Adenylosuccinate lyase deficiency

Molecular Genetics and Metabolism 89 (2006) 19–31 www.elsevier.com/locate/ymgme

Minireview

Adenylosuccinate lyase deWciency Erin K. Spiegel a,b,¤, Roberta F. Colman c, David Patterson a,b a

b

Eleanor Roosevelt Institute at University of Denver, Denver, CO, USA Human Medical Genetics Program, University of Colorado Health Sciences Center, Denver, CO, USA c Department of Chemistry and Biochemistry, University of Delaware, Newark, DE, USA Received 10 March 2006; received in revised form 26 April 2006; accepted 27 April 2006 Available online 12 July 2006

Abstract Adenylosuccinate lyase deWciency is a disease of purine metabolism which aVects patients both biochemically and behaviorally. The symptoms are variable and include psychomotor retardation, autistic features, hypotonia, and seizures. Patients also accumulate the substrates of ADSL in body Xuids. Both the presence of normal levels of ADSL enzyme activities in some patient tissues and the absence of a clear correlation between mutations, biochemistry, and behavior show that the system has unexplored biochemical and/or genetic complexity. It is unclear whether the pathological mechanisms of this disease result from a deWciency of purines, a toxicity of intermediates, or perturbation of another pathway or system. A patient with autistic features and mild psychomotor delay carries two novel mutations in this gene, E80D and D87E. The creation of a mouse model of this disease will be an important step in elucidating the in vivo mechanisms of the disease. Mice carrying mutations that cause ADSL deWciency in humans will be informative as to the eVects of these mutations both during embryogenesis and on the brain, possibly leading to therapies for this disease in the future. © 2006 Elsevier Inc. All rights reserved. Keywords: Autism; Psychomotor retardation; Purine biosynthesis; Enzyme; Succinylpurines

Introduction Adenylosuccinate lyase (ADSL, EC 4.3.2.2) deWciency is a defect of purine metabolism causing serious neurological and physiological symptoms. It was Wrst described in 1984 by Jaeken and Van den Berghe [1], who found succinylpurines in the cerebrospinal Xuid (CSF), plasma, and urine of three patients with severe psychomotor delay and autistic features. These succinylpurines, succinyladenosine (S-Ado) and succinylaminoimidazolecarboxamide riboside (SAICAr) are the dephosphorylated derivatives of ADSL substrates. This accumulation in their patients’ CSF suggested a deWciency in ADSL activity, and indeed, the investigators reported signiWcantly reduced ADSL enzyme activity in these patients. ADSL catalyzes two steps in the de novo purine biosynthetic pathway, which consists of 13 metabolic steps in the conversion of ribose-5-phosphate into *

Corresponding author. Fax: +1 303 333 8423. E-mail address: [email protected] (E.K. Spiegel).

1096-7192/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ymgme.2006.04.018

AMP or GMP (Fig. 1). These two steps are: the conversion of succinylaminoimidazolecarboxamide ribonucleotide (SAICAR) into aminoimidazolecarboxamide ribonucleotide (AICAR), and the conversion of succinyladenosine monophosphate (AMPS) to adenosine monophosphate (AMP). Regulation of the de novo purine biosynthetic pathway is highly controlled and occurs at multiple steps. Disruption of this regulation is known to cause other syndromes with neurodevelopmental abnormalities. For example, levels of PRPP, a substrate for the second step of de novo purine synthesis, are very important in the regulation of the pathway (Fig. 1). Mutations in PRPP synthetase leading to elevated activity of this enzyme cause elevated de novo purine synthesis and, in some families, neurodevelopmental impairment [2]. Lesch–Nyhan syndrome is caused by a deWciency of HPRT, an enzyme in the purine salvage pathway [3,4]. Since PRPP is a substrate in the reaction of HPRT, a deWciency of HPRT causes levels of free PRPP to rise and therefore stimulate Xux through the de novo purine

20

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31 R5P+ATP PRPPS PRPP GPAT PRA GAR GART

FGAR FGARAT FGAM AIR

AICR

CAIR

SAICAr

SAICAR AICAR

ADSL

FAICAR

ATIC

IMP ADSS AMPS

S-Ado

AMP

Fig. 1. A schematic representation of the pathway of de novo purine biosynthesis leading to the creation of AMP. Steps shown in red are catalyzed by ADSL. Abbreviations are as follows: ATP, adenosine triphosphate; R5P, ribose 5 phosphate; PRPP, phosphoribosylpyrophosphate; PRA, phosphoribosylamine; GAR, phosphoribosylglycineamide; FGAR, phosphoribosylformylglycineamide; FGAM, phosphoribosylformylglycineamidine; AIR, phosphoribosylaminoimidazole; CAIR, phosphoribosylcarboxyaminoimidazole; SAICAR, phosphoribosylsuccinylaminoimidazolecarboxamide; SAICAr, succinylaminoimidazolecarboxamide riboside; AICAR, phosphoribosylaminoimidazolecarboxamide; FAICAR, phosphoribosylformylaminoimidazolecarboxamide; IMP, inosine monophosphate; AMPS, succinyladenosine monophosphate; S-Ado, succinyladenosine; AMP, adenine monophosphate; PRPPS, PRPP Synthase; GPAT, phosphoribosylamidotransferase; GART, GAR synthase, GAR transformylase, AIR synthase (trifunctional protein); FGARAT, FGAR amidotransferase; AICR, AIR carboxylase, SAICAR synthetase (bifunctional protein); ADSL, adenylosuccinate lyase (bifunctional protein); ATIC, AICAR formyltransferase, IMP synthase (bifunctional protein). (For interpretation of the references to colour in this Wgure legend, the reader is referred to the web version of this paper.)

pathway [3,4]. The hallmark of this disorder is compulsive self-mutilation, although other characteristics include hyperuricemia, choreoathetosis, spasticity, aggression, and sometimes, mental retardation [3,4]. Another syndrome of purine synthesis is AICA-Ribosiduria, caused by a deWciency of the bifunctional enzyme AICAR transformylase/IMP cyclohydrase (ATIC) [5]. This enzyme acts in the two de novo pathway steps following the Wrst ADSL step (Fig. 1). A deWciency of this enzyme causes psychomotor retardation, epilepsy, congenital blindness, and dysmorphic features [5]. Like patients with ADSL deWciency, this patient accumulates SAICAr and S-Ado, as well as AICA-riboside (AICAr), the nucleoside derivative of AICAR, one substrate of ATIC [5]. Clearly, perturbations in the de novo purine pathway have profound eVects on many neurological functions, though the mechanisms are not understood.

Diagnosis and clinical features of ADSL deWciency In general, patients with ADSL deWciency are discovered during screens of children with unexplained developmental delay using the Bratton–Marshall assay for diazotizable amines [6]. This assay reveals accumulation of SAICAr in body Xuids, which is the main diagnosis criterion for this disease. However, since the Bratton–Marshall assay can cause false positive results if patients are taking certain medications, a diagnosis of ADSL deWciency is generally then made using HPLC analysis showing accumulation of both SAICAr and S-Ado. There is wide variation in the clinical presentation observed in patients with ADSL deWciency. In most documented cases, the disorder leads to profound psychomotor retardation (PMR), though there are some notable cases with only mild delay. This may reXect an ascertainment bias: as ADSL deWciency was thought to cause profound developmental delay, generally only patients with such delay were being tested for ADSL deWciency. Due to the heterogeneous nature of this disease, however, many clinical investigators are now calling for screening of patients with a wider range of PMR and behavioral phenotypes. Other, variable features of ADSL deWciency include severe hypotonia, abnormal brain glucose utilization, muscular wasting, and failure of muscle energy metabolism [7– 10]. Epilepsy and autistic features are also frequently seen [11–13]. The autistic features seen in patients with ADSL deWciency include failure to make eye contact, repetitive behaviors, agitation, temper tantrums, and autoaggressivity [14]. In most cases autistic behavior persists, except for occasional improvement of eye contact [14]. The genetics of ADSL deWciency The human ADSL gene has been mapped to chromosome 22q13.1-13.2 [15–17]. The gene is 23 kb in length, consists of 13 exons, and encodes a protein of 484 amino acids [18]. In 1992, Stone et al. [13] reported the Wrst sequence of a mutation in the ADSL gene leading to this syndrome. To date, 38 diVerent mutations have been reported in the ADSL gene that lead to ADSL deWciency (Table 1). Each mutation is a single base pair change that produces an altered ADSL protein. All are missense mutations with two exceptions: one exceptional mutation creates a new splice site and results in a 39 base pair deletion [19,20] and the other is a mutation in an NRF2 binding site in the promoter region of ADSL [21]. Most ADSL deWciency patients are compound heterozygotes and in cases in which the parents have been genotyped, each parent carries one mutant and one normal allele and is asymptomatic. No individuals with ADSL deWciency are completely lacking in enzyme activity [22,23]; complete lack of ADSL activity in humans is probably incompatible with life. Almost 50 cases of ADSL deWciency have been described to date, but attempts to correlate ADSL mutations or ADSL activity with the severity of the phenotype have uncovered no obvious patterns (Table 1).

Table 1 Characteristics of patients with ADSL deWciency Gender M

S-Ado/SAICAR .98

PMR

Autistic features

Severe

ND 1.8 (urine)

MR Present, at 1 year 3 month old, PMD < 6 weeks

F

1.3

Severe, at 6 years old, PMD < 6 weeks

F

0.7

Severe

F

0.9

Severe, lack of motor skills, such as voluntary prehension and sitting Severe

— Present, motor restlessness, no eye contact, movements of extremities Present, stereotyped movements of head and tongue, sensitive to touch Blank staring spells Severe, no eye contact, lack of language skills, repetitive behavior, stereotypic movements of hands —

Epilepsy, hypertonicity, thoracic scoliosis, contractures, spastic tetraplegia, severe cerebral and cerebellar hypotrophy Epilepsy, hypotonia, abnormal cortical function, Lissencephaly Epilepsy, microcephaly, muscle hypotonia

Convulsions

Epilepsy, muscle energy metabolism impairment

Hypotonia, hyperactivity, erethic oligofrenia, aggressiveness

Y114H R190Q

F

0.9

Severe

—

Epilepsy, apnoea, seizures, coma

F

0.52

—

F

0.97

Neonatal encephalopathy, seizures, hypotonia, death at 4 weeks Hypotonia, seizures, mental retardation, facial dysmorphism

M

0.9

M

2.4

Severe, standing at 15 months, walking at 2 years

M

2.2

Moderate, walking at 2 years, no speaking at 5 years

Present, motor restlessness, frequent crying attacks, no eye contact, exaggerated reaction to auditory stimuli, physical agitation of arms and legs, unintelligible speech —

[21,27]

[20], ADSLdb

Epilepsy, death at 5 months

1.2

[21,24]

I72V K246E P100A D422Y

—

F

t-49c P318L t-49c R426H

[20,56,64]

PMR

0.9

[26]

M26L R426H E80D D87E

1.0

Mild, washing movements of hands Severe, no eye contact, no social interaction, prefers solo play, early diagnosis of autism

[26] [26]

[21]

M

—

References [59]

t-49c S447P ATG to GTG in initiation codon S448P A2V S395R A3V R337X A3P R190X M1L R374W

ND

Severe, at 27 months old, could not sit unaided and had no speech Severe, at 15 months old, PMD at 3–5 months Mild, at 4 years old, Wne motor and social interactive skills at 3 year old level Severe

AA change

Severe muscular hypotonia, epilepsy Epilepsy

Dr. Stan Kmoch, personal communication [20,56,60] [18,61] [62] [63]

[24]

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

M F

Other symptoms Severe hypotonia, convulsions, clonus, cardio-respiratory arrest, death at 6 months Epilepsy, death at 13 years Cerebellar hypoplasia, hypertonicity

[39,65]

[18,61] (continued on next page) 21

22

Table 1 (continued) Gender

S-Ado/SAICAR

PMR

Autistic features

Other symptoms

AA change

References

F

2.1

Moderate, standing at 1 year, walking at 3 years speaking at 6 years

—

Hypotonia, hyperactivity, temper tantrums, erethic oligofrenia, aggressiveness

Y114H R190Q

[18,61]

M

0.7

Severe, Developmental arrest at 7 months

—

Hypotonia, epilepsy

Y114H R396H

M

1.5

Severe

—

Epilepsy, apnoea, acidosis, death at 6 months

F

1.8

Present, sitting at 2 years, walking at 3 years, speaking at 9 years Severe

Present, aggressive behavior, stereotypes present

R141W del 206-18 R190Q K246E

Dr. Stan Kmoch, personal communication [20,56,64]

Severe, PMD stopped at 9 months

—

1.0

Seizures, Death at 15 months

ND

F

1.5

Severe

3.0

Mild

F

3.7

Mild, at 4 years old, PMD at 2.5 years

F

1.1

F

1.6

—

Severe encephalopathy, West Syndrome, hypotonia, epilepsy

Slow growth (growth improvement with allopurinol treatment)

Severe

Mild, little eye contact, poor reaction to auditory stimuli —

Severe

—

Hypotonia, epilepsy

Hypotonia, epilepsy

Moderate F

2.5

M ND

L311V R396H R303C R303C R303C R303C V364M R452P V364M R452P R426H L423V R426H T450S R426H T450S R426H R426H

Moderate

—

Moderate

—

Severe, at 9 years old, able to stand with support but unable to walk Severe, at 5 years old, PMD at 6 months

Present

Epilepsy, hyperactivity

Severe, repetitive activities, grimacing, crying, teeth grinding, biting self, bouts of extreme agitation —

Hypotonia

R426H R426H

West syndrome, seizures, hypotonia

—

Profound muscle hypotonia

—

Seizures, hypotonia, deafness

R426H R426H R426H D430N R426H R426H

F

1.5

F

1.0 (urine)

F

2.6 (urine)

Severe, at 20 months old, PMD at 6 months Mild

M

1.2 (urine)

Severe

ADSLdb [18,61] Dr. Stan Kmoch, personal communication [59] [56] [20,26,56]

[59] [59] ADSLdb [56] [56] Dr. Stan Kmoch, personal communication [1,20,26]

[20,67] [7,18] [18]

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

F

R194C R396C R194C D268N D268N P467R

[20,66]

1.23

Severe, at 6 years old, could stand, scrawl, and speak few words. Psychomotor regression to bedridden state Severe, walk with aid and say few words at age 2, at age 15, can sit and walk few steps with aid

M

1.35

M

1.57

Severe, developmental milestones delayed, not able to stand at 11 years old

M

1.1

Severe, at 22 months old, PMD at 6 months

M

1.7

Severe, at 7 years old, PMD at 6 months

F

1.3

Severe, at 5 years old, PMD at 5 months

Mild, poor eye contact

Severe epilepsy, strabismus, spasticity, muscular wasting, severe growth retardation

R426H R426H

[30]

Severe, no eye contact, abnormal ocular movements, restlessness, continuous head and trunk rocking, screaming, aggressive tantrums Severe, no eye contact, erratic ocular movements, absence of speech, peculiar hand use (chaWng of hands in front of eyes) Present, no eye contact, motor restlessness, temper tantrums, moving hands before eyes, beating legs on bed, hyperXexion of feet

Epilepsy, spasticity, Xexion contractures of joints

R426H R426H

[30]

Past epileptic seizures, spastic paraparesis, brachycephaly

R426H R426H

[30]

R426H R426H

[20,26]

R426H R426H R426H R426H R426H R426H R426H R426H S438P S438P

ADSLdb

Severe, wandering gaze, no eye contact, rubbing hands and feet, clapping hands, moving hands before eyes, lying with knees raised, beating back on bed, hours handling same object, laughing to self, temper tantrums Severe, wandering gaze, no eye contact, rubbing hands and feet, clapping hands, moving hands before eyes, lying with knees raised, beating back on bed, hours handling same object, laughing to self, temper tantrums

Slow growth , muscle wasting, epilepsy

Slow growth, muscle wasting, epilepsy

S438P S438P

ADSLdb ADSLdb ADSLdb [1,13,26]

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

F

[1,13,26]

Unless otherwise noted, S-Ado/SAICAR ratios are from CSF. Abbreviations are as follows: PMR, psychomotor retardation; PMD, psychomotor development; ND, not determined; ADSLdb, ADSL database website, www.icp.ucl.ac.be/adsldb. 23

24

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

In 2000, Stathis et al. [24] described a patient who was provisionally diagnosed with autism at 2 years, 9 months. This boy engaged in little eye contact, no social interaction, and preferred solo play [24]. However, importantly, this child had only mild psychomotor delay and no physical abnormalities. At the age of 4 years, his Wne motor and social interactive skills were at the 3 year old level [24]. This patient was found to be a compound heterozygote for mutations causing amino acid changes E80D and D87E; these mutations in the ADSL gene had not been observed previously [25]. Although most ADSL mutations are simple missense mutations, Kohler et al. [19] identiWed a patient with a point mutation that created a 5⬘ consensus donor splice site mutation which results in a 39 base pair deletion in the mRNA and a deletion of amino acids 206–218 in the ADSL enzyme. The other ADSL allele of this patient carries the mutation R141W. The patient had very severe disease including infantile epileptic encephalopathy with reduced myelination and death at 6 months of age [19]. Eighteen patients with ADSL deWciency have been described who carry at least one copy of the ADSL mutation causing the amino acid change R426H (Table 1). This is the most commonly found mutation in ADSL deWciency, although these patients show signiWcant phenotypic heterogeneity. Four patients described as homozygous for the R426H mutation have severe psychomotor retardation and severe autistic features [1,20,26–29]. These childrens’ average psychomotor development (PMD) was that of a 6-month-old while their chronological age averaged 3 years. These patients displayed marked autistic behaviors, including repetitive behaviors, agitation and temper tantrums, crying, loss of eye contact, motor restlessness, and stereotypic behaviors [1,26,27,30]. One patient with severe PMR but only mild autistic features is a compound heterozygote with mutations resulting in amino acid changes M26L and R426H [19]. At 15 months old, this child’s psychomotor development was that of a 3-5 month old. However, the autistic features of this child were mild, manifesting as washing movements of the hands. Three ADSL deWcient patients with profound PMR and convulsions were found in whom one of the ADSL alleles revealed a normal coding sequence. Marie et al. [21] analyzed the “normal” allele in these three patients and found a mutation in the 5⬘ UTR of the ADSL gene. This mutation, ¡49T ! C, lies in the consensus sequence of a binding site for nuclear respiratory factor 2 (NRF2). NRF-2 is a positive stimulatory regulator of transcription of nuclear genes contributing to mitochondrial function [31]. Analysis of this mutation by RT-PCR and luciferase assay showed that less mRNA was transcribed from the ¡49T-C allele [21]. Less expression of the normal enzyme, combined with the mutation on the other ADSL allele, may explain the ADSL deWciency in these patients and indicate a role for NRF-2 in the regulation of purine biosynthesis.

Pathogenic mechanisms of ADSL deWciency The mechanism(s) whereby ADSL deWciency causes psychomotor delay and behavioral abnormalities is unknown. Possibilities include: deWcient synthesis of purine nucleotides; impairment of the purine nucleotide cycle; and toxic eVects of the intracellular and/or extracellular accumulation of the ADSL substrate derivatives. Synthesis of purine nucleotides Biochemical studies of Wbroblasts from ADSL deWcient patients suggest that Xux through the two ADSLcatalyzed steps of the purine biosynthetic pathway remains possible in cultured cells [23]. Another study found no signiWcant decrease in the rate of de novo purine synthesis or any irregularities in purine nucleotide concentrations in cultured Wbroblasts of patients with ADSL deWciency [22]. Moreover, Wbroblasts from ADSL deWcient patients also display normal growth rates and concentrations of ATP when grown in the presence of purine bases [22]. Because mammalian cells are known to synthesize purines by the salvage pathway when cultured in the presence of purine bases, these data suggest that adenine salvage is suYcient to sustain synthesis of adenine nucleotides [23]. In addition, normal in vivo concentrations of adenine and guanine nucleotides were found in liver, kidney, and muscle of ADSL deWcient patients [32]. It is possible that the ADSL defect is circumvented in these tissues by a supply of purines from non-aVected cell types (i.e., erythrocytes, granulocytes) via purine salvage enzymes [14], explaining the normal purine levels. However, little is known about the normal purine levels in the living brain. Technologies using focused microwaves on rat and mouse brains have been more successful in preservation of in vivo nucleotide concentrations than conventional methods of harvesting CNS tissues [33]. These measurements will be especially important in mouse models of ADSL deWciency. Moreover, little is known about the regulation of purine synthesis during embryogenesis, when a deWcit in purine metabolism may have its developmental eVect. Studies in rat and mouse suggest that early embryos depend primarily on de novo purine synthesis [34]. The defect caused by ADSL deWciency may occur during pre-natal development but be manifested later. Purine nucleotide cycle In addition to its role in de novo purine synthesis, ADSL participates in the purine nucleotide cycle along with AMP deaminase and adenylosuccinate synthetase (Fig. 2). This cycle regulates cellular metabolism by controlling both the levels of fumarate, a citric acid cycle intermediate, and of free AMP, which aVects the concentration of ATP [35]. Moreover, the purine nucleotide cycle is particularly important in muscle to maintain the

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

Fig. 2. A schematic representation of the purine nucleotide cycle.

ATP/AMP ratio [35]. In 1997, Salerno et al. [10] described a patient with ADSL deWciency whose muscle energy metabolism was impaired. The patient presented with a reduced muscle energy reserve which resulted from a reduced ability of mitochondria to respond to metabolic needs. The ATP levels in this patient were depleted during recovery after a relatively mild exercise; this shows that pathways maintaining adenylate energy charge are impaired in her skeletal muscle [10]. The impairment of ADSL could thus play a role in the muscle wasting and growth retardation seen in some patients. Recently, a patient has been described with a deWciency in fumarase and thus accumulates fumarate, one of the products of ADSL and a compound relevant to the purine nucleotide cycle. The researchers found elevated SAICAr and S-ado in cerebrospinal Xuid from this patient, but not in urine. They hypothesized that the excess fumarate inhibited ADSL resulting in accumulation of SAICAr, which may be responsible for the neuropathology seen in some individuals with fumarase deWciency [36]. Toxic eVects of intermediates Although no stringent correlation between genotype and phenotype has been observed, Van den Berghe et al. noted a correlation between the severity of psychomotor retardation in ADSL patients and the ratio of the two dephosphorylated intermediates (S-Ado and SAICAr) that accumulate in their body Xuids [23]. They identiWed two classes of patients with ADSL deWciency: Type I patients have severe psychomotor delay and S-ado: SAICAr ratios of »1; Type II patients have only mild psychomotor delay and have S-ado:SAICAr ratios of 2–4. Moreover, studies in Wbroblasts revealed that in Type I patients, the activities of ADSL with its two substrates, AMPS and SAICAR, were similar, both at about 30% of normal. However, in Type II patients, activity of ADSL with SAICAR was about 30% of normal while activity with AMPS was only 3% of normal. This could explain the high S-Ado/SAICAr ratio found in Type II individuals. The disproportionate decrease in ADSL activity with its two substrates might be due to defects that impair binding of S-AMP to a greater extent than the binding of SAICAR. From these Wndings, it has been hypothesized that SAICAr is a neurotoxic compound and S-Ado may be protective of SAICAr’s eVects. The Wnding that infusion

25

of SAICAr to rats induces neuronal damage in speciWc regions of the hippocampus is consistent with this hypothesis [37]. The structural similarities of SAICAr and S-Ado to adenosine suggested that the toxicity may involve an interference with adenosine receptors and/or re-uptake of adenosine [8]. However, the toxic eVect of SAICAr may not be due to its interference with adenosine receptors or with the re-uptake of adenosine as studies have failed to detect an interference by the two substrates with the binding and uptake of adenosine in crude membranes of rat brain cortex [38]. Moreover, although PET images reveal a profound decrease in Xuorodeoxyglucose (FDG) uptake in all cortical areas of ADSL deWcient patients, the pattern of impairment of FDG uptake in ADSL patients does not correspond to the distribution of adenosine A1 receptors and to adenosine re-uptake sites [8]. An important point here is that the dephosphorylated intermediates, SAICAr and S-Ado, are possibly not the compounds that are present inside the cell to cause a toxic eVect. It is perhaps more likely that the phosphorylated intermediate compounds, SAICAR and AMPS, are acting inside the cell and when accumulated are dephosphorylated and then able to exit the cell and accumulate in body Xuids. The biochemistry of ADSL deWciency and ADSL mutations Patient enzyme activities Measurements of ADSL activities in biopsy specimens show that the degree of enzyme deWciency is variable in patients and surprisingly is not present in all tissues [39,40]. Most patients have deWcient levels of enzyme activity in liver and kidney, but only some have deWcient levels in skeletal muscle [40,41]. Generally, ADSL activity in muscle is reduced in patients presenting with growth retardation and muscular wasting [14]. In some patients, low enzyme levels are also observed in erythrocytes and peripheral blood lymphocytes [14,39]. Taken together, these enzyme data demonstrate tissue heterogeneity of the ADSL defect, though the mechanism behind this phenomenon is unknown. In addition, many researchers have found the human enzyme to be diYcult to work with in the laboratory. It has been notoriously susceptible to inactivation upon freezing and it may be necessary to prepare the enzyme fresh for each experiment. Therefore, any inconsistencies in tissue handling and assay conditions may confound enzyme data obtained from patient tissues. ADSL protein structure The ADSL enzyme is a homotetramer whose subunits have a molecular weight of 52 kDa (Fig. 3). The crystal structures of ADSL enzymes from Thermotoga maritima and Pyrobaculum aerophilum have been determined at 1.8 and 2.1 Å resolution, respectively [42,43]. Based on these

26

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

Fig. 3. (A) An overlay of the B. subtilis adenylosuccinate lyase (cyan) and the human enzyme (white) models. Calculating from the  carbons, the two structures have an RMS value of 0.3 Å. (B) Homology model of B. subtilis adenylosuccinate lyase based upon the T. maritima crystal structure, PDB 1c3c [42,48]. On each of the four subunits (labeled I–IV and colored red, cyan, green, and yellow, respectively), the B. subtilis residues Met10, Ile123, and Thr367, corresponding to human enzyme mutations, have been displayed. One of the four active site regions has been encircled in white. There is an active site at each intersection of three subunits. Used with permission [58]. (For interpretation of the references to colour in this Wgure legend, the reader is referred to the web version of this paper.)

crystal structures and molecular modeling of the amino acid sequence, structural features of ADSL from Homo sapiens, Bacillus subtilis, T. maritima and P. aerophilum appear to be remarkably similar [42,43]. The B. subtilis ADSL shows 26% sequence identity plus 17% similarity with the human enzyme [44]. The ADSL monomer from T. maritima contains three domains: Domain 1 consists of residues 1–93 and is compact and largely helical; Domain 2 consists of residues 94– 349 and is elongated, with Wve -helices and one  sheet; and Domain 3 consists of residues 350–431 and, like D1, is compact and largely helical [42]. Four monomers come together to form an active tetramer with D2 symmetry [42]. The core of the tetramer is composed of four D2’s arranged as an elongated bundle of 20 -helices [42]. The active site cleft of the ADSL enzyme is formed by the intersection of three separate subunits: residues 19–93, 176–222, and 288– 313 of monomer 2; residues 237–257 and 263–283 of monomer 3; and residues 136–151 and 339–430 of monomer 4 [42]. As described below, His141, His68, and His89 from diVerent subunits are all found within the active site of the enzyme and all are required for catalysis [44,45]. The ADSL of P. aerophilum, a hyperthermic archaebacterium, contains only 403 amino acids, and the tetramer is quite compact, lacking a pronounced active site cleft [43]. Interestingly, this enzyme contains three disulWde bonds, while none have been found in T. maritima or B. subtilis [43]. These disulWde bonds appear to limit the mobility of peripheral domains D1 and D3, and may be needed to increase the stability of the enzyme in extreme environmental conditions [43].

ADSL is a member of a metabolic enzyme superfamily known as the fumarase enzymes. Members of this family catalyze reactions with fumarate as one of the products. Other members include aspartase (EC 4.3.1.1), fumarase (EC 4.2.1.2), argininosuccinate lyase (EC 4.3.2.1), -crystallin, and 3-carboxy-cis,cis-muconate lactonizing enzyme (CMLE; EC 5.5.1.5) [46,47]. Although the members of this enzyme superfamily exhibit low sequence homology, they have strikingly similar tertiary structures [45]. The structure of the core (D2) is highly conserved across the enzyme superfamily, whereas domains D1 and D3 diVer signiWcantly [42]. These structural diVerences correlate to diVerences in the substrates of the enzymes in the superfamily. For instance, the active site cleft of ADSL is wide and deep to accommodate its charged, bulky substrates, while fumarase C has a small substrate, and a correspondingly smaller active site cleft [42]. A highly conserved sequence found in residues 284–303 of human ADSL (257–276 in B. subtilis) is considered the “signature sequence” of fumarases, and is shared by members of this enzyme superfamily [18,48]. This signature sequence is in close proximity to the active site [48]. Both reactions of ADSL are -eliminations of a fumarate via a general acid–base catalytic mechanism in which the general base abstracts a methylene proton from the carbon in the -position relative to the leaving nitrogen. Likewise, the general acid donates a proton to the leaving nitrogen. Thus, a double bond is formed between the - and -carbons of the succinyl moiety and the bond between the -carbon and N6 of the adenosine ring of the leaving group is cleaved. AYnity labeling of B. subtilis ADSL with 2-[(4bromo-2,3-dioxobutyl)thio]adenosine 5⬘ monophosphate

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

(2-BDB-TAMP) or 6-(4-bromo-2,3-dioxobutyl)thioadenosine 5⬘ monophosphate (6-BDB-TAMP), which are both similar in structure to AMPS (and would therefore be expected to bind to the active site of ADSL), showed His68 and His141 (equivalent to His86 and His159 in human ADSL) to be the target residues of these aYnity labels, respectively [49,50]. Substitutions for His68 in B. subtilis caused signiWcant decreases in catalytic activity of the ADSL enzyme, and substitutions for His141 in B. subtilis completely abolished enzyme activity [44,50]. These data indicate that both His68 and His141 are important for catalysis. Moreover, studies on the pH-dependence of the B. subtilis enzyme support the hypothesis that these two histidines act as the general acid (His68) and the general base (His 141) in the catalytic reactions of ADSL [44]. The histidine residues at positions 68 and 141 are both conserved in the ADSL proteins of »100 species, as is His89. Mutants of His89 in B. subtilis caused a signiWcant (at least 70-fold) decrease in the enzyme activity compared to wild-type, and it was found that both the size and the charge of the residue are important for activity [45]. Indeed, it has been shown that His89 is important for binding the AMP portion of the substrate [45]. Moreover, complementation experiments in B. subtilis with mutant enzymes of the three diVerent histidine residues show that His89 and His68 contribute to the active site from one subunit, while His141 is contributed from a diVerent subunit of the tetramer [45]. Sequence similarities of ADSL among species have allowed structural models to be constructed for ADSL enzymes from human and B. subtilis based on the crystal structure of T. maritima ADSL [48,51] (Fig. 3). Using this model of B. subtilis ADSL and available biochemical data, AMPS was manually docked into the active site48]. In this assembly, conserved residues Gln212, Asn270, and Arg301 are close to the succinyl moiety of AMPS docked in the active site [52] (Fig. 5). Indeed, these residues were each found to be critical for catalysis by ADSL through binding and orienting the succinyl carboxylate groups of SAICAR and AMPS. Asn270 and Gln212 are likely important for

27

hydrogen bonding with the - and -carboxylate oxygens, while Arg301 interacts electrostatically with the other carboxylate oxygen [52]. Although ADSL catalyzes two diVerent reactions, it seems that no special adaptations by the protein are necessary to accommodate both substrates (Fig. 4) [42]. This is because in SAICAR, the carbon to which the succinyl moiety is attached is able to rotate 180° prior to ring closure [42]. Thus, the enzyme can bind SAICAR in the same orientation as AMPS, and the bond rotation allows ADSL to complete the insertion of a nitrogen in two diVerent places in the nucleotide, one before and one after ring closure [42]. In addition, it seems that the groups that diVer between the two substrates make no speciWc contacts to the enzyme, and therefore the enzyme accommodates both [42]. Two residues, Lys268 and Glu275, which are conserved in the “signature sequence,” lie in physical proximity to the other known active site residues, e.g. (Glu275 is positioned very close to His141). These were studied using site-directed mutagenesis of the B. subtilis ADSL enzyme [48]. Mutations of Lys268 caused signiWcant decreases in enzyme activity while mutations of Glu275 caused a complete lack of enzyme activity [48]. Mutation experiments and molecular modeling data suggest that both size and charge at Lys268 are important for binding the phosphate group of AMPS [48]. Moreover, based on conformational changes and substrate binding assays, it is possible that the Glu275– His141 interaction helps maintain enzyme conformation, and mutations at position Glu275 might “lock” the enzyme into a conformation unable to bind the substrate [48]. Another interesting Wnding from complementation experiments combining B. subtilis enzymes mutated at positions 268 and 275 with enzymes mutated at positions 141 or 89 was functional evidence that a third subunit contributes residues to the active site of ADSL [48]. Amino acid residues that are completely conserved in a protein family are considered to be important for stabilizing structure or for function. Likewise, mutations that cause disease are more likely to be found in amino acid

Fig. 4. Structures of AMPS and SAICAR, the two substrates of ADSL.

28

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

Fig. 5. Energy-minimized model of AMPS bound to the active site of wild-type ASL. AMPS (shown in “sticks”) is colored according to each atom, and each amino acid residue of ASL identiWed previously (also in “sticks”), as well as those that are the subject of the present study (shown in thicker “sticks” and labeled in bold type), is colored red, aqua, and yellow to denote separate subunits of origin. The R carboxylate of the succinyl moiety of AMPS is that which is closer to the N6 of the purine moiety, while the aˆ carboxylate is further from N6. Only the side chains of the enzyme active-site residues are shown for clarity. Distances between carboxylate oxygens on the succinyl moiety of AMPS and non-hydrogen atoms on R67, Q212, N270, and R301 are shown. The inset depicts the full tetrameric enzyme with the region containing the active site (enlarged to the left) encircled. The four subunits, denoted in red, aqua, green, and yellow, are numbered I, II, III, and IV, respectively. Used with permission [52]. (For interpretation of the references to colour in this Wgure legend, the reader is referred to the web version of this paper.)

residues important for structure and function. Even relatively conservative substitutions at these evolutionarily conserved residues may aVect enzyme function [53]. In general, disease-associated amino acid changes tend to be more radical than those tolerated by natural selection [54]. ADSL sequences are available from 119 organisms, 24 of which are from eukaryotes and 95 are from bacterial strains. The most commonly mutated amino acid leading to human ADSL deWciency is arginine (R), which is found mutated at Wve diVerent amino acid positions. Of patients with ADSL deWciency whose mutation is known, 31 have at least one mutation of an R residue, and 13 of these are the R426H mutation. In general, mutations at R residues constitute approximately 15% of disease mutations. This is due largely to the fact that it is a relatively common residue in human proteins, and that the arginine codon contains the highly mutable 5⬘-CpG dinucleotides [55]. Model systems to study ADSL deWciency Analyses of puriWed or recombinantly expressed enzymes have been presented as model systems to study ADSL. In one study, a set of ADSL—Trx (histidine tagged) fusion proteins were assayed [56]. Unfortunately, the Trx fusion protein has a diVerent activity proWle than the wild-type activity in Wbroblasts. SpeciWcally, the Trx

protein shows higher activity with SAICAR than with AMPS, while the wild-type (non-fusion) protein shows higher activity with AMPS than with SAICAR [56]. Also, the fusion protein appears to be more stable than the wild-type protein. Other studies of puriWed ADSL proteins have shown similar results [13,18,57]. The comparisons of the puriWed proteins in these studies still yielded interesting results. Some mutations causing ADSL deWciency aVected thermal stability while others did not, and unstable mutations seem to aVect the activity of the enzyme with both substrates while the stable mutations have a more marked decrease in activity with AMPS than with SAICAR [13,18,56,57]. This was especially apparent with the R303C mutation [56]. While it is still possible to compare the various mutant fusion and puriWed proteins, it is uncertain that the altered properties will quantitatively reXect the properties of the non-derivatized proteins intracellularly. In order to gain insight into the eVects of mutations in ADSL that lead to disease, disease-causing mutations were introduced into the B. subtilis ADSL. Sequence and structural alignments were used to determine the equivalent residues between B. subtilis and human ADSL enzymes. In patients with ADSL deWciency, mutations M26L, R141W, and S395R cause severe disease (these correspond to residues M10, I123, and T367 in the

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

B. subtilis enzyme, respectively) [58]. Since the latter two residues in the bacterial enzyme diVer from the human residues, I123R and T367S were constructed as controls to mimic the human sequence [58]. In all cases the altered enzymes retain considerable activity for both substrates, suggesting that these residues are not essential for catalysis [58]. Circular dichroism, thermostability, and native gel electrophoresis data indicate that unfavorable substitutions at all three positions yield structurally impaired enzymes at both the secondary and quaternary levels [58]. Enzyme instability, as monitored by thermal stability, was in some cases decreased by subunit complementation experiments using wild type enzyme [58]. In two mild cases of ADSL deWciency, a mutation was found in the ADSL gene causing the amino acid change R303C . This residue lies adjacent to the conserved signature sequence and appears to be located just out of the active site [19,51]. In these two patients it was hypothesized that the mutation causes a catalytic defect in the enzyme due to the proximity of the mutation to the active site [19]. However, mutations at this residue in B. subtilis (residue 276) have shown that amino acid replacements retain substantial substrate aYnity and enzymatic activity [51]. These observations suggest that this particular residue is not essential for catalysis, though it may be important for proper subunit interaction and tetramer assembly [51]. Interestingly, mutations at this site also cause diVerential activities of the enzyme with its two substrates, preferring SAICAR over AMPS [51]. This may reXect the greater Xexibility of SAICAR and its ability to adapt to a structurally impaired enzyme active site [51]. The two novel human ADSL mutations found in the autistic patient, E80D and D87E, were also investigated in B. subtilis (residues 62 and 69) to gain insight into their eVects on the enzyme [25]. The D87 residue is completely conserved among the 119 ADSL sequences and lies next to a residue required for catalysis in B. subtilis [50]. The residue at position 80, however, is variable in the diVerent organisms, although, interestingly, Glu does not naturally occur at this position. Importantly, B. subtilis has a negatively charged aspartate at position 65, whereas the human enzyme contains a positively charged arginine at that position [25]. This residue caused unfavorable interactions with other amino acids and was therefore replaced with an arginine (D65R). The 62E, 65R, 69D construct is the best representation of a normal “humanized” B. subtilis enzyme, and was used as the control in these experiments [25]. Interestingly, the 62D mutation causes a more signiWcant decrease in enzyme activity than does the 69E mutation although the residue at position 69 is conserved and lies closer to the active site [25,45]. Based on homology model comparisons of the wild type and mutant enzymes, mutations at 62 and 69 distort the enzyme–substrate complex and perturb subunit interactions in the enzyme, possibly causing the low activity and instability of the mutant enzymes [25]. These changes in the enzyme

29

may be responsible for the ADSL deWciency phenotype in this patient. Enzymatic studies of ADSL have been crucial in identifying residues in the protein that are important for stability, substrate binding, and catalysis. These studies will continue to be important in understanding the chemical properties of the enzyme. In addition, a more comprehensive approach is necessary to understand the biological repercussions of the mutations in ADSL and the links between the mutations and the phenotypes seen in patients. One attractive method will be to utilize mammalian cells with mutant ADSL. Such mutants have been isolated in CHO-K1 cells although the nature of the mutation has not yet been determined in these mutants [68,69]. These cells could be useful to study the mechanism(s) of toxicity of ADSL deWciency or to examine properties of the human ADSL wild-type and mutant proteins after transfection of appropriate clones. However, ADSL deWciency is characterized by developmental delay and behavioral features, and these may be best studied in animal models. A particularly attractive approach would be construction of mouse models of human ADSL deWciency. These could be constructed by introduction of normal and mutant forms of human ADSL into mice as transgenes, production of mice in which the mouse ADSL gene has been inactivated by targeted mutagenesis (ADSL knockouts), and breeding of the transgenic and knockout mice to produce mice expressing only the normal or mutant forms of human ADSL. These mice would be predicted to have the biochemical deWcits similar to those seen in humans with ADSL deWciency. Therefore, these mice could be used to examine the etiology of the pathogenic features of ADSL deWciency, for example by analysis of which intermediates actually accumulate in the brains of the mice during development, by dietary or pharmacologic manipulation of purine synthesis and levels, and by attempts to reverse the biochemical and pathogenic features of ADSL mutations. While it is not possible to predict with certainty whether the mice would have developmental delay and perhaps autistic features, tests are now being designed to detect autistic features in mice and could easily be applied to the study of mice with mutant ADSL [70,71]. Studies on ADSL deWcient mice, coupled with increased understanding of the biochemistry of the ADSL enzyme, will lead to new understanding of developmental delay and autism and to new approaches to therapies for individuals with this inborn error of metabolism. Acknowledgments We thank Dr. Miles Brennan for useful discussions about ADSL deWciency and for critical reading of the manuscript. This work was supported by NIH Grant MH65431 and grants from the BonWls–Stanton Foundation and the Ludlow-GriYth Foundation to D.P., and by NIH Grant DK60504 to R.F.C.

30

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

References [1] J. Jaeken, G. Van den Berghe, An infantile autistic syndrome characterised by the presence of succinylpurines in body Xuids, Lancet 2 (8411) (1984) 1058–1061. [2] Ahmed M., W. Taylor, P.R. Smith, M.A. Becker, Accelerated transcription of PRPS1 in X-linked overactivity of normal human phosphoribosylpyrophosphate synthetase, J. Biol. Chem. 274 (11) (1999) 7482–7488. [3] B.J.F. Rossiter, C.T. Caskey, Hypoxanthine–guanine phosphoribosyltransferase deWciency: Lesch–Nyhan syndrome and gout, in: C.R. Scriver, A.L. Beaudet, W.S. Sly, D. Valle (Eds.), The Metabolic and Molecular Bases of Inherited Disease, Sventh ed., McGraw Hill, New York, 1995, pp. 1679–1706. [4] A.A. Baumeister, G.D. Frye, The biochemical basis of the behavioral disorder in the Lesch–Nyhan syndrome, Neurosci. Biobehav. Rev. 9 (1985) 169–178. [5] S. Marie, B. Heron, P. Bitoun, T. Timmerman, G. Van Den Berghe, M.F. Vincent, AICA-ribosiduria: a novel, neurologically devastating inborn error of purine biosynthesis caused by mutation of ATIC, Am. J. Hum. Genet. 74 (6) (2004) 1276–1281. [6] P.K. Laikind, J.E. Seegmiller, H.E. Gruber, Detection of 5⬘-phosphoribosyl-4-(N-succinylcarboxamide)-5-aminoimidazole in urine by use of the Bratton–Marshall reaction: identiWcation of patients deWcient in adenylosuccinate lyase activity, Anal. Biochem. 156 (1) (1986) 81–90. [7] D. Valik, P.T. Miner, J.D. Jones, First U.S. case of adenylosuccinate lyase deWciency with severe hypotonia, Pediatr. Neurol. 16 (3) (1997) 252–255. [8] A.G. De Volder, J. Jaeken, G. Van den Berghe, A. Bol, C. Michel, M. Cogneau, A.M. GoYnet, Regional brain glucose utilization in adenylosuccinase-deWcient patients measured by positron emission tomography, Pediatr. Res. 24 (2) (1988) 238–242. [9] C. Salerno, M. Celli, R. Finocchiaro, P. D’Eufemia, P. Iannetti, C. Crifo, O. Giardini, EVect of D-ribose administration to a patient with inherited deWcit of adenylosuccinase, Adv. Exp. Med. Biol. 431 (1998) 177–180. [10] C. Salerno, S. Iotti, R. Lodi, C. Crifo, B. Barbiroli, Failure of muscle energy metabolism in a patient with adenylosuccinate lyase deWciency. An in vivo study by phosphorus NMR spectroscopy, Biochim. Biophys. Acta 1360 (3) (1997) 271–276. [11] L.Y. Tsai, Psychopharmacology in autism, Psychosom. Med. 61 (5) (1999) 651–665. [12] G. Wolterink, L.E. Daenen, S. Dubbeldam, M.A. Gerrits, R. van Rijn, C.G. Kruse, J.A. Van Der Heijden, J.M. Van Ree, Early amygdala damage in the rat as a model for neurodevelopmental psychopathological disorders, Eur. Neuropsychopharmacol. 11 (1) (2001) 51–59. [13] R.L. Stone, J. Aimi, B.A. Barshop, J. Jaeken, G. Van den Berghe, H. Zalkin, J.E. Dixon, A mutation in adenylosuccinate lyase associated with mental retardation and autistic features, Nat. Genet. 1 (1) (1992) 59–63. [14] G. Van den Berghe, M.F. Vincent, J. Jaeken, Inborn errors of the purine nucleotide cycle: adenylosuccinase deWciency, J. Inherit. Metab. Dis. 20 (2) (1997) 193–202. [15] M.L. Van Keuren, I.M. Hart, F.T. Kao, R.L. Neve, G.A. Bruns, D.M. Kurnit, D. Patterson, A somatic cell hybrid with a single human chromosome 22 corrects the defect in the CHO mutant (Ade-I) lacking adenylosuccinase activity, Cytogenet. Cell Genet. 44 (2–3) (1987) 142–147. [16] O. Delattre, C.J. Azambuja, A. Aurias, J. Zucman, M. Peter, F. Zhang, M.C. Hors-Cayla, G. Rouleau, G. Thomas, Mapping of human chromosome 22 with a panel of somatic cell hybrids, Genomics 9 (4) (1991) 721–727. [17] E.A. Fon, S. Demczuk, O. Delattre, G. Thomas, G.A. Rouleau, Mapping of the human adenylosuccinate lyase (ADSL) gene to chromosome 22q13.1 ! q13.2, Cytogenet. Cell Genet. 64 (3-4) (1993) 201–203. [18] S. Kmoch, H. Hartmannova, B. Stiburkova, J. Krijt, M. Zikanova, I. Sebesta, Human adenylosuccinate lyase (ADSL), cloning and charac-

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

[35]

[36]

terization of full-length cDNA and its isoform, gene structure and molecular basis for ADSL deWciency in six patients, Hum. Mol. Genet. 9 (10) (2000) 1501–1513. M. Kohler, B. Assmann, C. Brautigam, W. Storm, S. Marie, M.F. Vincent, G. Van den Berghe, H.A. Simmonds, G.F. HoVmann, Adenylosuccinase deWciency: possibly underdiagnosed encephalopathy with variable clinical features, Eur. J. Paediatr. Neurol. 3 (1) (1999) 3–6. S. Marie, H. Cuppens, M. Heuterspreute, M. Jaspers, E.Z. Tola, X.X. Gu, E. Legius, M.F. Vincent, J. Jaeken, J.J. Cassiman, G. Van den Berghe, Mutation analysis in adenylosuccinate lyase deWciency: eight novel mutations in the re-evaluated full ADSL coding sequence, Hum. Mutat. 13 (3) (1999) 197–202. S. Marie, V. Race, M.C. Nassogne, M.F. Vincent, G. Van den Berghe, Mutation of a nuclear respiratory factor 2 binding site in the 5⬘ untranslated region of the ADSL gene in three patients with adenylosuccinate lyase deWciency, Am. J. Hum. Genet. 71 (1) (2002) 14–21. F. Van den Bergh, M.F. Vincent, J. Jaeken, G. Van den Berghe, Functional studies in Wbroblasts of adenylosuccinase-deWcient children, J. Inherit. Metab. Dis. 16 (2) (1993) 425–434. F. Van den Bergh, M.F. Vincent, J. Jaeken, G. Van den Berghe, Residual adenylosuccinase activities in Wbroblasts of adenylosuccinase-deWcient children: parallel deWciency with adenylosuccinate and succinyl-AICAR in profoundly retarded patients and non-parallel deWciency in a mildly retarded girl, J. Inherit. Metab. Dis. 16 (2) (1993) 415–424. S.L. Stathis, D.M. Cowley, D. Broe, Autism and adenylosuccinase deWciency, J. Am. Acad. Child Adolesc. Psychiatry 39 (3) (2000) 274–275. S. Sivendran, D. Patterson, E. Spiegel, I. McGown, D. Cowley, R.F. Colman, Two novel mutant human adenylosuccinate lyases (ASLs) associated with autism and characterization of the equivalent mutant Bacillus subtilis ASL, J. Biol. Chem. 279 (51) (2004) 53789–53797. J. Jaeken, S.K. Wadman, M. Duran, F.J. van Sprang, F.A. Beemer, R.A. Holl, P.M. Theunissen, P. de Cock, F. van den Bergh, M.F. Vincent, et al., Adenylosuccinase deWciency: an inborn error of purine nucleotide synthesis, Eur. J. Pediatr. 148 (2) (1988) 126–131. M. Nassogne, B. Henrot, G. Aubert, C. Bonnier, S. Marie, C. SaintMartin, G. Van den Berghe, G. Sebire, M. Vincent, Adenylosuccinase deWciency: an unusual cause of early-onset epilepsy associated with acquired microcephaly, Brain Dev. 22 (6) (2000) 383–386. S. Kmoch, H. Hartmannora, J. Krijt, I. Sebesta, Adenylosuccinase deWciency—identiWcation of a new disease causing mutation, J. Inherit. Metab. Dis. 19 (suppl. 1) (1996) 13. S. Kmoch, H. Hartmannora, J. Krijt, D. Valik, J.D. Jones, I. Sebesta, Genetic heterogeneity in adenlyosuccinate lyase deWciency, Clin. Biochem. 30 (1997) 22. P. Edery, S. Chabrier, I. Ceballos-Picot, S. Marie, M.F. Vincent, M. Tardieu, Intrafamilial variability in the phenotypic expression of adenylosuccinate lyase deWciency: a report on three patients, Am. J. Med. Genet. 120A (2) (2003) 185–190. J.V. Virbasius, C.A. Virbasius, R.C. Scarpulla, Identity of GABP with NRF-2, a multisubunit activator of cytochrome oxidase expression, reveals a cellular role for an ETS domain activator of viral promoters, Genes Dev. 7 (3) (1993) 380–392. G. Van den Berghe, J. Jaeken, Adenylosuccinase deWciency, Adv. Exp. Med. Biol. 195 (Pt A) (1986) 27–33. S.M. Delaney, J.D. Geiger, Brain regional levels of adenosine and adenosine nucleotides in rats killed by high-energy focused microwave irradiation, J. Neurosci. Methods 64 (2) (1996) 151–156. M. Alexiou, H.J. Leese, Purine utilisation, de novo synthesis and degradation in mouse preimplantation embryos, Development 114 (1) (1992) 185–192. G. Van den Berghe, F. Bontemps, M.F. Vincent, F. Van den Bergh, The purine nucleotide cycle and its molecular defects, Prog. Neurobiol. 39 (5) (1992) 547–561. J. Zeman, J. Krijt, L. Stratilova, H. Hansikova, L. Wenchich, S. Kmoch, P. Chrastina, J. Houstek, Abnormalities in succinylpurines in

E.K. Spiegel et al. / Molecular Genetics and Metabolism 89 (2006) 19–31

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

fumarase deWciency: possible role in pathogenesis of CNS impairment, J. Inherit. Metab. Dis. 23 (4) (2000) 371–374. T.W. Stone, L.A. Roberts, B.J. Morris, P.A. Jones, H.A. Ogilvy, W.M. Behan, J.A. Duley, H.A. Simmonds, M.F. Vincent, G. van den Berghe, Succinylpurines induce neuronal damage in the rat brain, Adv. Exp. Med. Biol. 431 (1998) 185–189. M.F. Vincent, G. Van den Berghe, InXuence of succinylpurines on the binding of adenosine to a particulate fraction of rat cerebral cortex, Adv. Exp. Med. Biol. 253B (1989) 441–445. C. Salerno, C. Crifo, O. Giardini, Adenylosuccinase deWciency: a patient with impaired erythrocyte activity and anomalous response to intravenous fructose, J. Inherit. Metab. Dis. 18 (5) (1995) 602–608. J. Jaeken, S.K. Wadman, M. Duran, F.J. van Sprang, F.A. Beemer, R.A. Holl, P.M. Theunissen, P. de Cock, F. van den Bergh, M.F. Vincent, G. Van den Berghe, Adenylosuccinase deWciency: an inborn error of purine nucleotide synthesis, Eur. J. Pediatr. 148 (2) (1988) 126–131. B.A. Barshop, A.S. Alberts, H.E. Gruber, Kinetic studies of mutant human adenylosuccinase, Biochim. Biophys. Acta 999 (1) (1989) 19–23. E.A. Toth, T.O. Yeates, The structure of adenylosuccinate lyase, an enzyme with dual activity in the de novo purine biosynthetic pathway, Struct. Fold Des. 8 (2) (2000) 163–174. E.A. Toth, C. Worby, J.E. Dixon, E.R. Goedken, S. Marqusee, T.O. Yeates, The crystal structure of adenylosuccinate lyase from Pyrobaculum aerophilum reveals an intracellular protein with three disulWde bonds, J. Mol. Biol. 301 (2) (2000) 433–450. T.T. Lee, C. Worby, Z.Q. Bao, J.E. Dixon, R.F. Colman, His68 and His141 are critical contributors to the intersubunit catalytic site of adenylosuccinate lyase of Bacillus subtilis, Biochemistry 8 (1) (1999) 22–32. J.L. Brosius, R.F. Colman, A key role in catalysis for His89 of adenylosuccinate lyase of Bacillus subtilis, Biochemistry 39 (44) (2000) 13336–13343. P. Mazur, W.J. Henzel, S. Mattoo, J.W. Kozarich, 3-Carboxy-cis,cismuconate lactonizing enzyme from Neurospora crassa: an alternate cycloisomerase motif, J. Bacteriol. 176 (6) (1994) 1718–1728. J. Yang, Y. Wang, E.M. Woolridge, V. Arora, G.A. Petsko, J.W. Kozarich, D. Ringe, Crystal structure of 3-carboxy-cis,cis-muconate lactonizing enzyme from Pseudomonas putida, a fumarase class II type cycloisomerase: enzyme evolution in parallel pathways, Biochemistry 43 (32) (2004) 10424–10434. J.L. Brosius, R.F. Colman, Three subunits contribute amino acids to the active site of tetrameric adenylosuccinate lyase: Lys268 and Glu275 are required, Biochemistry 41 (7) (2002) 2217–2226. T.T. Lee, C. Worby, J.E. Dixon, R.F. Colman, IdentiWcation of His141 in the active site of Bacillus subtilis adenylosuccinate lyase by aYnity labeling with 6-(4-bromo-2,3-dioxobutyl)thioadenosine 5⬘-monophosphate, J. Biol. Chem. 272 (1) (1997) 458–465. T.T. Lee, C. Worby, Z.Q. Bao, J.E. Dixon, R.F. Colman, Implication of His68 in the substrate site of Bacillus subtilis adenylosuccinate lyase by mutagenesis and aYnity labeling with 2-[(4-bromo-2,3-dioxobutyl)thio]adenosine 5⬘-monophosphate, Biochemistry 37 (23) (1998) 8481–8489. J.B. Palenchar, R.F. Colman, Characterization of a mutant Bacillus subtilis adenylosuccinate lyase equivalent to a mutant enzyme found in human adenylosuccinate lyase deWciency: asparagine 276 plays an important structural role, Biochemistry 42 (7) (2003) 1831–1841. M.L. Segall, R.F. Colman, Gln212, Asn270, and Arg301 are critical for catalysis by adenylosuccinate lyase from Bacillus subtilis, Biochemistry 43 (23) (2004) 7391–7402. P.C. Ng, S. HenikoV, Predicting deleterious amino acid substitutions, Genome Res. 11 (5) (2001) 863–874.

31

[54] M.P. Miller, S. Kumar, Understanding human disease mutations through the use of interspeciWc genetic variation, Hum. Mol. Genet. 10 (21) (2001) 2319–2328. [55] D. Vitkup, C. Sander, G.M. Church, The amino-acid mutational spectrum of human genetic disease, Genome Biol. 4 (11) (2003) R72. [56] V. Race, S. Marie, M.F. Vincent, G. Van den Berghe, Clinical, biochemical and molecular genetic correlations in adenylosuccinate lyase deWciency, Hum. Mol. Genet. 9 (14) (2000) 2159–2165. [57] R.L. Stone, H. Zalkin, J.E. Dixon, Expression, puriWcation, and kinetic characterization of recombinant human adenylosuccinate lyase, J. Biol. Chem. 268 (26) (1993) 19710–19716. [58] J.B. Palenchar, J.M. Crocco, R.F. Colman, The characterization of mutant Bacillus subtilis adenylosuccinate lyases corresponding to severe human adenylosuccinate lyase deWciencies, Protein Sci. 12 (8) (2003) 1694–1705. [59] M. Castro, C. Perez-Cerda, B. Merinero, M.J. Garcia, J. Bernar, A. Gil Nagel, J. Torres, M. Bermudez, P. Garavito, S. Marie, F. Vincent, G. Van den Berghe, M. Ugarte, Screening for adenylosuccinate lyase deWciency: clinical, biochemical and molecular Wndings in four patients, Neuropediatrics 33 (4) (2002) 186–189. [60] F.A. vandenBergh, A.N. Bosschaart, G. Hageman, M. Duran, B. TienPoll-The, Adenylosuccinase deWciency with neonatal onset severe epileptic seizures and sudden death, Neuropediatrics 29 (1) (1998) 51–53. [61] I. Sebesta, J. Krijt, S. Kmoch, H. Hartmannova, M. Wojda, J. Zeman, Adenylosuccinase deWciency: clinical and biochemical Wndings in 5 Czech patients, J. Inherit. Metab. Dis. 20 (3) (1997) 343–344. [62] A.M. Marinaki, M. Champion, M.A. Kurian, H.A. Simmonds, S. Marie, M.F. Vincent, G. van den Berghe, J.A. Duley, L.D. Fairbanks, Adenylosuccinate lyase deWciency–Wrst British case, Nucleosides Nucleotides Nucleic Acids 23 (8–9) (2004) 1231–1233. [63] M. Holder-Espinasse, S. Marie, G. Bourrouillou, I. Ceballos-Picot, M.C. Nassogne, L. Faivre, J. Amiel, A. Munnich, M.F. Vincent, V. Cormier-Daire, Towards a suggestive facial dysmorphism in adenylosuccinate lyase deWciency? J. Med. Genet. 39 (6) (2002) 440–442. [64] M. Kohler, B. Assmann, C. Brautigam, W. Storm, S. Marie, M.F. Vincent, G. Van den Berghe, H.A. Simmonds, G.F. HoVmann, Adenylosuccinase deWciency: possibly underdiagnosed encephalopathy with variable clinical features, Eur. J. Paediatr. Neurol. 3 (1) (1999) 3–6. [65] D. Verginelli, B. Luckow, C. Crifo, C. Salerno, M. Gross, IdentiWcation of new mutations in the adenylosuccinate lyase gene associated with impaired enzyme activity in lymphocytes and red blood cells, Biochim. Biophys. Acta 1406 (1) (1998) 81–84. [66] J. Jaeken, F. Van den Bergh, M.F. Vincent, P. Casaer, G. VandenBerghe, Adenylosuccinase deWciency: a newly recognized variant, J. Inherit. Metab. Dis. 15 (3) (1992) 416–418. [67] P.D. Maaswinkel-Mooij, L.A. Laan, W. Onkenhout, O.F. Brouwer, J. Jaeken, B.J. Poorthuis, Adenylosuccinase deWciency presenting with epilepsy in early infancy, J. Inherit. Metab. Dis. 20 (4) (1997) 606–607. [68] D. Patterson, Biochemical genetics of Chinese hamster cell mutants with deviant purine metabolism. IV. Isolation of a mutant wich accumulates adenylosuccinic acid and succinylaminoimidazole carboxamide ribotide, Somatic Cell Genetics 3 (1976) 189–203. [69] A.S. Tu, D. Patterson, Biochemical genetics of Chinese hamster cell mutants with deviant purine metabolism.VI. Enzymatic studies of two mutants unable to convert inosinic acid to adenylic acid, Biochem. Genet. 15 (1–2) (1977) 195–210. [70] J.N. Crawley, Designing mouse behavioral tasks relevant to autisticlike behaviors, Mental Retardation and Developmental Disabilities Research Reviews 10 (2004) 248–258. [71] S.S. Moy, J.J. Nadler, A. Perez, R.P. Barbaro, J.M. Johns, T.R. Magnuson, J. Piven, J.N. Crawley, Sociability and preference for social novelty in Wve inbred strains: an approach to assess autistic-like behavior in mice, Genes, Brain, and Behavior 3 (2004) 287–302.