- Email: [email protected]
The aconitase family: three structural variations on a common theme
bgroups corresponding to the monomeric eukaryotic and the heterodimeric prokaryotic enzymes. The active-site cleft (Fig. la)6. This arrange ment provides the structural prototype a 73-residue insertion onitase family”which is essential for proteol tely sequenced isomer It is now apparent that the aconitase and a further nine pa characterized sequences in recent DNA family comprises several functionally or ylogenetically discrete subgroups, and atabases. The well-characterized memthat the overall mofecu~ar conformation
and d. ri& are at The Krebs Institute for Biomolecular Research, Department of Molecular Biology and Biotechnology, University of Sheffield, Western Bank, Sheffield, UK SIO 2TN.
. J.
*This hvpothesis explains how a symmetrical substrate like citric acid could be processed asymmetrically by contacting at least three sites on an enzyme. Thus, citrate was restored to the rank of intermediate in the citric acid cycle from whch ‘1 had been excluded on the basis of isotopic labelling studies for seven years. During this period &aconitate was regarded as the first intermediate of the renamed tricarboxylic cycle, with citrate as a side-product.
The ‘AconitaseB’group is represented arily by the second EscherkIM co/i
Neisserk gonorrhoeae genome contains two incomplete reading frames that are recognizable as acnA and acnB homologues. Curiously, no gene resembling eikr WL~ or acn!? could be detected in
Copyright 0 1997,Elsevier Science Ltd. All rights reserved. 0968-0004/97/517.00 PII: SO963-Or?Cl4(96)lQO69-4
3
(blue): note the position of the linker (yellow) and the 124residue amino terminus (red).
region of unknown structure at the
nome of ~~~~~0~~~1~s ~n~~~n~Q~, although many other citric-acid-cycle genes are represented. The ‘Homoaconitase’ group contains only one member(Fig.2).
e existence of different structural variants was first apparentfrom the distribution of domainsbetweenindependent subunits in the heterodimeric IPMI of prokaryotes.These enzymes are predicted to adopt the same overall structure as porcine aconitasewithout needing any polypeptidecontinuitybetween domains 3 and 4 (Pig.lb). The two IPMI subunits dltecoexpressed from a single
monomeric progenitor of the entire aconitase family.However,the dternative possibility that the monomeric forms have evolved from dimeric precursors by the fusion of two ancestral genes encoding domains 3 and 1, cannot be excluded. It will be interesting
conservation in the structural cores of each domain @!uein Pig. Id) and that the 18active-
periphery where external loops that are deleted in AcnB,are often replaced by
ucor racemosus
Lactffcotxw
la&s
Saccharomyces cerewisiae
aconitase family. The dendrogram was generated by GROWREE a multiple alignment (porcine aconitas sequences were obtained from G of Neisseria gonorrhoeae (412 res
domain acts as an
for the mitochondrial enzyme, because it lacks the extraordin~i~y long linker between domains 3 and 4.
Possible evolutionary mechanisms for the cyclic permutation of coding regi among the monomeric members of aconitase family include: (1) fusion of two separate ancestral genes encoding domains 1-3 and domain 4, which had previously come together in different relative orders corresponding to the current bacterial IPMI(leuCD) situation
lar region has been incorporated into the QC coding region durin shuffling of ancestral segment encoding different domains, presumably uring the apparent migration of domain to the i~~terrnin~ end of AcnB. Domain re~rangements are not uncommon in proteins. There are numerous examples of enzymes and enzyme subunits that are synthesized independently in one organism and fused in another. Specific nucleotide-binding do mains can be associated with differing domains in varient es. Several copies of ous
However, the aconitase-IPMS example is particularly interesting because the permuted domains (or independent subunits) are combined in different ways to single active sites that have essentially the same function. The existence of genetically distinct enzymes catalysing essentially the same metabolic reaction is not uncommon among the respiratory enzymes of E. colilg. In some cases, it ?,vouldappear that gene duplicationhas provided opportunities for functional specialization and
the acquisition of independent regulatory mechanisms associated with specific metabolic roles. This seems to be true for the structurally related aerobic succinate dehydrogenase and anaerobic fumarate reductase, and for the analogous Fe-Scontaining homodimeric fumarases (FumAand FumB).In other cases, a more distantly related iso-functional gene of either Gram-positive or eukaryotic origin, seems to have been imported to provide the same opportunities for catalytic and regulatory specialization. Included here are the iron and redoxstress induced fumarase C (FumC), which is a homotetramer and not an iron-sulphur protein, and possibly AcnA,which is not only induced by iron and redox-stress, but is also more like the eukarytrtic ‘IBP-aconitase A’grou of proteins than AcnB. The cryptic hexameric citrate synthasezOand the aerobic cytochrome oxidase (cyo) of E. co/i, likewise are quite distinct from the dimeric g~~~ncoded major citrate synthase and the microaerobic cytochrome oxidase (cyd). It is
now i~nportant to esta one of the aconitases ~arti~~lar~y AcnA) functions as responsive translatio gous to m~malia~ one, or both, of the enzymes is specifically adapted for operating in t acid or glyoxylate cycles. Cryst AcnA and AcnB have recently for structure determin ynin and f? J. Artymiuk lished) and it is hope confirm the predictions
S. C. Andrews, A. J. . and Lindsay, J. 6. (1992) Steinbuchel, A. (1994) J. Bacterial. 176,4394-4408
1 Beinert.
H. and Kennedy, fvl. C. (1993) FASEB.J. 7.1442-1449 2Ogstor1, A.G.(1948)Nalure162, 963 3 Klausner, R.0. and Rouault,T. A.(1993)Mol. Biol. Cell4,l-5 4 Rouault, T. A.and Klausner, R.D.(1996)rrends
The molten globule (MG) state is widely considere intermediate in protein folding and to have a polype native-like topology. The experimental evidence for t largely, however, with MG proteins containing native-like constraints. the four disulphide bonds of cl-lactalbumin were allowed to rearrange to those favoured by the MG, opposite conclusions were obtained. eration of all the experimental data indicates that any tendency of to be native-like is negligible relative to all the other topologies that it can adopt. Furthermore, the experimental data indicate that the MG is not the key to rapid protein folding. NSc to their native conformations by a random search of all conformations, so much effort has tions, unfolded proteins are observed to gone into identifying any stable inter- adopt rapidly a variety of conformations: mediates, which are assumed to be some remain largely unfolded, others become partly-folded, while many adopt a disordered, but compact ctate that T. E. &!#hten is at the European Molecular has come to be known as the ‘molten Biology Laboratory, Meyerhofstrasse 1, 069012 Heidelberg, Germany. globule’(MG)*J.These partly-folded and Email: [email protected] MC conformations are usually assumed 6
Biochem. Sci. 21,174-177 5 Hentze, M. W. and Kuhn. I_. C. (1996) proc. Nati. ,4cx?. L%%U. S. A. 93, 6175-8182 6 Robbins, A. H. and Stout, C. D. (1989) Proteins 5,289312 7 Nicholls, A., Sharp, K. A. and Honig, B. (1991) Proteins 11, 281-296 8 Bradbury, A. J.. Gruer, M. J., Rudd, K. E. and Guest, J. R. (1996) Microbiology 142. 389-400 9 Frishman, D. and Hentze, M. W. (1996) Eur. J. Biochem. 239,197-200 20 Rosenthal, E. R. and Calve, J. M. (1990) IVucleic Acids Res. 18,3872 11 Prodromou, C., Artymiuk, F?J. and Guest, J. R. (1992) Eur. J. Biochem. 204, 599-609
Ludwig, iv!. I. (1992) Science 258,1604-1610 995) Philos. Trans. R. Sot.
Danson, M. J. (1993)
the native conformation to minor structural rearrangements3. not been possible eit definitively the kinetic
intermediates, kinetic traps or simply the preferred co~formatio~af state of the unfolded protein under the particular circumstances. Characterizing kinetic folding intermediates is usually very difficult, as they are present only transiently, but the MG state has the advantage that it can be stable and present in a substantial
CopyrightO1997, ElsevierScienceLtd.All rightsreserved.0968-0004/97/$17.00 P11:S0968-0004(96)20030-1
and d. ri& are at The Krebs Institute for Biomolecular Research, Department of Molecular Biology and Biotechnology, University of Sheffield, Western Bank, Sheffield, UK SIO 2TN.
. J.
*This hvpothesis explains how a symmetrical substrate like citric acid could be processed asymmetrically by contacting at least three sites on an enzyme. Thus, citrate was restored to the rank of intermediate in the citric acid cycle from whch ‘1 had been excluded on the basis of isotopic labelling studies for seven years. During this period &aconitate was regarded as the first intermediate of the renamed tricarboxylic cycle, with citrate as a side-product.
The ‘AconitaseB’group is represented arily by the second EscherkIM co/i
Neisserk gonorrhoeae genome contains two incomplete reading frames that are recognizable as acnA and acnB homologues. Curiously, no gene resembling eikr WL~ or acn!? could be detected in
Copyright 0 1997,Elsevier Science Ltd. All rights reserved. 0968-0004/97/517.00 PII: SO963-Or?Cl4(96)lQO69-4
3
(blue): note the position of the linker (yellow) and the 124residue amino terminus (red).
region of unknown structure at the
nome of ~~~~~0~~~1~s ~n~~~n~Q~, although many other citric-acid-cycle genes are represented. The ‘Homoaconitase’ group contains only one member(Fig.2).
e existence of different structural variants was first apparentfrom the distribution of domainsbetweenindependent subunits in the heterodimeric IPMI of prokaryotes.These enzymes are predicted to adopt the same overall structure as porcine aconitasewithout needing any polypeptidecontinuitybetween domains 3 and 4 (Pig.lb). The two IPMI subunits dltecoexpressed from a single
monomeric progenitor of the entire aconitase family.However,the dternative possibility that the monomeric forms have evolved from dimeric precursors by the fusion of two ancestral genes encoding domains 3 and 1, cannot be excluded. It will be interesting
conservation in the structural cores of each domain @!uein Pig. Id) and that the 18active-
periphery where external loops that are deleted in AcnB,are often replaced by
ucor racemosus
Lactffcotxw
la&s
Saccharomyces cerewisiae
aconitase family. The dendrogram was generated by GROWREE a multiple alignment (porcine aconitas sequences were obtained from G of Neisseria gonorrhoeae (412 res
domain acts as an
for the mitochondrial enzyme, because it lacks the extraordin~i~y long linker between domains 3 and 4.
Possible evolutionary mechanisms for the cyclic permutation of coding regi among the monomeric members of aconitase family include: (1) fusion of two separate ancestral genes encoding domains 1-3 and domain 4, which had previously come together in different relative orders corresponding to the current bacterial IPMI(leuCD) situation
lar region has been incorporated into the QC coding region durin shuffling of ancestral segment encoding different domains, presumably uring the apparent migration of domain to the i~~terrnin~ end of AcnB. Domain re~rangements are not uncommon in proteins. There are numerous examples of enzymes and enzyme subunits that are synthesized independently in one organism and fused in another. Specific nucleotide-binding do mains can be associated with differing domains in varient es. Several copies of ous
However, the aconitase-IPMS example is particularly interesting because the permuted domains (or independent subunits) are combined in different ways to single active sites that have essentially the same function. The existence of genetically distinct enzymes catalysing essentially the same metabolic reaction is not uncommon among the respiratory enzymes of E. colilg. In some cases, it ?,vouldappear that gene duplicationhas provided opportunities for functional specialization and
the acquisition of independent regulatory mechanisms associated with specific metabolic roles. This seems to be true for the structurally related aerobic succinate dehydrogenase and anaerobic fumarate reductase, and for the analogous Fe-Scontaining homodimeric fumarases (FumAand FumB).In other cases, a more distantly related iso-functional gene of either Gram-positive or eukaryotic origin, seems to have been imported to provide the same opportunities for catalytic and regulatory specialization. Included here are the iron and redoxstress induced fumarase C (FumC), which is a homotetramer and not an iron-sulphur protein, and possibly AcnA,which is not only induced by iron and redox-stress, but is also more like the eukarytrtic ‘IBP-aconitase A’grou of proteins than AcnB. The cryptic hexameric citrate synthasezOand the aerobic cytochrome oxidase (cyo) of E. co/i, likewise are quite distinct from the dimeric g~~~ncoded major citrate synthase and the microaerobic cytochrome oxidase (cyd). It is
now i~nportant to esta one of the aconitases ~arti~~lar~y AcnA) functions as responsive translatio gous to m~malia~ one, or both, of the enzymes is specifically adapted for operating in t acid or glyoxylate cycles. Cryst AcnA and AcnB have recently for structure determin ynin and f? J. Artymiuk lished) and it is hope confirm the predictions
S. C. Andrews, A. J. . and Lindsay, J. 6. (1992) Steinbuchel, A. (1994) J. Bacterial. 176,4394-4408
1 Beinert.
H. and Kennedy, fvl. C. (1993) FASEB.J. 7.1442-1449 2Ogstor1, A.G.(1948)Nalure162, 963 3 Klausner, R.0. and Rouault,T. A.(1993)Mol. Biol. Cell4,l-5 4 Rouault, T. A.and Klausner, R.D.(1996)rrends
The molten globule (MG) state is widely considere intermediate in protein folding and to have a polype native-like topology. The experimental evidence for t largely, however, with MG proteins containing native-like constraints. the four disulphide bonds of cl-lactalbumin were allowed to rearrange to those favoured by the MG, opposite conclusions were obtained. eration of all the experimental data indicates that any tendency of to be native-like is negligible relative to all the other topologies that it can adopt. Furthermore, the experimental data indicate that the MG is not the key to rapid protein folding. NSc to their native conformations by a random search of all conformations, so much effort has tions, unfolded proteins are observed to gone into identifying any stable inter- adopt rapidly a variety of conformations: mediates, which are assumed to be some remain largely unfolded, others become partly-folded, while many adopt a disordered, but compact ctate that T. E. &!#hten is at the European Molecular has come to be known as the ‘molten Biology Laboratory, Meyerhofstrasse 1, 069012 Heidelberg, Germany. globule’(MG)*J.These partly-folded and Email: [email protected] MC conformations are usually assumed 6
Biochem. Sci. 21,174-177 5 Hentze, M. W. and Kuhn. I_. C. (1996) proc. Nati. ,4cx?. L%%U. S. A. 93, 6175-8182 6 Robbins, A. H. and Stout, C. D. (1989) Proteins 5,289312 7 Nicholls, A., Sharp, K. A. and Honig, B. (1991) Proteins 11, 281-296 8 Bradbury, A. J.. Gruer, M. J., Rudd, K. E. and Guest, J. R. (1996) Microbiology 142. 389-400 9 Frishman, D. and Hentze, M. W. (1996) Eur. J. Biochem. 239,197-200 20 Rosenthal, E. R. and Calve, J. M. (1990) IVucleic Acids Res. 18,3872 11 Prodromou, C., Artymiuk, F?J. and Guest, J. R. (1992) Eur. J. Biochem. 204, 599-609
Ludwig, iv!. I. (1992) Science 258,1604-1610 995) Philos. Trans. R. Sot.
Danson, M. J. (1993)
the native conformation to minor structural rearrangements3. not been possible eit definitively the kinetic
intermediates, kinetic traps or simply the preferred co~formatio~af state of the unfolded protein under the particular circumstances. Characterizing kinetic folding intermediates is usually very difficult, as they are present only transiently, but the MG state has the advantage that it can be stable and present in a substantial
CopyrightO1997, ElsevierScienceLtd.All rightsreserved.0968-0004/97/$17.00 P11:S0968-0004(96)20030-1