3-phosphate dehydrogenase gene promoter (12,16). Yeast transformants with the wild type L-CPTI gene and the mutants were grown in liquid medium supplemented with glucose. As previously reported (12), no CPT activity was found in the control yeast strain with the vector but without the CPTI cDNA insert.
Western blot analysis of wild type L-CPTI (88 kDa) and the mutants using a polyclonal antibody directed against a maltose-binding protein-L-CPTI fusion protein (12) is shown in Fig. 2. For the wild type and all the deletion and point mutants (⌬18, ⌬12, ⌬6, Glu 3 3 Ala, His 5 3 Ala, and Gln 6 3 Ala), proteins of predicted sizes were synthesized and were expressed at similar steady-state levels.
Effect of Mutations on L-CPTI Activity and Malonyl-CoA Inhibition-All of the mutants retained significant CPT activity, which was 60 -80% of that observed with the wild type yeast strain expressing L-CPTI (Table I). The IC 50 for malonyl-CoA inhibition of the wild type strain expressing L-CPTI was 2.0 M, in agreement with our previous report (12), whereas the IC 50 for ⌬6, ⌬12, and the point mutant Glu 3 3 Ala was 200 M, representing a 100-fold decrease in malonyl-CoA sensitivity compared with the wild type strain (Table I). The K i for malonyl-CoA inhibition of the Glu 3 3 Ala mutant L-CPTI was approximately 10-fold higher than that of the wild type (32 M versus 2.7 M), a trend similar to the increase in IC 50 for malonyl-CoA inhibition observed in the Glu 3 3 Ala mutant. Mutation of histidine residue 5 to alanine increased the IC 50 for malonyl-CoA inhibition to 25 M compared with the 2.0 M value of the wild type strain, representing a mild 12.5-fold decrease in malonyl-CoA sensitivity, whereas mutation of glutamine residue 6 to alanine had no effect on malonyl-CoA sensitivity (Table I). ⌬6 and point mutant Glu 3 3 Ala showed decreased malonyl-CoA sensitivity at all levels of the inhibitor tested compared with the wild type (Fig. 3), whereas only minor reduction in malonyl-CoA sensitivity was observed in the point mutant His 5 3 Ala compared with the control.
Kinetic Properties of Wild type and Mutant L-CPTIs-The Glu 3 3 Ala mutant exhibited normal saturation kinetics when the carnitine concentration was varied relative to a constant second substrate, palmitoyl-CoA, a property identical to that of wild type L-CPTI (Fig. 4A) and similar to that previously reported from our laboratory (12). For the Glu 3 3 Ala mutant, the calculated K m for carnitine was 74.5 M and the V max was 8.3 nmol/min/mg protein, which is similar to the 45 M and 12.6 nmol/min/mg for wild type L-CPTI (12). With respect to the second substrate, palmitoyl-CoA, the wild type, but not the Glu 3 3 Ala mutant, showed non-Michaelis-Menten saturation kinetics at a fixed concentration of albumin (1% W/V) (Fig. 4B), characteristics similar to our previous report (12). However, both the wild type and the Glu 3 3 Ala mutant exhibited normal saturation kinetics when the molar ratio of palmitoyl-CoA: albumin was fixed at 6.1:1 (Fig. 4C). The calculated K m for palmitoyl-CoA for the point mutant Glu 3 3 Ala was 69.4 M and the V max was 23.3 nmol/min/mg, which is similar to the wild type values of 104 M and 41.5 nmol/min/mg. Thus, deletion of the first 6 N-terminal amino acid residues or a substitution mutation of Glu residue 3 to Ala in L-CPTI abolishes malonyl-CoA sensitivity, but not catalytic activity. Mutation of glutamine residue 6 to alanine had no effect on malonyl-CoA inhibition or catalytic activity.   5 3 Ala was significantly lower compared with that observed in the mitochondria from the wild type strain but was saturable (Fig. 5). Malonyl-CoA binding clearly resolved into a high affinity and a low affinity site in the mitochondria from the wild type and point mutant His 5 3 Ala as shown by the Scatchard plots in Fig. 6 (A and B), but only very low affinity binding was observed in the mitochondria from point mutant Glu 3 3 Ala and ⌬6 (Fig. 6B). Deletion of the first 6 N-terminal residues or substitution of glutamic acid 3 with alanine completely abolished high affinity malonyl-CoA binding (K D1 ) and further decreased the low affinity malonyl-CoA binding (K D2 ) by 100-fold (Table I). Furthermore, the loss of high affinity malonyl-CoA binding observed with the Glu 3 3 Ala mutant correlates with the increase in IC 50 and K i for inhibition of L-CPTI by malonyl-CoA. Although the mutations increased the K D2 for the low affinity malonyl-CoA-binding site (⌬6 and Glu 3 3 Ala), there was no change in the calculated B max2 (which was 10-fold higher than the B max1 of the wild type), suggesting that the observed loss in malonyl-CoA sensitivity and binding is not due to decreased abundance of the second malonyl-CoA-binding entity of L-CPTI.
Substitution of histidine 5 with alanine showed a moderate loss in malonyl-CoA sensitivity of L-CPTI (12.5-fold), resulting in loss of the high affinity and a significant decrease in low affinity malonyl-CoA binding ( Table I). The low affinity binding site in mutant His 5 3 Ala resolved into two classes of binding sites (Fig. 6B), with the K D1 and K D2 for the mutant being 100and 16-fold higher, respectively, than the corresponding wild type values (Table I). The calculated B max1 for mutant His 5 3 Ala was close to the wild type value, but B max2 was 3-fold lower, indicating a decrease in the malonyl-CoA-binding entity of L-CPTI. These studies suggest that histidine 5 in L-CPTI is one of the residues involved in malonyl-CoA binding and inhibition of the enzyme. DISCUSSION To determine the role of the first 130 N-terminal amino acid residues of rat L-CPTI in malonyl-CoA sensitivity and binding, we previously constructed a series of deletion mutants and demonstrated that a mutant lacking the first conserved 18 N-terminal amino acid residues had activity and kinetic properties similar to those of wild type L-CPTI but had completely lost malonyl-CoA sensitivity and high affinity binding (14). Based on these previous studies (14), we report here on deletion mutations of the conserved first 12 and 6 N-terminal residues of L-CPTI. Like ⌬18, ⌬12 and ⌬6 had 60 -70% of the wild type activity and showed loss of both malonyl-CoA sensitivity and high affinity malonyl-CoA binding, indicating that residue(s) essential for malonyl-CoA binding and sensitivity reside within the conserved first 6 N-terminal amino acids. Of these conserved first 6 N-terminal amino acids, including the start codon Met, residues 2 and 4 are Ala, residue 3 is Glu, residue 5 is His, and residue 6 is Gln. Therefore, we constructed mutants with substitutions of Glu 3 with Ala, His 5 with Ala, and Gln 6 with Ala of L-CPTI.
The mutant L-CPTI with a replacement of Glu 3 with Ala had a phenotype similar to that of the N-terminal deletion mutants. The mutation resulted in complete loss of malonyl-CoA sensitivity and high affinity malonyl-CoA binding and a decrease in the low affinity malonyl-CoA binding. In contrast, substitution of Glu 3 with Ala did not have a significant effect on the kinetic properties of the enzyme, because there was no change in the K m value for palmitoyl-CoA and only a slight increase in the K m value for carnitine. The 29 -40% loss in catalytic activity observed with the deletion and point mutants compared with the wild type could be due to a reduction in the expression level or lack of interaction of the N-terminal domain with the catalytic domain as a result of the N-terminal mutations. A protein of the expected size (88 kDa) was detected in the mitochondria of the Glu 3 3 Ala mutant strain on immunoblotting with L-CPTI specific antibodies. These results demonstrate clearly that Glu 3 in the wild type L-CPTI is essential for malonyl-CoA inhibition and binding but not for catalysis, because the kinetic properties of the mutant enzyme are virtually indistinguishable from those of the wild type. This is the first report to demonstrate TABLE I Activity, malonyl-CoA sensitivity, and malonyl-CoA binding in yeast strains expressing wild type L-CPTI, N-terminal deletion, and point mutants Mitochondria were isolated from the yeast strains separately expressing L-CPTI, and the deletion and point mutants were assayed for CPT activity, malonyl-CoA sensitivity, and binding, as described under "Experimental Procedures." The IC 50 is the concentration of malonyl-CoA needed to inhibit 50% of the activity of the yeast-expressed L-CPTI, and the results are the means Ϯ S.D. of at least three independent experiments with different mitochondrial preparations.  the critical role of Glu 3 residue of L-CPTI for malonyl-CoA sensitivity and binding.
The high affinity site (K D1 , B max1 ) for binding of malonyl-CoA to L-CPTI was completely abolished in the Glu 3 3 Ala, ⌬6, and ⌬18 mutants, suggesting that the Ͼ100-fold decrease in malonyl-CoA sensitivity observed in these mutants was due to the loss of the high affinity binding entity of the enzyme. Although low affinity malonyl-CoA binding was weakened, there was no change in the B max2 value between wild type L-CPTI and mutants Glu 3 3 Ala, ⌬6, and ⌬18, suggesting that the residual malonyl-CoA sensitivity observed in the mutants was due to the low affinity malonyl-CoA-binding entity of the enzyme. The results of this study provide strong evidence implicating Glu 3 as one of the residues involved in high affinity malonyl-CoA binding. We hypothesize that the Glu 3 3 Ala substitution may disrupt a hydrogen bonding network or a salt bridge, perhaps to a residue near the active site of CPTI. As the high affinity site is abolished and binding to the low affinity site is weakened, the two sites may partially overlap. Alternatively, the possible loss of a salt bridge may weaken K D2 indirectly.
Replacement of His 5 with Ala had a much less drastic effect on the IC 50 for malonyl-CoA inhibition of L-CPTI but severely diminished both high and low affinity malonyl-CoA binding. The B max1 for this mutant showed a slight increase, but B max2 showed a significant decrease compared with the wild type value, suggesting that the 10-fold lower IC 50 for malonyl-CoA inhibition observed with this mutant, compared with mutants Glu 3 3 Ala, ⌬6, and ⌬18, may be due to a slight increase in abundance of the high affinity binding entity with a lowered (100-fold) affinity for malonyl-CoA. The decrease in low affinity malonyl-CoA binding observed for the His 5 3 Ala mutant (ϳ15-fold increase in K D2 ) may be due, in part, to the decreased abundance of the low affinity binding entity of the enzyme (ϳ3-fold decrease in B max2 ). Because mutation of His 5 3 Ala reduced the malonyl-CoA sensitivity and binding, L-CPTI may be affected by pH. A pH-induced shift in malonyl-CoA sensitivity has been reported for CPTI (25,26).
Our data clearly demonstrate that there are two classes of malonyl-CoA-binding sites in L-CPTI, namely, a high affinity and a low affinity binding site, similar to earlier studies in isolated rat liver and heart mitochondria (26,27). A previous attempt to express a mutant L-CPTI that lacked the first 82 N-terminal residues was described by Brown et al. (28), but results were inconclusive due to extremely low expression levels (12). The residual malonyl-CoA sensitivity shown by the deletion mutants is similar to that observed with yeast-expressed CPTII (12), suggesting that for these mutants malonyl-CoA may inhibit via direct interaction with the active site. Additional studies are needed to determine whether the active site acts as a low affinity malonyl-CoA-binding site, but our data suggest that there may be some overlap between the malonyl-CoA and palmitoyl-CoA binding sites. In the absence of malonyl-CoA, free CoA (50 M) and acetyl-CoA (500 M) inhibited the activities of both the wild type and the Glu 3 3 Ala mutant L-CPTI by 50%. 2 Because a total loss of the high affinity malonyl-CoA binding site was observed in the Glu 3 3 Ala mutant, the results suggest that CoA and acetyl-CoA inhibit by binding to the active site or the low affinity malonyl-CoA-binding site. At high concentrations, both CoA and the substrate palmitoyl-CoA reduce the inhibition of L-CPTI by malonyl-CoA (18,29), suggesting partial overlap between the malonyl-CoA and the substrate binding sites.
Based on limited proteolysis studies of intact and outer membrane rat liver mitochondria, a model for the membrane topology of L-CPTI has been proposed that predicts exposure of 90% of L-CPTI, including N and C termini domains crucial for activity and malonyl-CoA sensitivity of the enzyme on the cytosolic side of the outer mitochondrial membrane (30). A more recent detailed deletion mutation analysis study of the 129 N-terminal amino acid residues of the yeast-expressed L-CPTI from our laboratory clearly demonstrated that residues critical for malonyl-CoA inhibition and binding of L-CPTI are located within the conserved first 18 N-terminal amino acid residues of the enzyme (14). In this study, we demonstrate that glutamic acid residue 3 and histidine 5 are essential for malo-nyl-CoA binding and inhibition.
Limited proteolysis of intact and outer membrane preparations of rat liver mitochondria result in a marked loss in L-CPTI activity and malonyl-CoA sensitivity (30,31), accompanied by the cleavage of the extreme N terminus (Ͻ1 kDa) of L-CPTI (30). Mitochondria isolated from fasted and diabetic rat livers, metabolic conditions with increased fatty acid oxidation, exhibit increased L-CPTI activity and decreased malonyl-CoA sensitivity (32). Furthermore, insulin reverses the effects of diabetes on L-CPTI activity and malonyl-CoA sensitivity (32). Thus, fasting and diabetes, metabolic conditions that enhance protein degradation, reduce the sensitivity of CPTI to malonyl-CoA inhibition (32)(33)(34). The L-CPTI gene in INS-1 cells may be an early response gene like c-fos (35), suggesting that the enzyme may be subject to metabolic regulation by proteolysis (36), employing the cytosolic ubiquitin-proteasome system (36). Diabetes is a pathophysiologic condition associated with increased protein degradation, fatty acid oxidation, CPTI activity, and decreased malonyl-CoA inhibition of CPTI (33). Thus limited in vivo proteolysis of L-CPTI induced by diabetes, such as cleavage of the first 6 N-terminal residues, may decrease malonyl-CoA sensitivity and alter the normal control of hepatic fatty acid oxidation. Insulin inhibits proteasome activity resulting in decreased cellular protein degradation and con- trolled fatty acid oxidation (8). We are currently conducting in vitro partial proteolysis studies with yeast-expressed L-CPTI to determine the role of protein degradation on malonyl-CoA sensitivity.