Scholarly article on topic 'Mass spectrometric characterisation of a condensation product between porphobilinogen and indolyl-3-acryloylglycine in urine of patients with acute intermittent porphyria'

Mass spectrometric characterisation of a condensation product between porphobilinogen and indolyl-3-acryloylglycine in urine of patients with acute intermittent porphyria Academic research paper on "Chemical sciences"

0
0
Share paper
Academic journal
Journal of Mass Spectrometry
OECD Field of science
Keywords
{""}

Academic research paper on topic "Mass spectrometric characterisation of a condensation product between porphobilinogen and indolyl-3-acryloylglycine in urine of patients with acute intermittent porphyria"

Research article

Received: 27 January 2015 Revised: 1 April 2015 Accepted: 7 April 2015 Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jms.3603

Mass spectrometric characterisation of a condensation product between porphobilinogen and indolyl-3-acryloylglycine in urine of patients with acute intermittent porphyria

Josep Marcos,a,b Maria Ibanez,c Rosa Ventura,a,b Jordi Segura,a,b Jordi To-Figuerasd and Oscar J. Pozoa*

We document the presence of a previously unknown species in the urine of patients with acute intermittent porphyria (AIP). The compound was fully characterised by liquid chromatography tandem mass spectrometry. Interpretation of both full spectrum acquisition and product ion spectra acquired in positive and negative ionisation modes by quadrupole time of flight MS allowed for the identification of a condensation product arising from porphobilinogen (PBG, increased in the urine of AIP patients) and indolyl-3-acryloylglycine (IAG, derived from indolylacrylic acid and present in human urine). The structure was unequivocally confirmed through comparison between the selected reaction monitoring chromatograms obtained from the urinary species and the condensation product qualitatively synthesised in the laboratory. Owing to the large amounts of both PBG and IAG in urine of AIP patients, the possible ex vivo formation of PBG-IAG in urine samples was evaluated. The product was spontaneously formed at room temperature, at 4°C and even during storage at — 20°C when spiking a control sample with PBG. A positive correlation was found between PBG and PBG-IAG in samples collected from AIP patients. However, no correlation was found between PBG-IAG and IAG. Purified PBG-IAG did not form the characteristic chromogen after application of p-dimethylaminobenzaldehyde in HCl, thus suggesting that the current techniques used to measure PBG in urine of AIP patients based on Ehlrich's reaction do not detect this newly characterised PBG-IAG fraction. Copyright © 2015 John Wiley & Sons, Ltd.

Keywords: mass spectrometry; urine; acute intermittent porphyria;porphobilinogen; indolyl-3-acryloylglycine

Background

Porphobilinogen (PBG) is an intermediate in the biosynthesis of heme and porphyrins.111 It is produced in excess and excreted in the urine of patients with acute intermittent porphyria (AIP), a disorder that results from a deficiency of hydroxymethylbilane synthase (HMBS, EC 2.5.1.61), the third enzyme of the heme biosyn-thetic pathway.121 Carriers of mutations within the HMBS gene are at risk of presenting acute neurovisceral attacks. These result from inheritance of the HMBS enzymatic defect and induction in the liver of ¿-aminolevulinate synthase, the first rate-limiting enzyme of the heme synthesis pathway. Increase of the metabolic flux at the initial steps of the pathway combined with a reduced catalytic activity of HMBS leads to the accumulation of PBG, ¿-aminolevulinic acid (ALA) and porphyrins.131 Other types of porphyria as ALA-dehydratase deficiency, hereditary coproporphyria and porphyria variegata are also classified as acute porphyrias, because they may eventually also present overproduction of heme precursors.141 The biochemical landmark of an acute porphyria attack is a marked increase in the urinary concentration of PBG, a colourless pyrrole formed from the condensation of two molecules of ALA through the catalytic activity of the enzyme ALA-dehydratase.[5,6] Seminal studies in the 1950s lead to the identification and characterisation of PBG in the urine of porphyric patients.[2] The

compound was found to form a chromogen with Ehrlich's aldehyde reagent [p-dimethylaminobenzaldehyde (DMAB) in HCl] yielding a red pigment.[7] It was found, also, that fluorescent uroporphyrins were present in the porphyric urine as a result of spontaneous condensation of PBG in acid media.

Measurement of PBG in urine has been used for decades for the diagnosis of acute porphyrias. The preferred method is based on

* Correspondence to: Oscar J. Pozo, Bioanalysis Research Group, IMIM, Hospital del Mar Medical Research Institute, Doctor Aiguader 88, Barcelona 08003, Spain. E-mail: opozo@imim.es

a Bioanalysis Research Group, IMIM, Hospital del Mar Medical Research Institute, Doctor Aiguader 88, Barcelona 08003, Spain

b Department of Experimental and Health Sciences, Universitat Pompeu Fabra, Doctor Aiguader 88, Barcelona 08003, Spain

c Research Institute for Pesticides and Water, Universitat Jaume I, Avda. Sos Baynat, Castellón E-12071, Spain

d Biochemistry and Molecular Genetics Department, Hospital Clínic, IDIBAPS, University of Barcelona, Villarrroel 170, Barcelona 08036, Spain

Abbreviations: AIP, acute intermittent porphyria; HMBS, hydroxymethylbilane synthase; ALA, aminolevulinic acid; PBG, porphobilinogen.

J. Mass Spectrom. 2015, 50, 929-937

Copyright © 2015 John Wiley & Sons, Ltd.

the Mauzerall-Granicktest that involves separation of ALA and PBG from each other and from interfering substances in urine (notably urobilinogen and urea) by anion-exchange chromatography.181 Ehrlich's reagent reacts with pyrroles (e.g. PBG) forming a chromo-gen, and therefore, both ALA and PBG may be measured separately with a spectrophotometer.

The occurrence of mass spectrometric techniques coupled with both liquid chromatography (LC) and gas chromatography (GC) offered alternative tools for clinical studies.19-111 In the last decades, these techniques became the gold-standard approach in the detection, characterisation and elucidation of biomarkers.[12,13] In the case of AIP, several LC-MS/MS-based methods have been reported for the target quantification of PBG and ALA in human urine.[14-17] Additionally, both GC-MS and LC-MS/MS have been used in the evaluation of hormonal and metabolic imbalances in AIP patients.[18'19]

Inspection of the urine of AIP patients in our centre by LC-MS/MS leads to the observation of an unidentified chromatographic peak that appeared in all urines from patients with porphyria, whereas its presence was undetectable in all control samples. Mass spectra revealed a characteristic pattern different from that of PBG, ALA and porphyrins. The in-depth analysis led us to finally identify and characterise the compound as a condensation product of PBG and indolyl-3-acryloylglycine (IAG), an endogenous metabolite found in human urine.

Methods

Chemicals and materials

Porphobilinogen and IAG were obtained from Sigma-Aldrich (St. Louis, MO, USA). Ammonium formate, methanol (LC gradient grade), acetonitrile (LC gradient grade) and formic acid (LC-MS grade) were obtained from Merck (Darmstadt, Germany). Ehrlich's aldehyde reagent [DMAB (Sigma-Aldrich) in 6 N HCl] was freshly prepared prior to use (2%, w/v). Ultrapurified water was obtained using a Milli-Q purification system (Millipore Ibérica, Barcelona, Spain). The Sep-Pak Vac RC C18 (500 mg) cartridges were purchased from Waters (Milford, MA, USA).

Liquid chromatography

The LC separation was performed using an Acquity ultraperformance liquid chromatography (UPLC) BEH C18 column (2.1 x 100 mm i.d., 1.7 ^m) (Waters Associates), at a flow rate of 300 ^l/min. Water and methanol both with formic acid (0.01% v/v) and ammonium formate (1 mM) were selected as mobile phase solvents. A gradient programme was used; the percentage of organic solvent was linearly changed as follows: 0min, 1%;0.5min, 1%; 7 min, 60%; 8 min, 90%; 8.5 min, 1%; and 10 min, 1%.

High-resolution mass spectrometry (quadrupole time-of-flight MS)

An Acquity UPLC system (Waters) was interfaced to a quadrupole time-of-flight (QTOF) mass spectrometer (QTOF Xevo G2, Waters Micromass, Manchester, UK) using an orthogonal Z-spray electrospray interface. Nitrogen (Praxair, Valencia, Spain) was used as both the drying gas and the nebulizing gas. The gas flow rate was set at 1000 l/h. The resolution of the TOF mass spectrometer was approximately 20000 at full width half maximum at m/z 556. MS data were acquired over an m/z range of 50-1000 in a scan time

of 0.3 s.The microchannel plate detector potential was set to 3.7 kV. Capillary voltages of 0.7 and — 2.0 kV were used in positive and negative ionisation modes, respectively. A cone voltage of 15 V was applied. The collision gas was argon (99.995%, Praxair). The interface temperature was set to 600°C and the source temperature to 130°C. The column temperature was set to 40°C, and the samples were kept at 5°C before injection. MS/MS experiments were also performed, using a cone voltage of 15 Vand collision energies of 10, 15 and 20 eV.

Calibration of the mass axis from m/z 50 to 1000 was conducted daily with a 1 : 1 mixture of 0.05 M NaOH/5% (v/v) HCOOH diluted (1:25) with water/acetonitrile (2:8v/v). For automated accurate mass measurement, the lock-spray probe was performed, using as lock mass leucine enkephalin (10mg/l) in acetonitrile/water (1:1) containing 0.1% HCOOH, pumped at 20 ^l/min through the lockspray needle. The leucine enkephalin [M + H]+ ion (m/z 556.2771) for positive ionisation mode and [M — H] ion (m/z 554.2615) for negative ionisation were used for recalibrating the mass axis and to ensure a robust accurate mass measurement over time.

Mass spectrometry data were acquired in centroid mode and processed by the MetaboLynx XS application manager (within MassLynx v 4.1; Waters Corporation).

Low-resolution mass spectrometry (triple quadrupole)

The study was carried out using a triple quadrupole (Quattro Premier) mass spectrometer equipped with an orthogonal Z-spray-electrospray ionisation source interfaced to an Acquity UPLC system for the chromatographic separation (all from Waters Associates). Drying gas as well as nebulising gas was nitrogen. The desolvation gas flow was set to approximately 1200 l/h, and the cone gas flow was 50 l/h. A cone voltage of 25 V and a capillary voltage of 3.0 kV were used in both positive and negative ionisation modes. The nitrogen desolvation temperature was set to 450 °C, and the source temperature was 120°C.

PBG-IAG synthesis and purification

For the PBG-IAG synthesis, 2 mg of PBG and 2 mg of IAG were dissolved in 5 ml of water: methanol (1: 1) solution. The reaction mixture was maintained under agitation in a rocking mixer (20 movements/min) for 3 days at room temperature. The presence of PBG-IAG in the crude mix was confirmed by LC-MS/MS using the selected reaction monitoring (SRM) method described in Confirmation of the structure of the unknown species section.

In order to isolate PBG-IAG from the reaction mixture, a highperformance LC purification procedure was performed. A HP1090 liquid chromatograph with automatic injection system and diode array detection (Hewlett-Packard, Waldbronn, Germany) was used. The separation was achieved with an ACE 5 C18 column (250x4.6 mm i.d., 5 ^m) protected with a guard column (Advance Chromatography Technologies Ltd., Aberdeen, Scotland). The mobile phase was water and acetonitrile at a flow rate of 0.75 ml/min. Gradient elution was as follows: at 0 min, 5% of acetonitrile and at 20 min, 95%. The column effluent was monitored at 560 nm. An aliquot of 150 ^l of the reaction mixture was injected, and 0.5 min fractions were collected throughout the whole chromatogram. The fractions were evaporated to dryness and reconstituted in 1 ml of water.

The presence of PBG-IAG in the different fractions was tested by analysing 0.5 ml of these reconstituted fractions by LC-MS/MS using the method described in Confirmation of the structure of the unknown species section.

Sample preparation

For the LC-QTOF MS analysis, urine samples (2 ml) were passed through a C18 cartridge (Sep-Pak Vac RC C18, 500 mg, Waters), previously conditioned with 2 ml of methanol and 2 ml of water. The column was then washed with 2 ml of water, and the analytes were eluted with 2 ml of methanol. The eluate was evaporated to dryness under a nitrogen stream in a water bath at 50 °C. Then, the residue was dissolved in 150^l of a mixture of water and acetonitrile (9: 1 v/v), and 10 ^l was directly injected into the LC-QTOF MS system.

For the LC-MS/MS analysis using the triple quadrupole instrument, urine samples (150 ^l) were twofold diluted with acetic acid; 1 M and 10 ^l were directly injected into the LC-MS/MS system.

Urine samples

The unknown species was detected in the urine of 24 adult Caucasian Spanish patients with biochemical active AIP (22 women and 2 men, age ranging from 22 to 54years). All these AIP patients were high excretors of PBG and were either patients with recurrent attacks receiving heme arginate (n = 6, urine collected before treatment) or asymptomatic patients attending the hospital for clinical follow-up (n =18).

Urine samples from these patients were collected at the Hospital Clinic of Barcelona according to approved protocols. All patients were informed of the purpose of the study, and written consents were obtained.

Second morning urine samples after admission from all patients and controls were obtained between 09.00 and 10.00 h. Aliquots were immediately protected from light and frozen at — 80°C until analyses (interval time between collection and frozen ~1-2h). They were defrosted to analyse heme precursors and porphyrins, and the presence of the unknown peak was investigated. In agreement with the active AIP status, all samples were found to have increased concentrations of PBG (>0.8 mmol/mol creatinine), ALA (>5 mmol/mol creatinine) and porphyrins (>35 ^mol/mmol creatinine). PBG and ALA were analysed by ALA/PBG Column Test (BIORAD, based on the Mauzerall-Granick test) and porphyrins by fluorimetry and high-performance LC by standard methods.[20]

The absence of the unknown chromatographic peak in healthy population was checked by analysing samples collected from 24 healthy volunteers (20 women and 4 men; age 25-45 years) recruited from the laboratory staff.

The research was conducted in accordance with the Declaration of Helsinki Principles and was approved by the Hospital Clinic Ethic Committee.

to the synthesised PBG-IAG. The relative abundance of every transition was calculated in every sample and compared with the obtained in the synthesised material. Specific ion transitions were considered as those not showing any interference in the urine from healthy volunteers. Deviations lower than 20% in the ratio of the abundances of all specific transitions were considered as characteristics for unequivocal confirmation of the structure.

Quantification of PBG, IAG and PBG-IAG in urine

Urinary concentrations of PBG, IAG and PBG-IAG were calculated by LC-MS/MS using the triple quadrupole instrument. Urine samples were treated as described in Sample preparation section. Of the diluted sample, 1 ^L was injected into the system. The LC gradient described in Liquid chromatography section was used. The most abundant transition for each analyte was monitored (PBG: 210^94, cone 20V, collision energy 15eV;IAG: 245^ 170, cone 20V, collision energy 10eV;and PBG-IAG: 454^379, cone 20V, collision energy 10 eV). Quantification was performed against a calibration curve. Because of the absence of a certified material, PBG-IAG was quantified using the calibration curve for PBG.

Spontaneous formation of PBG-IAG in urine

The spontaneous formation of PBG-IAG in human urine under different storage conditions was checked by adding PBG to urine samples collected from two healthy volunteers to a final concentration of 1 mg/ml. Samples were aliquoted and stored at three different conditions: room temperature, 4°C and — 20 °C. Aliquots stored at room temperature and at 4 °C were analysed immediately before spiking and thereafter every day during a week and after 30days of storage. Samples stored at — 20 °C were analysed 0, 1, 2, 3, 7 and 30 days after storage.

In order to minimise the formation of PBG-IAG during sample preparation, the spiking process was performed at 4°C. The results of the aliquots analysed immediately after spiking were used as a reference to evaluate the potential formation during this process.

PBG-IAG response to the Ehrlich's reagent

In order to determine whether PBG-IAG reacts with the Ehrlich's reagent, we measured the ultraviolet spectra of the purified PBG-IAG before and after the addition of 0.5 ml of the Ehrlich's reagent. The measurements were carried out on an Agilent 8453 spectropho-tometer (Agilent Technologies, Palo Alto, CA, USA). Additionally, the same experiment was conducted with a 25 ^g/ml solution of PBG as a positive control for the test.

Confirmation of the structure of the unknown species

An SRM method including the maximum number of ion transitions was developed in the triple quadrupole instrument for the confirmation of the structure of the unknown species.1211 The SRM method monitored the ten ion transitions in positive ionisation mode: 454^379 (10eV), 454^333 (20eV), 454^319 (20eV), 454 ^ 291 (20 eV), 454 ^ 273 (20 eV), 454 ^ 168 (20 eV), 454 ^ 144 (20 eV), 454^ 245 (30 eV), 454 ^ 233 (30 eV) and 454 ^ 232 (30 eV); and seven ion transitions in negative mode: 469 ^ 425 (10 eV), 469 ^ 408 (20 eV), 469 ^ 333 (20 eV), 469 ^ 307 (20 eV), 469^ 289 (20 eV), 469 ^ 263 (30 eV) and 469 ^ 100 (30 eV).

The SRM method was applied to samples collected from patients (n = 24), to samples collected from healthy volunteers (n = 24) and

Statistical analysis

In order to reduce the variability due to urine dilution, the specific gravity of all samples was measured, and the urinary concentrations of the analytes were normalised to a specific gravity of 1.020 by applying the following formula:

C1.020 = Csample x (1.020 — 1)/(specific gravity^^ — 1)

Data were analysed using the spss software (v 18.0; IBM, New York, NY, USA). Correlations between the adduct PBG-IAG and its precursors (PBG and IAG) were evaluated. Because the low number of data analysed hampers the assumption of normality, the statistical analysis was conducted using nonparametric tests. A Spearman's correlation with a two-tailed test was used for the

evaluation of the correlation among all the compounds analysed. Statistical significance was set at p<0.01.

Results and discussion

Mass spectrometric behaviour of the unknown species

The LC-MS study of urine samples collected from AIP patients compared with those obtained from healthy volunteers revealed a prominent chromatographic peak in all patients' urine (Fig. 1(a)), which was absent in the control group (Fig. 1(b)). This unknown species was evaluated by LC-QTOF analysis.

In positive ionisation mode, the species showed a base peak at m/z 454.1609 corresponding to a formula of C23H24N3O7 with a mass error of —0.5 mDa (Fig. 1 (c)). Taking into account that the unknown compound could be related to PBG and that PBG is mainly ionised in positive mode as [M + H—NH3]+,[14-17] a more detailed study was necessary for establishing the molecular formula of the species.

The evaluation of minor adducts formed in positive ionisation mode revealed the presence of an ion at m/z 493.1678 with an associated formula of C23H26N4O7Na (mass error of 2.1 mDa), which could correspond to the [M + Na]+ of the analyte. Additionally, the evaluation of the results in negative ionisation mode showed an abundant ion at m/z 469.1723 corresponding to a formula of C23H25N4O7 (mass error of 0 mDa), which was associated to the [M — H]— (Fig. 1(d)). The combination of this information allowed concluding that the molecular formula of the unknown compound was C23H26N4O7.

The full-scan behaviour of the unknown species, i.e. ionisation as [M + H—NH3]+ in positive mode and as [M — H]— in negative mode, was similar to the behaviour observed for PBG. This fact suggested that the unknown compound showed a PBG-like structure maintaining the methylamino group responsible of the prominent [M + H—NH3]+ and at least one of the carboxylic acid moieties responsible for the [M — H]—.

In order to compile more structural information about the unknown compound, its product ion spectra were deeply studied.

Product ion spectra of the unknown species

Product ion spectra of the unknown species at different collision energies are summarised in Table 1. Similarly to the results extracted from the full-scan experiments, several of the product ions obtained in the MS/MS experiments in positive ionisation mode supported the fact that the unknown species had a PBG-like structure. Remarkably, after a neutral loss of 75 Da, neutral losses of formic acid, acetic acid and a combination of both were observed at m/z 333.1236, 319.1085 and 273.1023, respectively. These losses are related with the propanoic and ethanoic acid moieties present at C3' and C4' of the pyrrolic ring of the PBG and are similar to the MS/MS behaviour previously described for PBG-related compounds.[22] In negative ionisation mode, several fragments, which can be explained after losses of CO2 (469^425, 408^364 and 307 ^ 263), were observed indicating the presence of several carboxylic acids. All this information allowed us to propose the presence of an intact PBG unit in the unknown structure. Taking into account that a PBG unit counts for C10H14N2O4, the rest of the molecule should count for C13H12N2O3.

T, Time 0 111111 M 11111111111111

111 " 11 Time

3.50 4.00 4.50 5.00 5.50 3.50 4.00 4.50 5.00 5.50

454.1609

455.1641

493.1678

I ""I.....I...............................................|llll|llll|lll

400 420 440 460 480 500 520

469.1723

C3—'I'f■■ ■ II 'i-i II I f-i I-I-i /-f■■11-r I■■ ■■ I 11■>■■ f rI-r 11r-i■■ 11■■ ■ I-f I f■■ -r I f I -l-r■■ 11■■ I-II-i-I■■ ■ m/z 400 420 440 460 480 500 520

Figure 1. LC-MS chromatograms in the same scale for urine sample collected from (a) an acute intermittent porphyria patient and (b) a healthy volunteer. Full spectrum acquisition obtained for the unknown peak (c) in positive ionisation mode and (d) negative ionisation mode.

Table 1. Product ion spectra data for the unknown species Ionisation Collision Ion Abundance Formula mode energy (eV) (m/z) (%) Error (mDa)

ESI+ 10 379.1286 100 C21H19N2O5 0.8

333.1237 10 C20H17N2O3 0.2

319.1078 10 C19H15N2O3 0.5

291.1121 5 C18H15N2O2 1.3

262.0714 7 c13h12no5 0.1

20 379.1278 10 C21H19N2O5 1.6

333.1236 35 C20H17N2O3 0.3

319.1085 60 C19H15N2O3 0.2

291.1131 100 C18H15N2O2 0.3

273.1023 45 C,8H,3N2O 0.5

246.1132 15 C14H16NO3 0.2

245.1072 20 C17H13N2 0.7

233.1064 20 c13h15no3 1.2

218.0813 20 c12h12no3 0.4

168.0652 30 C8H,0NO3 0.9

144.0438 55 C9H6NO 1.1

30 319.1071 15 C19H15N2O3 1.3

291.1130 40 C18H15N2O2 0.4

273.1023 30 C^H^^O 0.5

247.1213 45 C14H17NO3 0.5

246.1136 55 C14H16NO3 0.6

245.1073 60 C17H13N2 0.6

233.1065 100 c13h15no3 1.3

232.0986 60 c13h14no3 1.2

168.0657 20 C8H10NO3 0.4

150.0542 30 C8H8NO2 1.3

144.0437 70 C9H6NO 1.2

ESI- 10 469.1726 100 C23H25N4O7 0.3

425.1819 20 C22H25N4O5 0.6

408.1565 40 C22H22N3O5 0.6

20 425.1819 20 C22H25N4O5 0.6

408.1559 85 C22H22N3O5 0.0

396.1562 25 C21H22N3O5 0.3

364.1651 10 C21H22N3O3 1.0

333.1239 50 C20H17N2O3 0.0

307.1454 75 C19H19N2O2 0.7

289.1340 95 C19H17N2O 0.1

263.1557 30 C18H19N2 0.9

247.1103 35 C16H13N3 0.6

100.0036 100 C3H2NO3 0.1

30 308.1375 15 C15H20N2O5 0.3

289.1341 45 C19H17N2O 0.0

263.1551 100 C18H19N2 0.3

247.1229 20 C,7H,5N2 0.6

189.1027 30 c„h13n2o 0.1

116.0498 55 C8H6N 0.2

100.0038 95 c3h2no3 0.3

74.0242 10 c2h4no2 3.1

ESI, electrospray ionisation source. Precursor ion at m/z 454 in ESI+ and at m/z 469 in ESI—.

Additional relevant information was extracted from the product ion spectra. The fragmentation ofthe [M + H—NH3]+ at low collision energy (10eV) resulted in a spectrum in which the precursor ion

was absent and which was clearly dominated by the ion at m/z 379.1286 (Table 1). This ion corresponds to a neutral loss of C2H5NO2, which can be associated with the amino acid glycine. This hypothesis was supported by the presence of the ion at m/z 74.0242 in negative ionisation mode corresponding to deprotonated glycine (Table 1). Additionally, a neutral loss of C2H5NO2 from the ion at m/z 408.1559 is the most suitable explanation for the formation of the ion at m/z 333.1239. All these results taken together suggested that the unknown species is a glycine conjugate.

Additionally, some ions suggested the presence of an indole moiety in the structure. Among them, in positive ionisation mode, the neutral loss of C8H7N (from m/z 379.1286 to m/z 262.0714) could correspond to the neutral loss of an indole group, and the presence ofthe ion at m/z 144.0438 (C9H6NO) could be associated to a keto-indole. In negative ionisation mode, the ion at m/z 116.0498 (C8H6N) could correspond with a deprotonated indole.

Based on these results, a bibliographic search was performed for endogenous indolic compounds, which are excreted as conjugated with glycine. The most likely structure was found to be IAG, a major component of the urine, which molecular formula of C13H12N2O3 was in agreement with the expected formula of the non-PBG part of the unknown species.

Based on the MS and MS/MS behaviours ofthe unknown species, the condensation product between PBG and IAG was proposed as the most feasible structure for the compound. This structure would explain the product ions observed in the MS/MS experiments (Fig. 2).

Confirmation of the proposed structure

Porphobilinogen contains several reactive groups, which would explain its relative low stability in urine. Among them, the C2' of the pyrrolic ring is a nucleophilic centre that usually reacts with the electrophilic C6' of an additional molecule of PBG to form uroporphyrinogen I. However, C2' can also potentially react with other electrophilic centres like the C3 of IAG by a Michael's 1,4-addition (Fig. 3).

In order to confirm this hypothesis, PBG and IAG were mixed and stirred for 3 days at room temperature, as described in PBG-IAG synthesis and purification section. The reaction product exhibited an abundant [M + H—NH3]+ at m/z 454.1609 in positive ionisation mode and a [M — H]— at m/z 469.1721 in negative ionisation mode similar to the behaviour observed for the unknown species. The mixture was purified by LC, and PBG-IAG was found only in fraction from 10 to 10.5 min, most likely corresponding to the peakeluting at 10.1 min, which exhibits a maximum absorption at 503 nm.

To unequivocally confirm that the condensation product synthe-sised and purified was the unknown species detected in urine of AIP patients, the relative abundances ofthe 17 selected ion transitions were compared. Because unexpected endogenous interferences could potentially interfere in the detection of some of the transitions, their specificity was tested by analysing 24 urines collected from healthy volunteers. None ofthe selected transitions showed any interference in the urine samples from healthy population. Therefore, all selected transitions were considered for the confirmation of the structure.

Results obtained for the confirmation ofthe structure are shown in Table 2. Relative abundances of all transitions were equivalent in all AIP patients' urine and similar to the purified condensation product. Therefore, it could be unequivocally confirmed that the unknown species found in urine is a condensation product between PBG and IAG.

Figure 2. Fragmentation pathway proposed for porphobilinogen-acute intermittent porphyria (PBG-IAG) in positive ionisation mode.

Correlation between PBG, IAG and PBG-IAG

The analysis of urinary PBG, IAG and PBG-IAG revealed that the amounts of urinary PBG-IAG were in the range of the microgram/millilitre accounting for approximately 10% of the PBG present (Table 3).

The application of the Spearman test showed a positive significant correlation (r=0.641, p = 0.001) between PBG and the adduct PBG-IAG. Although IAG is also a precursor of the adduct PBG-IAG, there was no correlation (r = 0.103, p = 0.632) between this urinary component and PBG-IAG (Table 3).

PBG-IAG

Figure 3. Reaction scheme for the formation of porphobilinogen-acute intermittent porphyria (PBG-IAG).

Table 2. Confirmation of porphobilinogen indolyl-3-acryloylglycine structure

ESI+ ESI—

Transition Synthesised material Real sample Transition Synthesised material Real sample

Average RSD Average RSD

454^379 100 100 — 469 ^ 425 54 55 22

454^333 40 39 13 469 ^ 408 100 100 —

454^319 35 33 11 469 ^ 333 50 47 12

454 ^ 291 13 14 6 469 ^ 307 43 48 13

454^273 6 6 8 469 ^ 289 68 70 16

454 ^ 245 14 14 13 469 ^ 263 61 66 16

454^233 16 15 9 469^100 98 102 21

454^232 7 6 11

454 ^ 168 2 2 13

454 ^ 144 13 12 8

ESI, electrospray ionisation source.

Relative abundances (%) of the transitions selected for porphobilinogen indolyl-3-acryloylglycine in the synthesised material and in the real urine samples collected from acute intermittent porphyria patients (n = 24).

Table 3. Concentrations (цд/ml) of urinary PBG, IAG and PBG-IAG and correlation between PBG-IAG and its precursors

Concentration (вд/ml) Correlation with PBG-IAG

Median Range Coefficient Significance

PBG IAG PBG-IAG 46 0.1 -366 10 3.1 -23 4.5 0.8-18 0.641** 0.103 1 0.001** 0.632

PBG, porphobilinogen; IAG, indolyl-3-acryloylglycine. ** Statistically significant p < 0.01.

Evaluation of the spontaneous formation of PBG-IAG in urine

Owing to the high reactivity of PBG and the large amounts of IAG found in urine, the potential spontaneous formation of PBG-IAG in urine was checked under different storage conditions. For this purpose, urine samples from two healthy volunteers were spiked with large amounts of PBG (1 mg/ml) and stored under three different conditions (room temperature, 4°C and — 20°C). Results were similar for both volunteers and are summarised in Fig. 4.

Samples analysed just after spiking (t = 0) showed undetectable amounts of PBG-IAG. This fact demonstrated that PBG-IAG was not produced in any step ofthe sample pretreatment. When storing

the sample at room temperature, a substantial and continuous increase of the amounts of PBG-IAG was observed during the first week. However, PBG-IAG seemed to be degraded after 1 month at room temperature. As expected, slower spontaneous formation of PBG-IAG was observed when storing the sample at 4°C. In this case, the increase was gradual during the first month of storage; at day 30, the observed abundance was comparable with those obtained after 1 week at room temperature (Fig. 4).

Porphobilinogen-IAG spontaneous formation was also observed at — 20°C. A gradual increase of the species was detected during the storage month (Fig. 4). This result has to be taken into account when determining PBG in old frozen samples from AIP patients, because even under these conditions, PBG is being gradually transformed to PBG-IAG.

Response of PBG-IAG to the Ehrlich's reagent

Porphobilinogen reacts with DMAB in acid solution (Ehrlich's reagent) to form a reddish-mauve-coloured compound with an absorption maximum at 553 nm. This colorimetric reaction is used to qualitatively detect increases of PBG in urine of patients with AIP (Hoesch test and Watson-Schwartz test). These tests may lack specificity because of interferences present in the urine;therefore, current quantitative tests use Ehrlich's reaction but introduce a previous anion-exchange chromatography procedure to remove urea and other urinary interferents.[8]

Figure 4. Evolution of the abundance of the porphobilinogen-acute intermittent porphyria peak along time under different storage conditions.

We investigated if a pure solution of PBG-AIG would react with the Ehrlich's reagent in order to know if the current PBG quantification tests would detect the PBG-IAG fraction.

An approximate concentration of PBG-IAG of 25 ^g/ml in the purified fraction was estimated by LC-MS in scan mode using PBG as reference material because both of them are mainly ionised as [M + H-NH3]+. In order to determine whether PBG-IAG reacts with the Ehrlich's reagent, we measured the ultraviolet/visible spectra of the purified PBG-IAG before and after the addition of 0.5 ml of the Ehrlich's reagent to 0.5 ml of the fraction containing the PBG-IAG. The collected PBG-IAG fraction remained colourless after the addition of the reagent, and the absorbance at 553 nm changed from 0.001 to 0.031 AU (data not shown).

The same experiment was conducted with a 25 ^g/ml solution of PBG. In this case, the characteristic cherry red colour was observed after reaction with the Ehrlich's reagent. As expected, a notable increase in the absorbance at 553 nm was observed, from 0.002 to 1.302 AU.

Implications of the PBG-IAG formation

Porphobilinogen is found in large amounts in the urine of AIP patients due to hepatic overproduction and high excretion of both heme precursors, PBG and ALA. Conversely, IAG is a regular constituent of human urine although its origin is not completely understood.[23,24] It is generally assumed that bacterial activity in the intestine could be a main source of IAG in human urine. Bacterial flora would catabolise the amino acid tryptophan to indole derivatives that would be absorbed and converted to indolylacrylic acid, which in turn can be transformed into IAG after conjugation with glycine.

However, the possibility that indolylacrylic acid may also have an endogenous origin arising from tryptophan metabolism, without intervention of intestinal microorganisms, cannot be excluded.[25] Moreover, indolylacrylic acid and/or IAG levels have been found to be modified in a number of disease states, including Hartnup disease,[26] light sensitive dermatitis,[27] phenylketonuria[28] and autism.[29] This would strengthen the hypothesis of IAG arising from endogenous tryptophan metabolism.

Liquid chromatography-MS/MS technology has allowed unequivocal identification of PBG-IAG in the urine of AIP patients. This shows that the state-of-the art technology may allow identification of previously unknown specific compounds in urine of AIP patients. Nevertheless, the existence of uncharacterised pigments in the porphyric urine was already anticipated in seminal studies that investigated the chemistry of PBG (i.e. porphobilin) but without a definitive characterisation.[30]

It is known that condensation of excess PBG in acidic urine could yield fluorescent uroporphyrins. Our study reports a new chemical entity, PBG-IAG, in the very complex porphyric urine. The conjugate seems to be spontaneously formed, not only in heating conditions but also at room temperature and even in freezing conditions. Therefore, a certain amount of PBG-IAG will inevitably be formed in urine of AIP patients during their storage. According to our results, the PBG-IAG conjugate does not form the red chromogen when exposed to Ehrlich's reagent. Consequently, conventional techniques for measuring PBG in urine (i.e. based on the Mauzerall-Granick test) will only detect free unconjugated PBG but not the PBG-IAG fraction.

Even if it is unlikely that a diagnosis of AIP would be missed because of the formation of this adduct (which may represent about 10% of total PBG in urine), good laboratory practices seeking to maximise accuracy in urinary PBG measurement should try to minimise its formation, by shortening the delay between sample collection and analysis. Moreover, long-time frozen urines used for, i.e. retrospective studies, may contain significant amounts of the adduct, which may lead to an underestimation of the total concentration of urinary PBG.

Conclusions

Our study confirms the identification in the urine of patients with AIP of a previously unreported compound, namely, PBG-IAG. State-of-the-art QTOF MS and LC-MS/MS technologies have allowed the unequivocal identification of PBG-IAG in the urine of AIP patients.

Most of this product seems to arise from the spontaneous condensation of PBG and IAG in human urine. However, our data do not allow to discard a fresh urinary emission already containing certain amounts of the adduct (i.e. formed in the urinary bladder or even in tissues).

We have shown that this ex vivo formation occurs even when urine samples are stored at — 20 °C. Because PBG-IAG do not react with the Ehrlich's reagent, the quantitation of PBG by the conventional colorimetric tests will only account for the PBG present in the free form and therefore may lead to an underestimation if samples have been stored for a long period. The possibility exists that a similar reaction may occur in vivo or even that PBG in tissues may react with other endogenous electrophilic molecules with similar structure to indolylacrylic acid (e.g. tryptophan metabolites). If this would be the case, the formation in vivo of these adducts could modify the metabolism of glycine or tryptophan both being possibly involved in acute porphyria attacks.[31] However, further research is needed to confirm these hypotheses.

Acknowledgements

This work was supported by grants from Instituto de Salud Carlos III FEDER, (CP/10/00576), the Spanish 'Fondo de Investigación Sanitaria' (PI11/00767) to Jordi To-Figueras and from the Generalitat de Catalunya (2014 SGR 692). Technical support of Nuria Renau and mass spectrometric discussions with Juan V. Sancho are acknowledged.

References

[1] F. T. Prunty. Acute porphyria: some properties of porphobilinogen. Biochem.J. 1945, 39,446.

[2] J. Waldenstrom. Studien uber porphyrie. Act. Med. Scand. Suppl. 1939, 82,1.

[3] H. Puy, L. Gouya, J. C. Deybach. Porphyrias. Lancet 2010, 375, 924.

[4] K. E. Anderson, S. S. Sassa, D. F. Bishop, R. J. Desnick. In The Metabolic Basis of Inherited Disease(C. R. Scriver, A. L. Beaudet, W. S. Sly, D. Valle Eds.). McGraw-Hill: New York, 2001.

[5] J. E. Falk, E. I. Dresel, C. Rimington. Porphobilinogen as a porphyrin precursor, and interconversion of porphyrins, in a tissue system. Nature 1953, 172, 292.

[6] E. I. Dresel, J. E. Falk. Conversion of alpha-aminolaevulinic acid to porphobilinogen in a tissue system. Nature 1953, 172,1185.

[7] C. J. Watson, S. Schwartz. A simple test for urinary porphobilinogen. Exp. Biol. Med. (Maywood) 1941,47, 393.

[8] D. Mauzerall, S. Granick. The occurrence and determination of 5-aminolevulinic acid and porphobilinogen in urine. J. Biol. Chem. 1956, 219,435.

[9] H. H. Maurer. Current role of liquid chromatography-mass spectrometry in clinical and forensic toxicology. Anal. Bioanal. Chem. 2007, 388,1315.

[10] K. S. Leung, B. M. W. Fong. LC-MS/MS in the routine clinical laboratory: has its time come? Anal. Bioanal. Chem. 2014,406,2289.

[11] D. H. Chace. Mass spectrometry in the clinical lab. Chem. Rev. 2001,101, 445.

[12] S. Neumann, S. Boecker. Computational mass spectrometry for metabolomics: identification of metabolites and small molecules. Anal. Bioanal. Chem. 2010, 398, 2779.

[13] W. B. Dunn, D. I. Ellis. Metabolomics: current analytical platforms and methodologies. TrAC-Trends Anal. Chem. 2005,24,285.

[14] J. Zhang, M.Yasuda, R.J. Desnick, M. Balwani, D. Bishop, C.Yu. A LC-MS/ MS method for the simultaneous quantification of 5-aminolevulinic acid and porphobilinogen. J. Chromatogr. B 2011, 879,2389.

[15] R. E. Ford, M. J. Magera, K. M. Kloke, P. A. Chezick, A. Fauq, J. P. McConnell. Quantitative measurement of porphobilinogen in urine by stable-isotope dilution liquid chromatography-tandem mass spectrometry. Clin. Chem. 2001,47,1627.

[16] C. M. Benton, L. Couchman, J. T. Marsden, D. C. Rees, C. Moniz, C. K. Lim. Direct and simultaneous determination of 5-aminolaevulinic acid and porphobilinogen in urine by hydrophilic interaction liquid chromatography-electrospray ionization/tandem mass spectrometry. Biomed. Chromatogr. 2012, 26, 1033.

[17] G. A. Lord, J. L. Luo, C. K. Lim. Capillary zone electrophoresis/mass spectrometry of 5-aminolaevunilic acid and porphobilinogen. Rapid Commun. Mass Spectrom. 2000, 14, 314.

[18] G.Casals,J.Marcos,O.J.Pozo,P.Aguilera,C.Herrero,J.To-Figueras.Gas chromatography-mass spectrometry profiling of steroids in urine of patients with acute intermittent porphyria. Clin. Biochem. 2013,46,819.

[19] O. J. Pozo, J. Marcos, A. Fabregat, R. Ventura, G. Casals, P. Aguilera, J. Segura, J. To-Figueras. Adrenal hormonal imbalance in acute intermittent porphyria patients: results of a case control study. Orphanet J. Rare Dis. 2014, 9, 54.

[20] A. C. Deacon, G. H. Elder. ACP best practice no 165: front line tests for the investigation of suspected porphyria. J. Clin. Pathol. 2001,54,500.

[21] O. J. Pozo, J. V. Sancho, M. Ibanez, F. Hernandez, W. M. A. Niessen. Confirmation of organic micropollutants detected in environmental samples by liquid chromatography tandem mass spectrometry: achievements and pitfalls. TrAC-Trends Anal. Chem. 2006,25,1030.

[22] M. Danton, C. K. Lim. High-performance liquid chromatography/ electrospray ionization tandem mass spectrometry of hydroxylated uroporphyrin and urochlorin derivates formed by photochemical oxidation of uroporphyrinogen I. Biomed. Chromatogr. 2007, 21, 534.

[23] E. Marklova, I. M. Hais. A method for the estimation of indolylacryloylglycine in urine. Clin. Chim. Acta 1972, 40, 455.

[24] E. Marklova. Where does indolylacrylic acid come from? Amino Acids 1999, 17,401.

[25] E. Marklova. Tryptophan metabolism by the indoleacrylic acid pathway. Cesk. Farm. 1988, 37, 275.

[26] M. D. Milne, M. A. Crawford, C. B. Girao, L. Loughridge. The metabolic disorder in Hartnup disease. Q. J. Med. 1960,29,407.

[27] J. Kimmig, W. Sticherlung, R.Tschesche, H. G. Urbach. Isolation of N-[b (3-indolyl)acryloyl]glycine from urine and its synthesis. Hoppe-Seyler's Z. Physiol. Chem. 1958, 311, 234.

[28] G. Strohmayer. Nachweis von indolylacrylsaure bei der Phenylke tonurie. Arch. Klin. Exp. Dermatol. 1964,219, 743.

[29] R. J. Anderson, D. J. Bendell, I. Garnett, P. W. Groundwater, W. J. Lough, M. J. Mills, D. Savery, P. E. G. Shattock. Identification of indolyl-3-acryloylglycine in the urine of people with autism. J. Pharm. Pharmacol. 2002, 54, 295.

[30] V. E. Hawkinson, C. J. Watson. The separation of porphobilinogen and an Ehrlich negative precursor of uroporphyrin. Science 1952, 115,496.

[31] M. Carichon, N. Pallet, C. Schmitt, T. Lefebvre, L. Gouya, N. Talbi, J. C. Deybach, P. Beaune, P. Vasos, H. Puy, G. Bertho. Urinary metabolic fingerprint of acute intermittent porphyria analyzed by (1)H NMR spectroscopy. Anal. Chem. 2014, 86, 2166.