An Artificial Gene for Human Porphobilinogen Synthase Allows Comparison of an Allelic Variation Implicated in Susceptibility to Lead Poisoning

Abstract

Porphobilinogen synthase (PBGS) is an ancient enzyme essential to tetrapyrrole biosynthesis (e.g. heme, chlorophyll, and vitamin B12). Two common alleles encoding human PBGS, K59 and N59, have been correlated with differential susceptibility of humans to lead poisoning. However, a model for human PBGS based on homologous crystal structures shows the location of the allelic variation to be distant from the active site with its two Zn(II). Previous microbial expression systems for human PBGS have resulted in a poor yield. Here, an artificial gene encoding human PBGS was constructed by recursive polymerase chain reaction from synthetic oligonucleotides to rectify this problem. The artificial gene was made to resemble the highly expressed homologous Escherichia coli hemB gene and to remove rare codons that can confound heterologous protein expression in E. coli. We have expressed and purified recombinant human PBGS variants K59 and N59 in 100-mg quantities. Both human PBGS proteins purified with eight Zn(II)/octamer; Zn(II) binding was shown to be pH-dependent; and Pb(II) could displace some of the Zn(II). However, there was no differential displacement of Zn(II) by Pb(II) between K59 and N59, and simple Pb(II) inhibition studies revealed no allelic difference. Porphobilinogen synthase (PBGS) is an ancient enzyme essential to tetrapyrrole biosynthesis (e.g. heme, chlorophyll, and vitamin B12). Two common alleles encoding human PBGS, K59 and N59, have been correlated with differential susceptibility of humans to lead poisoning. However, a model for human PBGS based on homologous crystal structures shows the location of the allelic variation to be distant from the active site with its two Zn(II). Previous microbial expression systems for human PBGS have resulted in a poor yield. Here, an artificial gene encoding human PBGS was constructed by recursive polymerase chain reaction from synthetic oligonucleotides to rectify this problem. The artificial gene was made to resemble the highly expressed homologous Escherichia coli hemB gene and to remove rare codons that can confound heterologous protein expression in E. coli. We have expressed and purified recombinant human PBGS variants K59 and N59 in 100-mg quantities. Both human PBGS proteins purified with eight Zn(II)/octamer; Zn(II) binding was shown to be pH-dependent; and Pb(II) could displace some of the Zn(II). However, there was no differential displacement of Zn(II) by Pb(II) between K59 and N59, and simple Pb(II) inhibition studies revealed no allelic difference. porphobilinogen synthase polymerase chain reaction isopropyl-β-d-thiogalactopyranoside phenylmethylsulfonyl fluoride Human genetic polymorphisms are increasingly being correlated with disease states through epidemiological analysis. These types of studies lay the groundwork for the extended practice of preventive medicine. A common genetic polymorphism in the human ALADgene, 1The naturally occurring human gene is called ALAD based on the commonly used alternative enzyme name 5-aminolevulinic-aciddehydratase. encoding the enzyme porphobilinogen synthase (PBGS2; EC 4.2.1.24), has been correlated with susceptibility to the environmental toxin lead (1.Wetmur J.G. Lehnert G. Desnick R.J. Environ. Res. 1991; 56: 109-119Crossref PubMed Scopus (142) Google Scholar, 2.Schwartz B.S. Lee B.K. Stewart W. Ahn K.D. Springer K. Kelsey K. Am. J. Epidemiol. 1995; 142: 738-745Crossref PubMed Scopus (103) Google Scholar, 3.Smith C.M. Wang X. Hu H. Kelsey K.T. Environ. Health Perspect. 1995; 103: 248-253PubMed Google Scholar, 4.Bergdahl I.A. Gerhardsson L. Schutz A. Desnick R.J. Wetmur J.G. Skerfving S. Arch. Environ. Health. 1997; 52: 91-96Crossref PubMed Scopus (65) Google Scholar, 5.Bergdahl I.A. Grubb A. Schutz A. Desnick R.J. Wetmur J.G. Sassa S. Skerfving S. Pharmacol. Toxicol. 1997; 81: 153-158Crossref PubMed Scopus (161) Google Scholar, 6.Fleming D.E. Chettle D.R. Wetmur J.G. Desnick R.J. Robin J.P. Boulay D. Richard N.S. Gordon C.L. Webber C.E. Environ. Res. 1998; 77: 49-61Crossref PubMed Scopus (78) Google Scholar). As lead poisoning is the most common preventable childhood neurological environmental disease, correlation of the epidemiological results with protein structure/function studies is in order. PBGS catalyzes the first common step in the biosynthesis of all tetrapyrroles (heme, chlorophyll, vitamin B12, cofactor F430, etc.). Human PBGS is a Zn(II) metalloenzyme unique in its sensitivity to inhibition by lead. Although all PBGSs appear to be metalloenzymes (7.Shemin D. Boyer P. The Enzymes. 3rd Ed. VII. Academic Press, New York1972: 232-237Google Scholar), metal ion usage varies dramatically between species (8.Petrovich R.M. Litwin S. Jaffe E.K. J. Biol. Chem. 1996; 271: 8692-8699Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar, 9.Frankenberg N. Jahn D. Jaffe E.K. Biochemistry. 1999; 38: 13976-13982Crossref PubMed Scopus (25) Google Scholar). One outcome of this variation is that microbial and plant PBGSs are poor models for studying the effect of lead on human PBGS function. Lead inhibition of human PBGS is one of the earliest physiological responses to lead intoxication and as such is believed to be related to the detrimental effects of low level lead poisoning. However, mapping the common human polymorphism onto the x-ray crystal structure of the related yeast PBGS protein does not indicate a structural variation that would obviously affect either metal binding or catalytic function. A model for human PBGS is presented in Fig.1 and shows the location of amino acid 59, which is lysine in the ALAD1 gene product and asparagine in the ALAD2 gene product. Production of human PBGS for study has been problematic. Purification of PBGS from human blood gives relatively low yield, is plagued by considerations of blood-borne diseases, and contains a mixture of the isozymes (10.Anderson P.M. Desnick R.J. J. Biol. Chem. 1979; 254: 6924-6930Abstract Full Text PDF PubMed Google Scholar, 11.Gibbs P.N. Chaudhry A.G. Jordan P.M. Biochem. J. 1985; 230: 25-34Crossref PubMed Scopus (34) Google Scholar). The human ALAD gene was cloned and sequenced more than a decade ago and found to express poorly inEscherichia coli (12.Wetmur J.G. Bishop D.F. Cantelmo C. Desnick R.J. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 7703-7707Crossref PubMed Scopus (103) Google Scholar). To generate a better expression system, the gene was cloned into yeast (13.Schauer W.E. Mattoon J.R. Curr. Genet. 1990; 17: 1-6Crossref PubMed Scopus (4) Google Scholar), but the levels of protein expression remain insufficient for thorough functional analysis. For a novel approach to heterologous expression of human PBGS, we chose to mimic a construct used for overexpression of the E. coli hemB gene in E. coli (14.Roessner C.A. Spencer J.B. Ozaki S. Min C. Atshaves B.P. Nayar P. Anousis N. Stolowich N.J. Holderman M.T. Scott A.I. Protein Expression Purif. 1995; 6: 155-163Crossref PubMed Scopus (31) Google Scholar, 15.Mitchell L.W. Jaffe E.K. Arch. Biochem. Biophys. 1993; 300: 169-177Crossref PubMed Scopus (53) Google Scholar). This system can generate hundreds of milligrams of E. coli PBGS (up to 30% of the soluble protein). The reasons for high level expression are not fully understood. In this case, the hemB gene is apparently downstream from a naturally strong E. coli promoter, and other poorly understood aspects of gene structure may contribute to the phenomenal levels of constitutive expression observed. A second consideration in artificial gene design is codon usage. The humanALAD gene was analyzed and found to contain clusters of codons that are rarely used by E. coli (see Fig.2 A). In contrast, the E. coli hemB gene contains only one rare codon. Kane (16.Kane J.F. Curr. Opin. Biotechnol. 1995; 6: 494-500Crossref PubMed Scopus (604) Google Scholar) has described how clusters of six specific rare codons can be detrimental to both the quality and quantity of heterologous proteins expressed in E. coli, and specific translational errors have been documented for such clusters (17.Forman M.D. Stack R.F. Masters P.S. Hauer C.R. Baxter S.M. Protein Sci. 1998; 7: 500-503Crossref PubMed Scopus (45) Google Scholar). Hence, in the design of an artificial gene for human PBGS, theE. coli hemB gene structure was mimicked to the greatest extent possible, and rare codons were replaced. Here we describe the design and synthesis of a PCR product (EJhum) containing an artificial gene coding for human PBGS.EJhum mimics a pUC119 version of pCR261 (14.Roessner C.A. Spencer J.B. Ozaki S. Min C. Atshaves B.P. Nayar P. Anousis N. Stolowich N.J. Holderman M.T. Scott A.I. Protein Expression Purif. 1995; 6: 155-163Crossref PubMed Scopus (31) Google Scholar), which gives high level constitutive expression of the E. coli hemB gene (encoding E. coli PBGS) in a variety of E. colihost strains. Successful large-scale protein production of human PBGS was obtained by expression under control of the T7 polymerase system using a plasmid named pMVhum, which encodes the K59 protein corresponding to the ALAD1 allele inserted into the pET3 vector. We present the expression and purification of human PBGS from this construct and a of the two human PBGS proteins K59 and N59, the a product of We the Zn(II) binding of the two proteins and the of Pb(II) to displace the essential Zn(II). were in the and used polymerase was obtained from and used of PCR and were from and were from and as as and were obtained from of and were obtained from The were from New were purified using plasmid purification and PCR were from using from was using of plasmid was using the of The E. coli gene construct pCR261 is an constitutive expression system (14.Roessner C.A. Spencer J.B. Ozaki S. Min C. Atshaves B.P. Nayar P. Anousis N. Stolowich N.J. Holderman M.T. Scott A.I. Protein Expression Purif. 1995; 6: 155-163Crossref PubMed Scopus (31) Google Scholar). The pUC119 is in Fig.2 The artificial gene for human PBGS was in the the plasmid is were in the design the of was from and in the of pUC119 and the of the E. The of containing E. coli is in are the first of are to a of the E. coli of to of of from the hemB by which a of the coding of The from A of the PCR product is shown in C. was to all of the of that is and downstream of the E. coli hemB gene as in D. The second step was to the codons of the human to mimic of the E. coli hemB gene in all the amino of the two proteins are This was and was by the and The step rare codons and that have been as in heterologous expression (16.Kane J.F. Curr. Opin. Biotechnol. 1995; 6: 494-500Crossref PubMed Scopus (604) Google Scholar). The of the human contains of which A shows are relatively and that some in The and were used to the of which were to common E. coli In the was for detrimental codons as described by J. Biol. Chem. 1995; Full Text Full Text PDF PubMed Scopus Google Scholar). codon were found with a of were to common codons to generate the artificial The E. coli hemB gene was found to contain codon and two The of codons The step and analyzed the of the plasmid using the This the first of by the coding of the artificial gene and of The plasmid the of the gene in the first no extended in the second and the artificial gene in the To a site in with of the we from A to The synthetic in Fig.2 and in in contains of and from a site to a unique were to the and to as the eight and eight were for the synthesis of recursive The were in with a that oligonucleotides would The were in were as as in the of with for the location of the eight and the eight and were used to of were by and The were and used as for the recursive PCR synthesis of two These and were by and the PCR product from the recursive PCR of and using The product contains the artificial gene encoding human However, this synthetic not generate an version of and a second was (see and the PCR product were with and The was and the and the PCR product were and with to this used to coli H. J. PubMed Scopus Google Scholar), the were not obtained under could be used to of However, of in E. coli were were by PCR using and were but human PBGS.EJhum was by PCR from of using and and the PCR were sequenced and found to contain and some To the errors from the or from the PCR was into the site and into for plasmid purification and In all the of but of could be obtained by PCR from or one of the of were used to and by which in were used to In this case, synthesis was and the was The of was not to of under could of was found to be with of the plasmid and no overexpression of the artificial We that constitutive expression of human PBGS from is to E. coli, and expression was To control the expression of human PBGS in E. coli, was to remove the and from the E. coli and to into one of the for T7 expression in a of Two oligonucleotides and were used as PCR to the artificial gene from to the site with the site of or and to site the codon The PCR product was into the and into were by PCR using the the artificial gene was and into the of both and to and of high between the artificial gene and the E. coli the and were as for both and In all were and overexpression was with using as the the expression and the all was with this of is in Fig.2 was in of and in an the from a from a the of which were and in of to of the were to and was to a of C.L. K. L. Protein Expression Purif. 1997; PubMed Scopus Google Scholar). under in an to a of and of the PBGS was in the soluble of the The N59 more into the soluble expression that protein the and used to expression for a of in the of Zn(II). were of in and fluoride the were and was to and the was for this an of and was with of and the was an of the were The mixture was through a The was for The was to a The was in and and to a that been with the The was with of the and to a that and Human PBGS the in a of the and was onto a that been with and The was with of this and was to a in the with a of The of the the was for with E. as shown by a of in by L.W. Jaffe E.K. Arch. Biochem. Biophys. 1993; 300: 169-177Crossref PubMed Scopus (53) Google Scholar). of high specific were and under to The protein was a a of The and Zn(II). of purified recombinant human PBGS were in and of and containing of Zn(II) of the protein was from the and analyzed for both protein by and for Zn(II) by using a were using the to Zn(II). was from the protein as of The binding were by using a model for two of as we found the Zn(II) binding for PBGS obtained E.K. Biochemistry. PubMed Scopus Google Scholar). Zn(II) binding were obtained in and which to the Zn(II) from the The were to and of containing of Zn(II) of Pb(II) in the These to a simple binding with a and a A model for the human PBGS was from Protein using the and N. S. G. Spencer P. P. P.M. J.B. Biol. 1997; PubMed Scopus Google Scholar, Protein Sci. 1997; 6: PubMed Scopus Google Scholar, J. Biol. 1997; PubMed Scopus Google Scholar, Genet. Scholar). the Zn(II) in yeast PBGS N. S. G. Spencer P. P. P.M. J.B. Biol. 1997; PubMed Scopus Google Scholar) are in human PBGS, the two Zn(II) are in the which is as a in However, the human PBGS a of eight Zn(II) that are to and and that are to and the two Zn(II) are only in one of the This mimics the metal binding in the crystal structure PBGS N. J.B. P.M. Jahn D. J. Biol. 1999; PubMed Scopus Google Scholar). were in and Zn(II) for to the of to a of were to for to with of acid and Porphobilinogen was by the of of The of the is PCR C. Protein PubMed Scopus Google Scholar) was used to an artificial gene encoding the protein by the more allele of the human gene for porphobilinogen The gene was to resemble the homologous E. coli hemB gene to the greatest extent and to remove codons rarely used by E. coli. As described under the synthetic and constitutive expression were problematic. We that was to synthetic in for recursive The gene may be related to alternative protein for PBGS proteins J. Biol. Chem. Full Text PDF PubMed Google Scholar, S. A. J. Biol. Chem. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar). The artificial gene encoding human PBGS in is for expression in coli and has been shown to high levels of expression as by for from in or its high expression of proteins that are found in as we have found for human The be as a purified of the protein that can be and into an active Human PBGS is a enzyme containing active and two types of and would have made the a we used active enzyme purified from the soluble of the in the of human PBGS expressed from The for the is based on the was in the first to expression of under the control of the such as the T7 polymerase in to the were to that not contain To the level of in the a low level was used to expression from as to the of the protein to C.L. K. L. Protein Expression Purif. 1997; PubMed Scopus Google Scholar). the low level of expression the first and expression the This can be on an using and to of The gives expression with better of the protein into the soluble The for human PBGS from the soluble of on and of E. coli, PBGSs (8.Petrovich R.M. Litwin S. Jaffe E.K. J. Biol. Chem. 1996; 271: 8692-8699Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar, 15.Mitchell L.W. Jaffe E.K. Arch. Biochem. Biophys. 1993; 300: 169-177Crossref PubMed Scopus (53) Google Scholar, E.K. J. Biol. Chem. Full Text PDF PubMed Google Scholar, P.M. 1986; PubMed Scopus Google Scholar, D. Biol. Chem. PubMed Scopus Google Scholar). In all human PBGS as from with a K59 of the and N59 of the The the of the and a or in specific Human PBGS the of the and protein of high specific of the The protein from the step for the protein product of of the a specific of which is than the for PBGS P.M. 1986; PubMed Scopus Google Scholar). The common N59 protein a specific of expressed under The in the specific of K59 and N59 was highly between and The of purified protein was for K59 and for the to the differential of isozymes between the soluble and the The protein purified in the of Zn(II) and was found by to contain eight as was found for PBGS E.K. J. Biol. Chem. Full Text PDF PubMed Google Scholar). A model for PBGS is a protein with functional active of which contains two and E.K. J. 1995; PubMed Scopus Google D. C. Biochemistry. 1990; PubMed Scopus Google Scholar). One of this model is the binding of Zn(II) to PBGS, which shows binding of and binding of a second is Zn(II) E.K. Biochemistry. PubMed Scopus Google Scholar). In the of recombinant human PBGS, A that this model for Zn(II) binding gives a to obtained under Here the shows a Zn(II) and a Zn(II) and A and obtained on PBGS for with the and model is that only are for of PBGS in the of E.K. J. Biol. Chem. Full Text PDF PubMed Google Scholar, D.R. P. D. J. Biol. Chem. Full Text PDF PubMed Google Scholar). shows a Zn(II) for N59, the to of The of Zn(II) from is as shown by the in The of the purified protein was found to be to the of as for human PBGS based on studies L.W. Jaffe E.K. Arch. Biochem. Biophys. 1993; 300: 169-177Crossref PubMed Scopus (53) Google binding for PBGS (see A and from from in a PBGS is purified with product L.W. Jaffe E.K. Arch. Biochem. Biophys. 1993; 300: 169-177Crossref PubMed Scopus (53) Google Scholar), and one model for PBGS the amino of porphobilinogen as a Zn(II) E.K. J. 1995; PubMed Scopus Google Scholar). can be used to both and product from a variety of microbial PBGSs (8.Petrovich R.M. Litwin S. Jaffe E.K. J. Biol. Chem. 1996; 271: 8692-8699Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar, N. D. Jahn D. Biochemistry. 1999; 38: PubMed Scopus Google Scholar). Hence, we to Zn(II) binding to human PBGS which was shown to all Zn(II) to low is to all product. The results in for K59 Zn(II) binding that not between the two types of Zn(II) that are in and that are in A. The for N59 are not but the binding are in was no in Zn(II) binding between human PBGS K59 and no Zn(II) to the the is and the of Zn(II) eight or more and and the is and the of Zn(II) is We that the low a of in human PBGS, but the structural for the between A and is Lead has been shown to be a of PBGS E.K. S. Biol. Res. 1991; PubMed Scopus (31) Google Scholar). studies using metal ion have shown that the for Pb(II) is than the for Zn(II) J. Biochem. 1995; PubMed Scopus Google Scholar). As a first step in the of Pb(II) to displace Zn(II) from the two isozymes of human PBGS, studies were Zn(II) of and and Pb(II) of and These with by low The results in that Pb(II) for of the Zn(II) was no differential displacement of Zn(II) by Pb(II) for K59 to results were in simple inhibition the were for the was was Zn(II) was with no and was with no Zn(II) Pb(II) is a E.K. S. Biol. Res. 1991; PubMed Scopus (31) Google Scholar), a more thorough of Pb(II) inhibition may some between K59 and in recombinant have results in heterologous protein expression in E. coli. In some such as PBGS, theE. coli can be expressed high expression of the heterologous human protein is of the with the poorly expressed protein may be clusters of codons rarely used in E. coli. for rare codons can be a and does not gene synthesis C.A. Protein Expression Purif. 1998; PubMed Scopus Google Scholar). may be to a coli to contain of the rare D.E. C.A. Arch. Biochem. Biophys. 1996; PubMed Scopus (34) Google Scholar). A understood is the high level and constitutive expression of the homologous E. The approach described of the that control gene expression in E. coli and simple to expression of the heterologous The artificial gene encoding human PBGS is to resemble theE. coli hemB gene to the greatest extent encoding the human protein and rarely used E. coli In the of human PBGS, the approach was and is for in protein The design of artificial is not a novel and has been for to in specific Res. 1993; PubMed Scopus Google Scholar). considerations in the homologous E. coli gene and rare codons were to both protein expression levels and protein quality by translational Production of heterologous proteins of quality is a for the and specific of errors in are in the (17.Forman M.D. Stack R.F. Masters P.S. Hauer C.R. Baxter S.M. Protein Sci. 1998; 7: 500-503Crossref PubMed Scopus (45) Google Scholar). Although the studies not the of the expression in with low levels of and a to levels can be used to translational or protein PBGS is an ancient and highly protein with specific in that crystal structures have shown to be essential to metal ion types of metal have been that to that have been called and E. Chem. 1993; Scopus Google Scholar). The two Zn(II) are in the yeast PBGS structure N. S. G. Spencer P. P. P.M. J.B. Biol. 1997; PubMed Scopus Google Scholar), and the is in the P. PBGS structure N. J.B. P.M. Jahn D. J. Biol. 1999; PubMed Scopus Google Scholar). PBGS as a with eight and the of the Zn(II) is not In contrast, the P. PBGS is of only one of the of contains only this has a the active Human PBGS has a to the yeast and contains the to to and in human PBGS that a is to one functional active two Zn(II) are shown in the model to be to only one of the is to that the to the as by extended x-ray structure D. C. Biochemistry. 1990; PubMed Scopus Google Scholar), are not the as the to the highly Zn(II) in the crystal structure of yeast PBGS with N. S. G. Spencer P. P. P.M. J.B. Biol. 1997; PubMed Scopus Google Scholar). We have shown that such as and binding can control the of metal between the PBGS R.M. Jaffe E.K. Biochemistry. 1997; PubMed Scopus Google Scholar), and may confound of studies on the These confound functional on Pb(II) inhibition of PBGS based on Pb(II) in the of We have an expression system for human PBGS that is for heterologous expression with a low translational This system was used for the purification of human PBGS by the two common alleles and found in human The purified proteins with other PBGSs and can be used for and structural of human PBGS and The only between the proteins by the two alleles is an variation in specific is no differential effect of the isozymes on Zn(II) Pb(II) for the Zn(II) or inhibition of by studies into differential effects of lead on Zn(II) and protein to a more for the genetic susceptibility of lead poisoning. epidemiological the allelic variation in human PBGS with lead poisoning (1.Wetmur J.G. Lehnert G. Desnick R.J. Environ. Res. 1991; 56: 109-119Crossref PubMed Scopus (142) Google Scholar, 2.Schwartz B.S. Lee B.K. Stewart W. Ahn K.D. Springer K. Kelsey K. Am. J. Epidemiol. 1995; 142: 738-745Crossref PubMed Scopus (103) Google Scholar, 5.Bergdahl I.A. Grubb A. Schutz A. Desnick R.J. Wetmur J.G. Sassa S. Skerfving S. Pharmacol. Toxicol. 1997; 81: 153-158Crossref PubMed Scopus (161) Google Scholar), such not a and one may not