David M. Young is a graduate student in Molecular, Cellular and Developmental Biology at Yale University, Donna Parke is research scientist in Molecular, Cellular and Developmental Biology at Yale University, David A. D'Argenio is a postdoctoral fellow in Genetics at the University of Washington, Seattle, Michael A. Smith is a graduate student in Molecular, Cellular and Developmental Biology at Yale University, and L. Nicholas Ornston is Professor of Molecular, Cellular and Developmental Biology at Yale University, New Haven, Conn.
Links to Other ASM Pages:
Evolution of a Catabolic Pathway
Analysis of theb -ketoadipate pathway provides evolutionary insights, including how genes behave when environments change
David M. Young, Donna Parke, David A. D'Argenio, Michael A. Smith, and L. Nicholas Ornston
"When Coleridge tried to define beauty, he returned always to one deep thought: beauty, he said, is ‘unity in variety’. Science is nothing else than the search to discover unity in the wild variety of nature - or more exactly in the variety of our experience."
Albert J. Kluyver
Kluyver deserves much credit for the concept of biochemical unity, recognition that a relatively small set of coenzymes and essential biochemical reactions are shared by all organisms. The unifying concept emerged from an appreciation of the breadth of microbial diversity revealed by the Delft school of microbiology, a group that Kluyver once headed.
Microbial diversity has allowed biochemists to choose the microorganism best suited for analysis of a particular biochemical pathway. Here we describe a different approach, the use of a single metabolic pathway to study the diversity of microorganisms. The b -ketoadipate pathway provides such a road to understanding, and its analysis has provided a rich record of mechanisms contributing to microbial diversity.
Issues that have been addressed include the genetic material that was called upon during evolution of theb -ketoadipate pathway, mechanisms of genetic variation adapting this trait to its different biological hosts, and the potential plasticity of biological molecules even after extensive evolutionary divergence. Studying how microorganisms acquire novel traits in the laboratory teaches us how structure influences function in proteins and opens questions about how evolutionary variation actually was achieved in the environment.
Convergent and Divergent Evolution in the b -Ketoadipate Pathway
The b -ketoadipate pathway acquired its name from compound that accumulates during bacterial catabolism of phenol, an effluent in industrial wastes. In many microorganisms, phenol is converted to catechol which supports microbial growth after its metabolism through b -ketoadipate to the common metabolites succinyl-CoA and acetyl-CoA. There are two branches of the b -ketoadipate pathway used by bacteria for growth with catechol and protocatechuate (Fig. 1); fungi use a metabolic variation for growth with protocatechuate.
The two terminal steps giving rise to succinyl-CoA and acetyl-CoA are conserved in the different metabolic variants of the b -ketoadipate pathway, but the earlier steps are not (Fig. 1). For example, the pathways used by fungi and bacteria diverge after cleavage of protocatechuate and converge again as b -ketoadipate is formed. This finding can be taken as evidence of independent evolutionary origins of the pathway from different genetic resources in the two biological groups.
Perhaps more remarkable is the separate ancestry of the respective pca and cat genes for the protocatechuate and catechol branches of the pathway in bacteria. As the biochemical outline of each branch was revealed, it was apparent that they proceeded by seemingly similar biochemical transformations. This metabolic parallelism made it seem obvious that, following an ancient genetic duplication, copies of a single set of ancestral genes were adapted for the protocatechuate and catechol branches of the pathway. As is often the case in biology, what seemed obvious turned out to be incorrect; some of the biochemical similarities are illusory. Enzymes encoded by pcaB and catB catalyze stereochemically different reactions, and the genes were recruited from different ancestors.
Separate ancestral genes also gave rise to pcaC and catC. The enzymes that act upon protocatechuate and catechol—encoded by pcaHG and catA, respectively—share a similar segment of primary sequence. However, the overall three-dimensional structures of the proteins are so dissimilar that they can be regarded as different molecules that acquired homologous segments from the same ancestor through recombination.
Different Mechanisms of Divergence in Genes for Parallel Step Reactions
Only at the steps where biochemical reactions are identical is evolutionary homology of enzymes associated with protocatechuate and catechol catabolism evident. Even here, evolutionary divergence does not fit a simple pattern of steady nucleotide substitution within a common template over time.
Although they encode nearly identical enzymes, the catIJF and pcaIJF gene sets (Fig. 1) of Acinetobacter sp. strain ADP1 are maintained at distant positions in the chromosome because they are expressed under different growth conditions. The genes are unusual in that they share nucleotide sequence identity in excess of 98%, and one set of genes can serve as a template for repair through gene conversion of mutations inactivating the other set.
The catIJF and pcaIJF genes also possess an unusually high G+C content, roughly 10% above the 38 to 44% G+C content characteristic of almost all other known genes in the Acinetobacter chromosome. One interpretation of this finding is that the paired catIJF and pcaIJF genes are recent evolutionary acquisitions that have not yet acquired the G+C content characteristic of the host chromosome. Another interpretation, not inconsistent with the first, is that ongoing gene conversion between catIJF and pcaIJF has buffered these genes against forces that caused the other Acinetobacter genes to achieve a lower G+C content.
Like the catIJF and pcaIJF genes, Acinetobacter catD and pcaD (Fig. 1) are separately regulated homologous genes that encode enzymes with identical biochemical functions. In contrast to the situation with catIJF and pcaIJF, evolutionary divergence of catD and pcaD has been extreme. For example, the G+C contents of catD and pcaD differ by 6%, a separation that if found in different genomes would put their hosts at the limit acceptable for inclusion of strains in the same genus. In addition, the two genes are organized very differently, with pcaB interposed and thus distinguishing the pcaIJFBD chromosomal order from that of catIJFD.
Selection of DNA Arrangements and Rearrangements
What accounts for the near absence of sequence divergence in catIJF and pcaIJF as opposed to the extensive divergence of catD and pcaD? One possibility is that the differences illustrate extremes of evolutionary advantages and hazards associated with maintenance of genes with similar nucleotide sequences in the same chromosome. On one hand, nonreciprocal sequence exchange can stabilize DNA sequences by providing a spare copy as a source of information for repair.
On the other hand, in mutants engineered to bring catIJF and pcaIJF close together in the chromosome, reciprocal recombination between similar chromosomal segments can cause deletions removing intervening DNA. Thus bacterial genes with common ancestry usually are divergent in sequence if they occupy the same chromosome. The demand for divergence increases after genes pass a threshold of divergence causing their protein products to assume different conformations. After this threshold is passed, recombination leading to hybrid conformations in gene products is likely to be selected against.
When the advantage in possessing a nearly identical spare copy is counterbalanced by the demand for sequence divergence, selection for mutations causing multiple nucleotide substitutions would be imposed. Such mutations have been observed: a short segment of nucleotide sequence serves as a template for sequence transfer converting a partially repeated sequence into a more perfect sequence repetition. DNA that has undergone selective pressure for rapid sequence divergence might be expected to be rich in nucleotide repetitions, and this is the pattern observed in catD and pcaD.
An additional mechanism for avoiding deleterious recombination is gene rearrangement as exemplified by the different gene orders catIJFD and pcaIJFBD. Comparative studies show that gene rearrangements were the rule rather than the exception during divergence of the b -ketoadipate pathway in different bacterial genera. Extensive rearrangements might be expected to distribute independently regulated genes randomly around the bacterial chromosome. Frequently the opposite pattern is observed.
An Acinetobacter chromosomal cluster contains more than 40 genes associated with utilization of chemicals produced by plants (Fig. 2). These compounds, growth substrates for the bacteria, often serve structural and protective functions in plants, notably as components of the polymers lignin and suberin. Lignin is formed largely by aromatic structures (such as those depicted in Fig. 2) held together by relatively intractable ether bonds. Similar aromatic compounds as well as dicarboxylic acids (Fig. 2) contribute to the structure of the suberin network, which is dependent on ester bonds. These bonds, unlike the ether bonds of lignin, are amenable to breakage by hydrolysis.
Chlorogenate is an additional ester formed by plants for protection, and one of the clustered Acinetobacter genes (clgA) encodes an esterase that cleaves this compound to quinate and caffeate, substrates for further metabolism by other enzymes encoded in the cluster (Fig. 2). Thus the Acinetobacter chromosomal segment, termed a supraoperonic gene cluster because its operon components are subject to independent transcriptional control, seems suited for metabolism of compounds that plants bring to the table in the form of chlorogenate and suberin. In light of the apparent ecological function of the gene cluster, it has been provisionally designated a suberon.
Natural selection apparently favored supraoperonic clustering of the Acinetobacter genes, yet the selective forces remain to be defined. One possibility is that clustering of the genes allows their co-amplification by tandem duplication. Benefits that might be derived from such a mechanism include augmented expression through increased gene dosage. Multiple gene copies formed by duplication would increase the probability of their horizontal transfer by a natural transformation system that is unusually effective in Acinetobacter sp. strain ADP1. Evidence for or against amplification of DNA containing the Acinetobacter dca-pca-qui-pob-clg-ppa cluster is not available, but DNA containing another Acinetobacter supraoperonic cluster, the sal-are-ben-cat genes, undergoes tandem duplications resulting in up to 20 copies of a 30-kb DNA segment, according to Andrew B. Reams and Ellen L. Neidle of the University of Georgia, Athens.
What Selective Benefits Counterbalance Formation of Deleterious Mutations?
A propensity for expansion and contraction of chromosomal DNA may help to account for the frequency with which selection for loss of function produces strains containing large deletions. About 20% of mutants selected on the basis of an inability to express pcaHG, the genes encoding the enzyme that acts upon protocatechuate (Fig. 1), contained deletions extending at least 7 kb into the dca region required for growth with dicarboxylic acids (Fig. 2). At first glance, a selective benefit that might be associated with such large deletions is not evident. In a broader sense, a system that gives rise to the deletion mutations as occasional accidents in some cells may be tolerated if it allows other members of the population to contain multiple tandem copies of genes that can be called upon by intermittent selective forces.
This concept of evolutionary tradeoffs also provides an interpretation for shorter deletions that are frequently recovered after selection for spontaneous mutations causing loss of function. A clear example is a 20-bp deletion repeatedly recovered in strains defective in an operator governing transcription of the Acinetobacter pca genes. This deletion appears to be guided by a 10-bp direct nucleotide sequence repetition that serves as a binding site for a transcriptional regulator. Misalignment of complementary DNA strands containing the repetition is likely to guide the mechanism that creates the deletion. Here a benefit, namely effective functioning of an operator that demands the 10-bp sequence repetition, offsets a potential flaw, the occasional loss of the 20 bp of DNA between the sequence repetitions.
Less obvious are selective pressures leading to duplications or deletions that impair the function of a structural gene. A case in point is presented by Acinetobacter pcaH and pcaG, contiguous genes that encode protocatechuate 3,4-dioxygenase, the enzyme that catalyzes the first step in protocatechuate utilization (Fig. 1). The enzyme is formed by association of two evolutionarily homologous proteins, an a subunit designated PcaG and a b subunit designated PcaH (Fig. 3). Conserved amino acid sequences in the homologous subunits provide a structural basis for association of the proteins in a symmetrical a b array that confers catalytic function.
However, a hazard is presented by this demand for identical amino acid sequences because extensive nucleotide sequence identity, essential for coding of the conserved amino acid sequence, could serve as a basis for misalignment of DNA strands leading to deletion of intervening segments of pcaH and pcaG. Therefore, it is of interest that a region of pcaH where sequence repetition is demanded at the level of amino acid—and thus at the nucleotide level—is flanked by a region that exhibits a pronounced tendency to undergo deletion and duplication mutations (Fig. 3). A hypothesis prompted by these findings is that slippery DNA strands, the basis for the deleterious deletion and duplication mutations within pcaH (Fig. 3), might confer a selective benefit by impairing perfect alignment of oligonucleotides conserved in pcaH and pcaG (Fig. 3). Thus the strand slippage would deter large-scale recombination events that may cause big deletions removing portions of pcaH and pcaG.
The high frequency of spontaneous insertion and deletion mutations is a source of fascination because of the evidence it provides about how DNA changes. An example is PobR, the transcriptional activator that governs expression of pobA, the structural gene for the enzyme that converts p-hydroxybenzoate to protocatechuate (Fig. 2). Of spontaneous mutations causing loss of PobR function, 80% are caused by random introduction of an insertion sequence, IS1236, into its structural gene, pobR. Spontaneous mutations in other genes relatively rarely are caused by insertion of IS1236, so the results imply that pobR DNA has an unusual conformation rendering it susceptible to IS1236 insertion throughout the gene.
Plasticity of Proteins Revealed by Polymerase Chain Reaction Mutations
As intriguing as the spontaneous insertion and deletion mutations are, they hinder isolation of mutants containing amino acid substitutions that could be informative about how proteins such as PobR function. An avenue around this difficulty was created by development of a procedure for recovering Acinetobacter transformants that had acquired nucleotide substitution mutations during PCR amplification of a DNA fragment containing pobR.
This procedure combines the powerful natural transformation system of Acinetobacter sp. strain ADP1 with a counterselection procedure allowing recovery of strains that have lost pob function. In essence, successful transformants obtain mutated DNA fragments created by PCR, and selected colonies of these bacteria contain predominantly nucleotide substitutions. Since mutations are observed at the level of phenotype, it is possible to identify colonies containing conditional mutations causing subtle changes. These include unusual mutations. For example, two cold-sensitive pob mutants were not in the PobR protein itself but in the ribosomal binding site for the pobR transcript. Something like a car engine on a cold morning, the mutant pobR messengers do not turn over the translational apparatus at low temperatures.
Sequence analysis of PCR-generated mutations identifies amino acid residues that are essential for activity. Thus far, PCR mutagenesis has been applied successfully to structure-function analysis of PobR and VanAB, the dimeric enzyme that converts vanillate to protocatechuate (Fig. 2). Once a loss of function mutation is generated by PCR mutagenesis, an additional round of PCR mutagenesis can provide information about the tolerance for amino acid substitution at a given position. For example, the PCR-generated amino acid substitution Asp-228® Val blocks the activity of the VanA subunit. Selection for strains that have regained activity after PCR mutagenesis of the mutant gene invariably produces true revertants in which aspartate has replaced the mutant valine.
A contrasting situation is observed with Gln-328, an amino acid conserved in VanA and evolutionarily related proteins. Not surprisingly, substitution of Gln-328 disrupts VanA activity. More remarkably, activity can be restored by PCR-generated substitutions of chemically dissimilar lysyl or leucyl residues for the wild type glutaminyl residue at position 328. Therefore conservation of Gln-328 cannot be attributed to strict demand for its chemical properties at position 328 in VanA.
PCR mutagenesis followed by natural transformation made it possible to select Acinetobacter strains in which a gene that had undergone random PCR mutagenesis was selected to replace the function of a different, defective gene. In this procedure, the operator-binding functions of the transcriptional activators PobR and PcaU are swapped (Fig. 4). A similar approach was used to obtain mutants in which a single amino acid substitution allowed PcaHG to substitute for CatA in its activity towards catechol (Fig. 1).
These results are gratifying in that they suggest a flexibility that allows highly specialized proteins to assume different functions. Yet it is worth noting that repeated experiments always produced a single amino acid substitution: Gly-222® Val in PcaU, Thr-57® Ala in PobR, and Arg-133® His in PcaG. The appearances of flexibility actually were highly constrained responses to narrowly targeted selections. Largely unexplored, for example, are amino acid residues that allow proteins to assume effective conformations under a range of temperatures and other variables in the environment. Mutational analysis of changes in protein function can provide valuable insights, but these insights must be recognized as genetic reporting rather than as evolutionary history.
Application of Acinetobacter Genetics to DNA from Other Organisms
The combination of PCR mutagenesis and natural transformation offers many opportunities for genetic analysis of Acinetobacter genes, and in some instances it would be worthwhile to apply the process to a gene from other biological origins.
This goal can be achieved by inserting the gene into a plasmid containing a multiple cloning site interposed between DNA segments from the Acinetobacter chromosome. PCR amplification provides linear DNA containing randomly distributed mutations, and natural transformation introduces the mutated DNA into the chromosome. Mutations introduced in the foreign gene during PCR amplification then can be identified phenotypically from the pool of transformants by a screen or a selection specific to the foreign gene.
The Central Dilemma: How the Habitat Shapes the Genome
Biologists are familiar with the "Central Dogma"—information flows from DNA through RNA to protein and ultimately to an expressed phenotype. Microbiologists must face a greater challenge, which we call the "Central Dilemma"—how does information flow from the habitat to shape the individual genomes of different microbial species? To be sure, trial and error have much to do with the process, but what is tried and what errors are tolerated?
With respect to the bacterial protocatechuate pathway (Fig. 1), structural genes for enzymes have stood the test of time. In every known case, a common ancestor was the source of genes encoding enzymes with identical function. Dramatic differences are apparent in the organization of the pca and cat genes in different bacterial genera. It is evident that evolutionary divergence was punctuated by major gene rearrangements that subsequently were conserved in different biological groups.
Often correlated with differences in gene organization are distinctive patterns of transcriptional control. For example, b -ketoadipate, which serves no known regulatory role in Acinetobacter, is an inducer of many enzymes in Pseudomonas species. Intriguingly, the Pseudomonas transcriptional activator that responds tob -ketoadipate is a close evolutionary homolog of the Acinetobacter regulators PobR and PcaU that respond to p-hydroxybenzoate and protocatechuate respectively; all three activators bind to operators that are similar in nucleotide sequence. An exception to the general pattern of inducible control of enzyme synthesis is found in Bradyrhizobium species which constitutively express both enzymes for catabolism of b -ketoadipate and a chemotactic system that attracts the bacteria to the compound.
As nucleotide sequences become available, genes encoding transporters and porins appear as increasingly evident contributors to the biological individuality of organisms that employ theb -ketoadipate pathway. The importance of the membrane proteins was heralded by phenotypic characterization of transport systems that communicate from the habitat to the interior of the cell. For example, PcaT, a transporter for b -ketoadipate in fluorescent Pseudomonas species, appears to be a scavenger system because its expression is manifested as cells enter starvation. In another case, evolutionarily homologous transporters, PcaK and VanK, combine through inducible expression and readily reversible mutation to produce Acinetobacter cells that respond to the levels of protocatechuate that are presented by the environment. Improved knowledge of the functions of such membrane proteins will increase our understanding of the distinctive strategies that shaped the properties of different microbial groups.
ACKNOWLEDGMENTS
We thank Andrew B. Reams and Ellen L. Neidle for making information available to us prior to publication. Work in our laboratory has been supported by the Army Research Office, the Department of Energy and the National Science Foundation. This is publication 27 from the Biological Transformation Center in the Yale Institute for Biospheric Studies.
SUGGESTED READING
D'Argenio, D. A., M. W. Vetting, D. H. Ohlendorf, and L. N. Ornston. 1999. Substitution, insertion, deletion, suppression, and altered substrate-specificity in functional protocatechuate 3,4-dioxygenases. J. Bacteriol. 181:6478-6487.
Kok, R. G., D. A. D'Argenio, and L. N. Ornston. 1998. Mutation analysis of PobR and PcaU, closely related transcriptional activators in Acinetobacter. J. Bacteriol. 180:5058-5079.
Kok, R. G., D. M. Young, and L. N. Ornston. 1999. Phenotypic expression of PCR-generated random mutations in a Pseudomonas putida gene after its introduction into an Acinetobacter chromosome by natural transformation. Appl. Env. Microbiol. 65:1675-1680.
Kowalchuk, G. A., G. B. Hartnett, A. Benson, J. E. Houghton, K. L. Ngai, and L. N. Ornston. 1994. Contrasting patterns of evolutionary divergence within the Acinetobacter calcoaceticus pca operon. Gene 146:23-30.
Morawski, B., A. Segura, and L. N. Ornston. 2000. Substrate range and genetic analysis of Acinetobacter vanillate demethylase. J. Bacteriol. 182:1383-1389.
Parke, D. 1997. Aquisition, reorganization, and merger of genes: novel management of theb -ketoadipate pathway in Agrobacterium tumifaciens. FEMS Microbiol. Lett. 146:3-12.
Parke, D., 2000. Positive selection for mutations affecting bioconversion of aromatic compounds in Agrobacterium tumifaciens: analysis of spontaneous mutations in the protocatechuate 3,4-dioxygenase gene. J. Bacteriol. 182:6145-6153.
Parke, D., D. A. D'Argenio, and L. N. Ornston. 2000. Bacteria are not what they eat: that is why they are so diverse. J. Bacteriol. 182:257-263.
Parke, D., and L. N. Ornston. 1986. Enzymes of the b -ketoadipate pathway are inducible in Rhizobium and Agrobacterium spp. and constitutive in Bradyrhizobium spp. J. Bacteriol. 165:288-292.
Vetting, M. W., and D. H. Ohlendorf. 2000. The 1.8Å crystal structure of catechol 1,2-dioxygenase reveals a novel hydrophobic helical zipper as a subunit linker. Structure 8:429-440.
Email: webmaster@asmusa.org