The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240

Rat kidney porphobilinogen deaminase kineticsDetection of enzyme–substrate complexes

Guillermo Noriegaa, Guillermo Matteib, Alcira Batllea, Adela Ana Juknata,∗a Centro de Investigaciones sobre Porfirinas y Porfirias (CIPYP), CONICET, Departamento de Quı́mica Biológica,

1428 Buenos Aires, Argentinab Departamento de Fı́sica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria,

1428 Buenos Aires, Argentina

Received 23 January 2002; received in revised form 21 March 2002; accepted 25 March 2002

Abstract

Background and aims: Acute intermittent porphyria (AIP) is an inherited disease resulting from a reduced activity of theenzyme porphobilinogen deaminase (PBG-D). The kidney is an important target for numerous porphyrinogenic drugs and itmay contribute to the clinical manifestations of porphyric attacks. An evaluation of kidney PBG-D role in the AIP pathophys-iology requires detailed information on kidney PBG-D properties, under normal conditions.Methods: Rat kidney PBG-Dwas purified to homogeneity and initial reaction velocities were calculated by measuring uroporphyrinogen I formation atpH 8.2 for different incubation times (0–20 min) and over a wide range of substrate concentrations (0.8–66�M). Results:Purified rat kidney PBG-D is a monomeric enzyme showing only a single protein band after SDS–PAGE, Western blot andisoelectric focusing (pI 4.9). Its molecular mass is 40±2.3 kDa, determined by SDS–PAGE and 39.8±2 kDa by gel filtrationchromatography. Rat kidney PBG-D has an unusual kinetic behaviour, exhibiting a deviation from the Michaelis–Mentenhyperbola. PBG-D kinetic data required a fitting to an equation of higher degree, leading to the following apparent kineticconstants:K1 = 2.08±0.01�M andK2 = 0.102±0.003�M. Conclusion: The values of these constants fulfil the restriction4K2 ≤ K2

1, necessary for the occurrence of isoenzymes, interpreted in this work as enzyme–substrate intermediates. Theinitial reaction velocity expression here defined, correlates with an enzyme carrying only one active site but allowing, throughconformational changes, the detection of at least two enzyme–substrate intermediates formed during PBG-D reaction.© 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Porphobilinogen deaminase; Kinetic mechanism; Enzyme–substrate complexes; Isoenzymes; Rat kidney

Abbreviations: AIP, acute intermittent porphyria; DTNB, 5,5′-dithiobis (2-nitrobenzoic acid); DPC, diethyl pirocarbonate; HMB,hydroymethylbilane; NEMI,N-ethylmaleimide; PBG, porphobili-nogen; PBG-D, porphobilinogen deaminase; PCMB,p-chloromer-curibenzoate

∗ Corresponding author. Present address: Beruti 3244, 4 Piso-Dto.B, 1425 Buenos Aires, Argentina. Fax:+54-11-4-345-1203.

E-mail addresses: aajuknat@mail.retina.ar, aajuknat@qb.fcen.uba.ar (A.A. Juknat).

1. Introduction

Porphobilinogen-deaminase (PBG-D; EC 4.3.1.8)also known as hydroxymethylbilane synthase, catal-yses the head to tail polymerization of the monopy-rrol porphobilinogen (PBG) to synthesize the highlyunstable linear tetrapyrrole 1-hydroxymethylbilane(HMB) or pre-uroporphyrinogen[1–3]. HMB is trans-formed by uroporphyrinogen III synthase (Isomerase;EC 4.2.1.75) into uroporphyrinogen III[3,4], the first

1357-2725/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved.PII: S1357-2725(02)00051-1

G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240 1231

cyclic precursor of a family of tetrapyrroles includinghaem, siroheams, chlorophylls, bacteriochlorophylls,cytochromes, factor F430 and Vitamin-B12.

PBG-D has been purified to homogeneity from awide variety of prokaryotic and eukaryotic sourcesand their properties have been well established[5–11].A dipyrromethane cofactor, unique prosthetic groupmade up of two PBG-derived pyrroles linked to-gether, has been first described for theEscherichiacoli enzyme where it has been found to be cova-lently bound to cysteine-242[12,13]. Two isoformsof PBG-D, which differ in their N-terminal aminoacidsequence, have been first identified in humans: anerythroid-specific isoenzyme of 42 kDa[14], ex-pressed only in erythroid cells and the so-calledhousekeeping isoform of 44 kDa[14] present in allcell types, both encoded by a single structural gene[15]. The PBG-D gene is transcribed into two differentmRNAs through alternative splicing of two primarytranscripts arising from two promoters[15,16]: theerythroid promoter, induced during erythropoyesis andthe ubiquitous PBG-D gene promoter. However, theexistence of an additional erythroid-specific isoformof PBG-D (PBG-D-erythroid alternative: PBG-D-EA)mRNA in primary erythroid cell cultures, bone mar-row and fetal liver, has been recently reported[17].

Sequences of cloned cDNAs and nuclear genesencoding PBG-D have been reported from a varietyof organisms such as human erythrocytes[15,16],Euglena gracilis [18], yeast [19], E. coli [20,21],mouse[22] and rat [23]. In all cases, the PBG-Dsequence obtained showed considerable homology,indicating the existence of a structurally conservedenzyme throughout the different species.

Because PBG-D activity has been found reducedin different tissues of patients with acute intermit-tent porphyria (AIP) and PBG-D assay in red bloodcells is the method of choice for its diagnosis[24],extensive information exists on PBG-D from humanerythrocytes and its regulation in liver and erythro-poietic tissue[6,15,16,24–26]. However, consideringthe fact that symptoms can be present in many tissuesduring AIP attacks, so far, less attention has beenpaid to porphyrin biosynthesis and heme metabolismin other than those above-mentioned tissues.

As it is known, the kidney is a very important targetorgan for numerous porphyrinogenic drugs and it maycontribute to the clinical manifestations of porphyric

attacks[27]. The fact that porphyrin biosynthesis isnot homogenously distributed throughout all renalcells and that heme regulation in the kidney is morerefractory to modulation than in the liver[27], stimu-lated our interest in studying tetrapyrrole metabolismin the kidney. Moreover, to the best of our knowledge,no studies on purified PBG-D had been reported yet.

PBG-D isoenzymes separated by DEAE–Sephadexchromatography and gel electrophoresis have beendescribed in human erythrocytes[28]. In subse-quent research, multiple forms representing stableenzyme–substrate covalent intermediates have beenreported for human erythrocytes[6] and rat spleen[29]. Recent work in our laboratory showed an un-usual kinetic behaviour for rat kidney PBG-D, whoseexperimental data can be interpreted considering theoccurrence of different enzyme–substrate complexes.

2. Materials and methods

2.1. Materials

PBG was purchased from Sigma (St. Louis, MO)and Sephadex G-25, Sephadex G-100, Q-SepharoseFast Flow and Blue Sepharose CL-4B were obtainedfrom Pharmacia Fine Chemicals (Uppsala, Sweden).All other chemicals were of the highest purity avail-able from different commercial sources.

2.2. Animals

The present study complied with the Argentine reg-ulations for care and use of laboratory animals. MaleChbb Thom rats (250–300 g) maintained on Purinapellet food and water ad libitum, were housed undera 12 h light/12 h dark cycle (lights on from 6:00 to18:00 h). Rats were killed between 10:00 and 11:00a.m., under light ether anesthesia by cardiac puncturebleeding. The kidneys were immediately removed,previous perfusion in situ with ice cold 0.9% NaCl.They were rinsed with the same solution and blottedon filter paper. Kidneys were stored at−20◦C untilbeing used.

2.3. Enzyme purification

All operations were performed at 4◦C, unless oth-erwise stated. Homogenate preparation and PBG-D

1232 G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240

purification steps were carried out as already described[30,31] with some modifications. Kidneys were ho-mogenized in 50 mM Tris–HCl buffer pH 8.2 (10%; gwet tissue/ml buffer) by using a Potter–Elvehjem glasshomogenizer equipped with a motor driven Teflon pes-tle. The suspension was then centrifuged at 24 000×g

for 20 min. The ensuing supernatant was kept at 70◦Cfor 5 min and centrifugated again at 24 000× g for10 min. Solid ammonium sulfate was added to the fi-nal supernatant until 35% saturation was reached. Theprecipitated protein was centrifuged and discarded.The fraction precipitating between 35 and 70% satura-tion containing PBG-D activity was collected, desaltedthrough a Sephadex G-25 column (2 cm× 16 cm)and concentrated using Centripep-10 concentrators(Amicon, Beverly, USA). The concentrated fractionfrom the previous step was applied to a SephadexG-100 column (2.5 cm× 50 cm; 200 ml bed volume)equilibrated with 50 mM Tris–HCl buffer pH 8.2(buffer A) and eluted with the same buffer at a flow rateof 9 ml/h. Fractions containing PBG-D activity werecombined, concentrated and applied to a column ofQ-Sepharose FF (2.0 cm× 25 cm; 80 ml bed volume)previously equilibrated with buffer A. The enzyme waseluted from the column at a flow rate of 45 ml/h, witha linear gradient of 0–1.0 M NaCl in the same buffer.Active fractions collected between 0.5 and 0.6 M NaClwere pooled and dialyzed overnight against 20 mMTris–HCl buffer pH 8.2 (buffer B). This fraction wasapplied to a Blue-Sepharose column (1.0 cm× 16 cm;12 ml bed volume) equilibrated in buffer B. The col-umn was washed with buffer B containing 0.3 MNaCl and the enzyme was eluted at a flow rate of24 ml/h, with a linear gradient of 0.3–1.3 M NaClin the same buffer. The active fractions were pooledand dialyzed overnight against buffer A. The dialyzedfraction from the preceeding step was subjected to apreparative polyacrylamide gel electrophoresis undernon-denaturing conditions. The gel was cut into 5 mmlength sections and enzyme activity was determinedby incubating each of the gel segments, using thestandard incubation system but at 60◦C. Fluorescencedue to uroporphyrin I formation, was measured in aShimadzu spectrofluorometer RF-510 as previouslydescribed[32]. The active band was electroelutedwith a Bio-Rad model 422 device. During PBG-Dpurification, identification of porphyrins formed[33]as well as determination of the isomeric composition

of uroporphyrin fractions[34] were performed byHPLC. Protein concentration was determined by themethod of Bradford[35] employing bovine serumalbumin (BSA) as standard or by measuring the ab-sorbance at 280 nm. Here it is important to emphasizethat only very small amounts of the highly purifiedenzyme have been isolated and that the N-terminalsequence of the protein was found to be blocked.

2.4. Studies on initial reaction velocities

For studies on the initial reaction velocities, thestandard incubation system contained 500�l PBG-Dpreparation, 50 mM Tris–HCl buffer pH 8.2 and sub-strate PBG (0.8–66�M) in a final volume of 1.0 ml.The reaction was started by substrate addition andincubations were carried out in the dark under aer-obic conditions, with mechanical shaking, at 37◦Cfor different times (0–20 min). Blanks were alwaysrun omitting either PBG or the enzyme preparationand uroporphyrins from the blanks were substracted.The reaction was stopped by addition of concentratedHCl (final concentration 5%). The uroporphyrinogensformed were oxidized by exposing them to white light(60 W) for 20 min. The precipitated protein was dis-carded by centrifugation and total uroporphyrin I wasdetermined in acid solution[36] in a Hewlett-PackardDiode Array 8452A spectrophotometer. The productformation proceeded linearly with time under all thekinetic experimental conditions. Only measurementswith an optimal linear regression (r ≥ 0.99) were usedfor calculations.

2.5. SDS-polyacrylamide gel electrophoresis

Homogeneity of purified rat kidney PBG-D wasdetermined by SDS–PAGE as described by Laemmli[37]. Runs were performed in a Bio-Rad Mini-ProteanII Dual Vertical Electrophoresis System at 170 V forapproximately 60 min. Protein bands on the gels werevisualized by Coomassie brilliant blue R-250 or bysilver staining.

2.6. Western blots

After SDS–PAGE, proteins were transferred tonitrocellulose membrane in 20 mM Tris, 192 mMglycine pH 8.3, 20% methanol (v/v), at 100 V for 1 h

G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240 1233

using the Bio-Rad Mini Trans-Blot ElectrophoreticTransfer System. The membrane was blocked with5% BSA in Tris-buffered saline (150 mM NaCl,50 mM Tris–HCl pH 8.0) containing 0.1% SDS(TBS buffer) and incubated with anti-PBG-D anti-body (dilution 1: 1000) (Juknat, Doernemann andSenger, unpublished data) for 1 h at room tempera-ture. After extensive washing with 2% BSA in TBSbuffer, the membrane was further incubated withgoat anti-rabbit IgG (dilution 1: 1000; Sigma, St.Louis, MO) for 1 h and again thoroughly washedin the same buffer. The last incubation was carriedout with the alkaline–phosphatase conjugated rabbitanti-goat IgG (dilution 1: 1000; Sigma, St. Louis,MO) for 1 h followed by similar washing as describedabove. Immunoreactive bands were revealed usingp-nitrotetrazolium blue (NBT; Sigma; St. Louis, MO)and 5-bromo-4-chloro-3-indolyl phosphate (BCIP;Sigma; St. Louis, MO) as chromogenic substratesin developing buffer (100 mM Tris–HCl buffer pH9.0/100 mM NaCl/5 mM MgCl2).

2.7. Isoelectric focusing

Isoelectric focusing was performed in a 4.5% (w/v)polyacrylamide gel containing carrier ampholytes pH3–10 (Merck, Darmstadt, Germany) following themanufacturer’s procedure of Pharmacia (Uppsala,Sweden). Before sample application, a pre-focusingwas carried out for 30 min. Gels were focused at 8◦Cand 6 W for 1 h. After running, one gel was stainedwith Coomassie brilliant blue R-250 and the otherwas cut into 5 mm length sections and assayed for pH.

2.8. Relative molecular mass determination

The relative molecular mass (Mr) of PBG-D wasdetermined by gel filtration as well as by SDS–PAGE.Concerning gel filtration, the enzyme and the stan-dards were applied to a Sephadex G-100 column(2.5 cm × 50 cm; bed volume, 200 ml) and to aSephacryl S-200 column (2 cm× 40 cm; bed vol-ume, 70 ml), both previously equilibrated with bufferA. Protein content and enzyme activity were deter-mined in each column eluate. The standards usedwere�-globulin (160 kDa), BSA (67 kDa), ovalbumin(45 kDa), carbonic anhydrase (29 kDa), ribonuclease(25 kDa) and cytochrome c (12.5 kDa). TheMr for

PBG-D was calculated from linear regression analy-sis of a plot of logMr against the partition coefficientKav. To determine theMr of the denatured enzymeby SDS–PAGE, the gel was run as described above,until the bromophenol blue reached the bottom of thegel. The electrophoretic movilities (m) of the markers(MW-SDS-70; Sigma, St. Louis, MO, USA) werecalculated and logMr against m was plotted.

2.9. Analysis of data

All experiments were carried out at least four timesin triplicate unless otherwise stated, with different ratkidney enzyme preparations, in order to obtain a meanvalue and a standard deviation for each initial rate.Residual patterns and analysis of Michaelis–Mentenkinetic data were performed by using the Leonoraprogram for enzyme data[38].

2.10. Curve fitting of experimental data

We assumed a steady state with product con-centration zero and absence of significant enzymeassociation–dissociation. A linear relationship be-tween the initial reaction velocity (v) and the enzymeconcentration (E0) was found in all experiments, in-dicating absence of enzyme aggregation phenomena.Curve fitting and statistical treatment were performedusing the Microcal Origin 4.1 interactive programand the Statistica Software. Kinetic sets of data werefitted by non-linear regression methods to rate equa-tions of degree higher than that of Michaelis–Menten.In all cases, parameters were constrained to be pos-itive as they are required for physical meaning.Curve fitting was carried out iteratively using theχ2-test until the weighted sum of squares becameminimized:

χ2J =

[∑ni=1wi(vi − viJ)

2]

n − pJ

where wi is the statistical weight,vi refers to theinitial rate obtained in theith experimental point,viJ

the theoretical velocity calculated according to thejthmodel,n the number of experimental points andpJ thenumber of parameters included in model J. To discrim-inate between fittings to rate equations of differentsuccessive degrees, the statisticalF-test was applied,as described by Burguillo et al.[39] and F values

1234 G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240

were referred to the corresponding confidence levelof 95%. The choice of the best mathematical modelwas also followed by examination of residual plots[38].

3. Results

3.1. Electrophoretic purity and identificationof PBG-D

The purity of the enzyme was assessed at vari-ous stages of purification, under both non-denaturing(Fig. 1A) and denaturing conditions (Fig. 1B). Afterthe final preparative gel electrophoresis, the proteinmigrated as a single band in SDS–PAGE and innon-denaturing conditions (Fig. 1C). The staining foractivity allowed an unequivocal identification of theprotein band as PBG-D. Additional proof of puritycame from Western blots (Fig. 1D) and isoelectric fo-cusing (data not shown), where a single protein bandwas also obtained.

Results of SDS–PAGE revealed only a band of40 ± 2.3 kDa, which is in agreement with datafrom the native enzyme determined by gel filtration(39.8 ± 2 kDa), showing that PBG-D is a monomericprotein. These results are consistent with those re-ported for PBG-D purified from different sources

Fig. 1. Purity of PBG-D from rat kidney. PBG-D was analysedon a 7.5% polyacrylamide gel, under non-denaturing (lanes Aand C) and denaturing conditions (lanes B and D). Lanes A andB, Blue-Sepharose fraction; lane C, homogeneous PBG-D afterpreparative electrophoresis; lane D, Western blot corresponding tofraction from lane C. Protein band identified as PBG-D is denotedby an arrow.

(34–44 kDa) ([10] and references therein) and withthe molecular mass of 39.361 kDa, calculated fromthe 361 aminoacid sequence corresponding to the ratubiquitous isoform[23]. The isoelectric point (pI) ofPBG-D, determined by isoelectric focusing on poly-acrylamide gels, was found to be 4.9, in good agree-ment with pI values ranging from 4.2 to 5.2, reportedby other authors ([11] and references therein).

3.2. Properties of PBG-D

In order to obtain the optimum pH, enzyme ac-tivity was determined in 50 mM citrate–phosphate,Tris–HCl and glycine–NaOH buffers in the range ofpH 5.0 to 10.6. A typical bell-shaped curve was ob-tained with an optimum pH of 8.2 in 50 mM Tris–HClbuffer (data not shown). Activity was indetectablebelow pH 5.0 and above pH 10. This optimumpH obtained is within the range (pH 7.4–8.2) re-ported for other sources ([11] and references therein,[40]).

Rat kidney PBG-D showed a remarkable thermalstability with little loss of activity even after heatingat 75◦C for 20 min. Heat-stability has already beenreported for this enzyme in many other sources andit could be explained by the large number of exist-ing protein-cofactor interactions[11]. PBG-D couldbe stored with no loss of activity for 35 days at−20◦Cbut after 4 months a 30% of the initial activity waslost.

Enzyme activity reached its maximum at 65◦Cand the activation energy calculated from theArrhenius plot was found to be 19.3 kcal/mol (datanot shown), a value similar to those reported forE. gracilis (19.7 kcal/mol) [41] and rat harderiangland (18.2 kcal/mol)[31].

Activity was also assayed in the presence ofprotein-modification reagents. Diethyl pirocarbonate(DPC) (1.0 mM) reduced PBG-D activity (20 minincubation with 66�M PBG) by 62% suggestingthat histidine residues exist at or near the activesite of the enzyme. The sulfhydryl reagents, such asN-ethylmaleimide (NEMI),p-chloromercuribenzoate(PCMB) and 5,5′-dithiobis (2-nitrobenzoic acid)(DTNB) at 1.0 mM did not significantly affect PBG-Dactivity, however higher concentrations such as5.0 mM were shown to inhibit the enzyme activity(40–100%) (data not shown).

G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240 1235

3.3. Studies on PBG-D initial reaction velocities

Experiments were performed by measuring prod-uct formation at pH 8.2 for different incubation times(0–20 min) and over a wide range of PBG concen-trations (0.8–66�M). Initial reaction velocities werecalculated from the typical progress curves, underthe incubation conditions indicated inSection 2.Activity measured below 0.8�M PBG was too lowto be detected reproducibly. No substrate inhibitionwas observed with concentrations of PBG up to130�M.

As it is shown in Fig. 2, initial reaction veloc-ity rapidly increases with PBG concentration butit does not obey the Michaelis–Menten equationwhen compared with the theoretical curve predictedby the hyperbolic model using Leonora program.This deviation was confirmed making residual plotswith different combinations of variables, one ofthem can be seen in the inset ofFig. 2. As ex-pected, Lineweaver-Burk and Eadie-Hofstee (v/[S]versusv) plots were non-linear supporting further anon-hyperbolic behaviour for rat kidney PBG-D (datanot shown).

Fig. 2. Michaelis–Menten plot for rat kidney PBG-D. Initial re-action velocity was measured as described inSection 2and itwas expressed as pmol of uroporphyrin I per min per mg protein.Data are means of four independent experiments, carried out intriplicate assays. Theoretical curve results from fitting experimen-tal data (�) to the expression corresponding to Michaelis–Mentenpredicted by the Leonora program (χ2 = 31.6; P < 0.009). Inset:plot of residuals vs. predicted velocities.

Fig. 3. Kinetic behaviour plot for rat kidney PBG-D. Initial reactionvelocity was measured as indicated inSection 2. Theoretical curveresults from fitting experimental data (�) to the Eq. (1) definedin Results (χ2 = 0.2062; P > 0.05), which in theF-test hada significance of 99.5% against the model of Michaelis–Menten.Inset: Plot of residuals vs. predicted velocities.

3.4. Model fitting to kinetic data

Deviations from Michaelis–Menten kinetics shownin Fig. 2are statistically significant. Thus, a rate equa-tion with higher degree is required in order to describekidney PBG-D kinetics. The best fit for our experi-mental data (Fig. 3) was obtained employing thevversus [S] expression corresponding toEq. (1)

v = V1[S] + V2[S]2

1 + K1[S] + K2[S]2(1)

which led to the following apparent kinetic constants:V1 = 1017.41 ± 2.87 pmol/min mg,K1 = 2.08 ±0.01�M; V2 = 86.10 ± 0.19 pmol/min mg,K2 =0.102± 0.003�M.

Fitting to rate equation I was statistically better thanfitting to the Michaelis–Menten equation, at a signifi-cance level of 99.5% according to theF-test, whereasin no case in different experiments, was such an im-provement produced with higher degrees with respectto Eq. (1).

The apparentKm values for rat kidney PBG-Dare similar to those reported for human erythrocytes(6.2�M at pH 8.2) [40], rat spleen (1�M at pH7.5) and rat harderian gland (1.1�M at pH 8.0),but lower than those corresponding to PBG-D from

1236 G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240

Scenedesmus obliquus (79�M at pH 7.4),Chlorellaregularis (85–90�M at pH 7.4) and E. gracilis(195�M at pH 8.0) ([31] and references therein).

3.5. Kinetic studies on different purifiedpreparations

Kinetic studies were also performed during thePBG-D purification procedure. So, fractions of dif-ferent degree of purification, arising from SephadexG-100, Q-Sepharose and Blue-Sepharose wereemployed. Initial reaction velocities were mea-sured as indicated above and Lineweaver-Burk andEadie-Hofstee (v/[S] versus v) plots were ana-lyzed. Incubations of Sephadex G-100 prepara-tions and dialyzed fractions of Q-Sepharose andBlue-Sepharose yielded non-linear double recipro-cal plots and non-hyperbolic response following thesame kinetic behaviour described byEq. (1). How-ever, when Q-Sepharose and Blue Sepharose fractionswere incubated in the presence of 1.0 M NaCl, be-fore dialysis, PBG-D reaction proceeded by way ofMichaelis–Menten kinetics. The Eadie-Hofstee plotwas linear with a regression line of 0.99369 (Fig. 4).ApparentKm andVmax values determined at pH 8.2for the Blue-Sepharose PBG-D fraction and calcu-lated with the Leonora program, were found to be

Fig. 4. Eadie–Hofstee plot for rat kidney PBG-D. Initial reactionvelocity was measured as indicated inSection 2, in the presenceof 1.0 M NaCl, employing the Blue-Sepharose fraction. Theoreti-cal curve results from fitting experimental data (�) to expressioncorresponding to Michaelis–Menten predicted by the Leonora pro-gram (χ2 = 83.12; P > 0.05).

0.855± 0.013�M and 439.73 ± 1.01 pmol/min mg,respectively. These results were confirmed by incu-bating both dialyzed fractions with 1.0 M NaCl.

It has been shown that heme biosynthesis is differ-entially distributed in the kidney, with greater activityin the cortex than in the medulla[27] and that PBG-Dimmunoreactivity is more pronounced in distal andcollecting tubules than in the rest of the kidney[42].However, we did not find any differences in the activityand kinetic behaviour of PBG-D purified from eithercortex or medulla. On the other hand, NaCl has beenfound to be distributed in the kidney in a gradient wayincreasing from the cortex to the medulla and reach-ing concentrations as high as 1200 mM in the Henle’sloop. Although, the 1.0 M NaCl concentration testedin this work is very close to that found in vivo, wedo not know yet, whether or not this PBG-D complexkinetic behaviour has any physiological significance.

3.6. Conformational flexibility of PBG-Dactive site

As it was reported, a local flexibility in PBG-Dactive site is necessary to allow cofactor attachment,substrate binding and pyrrole chain polymerization[43]. In an attempt to analyze this possibility, confor-mational changes in rat kidney PBG-D were moni-tored by the extent of histidine (Fig. 5A) or cysteine(Fig. 5B) exposure, which were measured by fluores-cence changes, in the absence or presence of differentPBG concentrations, with and without 1.0 M NaCl.Initially, there were no significant changes in histi-dine fluorescence with PBG up to 7.0�M, but it de-creased at higher PBG concentrations (75% at 66�M)(Fig. 5A). In the presence of 1.0 M NaCl, histidinefluorescence decreased only 29% at 0.8�M PBG,reaching a plateau at higher substrate concentrations.When PBG-D was pre-incubated with 1.0 mM DPC,no further changes in histidine fluorescence in thepresence of PBG and/or 1.0 M NaCl were found (datanot shown). These observations lead us to concludethat the addition of substrate molecules produces aconformational change that finally buries the histidinegroup in the native enzyme, a situation that can notbe developed in the presence of NaCl or after DPCmodification.

In marked contrast, cysteine residue exposure in-creased as a function of PBG concentration specially

G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240 1237

Fig. 5. Relative fluorescence of histidine and cysteine residues.(A) Histidine fluorescence: excitation and emission wavelenghtswere 247 nm and 490 nm respectively, with a spectral bandwidth of10 nm. (B) Cysteine fluorescence: excitation and emission wave-lenghts were 278 nm and 561 nm respectively, with a spectralbandwidth of 10 nm. Purified PBG-D was pre-incubated for 15 minwith different PBG concentrations, as indicated in the figure, inthe absence (�) or presence of 1.0 M NaCl (). All data are fromat least four independent experiments.

in the presence of 1.0 M NaCl (two-fold enhancementat 66�M) (Fig. 5B). All of these results are consistentwith conformational changes occurring in the pres-ence of different PBG concentrations.

4. Discussion

Rat kidney kinetic studies showed that thePBG-D reaction has an unusual mechanism anddoes not obey Michaelis–Menten behaviour (Fig. 2).This non-hyperbolic response was confirmed byLineweaver-Burk and Eadie-Hofstee plots, where thepoints followed a curved shape instead of the knownstraight line predicted by the Michaelis–Mentenequation. These findings were not expected since ahyperbolic relationship between initial velocity andPBG concentration has been reported for PBG-Dfrom most sources ([10,25] and references therein).However, Eadie-Hofstee plots for human erythro-cytes PBG-D[25] were bell-shaped indicating theexistence of weakly positive cooperative effects andPBG-D from soybean callus exhibited many of thecharacteristics of negative cooperativity[44].

It is known that enzymes which catalyze reactionsinvolving four substrate molecules, lead to com-plex initial velocity rate equations and non-linearreciprocal plots[45]. On the other hand, it is wellestablished that HMB is formed through a reactionwhere PBG-D utilizes four molecules of the samesubstrate. The proposed mechanism involves a singleactive center which catalyzes the deamination of thesubstrate PBG, the condensation of the deaminatedsubstrate with either the dipyrromethane cofactor oran enzyme–substrate complex (ES1, ES2 or ES3) andthe final hydrolitic cleavage of ES4 to yield the HMBproduct. Furthermore, two pyrrole recognition siteshave been defined in the catalytic center, site S forPBG binding and its deamination, and site C, whichis occupied by the permanently bound cofactor andready to react with the deaminated substrate translo-cated from site S[46]. Interaction of apo-deaminasewith pre-uroporphyrinogen originates directly theholoenzyme yielding the ES2 intermediate complex.So, it shows that the cofactor rings remain perma-nently attached to the holoenzyme[47]. On the otherhand, a local flexibility, allowing little conformationaladjustments associated with PBG binding, with attach-ment of the deaminated substrate to the cofactor, withelongation of the pyrrole chain and with release of theHMB product, is necessary for enzyme activity[43].

Purified rat kidney PBG-D revealed the presenceof a single protein band in native polyacrylamide gelelectrophoresis and SDS–PAGE. Additional proofs

1238 G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240

came from Western blots and isoelectric focusing. Itis therefore obvious from our results that deviationsfrom Michaelis–Menten kinetic cannot be due to en-zyme subunit interactions or generated by site-siteinteractions derived from different subunits, becauserat kidney PBG-D is made up of only one polypeptidechain and exhibits only one catalytic site per molecule.

Jordan et al.[8] reported the detection of five activepeaks in purified PBG-D fromE. coli after Mono Qchromatography. These multiple species were inter-preted to be different structural forms of the enzymepossible due to deamination of asparagine residues.PBG-D isoenzymes and multiple forms representingstable enzyme–substrate intermediates have been de-scribed for human erythrocytes[6,28], E. coli [13],rat spleen[29], bovine liver[48], Rhodopseudomonasspheroides [49] and Arabidopsis thaliana [50]. Un-fortunately we could not isolate enzyme-intermediatecomplexes from rat kidney purified PBG-D, after fplcMono Q HR 5/5 chromatography. In this regard, itshould be pointed out that in order to analyse thepresence of enzyme-intermediate complexes, PBG-Dwas incubated and eluted under different conditions,with and without PBG (data not shown). On the otherhand, we did not find any differences in the activityand kinetic behaviour of PBG-D purified from eitherthe kidney cortex or the medulla.

Neither different protein species nor enzyme–substrate complexes could be detected through ourchromatographic and electrophoretic studies, but thevalues of our apparent kinetic constants fulfil the re-striction 4K2 ≤ K2

1, which is strong evidence of thepresence of isoenzymes[51]. So, our kinetic resultsconfirm the occurrence of rat kidney PBG-D isoen-zymes, which should be interpreted to be enzyme–substrate intermediates as they were here noticedduring incubations with the substrate PBG.

According to our experimental data, a mechanismwith a different input for PBG molecules generatingseparate elementary steps, has to be considered, asindicated in the following scheme:

where the enzyme–substrate intermediates ES1, ES2,ES3 and ES4 are formed on the same active site, dur-

ing the stepwise PBG binding to the enzyme-boundcofactor (Eholo) as shown above.

Evidence for the existence of conformationalchanges occurring in the presence of different PBGconcentrations is also here presented. Rat kid-ney PBG-D is remarkably unreactive to sulfhydrylreagents such as NEMI, PCMB and DTNB, but inthe presence of NaCl, addition of substrate inducesa conformational change leading to the exposure ofsulfhydryl groups which were previously buried inthe enzyme (Fig. 5B). These results are in agreementwith those reported forE. coli PBG-D where the na-tive enzyme was unreactive to sulfhydryl reagents butwhen they reacted with enzyme–substrate complexes,a time dependent activation was observed, speciallywith ES2 and ES3 [46].

In summary, we can conclude that the presentkinetic behaviour may be due to the detection ofenzyme–substrate intermediates, generated duringPBG-D reaction. Thev versus [S] expression (Eq. (1))here defined, can correlate with an enzyme carryingonly one active site. But, through limiting steps relatedto defined conformational changes, PBG-D kineticsallow the detection of at least two enzyme–substrateintermediates, which were not able to be identified yet.

Our present in vitro results complement those ofWolansky et al.[52] who reported that circulatingALA, taken up by rat kidney, strongly affects invivo PBG-D activity, due to the high endogenousPBG levels in the organ. These authors concludedthat at the time when ALA exerts its neurotoxic ef-fects, PBG might modulate PBG-D kinetics leadingto inhibition of porphyrin biosynthesis, increase ofenzyme–substrate intermediates and further PBG andALA accumulation. Enhanced levels of precursors inkidney might be one of the underlying causes of theneurological signs associated with AIP.

Acknowledgements

We warmly thank Dr. Noemi Steyerthal for experttechnical help in performing the isoelectric focusingand Dr. M. Wolansky for his skillful assistance incomputer data processing. Grants from the ConsejoNacional de Investigaciones Cientı́ficas y Tecnológ-icas (CONICET) and Universidad de Buenos Aires(Argentina) are gratefully acknowledged.

G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240 1239

References

[1] G. Burton, P.E. Fagerness, S. Hosozawa, P.M. Jordan,A.I. Scott, 13C NMR: evidence for a new intermediate,pre-uroporphyrinogen, in the enzymic transformation ofporphobilinogen into uroporphyrinogens I and III, J. Chem.Soc. Chem. Commun. (1979) 202–204.

[2] A.R. Battersby, C.J.R. Fookes, G.W.J. Matcham, E.McDonald, Order of assembly of the four pyrrole rings duringbiosynthesis of the natural porphyrins, J. Chem. Soc. Chem.Commun. (1979) 539–541.

[3] A.I. Scott, G. Burton, P.M. Jordan, H. Matsumoto, P.E.Fagerness, L. Pryde, J. Chem. Soc. Chem. Commun. (1980)384–387.

[4] P.M. Jordan, G. Burton, H. Nordlov, M.N. Schneider, L.Pryde, A.I. Scott, Pre-uroporphyrinogen: a substrate foruroporphyrinogen III cosynthetase, J. Chem. Soc. Chem.Commun. (1979) 204–205.

[5] A.M. del C. Batlle, M.V. Rossetti, Enzymic polymerizationof porphobilinogen into uroporphyrinogens, Review, Int. J.Biochem. 8 (1977) 251–267.

[6] P.M. Anderson, C.J. Desnick, Purification and properties ofuroporphyrinogen I synthase from human erythrocytes, J.Biol. Chem. 255 (1980) 1993–1999.

[7] G.J. Hart, C. Abell, A.R. Battersby, Purification, N-terminalamino acid sequence and properties of hydroxymethylbilanesynthase (porphobilinogen deaminase) fromEscherichia coli,Biochem. J. 240 (1986) 273–276.

[8] P.M. Jordan, S.D. Thomas, M.J. Warren, Purification,crystallization and properties of porphobilinogen deaminasefrom a recombinant strain ofEscherichia coli K12, Biochem.J. 254 (1988) 427–435.

[9] A.J. Spano, M.P. Timko, Isolation, characterization andpartial amino acid sequence of a chloroplast-localizedporphobilinogen deaminase from pea, Biochim. Biophys. Acta1076 (1991) 29–36.

[10] A.A. Juknat, D. Dörnemann, H. Senger, Purification andkinetic studies on a porphobilinogen deaminase from theunicellular green algaScenedesmus obliquus, Planta 193(1994) 123–130.

[11] R.M. Jones, P.M. Jordan, Purification and propertiesof porphobilinogen deaminase fromArabidopsis thaliana,Biochem. J. 299 (1994) 895–902.

[12] P.M. Jordan, M.J. Warren, H.J. Williams, N.J. Stolowich, C.A.Roessner, S.K. Grant, A.I. Scott, Identification of a cysteineresidue as the binding site for the dipyrromethane cofactor atthe active site ofEscherichia coli porphobilinogen deaminase,FEBS Lett. 235 (1988) 189–193.

[13] G.J. Hart, A.D. Miller, A.R. Battersby, Evidence thatthe pyrromethane cofactor of hydroxymethylbilane synthase(porphobilinogen deaminase) is bound through the sulphuratom of a cysteine residue, Biochem. J. 252 (1988) 909–912.

[14] B. Grandchamp, H. De Verneuil, C. Beaumont, S. Chretien,O. Walter, Y. Normann, Tissue-specific expression ofporphobilinogen deaminase. Two isoenzymes from a singlegene, Eur. J. Biochem. 162 (1987) 105–110.

[15] N. Raich, P.H. Romeo, A. Dubart, D. Beaupain, M.Cohen-Solal, M. Goossens, Molecular cloning and completeprimary sequence of human erythrocyte porphobilinogendeaminase, Nucl. Acids Res. 14 (1986) 5955–5968.

[16] S. Chretien, A. Dubart, D. Beaupain, N. Raich, B.Grandchamp, J. Rosa, M. Goossens, P.H. Romeo, Alternativetranscription and splicing of the human porphobilinogendeaminase gene result either in tissue-specific or inhousekeeping expression, Proc. Natl. Acad. Sci. U.S.A. 85(1988) 6–10.

[17] A.N. Gubin, J.F. Miller, Human erythroid porphobilinogendeaminase exists in two splice variants, Blood 97 (2001)815–817.

[18] A.L. Sharif, A.G. Smith, C. Abell, Isolation and charac-terization of cDNA clone for a chlorophyll synthesis enzymefrom Euglena gracilis, Eur. J. Biochem. 184 (1989) 353–359.

[19] P.L. Gellerfors, J. Saltzgaber-Müller, M.G. Douglas, Selectionby genetic complementation and characterization of the genecoding for the yeast porphobilinogen deaminase, Biochem. J.240 (1986) 673–677.

[20] S.D. Thomas, P.M. Jordan, Nucleotide sequence of the hemClocus encoding porphobilinogen deaminase ofEscherichiacoli K12, Nucl. Acids Res. 14 (1986) 6215–6226.

[21] P.A. Alefounder, C. Abell, A.R. Battersby, The sequence ofhemC, hemD and two additionalE. coli genes, Nucl. AcidsRes. 16 (1988) 9871.

[22] C. Beaumont, C. Porcher, C. Picat, Y. Nordmann, B.Grandchamp, The mouse porphobilinogen deaminase gene,J. Biol. Chem. 264 (1989) 14829–14834.

[23] C.A. Cardalda, A.M. del C. Batlle, A.A. Juknat, Sequence andstructure of the rat housekeeping PBG-D isoform, Biochem.Biophys. Res. Commun. 249 (1998) 438–443.

[24] A.M. del C. Batlle, E.A. Wider, A.M. Stella, A simple methodfor measuring erythrocyte porphobilinogenase and its use inthe diagnosis of acute intermittent porphyria, Int. J. Biochem.9 (1978) 871–875.

[25] S.A. Fumagalli, M.L. Kotler, M.V. Rossetti, A.M. del C.Batlle, Human red cell porphobilinogen deaminase. A simplermethod of purification and some unusual properties, Int. J.Biochem. 17 (1985) 485–494.

[26] E. Smythie, D.C. Williams, A simple rapid purificationscheme for hydroxymethylbilane synthase from humanerythrocytes, Biochem. J. 251 (1988) 237–241.

[27] J.S. Woods, Regulation of porphyrin and heme metabolismin the kidney, Sem. Hematol. 25 (1988) 336–348.

[28] K. Miyagi, M. Kaneshima, J. Kawakami, F. Nakada, Z.J.Petryka, C.J. Watson, Uroporphyrinogen I synthase fromhuman erythrocytes: separation, purification and propertiesof isoenzymes, Proc. Natl. Acad. Sci. U.S.A. 76 (1979)6172–6176.

[29] D.C. Williams, Characterization of the multiple forms ofhydroxymethylbilane synthase from rat spleen, Biochem. J.217 (1984) 675–683.

[30] G.O. Noriega, A.A. Juknat, A.M. del C. Batlle, Non-essentialactivation of rat liver porphobilinogen-deaminase by folicacid, Z. Naturforsch. 47c (1992) 416–419.

[31] C.A. Cardalda, A.A. Juknat, F.G. Princ, A.M. del C. Batlle,Rat harderian gland porphobilinogen deaminase: characteriza-

1240 G. Noriega et al. / The International Journal of Biochemistry & Cell Biology 34 (2002) 1230–1240

tion studies and regulatory action of protoporphyrin IX, Arch.Biochem. Biophys. 347 (1997) 69–77.

[32] F.G. Princ, A.A. Juknat, A. Batlle, Porphyrinogenesis in ratcerebellum. Effect of high�-aminolevulinic acid concen-tration, Gen. Pharmacol. 25 (1994) 761–766.

[33] S. Seubert, A. Seubert, Porphyrin Bestimmung in Harn, MerckSpectrum 2 (1989) 8–9.

[34] C.K. Lim, J.M. Rideout, D.J. Wright, Separation of porphyrinisomers by high-performance liquid chromatography, Bio-chem. J. 211 (1983) 435–438.

[35] M.M. Bradford, A rapid and sensitive method for thequantification of microgram quantities of protein utilizing theprinciple of protein-dye binding, Anal. Biochem. 72 (1976)248–254.

[36] F.G. Princ, A.G. Maxit, C.A. Cardalda, A.M. del C.Batlle, A.A. Juknat, In vivo protection by melatonin against�-aminolevulinic acid-induced oxidative damage and itsantioxidant effect on the activity of haem enzymes, J. Pineal.Res. 24 (1998) 1–4.

[37] U.K. Laemmli, Cleavage of structural proteins duringassembly of the head of bacteriophage T4, Nature 227 (1970)680–685.

[38] A. Cornish-Bowden, Least-squares analysis: basic principles,in: A. Cornish-Bowden (Ed.), Analysis of Enzyme KineticData, Oxford University Press, New York, 1995, pp. 3–26.

[39] F.J. Burguillo, A.J. Wright, W.G. Bardsley, Use of theF test for determining the degree of enzyme-kinetic andligand-binding data, Biochem. J. 211 (1983) 23–34.

[40] E.J. Erlandsen, P.E. Jørgensen, S. Markussen, A. Brock,Determination of porphobilinogen deaminase activity inhuman erythrocytes: pertinent factors in obtaining optimalconditions for measurements, Scand. J. Clin. Lab. Invest. 60(2000) 627–634.

[41] D.C. Williams, G.S. Morgan, E. McDonald, A.R. Battersby,Purification of porphobilinogen deaminase fromEuglenagracilis and studies of its kinetics, Biochem. J. 193 (1981)301–310.

[42] P.E. Jørgensen, E.J. Erlandsen, S.S. Poulsen, S. Markussen,C. Koch, A. Brock, Activity and immunohistochemicallocalization of porphobilinogen deaminase in rat tissues,Scand. J. Clin. Lab. Invest. 60 (2000) 635–641.

[43] G.V. Louie, P.D. Brownlie, R. Lambert, J.B. Cooper, T.L.Blundell, S.P. Wood, M.J. Warren, S.C. Woodcock, P.M.Jordan, Structure of porphobilinogen deaminase reveals aflexible multidomain polymerase with a single catalytic site,Nature 359 (1992) 33–39.

[44] E.B.C. Llamb́ıas, A.M. del C. Batlle, Porphyrin biosynthesisin soybean callus. V. The porphobilinogen deaminase-uropor-phyrinogen cosynthetase system. Kinetic studies, Biochim.Biophys. Acta 220 (1970) 552–559.

[45] K.R.F. Elliott, K.F. Tipton, A kinetic analysis of enzymesystems involving four substrates, Biochem. J. 141 (1974)789–805.

[46] M.J. Warren, P.M. Jordan, Investigation into the natureof substrate binding to the dipyrromethane cofactor ofEscherichia coli porphobilinogen deaminase, Biochemistry 27(1988) 9020–9030.

[47] P.M. Shoolingin-Jordan, M.J. Warren, S.J. Awan, Discoverythat the assembly of the dipyrromethane cofactor ofporphobilinogen deaminase holoenzyme proceedes initiallyby the reaction of pre-uroporphyrinogen with the apoenzyme,Biochem. J. 316 (1996) 373–376.

[48] M.L. Kotler, A.A. Juknat, S.A. Fumagalli, A.M. del C.Batlle, Involvement of free and enzyme-bound intermediatesin the reaction mechanism catalyzed by the bovine liverimmobilized porphobilinogen deaminase. Proof that theyare substrates for cosynthetase in uroporphyrinogen IIIbiosynthesis, Biotechnol. Appl. Biochem. 13 (1991)173–180.

[49] P.M. Jordan, A. Berry, Mechanism of action of porpho-bilinogen deaminase, Biochem. J. 195 (1981) 177–181.

[50] M. Witty, R.M. Jones, M.S. Robb, P.M. Shooligin-Jordan,A.G. Smith, Subcellular location of the tetrapyrrole synthesisenzyme porphobilinogen deaminase in higher plants: animmunological investigation, Planta 199 (1996) 557–564.

[51] K.-B. Augustinsson, T. Bartfai, B. Mannervik, A steady-statekinetic model of butyrylcholinesterase from horse plasma,Biochem. J. 141 (1974) 825–834.

[52] M.J. Wolansky, A.A. Juknat, M.L. Kotler, A.M. del C.Batlle, Effect of delta-aminolevulinic acid administration onporphobilinogen levels and porphyrin metabolism in the rat,Pharmacol. Commun. 7 (1995) 51–59.