Main

With the exception of β-ketothiolase deficiency (McKusick 203750), there are no reported human genetic abnormalities specific to L-isoleucine degradation (1). During routine metabolic work-up of a 4-mo-old male infant who presented with mild hypoglycemia, lethargy, and apnea, we detected increased plasma short-chain acylcarnitine (AC) using standard methods (2). Tandem mass spectrometric analysis of plasma AC revealed an elevated five-carbon species, representing the sum of isovaleryl- and/or 2-methylbutyrylcarnitine, isomers which are not differentiated by this methodology (3–5). Separation of isomers using a different method subsequently confirmed 2-MBC as the predominant species. Increased quantities of 2-MBG were routinely detected in the proband's urine, without other diagnostic abnormalities. These results suggested an isolated defect of 2-MBCDase (Fig. 1). Subsequent metabolic, enzymatic, immunoblot, molecular and prenatal diagnostic investigations verified 2-MBCDase deficiency.

Figure 1
figure 1

Catabolic pathway for L-isoleucine metabolism. The site of the defect in the proband is represented by the cross-hatched box. Numbered reactions include: 1, branched chain 2-oxo acid dehydrogenase; 2, 2-methylbutyryl-CoA dehydrogenase; and 3, β-ketothiolase.

Full size image

The male proband was delivered at 39 wk gestation to a gravida 1, para 1 mother of Northern European descent. The father was Eritrean. There was no significant family history. Birth and delivery were uneventful. He was alert at 20 h, with weak suck and decreased tone. Discharge at 24 h revealed blood glucose at 50 mg/dL (2.8 mmol/L) (nl 55–100 mg/dL; 3.1–5.6 mmol/L). The patient was hospitalized at day 3 of life for poor feeding, lethargy, and hypothermia. Mother reported gagging without emesis in the preceding 24 h. At admission, weight was 2.42 kg, temperature 36.7°C, heart rate 206, blood glucose 10 mg/dL (0.6 mmol/L), and venous pH 7.31 (anion gap −5.0). Administration of 20 mL 10% dextrose improved glucose to 64 mg/dL (3.6 mmol/L); within 2 h, however, the value had dropped to 18 mg/dL (1 mmol/L), and he was continued on 10% dextrose at 18 mL/h to maintain blood glucose between 70–100 mg/dL (3.9–5.6 mmol/L). He was transferred to intensive care following two episodes of apnea, where he was described as hypothermic, tachycardic, lethargic, and significantly dehydrated with cool extremities and barely palpable peripheral pulse. EEG revealed mild attenuation of background activity and intermittent frequent bilateral sharp transients in the central and temporal regions, suggesting diffuse cerebral disturbance. MRI revealed altered signal intensities in the gray and white matter of the cerebral hemispheres, with some cortical enhancement in the post Gadolinium series mainly involving the parietal and occipital lobes bilaterally. Foci of slightly increased T1 signal intensity were observed in the lentiform nuclei bilaterally. These findings were consistent with global hypoxia. A full description of this patient will be presented elsewhere. Current therapeutic regimen includes protein restriction (1.4 g/kg/day) and L-carnitine supplementation (600 mg/day at 71 mg/kg, 3 times a day). All investigations were performed with informed consent and approval by the applicable Institutional Review Board.

METHODS

Carnitine in plasma and 2-MBG/2-MBC in urine were determined using established methods (3, 4, 6). Quantitation of 2-MBG in urine and amniotic fluid by stable-isotope dilution used D3-isovalerylglycine as internal standard with gas chromatography-mass spectrometric quantitation (7). The latter assumed roughly equal extraction efficiency for the different glycine conjugates. Quantitation of C5 and C3 AC in amniotic fluid were determined by MSMS (3–5). Oxidation of U-14C-isoleucine to 14CO2 in intact fibroblasts was performed as described (8). Intact cell oxidation of 13C6-leucine, 13C5-valine, and 13C6-isoleucine in fibroblasts and amniocytes, incubated in the presence of 0.4 mmol/L L-carnitine, followed established methods with determination of AC products by MSMS (9). Western blot analysis of 2-MBCDase in extracts of fibroblasts was performed as described (10).

Characterization of 2-MBCDase cDNA and genomic sequences was achieved using established molecular biologic methods (numbering refers to the published sequence, accession NM 001609.1). For reverse-transcription/polymerase chain reaction (RT-PCR) of cDNA, the following primers were used: nucleotides (nt) 90–113, 5′-TTGGAAGATTCCTCCTCATGTCTC-3′ forward primer 1; nt 844–822, 5′-CCAAGATATTGGCTTCTGGAACC-3′ back primer 1, producing a 755 nt product;nt 670–693, 5′-GTGATGGCAAATGTAGACCCTACC-3′ forward primer 2; nt 1744–1723, 5′-AATGAACCAGGCATGGTGGTGC-3′ back primer 2, producing a 1075 nt product. The total region amplified and sequenced spanned nt 90 to 1744; the first 90 bp could not be readily amplified using standard PCR methodology. For genomic analysis of the mutation found by RT-PCR, the following primers were used: 5′-TCTTAGTAGATCGTGATACTCCG-3′, nt 741–719 forward primer 3 and 5′-GGCAGGTGGAAGAAGCTCTG-3′, nt 799–780 back primer 3, producing an 80 bp fragment (apparently all within one exon); sequence analysis revealed a product of the expected sequence. The mutation detected in the proband was directly next to back primer 3, deleting a Dde I restriction site. Restriction digestion of the 80 bp amplified genomic fragment with Dde I produced a 22-bp product. In the region of the restriction site, wild-type sequence is GGGGcTCAGAGCT, while the same region in the patient was GGGGtTCAGAGCT (Dde I recognizes the sequence CTNAG, in which N represents any base; CAGAGCT is complimentary to the end of back primer 3). Expression of mutant and wild-type 2-MBCDase was performed in an E. coli system as previously described (11). Extracts from cells containing wild-type and mutant 2-MBCDase plasmids were tested for enzyme activity using the electron transferring flavoprotein assay previously described (11).

RESULTS

Short-chain AC in proband plasma were increased (range 8–22 μM, n = 7; nl, 3–10); free carnitine and the AC/free carnitine ratio were normal. MSMS analysis revealed elevated plasma C5-AC (1.4–2.4 μM; nl < 0.4), representing the sum of 2-methylbutyrylcarnitine and isovalerylcarnitine. Qualitative acyloxylactone analysis of urine AC by gas chromatography-mass spectrometry verified the presence of 2-MBC. Urine 2-MBG was consistently elevated (range 8–30 mg/g creatinine, n = 6; nl undetectable using routine methods;Fig. 2). Protein restriction and L-carnitine supplementation led to normalization of 2-MBG excretion in urine; urine 2-MBC was not routinely determined. Plasma amino acids were repeatedly normal.

Figure 2
figure 2

Ion chromatogram of C5-acylglycine species in urine as analyzed by stable-isotope dilution gas chromatography-mass spectrometry. (A) control; (B) proband. Both specimens were equivalent to 0.50 mg creatinine. Peak legend: 1) 2-methylbutyrylglycine (2-MBG); 2) D3-isovalerylglycine (internal standard, 35 nmoles); 3) isovalerylglycine. Monitored ions were m/z 158 for 2-MBG and m/z 161 for D3-isovalerylglycine.

Full size image

In a subsequent pregnancy, the concentration of 2-MBG in amniotic fluid obtained at 15 wk gestation was 0.27 μmol/L (nl < 0.03, n = 4). C5-AC concentration determined by MSMS was significantly elevated at 1.93 μmol/L (mean control 0.37 ± 0.18 (SD), n = 27), and the ratio of C5:C3 AC (C3 = propionylcarnitine, normal in at-risk amniotic fluid) was elevated at 1.68 (mean control 0.43 ± 0.18 (SD), n = 27). During the second pregnancy, the mother was prescribed L-carnitine, and 2-MBC was detected in her urine using qualitative acyloxylactone analysis. Urine before L-carnitine supplementation was not available. The second pregnancy resulted in birth of a healthy female. Urine 2-MBG was increased in two urine samples (4–8 mg/g creatinine; nl undetected), and plasma C5-AC (again representing the sum of 2-methylbutyrylcarnitine and isovalerylcarnitine) was elevated at 1.8 μM (nl < 0.4).

Intact fibroblast conversion of U-14C-isoleucine to 14CO2 in cultured fibroblasts was: control (n = 7) 1,250 ± 427 dpm 14CO2/hr/mg protein, range 465–1,852, mean of duplicate determinations for each cell line); proband 322 (26% of control, > 2 SD below control mean); branched chain 2-oxo acid dehydrogenase deficiency, 11 (1% of mean control); β-ketothiolase deficient fibroblasts, 407 (33% of mean control) (see Fig. 1 for location of enzymes). The conversion of U-[13C]-branched chain amino acids to AC intermediates, in the presence of 0.4 mmol/L L-carnitine, was investigated in fibroblasts derived from proband and controls. The conversion of 13C6-leucine and 13C5-valine in proband fibroblasts revealed no abnormal accumulation of AC; conversely, there was a 10-fold accumulation of C5-AC (control 2.9 nmol 13C5 AC/mg protein/72 h incubation; patient, 29.8) when incubated with 13C6-isoleucine and L-carnitine. Comparable data were obtained using cultured amniocytes derived from the subsequent pregnancy. Oxidation of 13C6-leucine and 13C5-valine in intact at-risk amniocytes was normal; conversely, at-risk amniocytes accumulated 13C5-AC when incubated in the presence of 13C6-isoleucine (control, 2.4 and 6.9 (n = 2) nmol 13C5-AC/mg protein/96 h; at-risk amniocytes, 22.5). Western-blot analysis of proband fibroblasts, using antisera raised in rabbit against purified 2-MBCDase, revealed absence of 2-MBCDase cross-reactive material (11) (Fig. 3).

Figure 3
figure 3

Western blot analysis of crude fibroblast extracts from control and patient for the presence of 2-MBCDase cross-reactive material (lower panel; SBCAD, for short branched-chain acyl-CoA dehydrogenase; the upper panel is a control blot, isovaleryl-CoA dehydrogenase (IVD)). In both panels, the left lane shows purified enzymes (ACD, acyl-CoA dehydrogenase).

Full size image

Screening of 2-MBCDase sequences amplified from cDNA synthesized from proband fibroblasts revealed a 778 C>T substitution in the coding region, leading to substitution of phenylalanine for leucine at amino acid 222 (L222F); the patient's mother also carried this mutation; analysis of 33 control genomic DNA samples (66 chromosomes) revealed a C at position 778. The patient appeared homozygous for this allele in RT-PCR cDNA analyses. However, he (and his mother) were clearly heterozygous in genomic PCR and PCR/restriction analyses. Lack of availability of the genomic structure of the 2-MBCDase has hampered our search for the paternal mutation, but these studies are in progress. The L222F protein is inactive when produced in an E. coli expression system, while wild-type enzyme activity is readily detected (61.9 ± 1.9 mU/mg protein using 2-methylbutyryl-CoA as substrate) (11). The L222F protein was still inactive upon addition of 50–100 times more protein to the assay. Immunoblotting of bacterial extracts showed that the mutant protein was efficiently expressed in this system, but was less soluble than the wild-type enzyme (data not shown).

DISCUSSION

Our results provide evidence for a defect in 2-MBCDase activity in our patient, a new inborn error of L-isoleucine metabolism. In contrast to patients with isovaleryl-CoA dehydrogenase deficiency, who manifest a defect at the corresponding step of L-leucine degradation, our patient did not demonstrate overwhelming metabolic decompensation nor ketosis, which may often accompany this disorder. One puzzling feature of our patient was the low levels of 2-MBG excreted, which contrasts with isovaleryl-CoA dehydrogenase deficient patients who may excrete hundreds of mg of isovalerylglycine per g of creatinine. Perhaps 2-methylbutyryl-CoA is a poor substrate for glycine conjugation, or much of the accumulated 2-methylbutyryl-CoA is hydrolyzed in lieu of conjugation. Our proband was also clinically and metabolically distinct from a patient with isolated isobutyryl-CoA dehydrogenase deficiency, which is the corresponding enzyme defect on the L-valine catabolic pathway. That patient manifested anemia, dilated cardiomyopathy and L-carnitine insufficiency (9). The clinical picture in our patient was that of neurologic disease, as the proband (currently 12 mo old) carries the diagnosis of athetoid cerebral palsy, and continues to manifest impaired visual, motor and cognitive skills but without persistent hypoglycemia. Brain lesions may have been due to early neonatal hypoglycemia or may be unrelated to the underlying enzyme deficiency.

Direct enzyme determination of 2-MBCDase in human fibroblasts is hampered by low levels of activity, and significant overlap with short-chain acyl CoA dehydrogenase (11). More optimal enzyme studies can be performed in biopsied liver or muscle, but this was not ethically possible in our proband. In lieu of such analyses, production of 2-MBCDase carrying the L222F mutation in bacteria verified that this alteration resulted in complete inactivation of enzyme activity. The identification of an affected female fetus in the subsequent pregnancy, who manifests the same metabolite pattern as her brother, in addition to healthy parents suggests that inheritance is autosomal recessive. We are currently assessing levels of 2-MBC in maternal urine (still on L-carnitine intervention) to determine whether this metabolite derived from the female fetus. Definitive demonstration of the autosomal recessive nature of the disease will require assay of 2-MBCDase using authentic CoA substrate in cells derived from the affected siblings and their parents. The presentation of patients with isolated deficiencies of 2-MBCDase, isovaleryl-CoA dehydrogenase and isobutyryl-CoA dehydrogenase indicates that separate enzymes are involved in L-leucine, L-valine and L-isoleucine metabolism in humans.