Deficient DNA-ligase activity in the metabolic disease tyrosinemia type I

Abstract

Hereditary tyrosinemia type I (HT1) is an autosomal recessive inborn error of metabolism caused by the deficiency of fumarylacetoacetate hydrolase, the last enzyme in the tyrosine catabolism pathway. This defect results in accumulation of succinylacetone (SA) that reacts with amino acids and proteins to form stable adducts via Schiff base formation, lysine being the most reactive amino acid. HT1 patients surviving beyond infancy are at considerable risk for the development of hepatocellular carcinoma, and a high level of chromosomal breakage is observed in HT1 cells, suggesting a defect in the processing of DNA. In this paper we show that the overall DNA-ligase activity is low in HT1 cells (about 20% of the normal value) and that Okazaki fragments are rejoined at a reduced rate compared with normal fibroblasts. No mutation was found by sequencing the ligase I cDNA from HT1 cells, and the level of expression of the ligase I mRNA was similar in normal and HT1 fibroblasts, suggesting the presence of a ligase inhibitor. SA was shown to inhibit in vitro the overall DNA-ligase activity present in normal cell extracts. The activity of purified T4 DNA-ligase, whose active site is also a lysine residue, was inhibited by SA in a dose-dependent manner. These results suggest that accumulation of SA reduces the overall ligase activity in HT1 cells and indicate that metabolism errors may play a role in regulating enzymatic activities involved in DNA replication and repair.

Hereditary tyrosinemia type 1 (HT1) is an autosomal recessive disease that has been reported worldwide but with variable frequency. Two forms of the disease have been described. The acute form is characterized by liver failure, and patients usually die within the first year of life. In the chronic form, patients show renal tubular dysfunction, hypophosphatemic rickets, progressive liver disease, and a high incidence of hepatocellular carcinoma (1).

It has been shown that HT1 is caused by a deficiency of fumarylacetoacetate hydrolase (FAH), the enzyme involved in the last step of the catabolic pathway of tyrosine (2). The human cDNA encoding FAH has been cloned and the FAH gene has been located to region q23–q25 of chromosome 15 (3). The gene contains 14 exons and spans approximately 35 kb of DNA (4). The enzyme is expressed predominantly in the liver but is also found in a wide range of tissues and cell types (5). The FAH deficiency results in an accumulation of two abnormal tyrosine metabolites, fumarylacetoacetate (FAA) and maleylacetoacetate (MAA), which, after reduction and decarboxylation, are transformed to succinylacetone (SA). SA is excreted in the urine and its finding is part of the HT1 diagnosis. It has been suggested that FAA and MAA possibly act as natural alkylating agents and/or disrupt sulfhydryl metabolism. FAA is mutagenic in Chinese hamster V79 cells (6), and free glutathione concentration in HT1 liver specimens is reduced to about half of normal levels (7). The final metabolite, SA, reacts nonenzymatically with proteins and free amino acids, and the majority of urinary SA in HT1 patients is found in the form of such adducts (8).

As SA reacts with amino acids, and specially with lysine, generating adducts via Schiff base formation (8), we made the assumption that SA could react with proteins involved in DNA metabolism and preferentially with proteins whose active site includes a lysine residue. A lysine residue is present in the active site of DNA-ligases that are involved in DNA metabolism to rejoin strand interruptions formed transiently during replication, repair, and recombination (9). The assumption that DNA-ligases could be a target for SA was supported by the observation that fibroblasts from HT1 patients exhibit a high level of chromosomal breakage (10).

To study whether the accumulation of SA could play a role in the DNA metabolism of HT1 cells, we have measured the DNA-ligase activity in fibroblasts from several HT1 patients. The influence of SA on the activity of purified T4 DNA-ligase also was investigated, as this enzyme and mammalian ligases act by an identical mechanism, with the formation of a covalent lysine–AMP and DNA–AMP reaction intermediates, using ATP as the source of AMP (11).

MATERIALS AND METHODS

Chemicals.

Poly(A) and oligo(dT)₁₆ were from Pharmacia. [α-³²P]ATP and [γ-³²P]ATP (3,000 Ci/mmol; 1 Ci = 37 GBq) were obtained from Amersham. Succinylacetone (Sigma) was dissolved at a concentration of 500 mM in 1 M Hepes, pH 7.5, then diluted to the appropriate concentrations in the incubation buffers. T4 DNA-ligase was from Boehringer Mannheim.

Cell Culture.

Fibroblasts from five unrelated HT1 patients were obtained from skin biopsies. One of these HT1 cell lines was designated as ALV fibroblasts. Normal (MRC 5) and HT1 fibroblasts were maintained in Ham’s F12 (GIBCO) supplemented with 10% fetal calf serum and 5% horse serum, penicillin, and streptomycin. The doubling times were about 30 and 50 hr for normal and HT1 fibroblasts, respectively. LICH cells (derived from a human hepatoma) (12) were grown in Dulbecco’s medium supplemented with 5% fetal calf serum and 5% horse serum in the presence of antibiotics.

To prepare cell extracts, exponentially growing cells were harvested by trypsinization, washed in Earle’s balanced salt solution (EBSS), and suspended (5 × 10⁷ cells per ml) in a buffer containing 66 mM Hepes at pH 7.5, 5 mM MgCl₂,1 mM DTT, and 10% glycerol. After addition of protease inhibitors, cells were disrupted by sonication at 0°C, and cell debris was removed by centrifugation (13).

Reverse Transcription, PCR, and Sequencing.

Five micrograms of total RNA, isolated from normal or HT1 fibroblasts (14), was reverse-transcribed by using the First-strand cDNA synthesis kit (Boehringer Mannheim) according to the instructions of the manufacturer. The FAH and ligase I cDNA were amplified by PCR in a DNA thermal cycler (Perkin–Elmer/Cetus) by using the PCR Expand kit (Boehringer) and specific primers (15). They were sequenced by using the Applied Biosystems (ABI Prism 310 Genetic Analyzer) Taq DyeDeoxy Terminator Cycle Sequencing procedure, following the manufacturer’s specifications. To determine the amount of ligase I mRNA, increasing amounts of total RNA were reverse-transcribed and then amplified by using specific primers. PCR conditions were determined to obtain an exponential amplification. The PCR products were run on a 1.8% agarose gel and quantified.

Formation of DNA-Ligase AMP Intermediates.

T4 DNA-ligase (2 units, 0.150 μg of proteins) was incubated (final volume, 10 μl) in a buffer containing 66 mM Hepes at pH 8.0, 10 mM MgCl₂, and increasing SA concentrations for 15 min at 25°C. The protein was incubated further for 15 min at 25°C after addition of 10 μl of the same buffer containing [α-³²P]ATP diluted with nonradioactive ATP to a specific activity of 250 Ci/mmol and added to a final concentration of 2 μM ATP (16). The reaction products were analyzed by electrophoresis in SDS/polyacrylamide gels and quantified by densitometry.

DNA-Ligase Activity Determination.

The substrate was [5′-³²P]oligo(dT)₁₆⋅poly(dA), prepared as described (17). The substrate (0.25 μg, about 20,000 cpm) was incubated in 66 mM Hepes, pH 7.5/5 mM MgCl₂/35 μM ATP, with the T4 DNA-ligase (0.5 unit) in a final volume of 20 μl. Samples were incubated at 37°C for 5 min. The reaction was stopped by the addition of formamide (10 μl) and heating at 100°C for 5 min. The reaction products were separated in 15% polyacrylamide/7 M urea gels and quantified by using an InstantImager (Packard). In the case of cellular extracts, incubation was for 30 min at 37°C, and then the samples were treated with an equal volume of phenol/chloroform and centrifuged, and the aqueous phase was analyzed by electrophoresis.

Analysis of Okazaki Fragments.

Cells were synchronized by isoleucine deprivation for 48 hr followed by release in complete medium containing aphidicolin (10 μg/ml) for 12 hr, washed, and cultured in normal medium to transit to S phase. They were collected, suspended in replication mixture, permeabilized by addition of Nonidet P-40, and incubated in the presence of [α-³²P]dATP as described (16). DNA replication was allowed for 1 min, and then the reaction was chased for increasing periods of time by adding nonradioactive dATP (10 mM final concentration). Samples were treated as described (16) and analyzed by electrophoresis in 10% polyacrylamide/8 M urea gels, and the formation of high molecular DNA was quantified by densitometry.

RESULTS

DNA-Ligase Activity in HT1 Fibroblasts.

To confirm the clinical diagnosis of tyrosinemia type I in the ALV fibroblasts used in this study, the FAH cDNA (1,396 bp) was amplified by reverse transcription–PCR and sequenced. Two patterns were superposed for exon 12 of the ALV cDNA. One pattern contained a C-to-G transversion changing serine to glycine in codon 348. The second pattern showed a deletion of the first 50 bp of exon 12. This deletion was confirmed by the loss of the PstI site located in exon 12 (data not shown). These results characterize the FAH defect in ALV cells and confirm the clinical diagnosis of tyrosinemia type I.

The overall DNA-ligase activity was measured by the joining of single-strand interruptions in the oligo(dT)₁₆⋅poly(dA) substrate, in extracts from ALV (Fig. 1A) and normal human fibroblasts (MRC5 cells) (Fig. 1B). In the same experimental conditions, the kinetics experiments showed a low activity in ALV fibroblasts, reduced about 5-fold in comparison with the activity present in the normal fibroblasts extracts (Fig. 1C). Similar results were obtained when the DNA-ligase activity was measured in fibroblasts from five unrelated HT1 patients and compared with that present in normal human fibroblasts (MRC5 cells) and human hepatoma cells (LICH cells) (Fig. 2). In the five HT1 cell lines, the activity was low and corresponded to about 20–30% of the level measured in non-HT1 cells.

Figure 1.

DNA-joining activity in ALV fibroblasts. Extracts (25 μg of proteins) from ALV (A) or normal (B) fibroblasts were incubated with the polynucleotide substrate [5′-³²P]oligo(dT)₁₆⋅poly(dA) at 37°C for 0 (lane 1), 15 (lane 2), 30 (lane 3), 60 (lane 4), 90 (lane 5), or 120 min (lane 6). The oligo(dT)₁₆ multimers were separated in polyacrylamide/urea gels. The films were overexposed to detect the small amounts of multimers in A. (C) The data from A and B are expressed in femtomoles of oligo(dT)₁₆ ligated. (•) ALV and (▴) normal fibroblasts.

Figure 2.

DNA-ligase activity in fibroblasts from various HT1 patients. Extracts (25 μg of proteins) from normal fibroblasts (lane 1), LICH cells (lane 2), or fibroblasts from five unrelated HT1 patients (lanes 3–7) were incubated with the polynucleotide substrate [5′-³²P]oligo(dT)₁₆⋅poly(dA) for 120 min at 37°C. The oligo(dT)₁₆ multimers were detected by electrophoresis in polyacrylamide/urea gels and autoradiography. For details, see Materials and Methods.

To check that this low DNA-ligase activity was not a result of an unexpected lability of the enzymes present in HT1 cells extracts, two other activities were measured. The glyceraldehyde-3-phosphate dehydrogenase (18) and the N3-methyladenine DNA glycosylase (19) had identical activities in HT1, normal fibroblasts, and human hepatoma cells (LICH cells) (data not shown).

Joining of Okazaki Fragments in HT1 Cells.

To investigate whether this low overall ligase activity could impair DNA replication in HT1 cells, synthesis and joining of Okazaki fragments were measured. Normal and ALV cells were synchronized at the G₁/S boundary and pulse-labeled with [α-³²P]dATP, and the newly synthesized DNA fragments were chased with nonradioactive dATP for different time lengths. There was a slight difference in the conversion of the fragments to high-molecular-weight DNA after a 2-min chase period, but about 35% of the DNA remained in low molecular weight in ALV cells during the period studied (20 min), whereas no such fragments were observed in normal cells (Fig. 3).

Figure 3.

Joining of Okazaki fragments in normal and ALV fibroblasts. Cells synchronized in S phase were incubated in the presence of [α-³²P]dATP for 1 min. The reaction then was chased by incubation with nonradioactive dATP for different times. After DNA isolation, replication intermediates were analyzed by gel electrophoresis and detected by autoradiography. The amount of low-molecular-weight DNA corresponding to Okazaki fragments was estimated by densitometry. (▴) Normal and (•) ALV fibroblasts.

It should be noted that, in our experimental conditions, the doubling time for ALV fibroblasts (about 50 hr) is longer than that of normal cells (30 hr) and this could be a consequence of the impaired DNA replication.

Sequencing and Expression of the Ligase I cDNA from ALV Fibroblasts.

DNA-ligase I represents an important part of the total activity present in human cells and is involved primarily in DNA replication. The low overall ligase activity and the reduced rate of rejoining of Okazaki fragments suggested that ligase I activity was defective in ALV cells. Therefore, the ligase I cDNA from ALV fibroblasts was sequenced. The results showed the absence of mutations in this gene. Furthermore, the level of expression of ligase I mRNA, measured by reverse transcription–PCR, was identical in normal and ALV fibroblasts (data not shown).

Influence of Succinylacetone on the T4 DNA-Ligase Activity.

The results described above suggested the presence of a DNA-ligase inhibitor in ALV cells. An inhibitor, which is a heat-labile protein, has been described in extracts from human cells (20). To test whether the low activity measured in ALV cells was a result of the presence of this protein, extracts from ALV cells were heated to 65°C for 30 min and then added to normal cell extracts. The activity of normal cell extracts was decreased by 40% when ALV extracts, heated or not, were added to the incubation medium, showing that these extracts contain an inhibitor that is not a heat-labile protein. A possible inhibitor was SA, which accumulates in HT1 cells and forms lysine adducts. To test this hypothesis, the influence of this compound was measured, either on the activity of purified T4 DNA-ligase or on the activity present in normal cell extracts.

T4 DNA-ligase was incubated for 15 min at 25°C with increasing SA concentrations, and its activity was measured by the joining of single-strand interruptions in the double-stranded [5′-³²P]oligo(dT)₁₆⋅poly(dA) substrate. As shown by analysis in polyacrylamide gels, the T4 DNA-ligase activity decreases after incubation with SA (Fig. 4A). This decrease is dose-dependent with an IC₅₀ of about 15 μM (Fig. 4B). When the protein was incubated with SA (25 μM) in the presence of free lysine (25 μM), 70% of the activity was recovered. This significant protection of the activity suggests that the inhibition is due to the reaction of SA with the lysine residue within the active site of the protein. Increasing the incubation times with SA did not enhance further the inhibitory effect of this compound on the enzyme activity (data not shown). The first step of the ligation reaction is the formation of a covalent ligase–AMP intermediate with a lysine–adenylate phosphoamide bond. The influence of SA on this step of the reaction was measured by incubating the T4 DNA-ligase for 15 min at 25°C with increasing SA concentrations before incubation with [α-³²P]ATP. The formation of the DNA-ligase adenylate complex is inhibited in the presence of SA (Fig. 4C). The amount of adenylylated protein decreases with the SA concentration with an IC₅₀ of about 20 μM, showing that the inhibition occurs at the first step of the ligation reaction, the enzyme–AMP formation. For comparison, the protein was incubated in the same conditions with pyridoxal phosphate that is known to interact with lysine residues and to inhibit the DNA-ligase activity (21). The formation of T4 DNA-ligase adenylate was inhibited by 70% in the presence of 50 μM pyridoxal 5′-phosphate (not shown).

Figure 4.

T4 DNA-ligase activity in the presence of SA. T4 DNA-ligase was treated or not with increasing SA concentrations (15 min at 25°C) and then incubated at 37°C for 1 min with the oligo substrate. (A) The oligo(dT)₁₆ multimers were separated in polyacrylamide/urea gels: T4 DNA-ligase without SA (lane 1) or incubated with increasing SA concentrations of 2.5(2), 5(3), 10(4), and 20(5) μM. (B) The activity was quantitated using an InstantImager (Packard). (C) Inhibition of enzyme-adenylate formation by SA. T4 DNA-ligase was incubated or not with increasing SA concentrations (15 min at 25°C) before the addition of [α-³²P]ATP. The enzyme adenylate complexes were separated by electrophoresis and detected by autoradiography.

Incubating normal fibroblast extracts (25 μg of proteins) with SA (100 μM) for 15 min reduced the overall ligase activity to 50% of the control value (Fig. 5). This shows that SA is able to inhibit the ligase activity even in presence of the large amount of proteins contained in this crude extract.

Figure 5.

Influence of SA on the DNA-ligase activity present in normal cell extracts. Normal fibroblasts extracts (25 μg proteins) were preincubated for 15 min at 25°C without (▴) or with (•) SA (100 μM). The ligase activity was then measured as described in Materials and Methods.

DISCUSSION

Tyrosinemia type I is caused by a deficiency of FAH, and different mutations have been identified in the FAH gene, including missense, nonsense, and splice consensus site mutations (15, 22–24). However, there is no strict correlation between genotype and phenotype, i.e., the acute, subacute, or chronic form of the disease (24). The lack of FAH results in the formation of FAA and MAA, but these two compounds have never been isolated as circulating or excreted metabolites (25). This suggests that they are transformed rapidly to SA, and SA concentrations from 6 to 43 μM were measured in the plasma of HT1 patients (26). Attempts have been made in vitro to express the FAH cDNA in cultured HT1 cells by using retroviral-mediated gene transfer (27), but liver transplantation and 2-(2-nitro-4-trifluoromethylbenzoyl)-1,3-cyclohexanedione treatment (28), which blocks the accumulation of metabolites upstream of FAH, are, to date, the only therapies for this disease. In the chronic form of HT1, when patients survive beyond 2 years of age, there is a high incidence (37%) of hepatocellular carcinoma, which is an uncommon tumor in children (29).

DNA damage occurs by spontaneous base deamination, alkylation, or oxidation by endogenous or environmental exposure to various compounds. Repair of these damages and correction of replicative errors are critical to maintain the genome integrity, and DNA-repair deficiency syndromes have an increased risk of developing cancer (reviewed in ref. 30). Therefore, the high cancer incidence occurring in HT1 patients, the high level of chromatid breaks observed in HT1 cells (10), and the hypersensitivity of established HT1 cells to several DNA-damaging agents (F.L., unpublished results) lead to the hypothesis that enzymes involved in DNA repair and/or replication could be altered in HT1 cells.

Different DNA-ligases have been described in mammalian cells (31, 32) and are implicated in DNA replication and in recombination and are required in the final step of base excision and nucleotide excision repair (33). As diketones react with the ɛ-amino group of lysine (34), among the possible targets for the formation of a Schiff base with SA was the lysine residue present in the active site of mammalian ligases. Our results show that the overall DNA-ligase activity measured in cells from unrelated HT1 patients corresponds to about 20% of the activity present in normal human fibroblasts. SA reduces the activity present in normal cells extracts as well as the activity of purified T4 DNA-ligase, whose active site is also a lysine residue, strongly suggesting that this compound is responsible for the low activity measured in HT1 cells. This deficiency results in a slow rejoining of Okazaki fragments in HT1 cells, probably plays a role in the completion of DNA repair processes, and might result in genomic instability. It has been suggested that the extent of DNA-ligase I deficiency tolerable to mammalian cells is low and that ligases I, II, and III are unable to compensate for each other (35, 36). Hence, the overall ligase deficiency observed in HT1 cells probably plays a key role in the symptoms associated with this disease.

A human cell line (46BR) is sensitive to killing by DNA-damaging agents (37), shows defective rejoining of Okazaki fragments, and possesses reduced DNA-ligase I activity (16). Sequencing of PCR-amplified ligase I cDNA revealed mutations in 46BR cells, carried in different alleles (38). As contrasted to 46BR cells, no mutation was detected in the ligase I cDNA of HT1 cells, and the level of transcription of this gene was identical in normal and HT1 cells. Although this activity is reduced in both 46BR and HT1 cells, the clinical symptoms associated with the diseases are different (38), suggesting that besides its inhibitory effect on DNA-ligases, SA may have other deleterious effects on the cellular metabolism.

The precise influence of SA on the different mammalian DNA-ligases now should be determined to correlate the level of each activity with specific defects observed in HT1 cells. The influence of SA on other enzymes also has to be tested, especially on enzymes whose active site includes a lysine residue, e.g., the mRNA capping enzyme (39). However, HT1 cells, established from a defined metabolic deficiency, represents, to our knowledge, the sole example of a disease characterized by a mutation in a gene with a clearly defined role in amino acid metabolism, which results in a defect in the processing of DNA.

ABBREVIATIONS

HT1

hereditary tyrosinemia type I

SA

succinylacetone

FAH

fumarylacetoacetate hydrolase

FAA

fumarylacetoacetate

MAA

maleylacetoacetate