Evolutionary Potential
Determining the of a Gene
Barry G. Hall and Harmit S. Malik
Biology Department, University of Rochester
In addition to information for current functions, the sequence of a gene includes potential information for the evolution of new functions. The wild-type ebgA (evolved b-galactosidase) gene of Escherichia coli encodes a virtually inactive b-galactosidase, but that gene has the potential to evolve sufficient activity to replace the lacZ gene for growth on the b-galactoside sugars lactose and lactulose. Experimental evidence, which has suggested that the evolutionary potential of Ebg enzyme is limited to two specific amino acid replacements, is limited to examining the consequences of single base-substitutions. Thirteen b-galactosidases homologous with the Ebg b-galactosidase are widely dispersed, being found in gram-negative and gram-positive eubacteria and in a eukaryote. A comparison of Ebg b-galactosidase with those 13 b-galactosidases shows that Ebg is part of an ancient clade that diverged from the paralogous lacZ b-galactosidase over 2 billion years ago. Ebg differs from other members of its clade at only 2 of the 15 active-site residues, and the two mutations required for full Ebg b-galactosidase activity bring Ebg into conformity with the other members of its clade. We conclude that either these are the only acceptable amino acids at those positions, or all of the single-base-substitution replacements that must arise as intermediates on the way to other acceptable amino acids are so deleterious that they constitute a deep selective valley that has not been traversed in over 2 billion years. The evolutionary potential of Ebg is thus limited to those two replacements.
Introduction
Encoded within the genome of each organism is the information for all of the functions necessary for survival and reproduction in that organism’s current environment. Also present within that genome is the potential for that organism’s evolution of novel functions for success in new or alternative environments. In a trivial sense, each genome’s potential is infinite, because given enough additions, deletions, rearrangements, and base substitutions, any sequence can evolve into any other sequence. In reality, however, evolution is subject to a variety of constraints that limit this potential, and understanding evolutionary processes amounts to understanding those constraints.
Evolutionary biologists are usually forced to infer historical evolutionary processes by examining the present-day outcomes of those processes. That is an unsatisfactory means of understanding a dynamic process, partly because neither the historical selective constraints nor the detailed molecular functions of ancestral states are well understood.
Experimental evolution of novel enzyme functions in microbial populations provides a powerful alternative approach to understanding evolutionary processes. The typical strategy is to apply strong selection for catabolism of a novel resource to a large population of the model microorganism, and to determine from the resulting spontaneous mutants the detailed changes that have given rise to the newly selected phenotype. Even given the enormous sizes of experimental microbial populations, often in excess of 1011 individuals, in vivo systems are limited to examining the outcomes that can be produced by one or two mutations. Within that limitation, the array of successful mutants defines the evo-
Key words: Ebg, b-galactosidase, evolutionary potential.
Address for correspondence and reprints: Barry G. Hall, Biology Department, Hutchison Hall, River Campus, University of Rochester, Rochester, New York 14627. E-mail: drbh@uhura.cc.rochester.edu.
Mol. Biol. Evol. 15(8):1055–1061. 1998 q 1998 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038 |
lutionary potential of the selected gene for the chosen novel function.
For over 25 years, the Ebg (evolved b-galactosidase) operon of Escherichia coliK12 has been used as a model system to study the evolution of new metabolic functions (reviewed in Hall 1990). More recently, those genetic studies have been complemented by detailed biochemical studies of catalysis mediated by the Ebg enzyme (Hall 1981; Burton and Sinnott 1983; Hall et al. 1983; Li, Osborne, and Sinnott 1983; Elliott et al. 1992; Srinivasan et al. 1993; Calugaru, Hall, and Sinnott 1995, Srinivasan, Hall, and Sinnott 1995; Calugaru et al. 1997). The operon consists of four genes: ebgR, which encodes a repressor that controls expression of ebgACB (Hall and Hartl 1975; Hall and Clarke 1977; Hall 1978b; Hall, Betts, and Wootton 1989); ebgA, which encodes the 118-kDa a subunit of Ebg enzyme (Hall, Betts, and Wootton 1989); ebgC, which encodes the 20-kDa b subunit of Ebg enzyme (Hall, Betts, and Wootton 1989); and ebgB, whose function is unknown (Hall and Zuzel 1980), but whose sequence (YGJI ECOLI gene product, GenBank/SwissProt accession number P42590) suggests that it is a transport protein. The wildtype Ebg enzyme is such a feeble b-galactosidase that even when the operon is expressed constitutively (ebgR2) such that Ebg enzyme constitutes 5% of the soluble cell protein, DlacZ cells cannot utilize b-galactoside sugars as sole sources of carbon and energy (Hall 1982a). Studies that have examined 154 independent spontaneous mutants, 119 UV-induced mutants, 45 mutS-induced mutants, 52 mutY-induced mutants, and 99 mutD-induced mutants have identified only two sites in ebgA where mutations occur that can enhance the activity of Ebg enzyme sufficiently to permit utilization of lactose (galactosyl-1,4-b-D-glucose) or lactulose (galactosyl-1,4-b-D-fructose) (Hall 1995; Hall, Betts, and Wootton 1989). Numbering the Ebg operon sequence as in Hall, Betts, and Wootton (1989) (GenBank accession M64441), a G1566→A mutation (previously called the Class I site) results in an asp92 (D)→asn (N) replacement. That replacement increases the activity (kcat/km) of
1055
Table 1 Protein Sequences Used in this Study Organism and Enzyme Accession Number Classification Actinobacillus pleuropneumoniae b-galactosidase. . . . . . . . . . . 1655857 Eubacteria/Proteobacteria (gamma subdivision)/Pasteurellaceae Arthrobacter sp. b-galactosidase. . . . . . . . . . . . . . . . . . . . . . . . . 2127398 Eubacteria/Firmicutes/Actinomycetes Clostridium acetobutylicum b-galactosidase. . . . . . . . . . . . . . . . 1352076 Eubacteria/Firmicutes/low GC gram positive/Clostridiaceae Enterobacter cloacae b-galactosidase. . . . . . . . . . . . . . . . . . . . . 1091877 Eubacteria/Proteobacteria (gamma subdivision)/Enterobacteriaceae Escherichia coli b-glucuronidase . . . . . . . . . . . . . . . . . . . . . . . . 584839 Eubacteria/Proteobacteria (gamma subdivision)/Enterobacteriaceae Escherichia coli lacZ b-galactosidase. . . . . . . . . . . . . . . . . . . . . 2623984 Eubacteria/Proteobacteria (gamma subdivision)/Enterobacteriaceae Escherichia coli Ebg b-galactosidase . . . . . . . . . . . . . . . . . . . . . 114935 Eubacteria/Proteobacteria (gamma subdivision)/Enterobacteriaceae Homo sapiens b-glucuronidase . . . . . . . . . . . . . . . . . . . . . . . . . . 114963 Eukaryotae/Metazoa/Vertebrata/Primates Klebsiella pneumoniae b-galactosidase. . . . . . . . . . . . . . . . . . . . 114941 Eubacteria/Proteobacteria (gamma subdivision)/Enterobacteriaceae Kluyveromyces lactis b-galactosidase . . . . . . . . . . . . . . . . . . . . . 399112 Eukaryotae/Fungi/Ascomycota/Saccharomycetaceae Lactobacillus delbreuckii bulgaricus. . . . . . . . . . . . . . . . . . . . . . 114943 Eubacteria/Firmicutes/low GC gram positive/Lactobacillaceae Lactobacillus sake b-galactosidase . . . . . . . . . . . . . . . . . . . . . . . 1223762/1223763a Eubacteria/Firmicutes/low GC gram positive/Lactobacillaceae Lactococcus lactis b-galactosidase . . . . . . . . . . . . . . . . . . . . . . . 1556406 Eubacteria/Firmicutes/low GC gram positive/Streptococcaceae Staphylococcus xylosus b-galactosidase . . . . . . . . . . . . . . . . . . . 2462706 Eubacteria/Firmicutes/low GC gram positive/Bacillaceae Streptococcus salivarius thermophilus b-galactosidase . . . . . . . 153688 Eubacteria/Firmicutes/low GC gram positive/Streptococcaceae Thermotoga maritima b-galactosidase . . . . . . . . . . . . . . . . . . . . 473272 Eubacteria/Thermotogales a The alignment utilized both the large and small subunits of the Lactobacillus sake b-galactosidase, which consists of two subunits as the result of splitting the gene. For purposes of this alignment, the N-terminus of the small subunit was joined to the C-terminus of the large subunit. |
Ebg enzyme toward lactose 46-fold (Hall 1981) and allows the mutant to utilize lactose. However, because that mutation increases activity toward lactulose only 1.4fold, those mutants cannot utilize lactulose. An alternate G4223→(T or C) mutation (previously called the Class II site) results in a trp977 (W)→cys (C) replacement. That substitution increases the activity of Ebg enzyme 11fold toward lactose, increases the activity 49-fold toward lactulose (Hall 1981), and allows those mutants to utilize both lactose and lactulose.
Because all of the spontaneous and induced singlestep mutants selected for growth on lactose or lactulose have one of the above mutations, it was previously suggested that the evolutionary potential of the ebgA gene to promote effective b-galactosidase activity is limited to those two changes (Hall 1995).
Because the experimental system is limited to investigating the consequences of single base-substitutions, it was not possible to determine whether multiple base-substitutions might yield equal or better improvements in catalytic efficiency. If such multiple changes are possible, then the limitations of the experimental system might have given a false impression of a very limited evolutionary potential for Ebg enzyme.
One approach to resolving this issue would be to use in vitro site-directed mutagenesis, in which multiple base-substitutions are introduced to individual codons, or in which saturation mutagenesis is applied to a series of two or three adjacent codons to generate all possible combinations of amino acid substitutions in candidate regions. The latter approach has been used very successfully to define the set of amino acid replacements that can greatly improve the activity of the TEM-1 blactamase (Huang et al. 1996). However, the results of site-directed mutagenesis experiments cannot provide the necessary information about ‘‘real’’ evolutionary potential. Even if some multiple mutations were found to give equal or better activity than the single mutations already identified, we would not know if those changes could evolve naturally, i.e., if they were within the natural evolutionary potential of the genome. Because multiple mutations require that the organism traverse one or more intermediate steps, evolutionary potential is limited by the fitness associated with those intermediate steps. For some possible protein changes, it may well be the case that ‘‘you can’t get there from here’’ because the intervening single-step mutations are too deleterious.
Because neither in vivo nor in vitro experimental evolution reveals whether the observed evolutionary potential for b-galactosidase activity is actually limited to those two amino acid replacements, we have returned to comparisons of the outcomes of nature’s 3.8-billion-year experiment with life to shed light on the issue.
Materials and Methods
The BLAST program (Altschul et al. 1990) was used to search the non-redundant protein database for sequences related to that of the EbgA protein. Fourteen b-galactosidases were identified in which the probability of a match by chance alone was ,10290. Table 1 lists those sequences and associated accession numbers. The
FIG. 1.—Anchor sites used in the alignment of b-galactosidases. The alignment of the most conserved stretches among the different bgalactosidases and b-glucuronidases is shown. The numbers on top of each segment are based on the EbgA protein numbering, with the beginning and end of each anchor site indicated. Numbers at the bottom refer to the E. coli b-glucuronidase numbering. Previously identified active-site residues are highlighted in boldface. The Lactobacillus sake lacZ indicated here is a combination of the large and small subunits (table 1). ‘‘*’’ |
and ‘‘.’’ indicate identical and similar amino acid residues, respectively.
14 sequences were initially aligned by the CLUSTAL W 1.6 program (Thompson, Higgins, and Gibson 1994) using the BLOSUM similarity matrix, a multiple-alignment gap opening penalty of 10, and a gap extension penalty of 0.05 (default parameters). That initial alignment permitted the identification of 19 strongly conserved regions that we term ‘‘anchor sites’’ (fig. 1). Each of the regions between those sites, together with the flanking anchor sites, were then independently aligned using a multiple-alignment gap opening penalty of 20
FIG. 2.—Phylogeny of the b-galactosidases. A, Distance tree based on the Neighbor-Joining method. Bootstrap values are indicated as percentages based on 1,000 replicates. B, Maximum-parsimony tree for the b-galactosidases, with bootstrap values indicated. In both A and B, the b-glucuronidases are used to conceptually root the trees as described in the text. Actual bootstrap values are based on analysis of entire bgalactosidase sequences. By both methods, the b-galactosidases split into three distinct clades. The branches of the b-glucuronidases are shown in gray to indicate that the rooting is conceptual and that the lengths of the b-glucuronidase branches are unrelated to actual distances. |
and a gap extension penalty of 0.5. The objective of this approach was to ensure an optimal alignment of the intervening sequences between the conserved ‘‘anchor sites.’’ The inclusion of the anchor sites ensures that the ends of each of those regions remain aligned. Finally, the individually aligned segments were reassembled with the aligned anchor sites. The resulting final alignment has been deposited with the EMBL under accession number DS32829.
To provide an outgroup for phylogenetic tree construction a human b-glucuronidase was aligned with an E. coli b-glucuronidase (table 1), and the b-glucuronidase alignment was aligned to the b-galactosidase alignment using the ‘‘profile alignment’’ option of CLUSTAL W. That option retains all of the gaps within each alignment, but permits the introduction of new gaps between the alignments. Only a segment of b-glucuronidase is homologous to the b-galactosidases. Thus, to maximize phylogenetic resolution, we use that homologous segment of the b-glucuronidases only to root the tree.
The final alignment was used to construct a distance tree (fig. 2A) by the Neighbor-Joining method (Saitou and Nei 1987) as implemented by CLUSTAL W 1.6 with 1,000 bootstrap replications and to construct a parsimony tree (fig. 2B) with PAUP 3.1.1 (Swofford 1993) using the heuristic search method of branch swapping with stepwise addition of closest neighbors with 100 bootstrap replications.
Results and Discussion
By far, the most thoroughly studied of these b-galactosidases is that encoded by the lacZ gene of E. coli. That enzyme has now been crystallized, and the 15 active site residues identified by X-ray crystallography (Jacobson et al. 1994) are highlighted by boldface in figure 1. Twelve of these residues are conserved across all bgalactosidases. Phylogenetic analyses indicate that the b-galactosidases fall into three distinct clades, by both maximum-parsimony and Neighbor-Joining methods (fig. 2). Clade A includes representatives from E. coli (ebgA), Arthrobacter sp., Actinobacillus pleuropneumoniae, Staphylococcus xylosus, and Kluyveromyces lactis. Clade B includes Thermotoga maritima, Lactobacillus sake, Lactobacillus delbrueckii, Clostridium acetobutylicum, and Streptococcus salivarius (subspecies thermophilus), while clade C includes b-galactosidases from E. coli (lacZ), Lactococcus lactis, Klebsiella pneumoniae, and Enterobacter cloacae. The observation that several amino acid replacements in active-site-containing anchor sites occurred along the branch that leads to clade BC supports the existence of clade A either as a real clade or as an arbitrary group that includes everything that is not clade BC.
Resolving the apparent trichotomy of the three clades of b-galactosidases to assign the correct rooting of the trees was achieved by utilizing the b-glucuronidases, a related family of sugar-hydrolyzing enzymes. Amino acids 269–445 of the b-glucuronidases are homologous to residues 315–476 of the b-galactosidases. The inherent assumption is of related ancestry, a premise justified by a high degree of sequence conservation. For example, four of the six active-site residues are conserved over this homologous stretch (fig. 1). Phylogenies based on that region of homology clearly show that the b-glucuronidases form an outgroup with respect to all of the b-galactosidases. The possibility that b-glucuronidases are merely a recently derived subclade of the b-galactosidases (and thereby not an appropriate outgroup) is belied by the presence of very divergent bglucuronidases in E. coli and mammals. Rooting the tree conceptually on the b-glucuronidases gives us a rooted tree of the 14 b-galactosidases, shown in figure 2. While it may appear that the addition of the b-glucuronidases ‘‘breaks up’’ clade A, it should be noted that this is merely a lack of the ability to resolve clade A. The bgalactosidases emerge as an ancient clade of enzymes, with their origin well before the inception of mammals, the major extant source of lactose. Clearly, the suggestion emerges that this ancient class of enzymes may have historically been preserved to hydrolyze a b-galactoside other than lactose. The lacZ and ebgA genes of E. coli are paralogs that resulted from a very ancient gene duplication event. Their divergence dates back
FIG. 3.—Some amino acid replacements that are common to a lineage. A schematic tree of the b-galactosidases is shown, with clade A as the ancestral clade. Amino acid replacements are within anchor sites that include active-site residues. Active-site residues are highlighted by outlining. Note that the C977→W substitution that occurred along the branch to clade BC also occurred in Ebg. Note that position 975 is a gap in all members of clade A, while it is a conserved aspartic acid (D) in all members of clade BC.
close to the root of all b-galactosidases, predates the 2.2billion-year-old divergence of gram-positive and gramnegative Eubacteria (Feng, Cho, and Doolittle 1997), and may even predate the divergence of Eubacteria and Eukaryota, between 3 and 4 billion years ago (Feng, Cho, and Doolittle 1997). While a rigorous maximumlikelihood analysis might slightly modify branching orders that are weakly supported by distance and parsimony, it is both unlikely and without precedent that maximum-likelihood would give contrary results to well-supported nodes. Thus, a time-consuming and resource-intensive maximum-likelihood analysis would not alter the conclusions of this work.
Within the conserved anchor regions that include the active-site residues (see Materials and Methods), the sequences of the enzymes from clade A share several features that further support the notion that they belong to a the same clade (summarized in fig. 3). Thus, for the ebgA enzyme, cys977 (C) is acceptable only within the context of clade A, with the trp (W)977→cys (C) mutation bringing ebgA into close conformity with the ancestral clade to which it clearly belongs. The asp (D)92→asn (N) replacement, on the other hand, brings ebg into conformity with all the other enzymes. Given that uniformity, it might well be supposed that an asparagine at that position is an absolute requirement for effective activity, but that is not the case. The Ebg trp977→cys mutant enzyme, with sufficient activity to permit growth on lactose and lactulose, retains the aspartic acid (D) at position 92.
At position 977, most enzymes have a tryptophan; only clade A enzymes have a cysteine at that site. Ebg is an exception to the clade A enzymes at position 977; it has a tryptophan at that site. The trp977→cys substitution, however, brings Ebg enzyme into conformity with the other members of that group.
The asp92→asn 1 trp977→cys double mutant Ebg enzyme is identical to the other clade A enzymes at all 15 active site residues. That double-mutant enzyme (designated Class IV [Hall 1978a]) differs profoundly from either of the single-mutant enzymes. Its activity toward lactose is 466 times greater than that of the wildtype enzyme, with most of that increase resulting from a dramatic drop in the km (Hall 1981). The double-mutant enzyme hydrolyzes galactosyl-b-D-arabinose well enough to permit growth on that sugar, while neither of the single-mutant enzymes permits such growth (Hall 1978a, 1981). The double mutant enzyme, unlike either of the single-mutant enzymes, converts lactose into an inducer of the lac operon (Rolseth, Fried, and Hall 1980; Hall 1982b).
Mutagenesis with UV light and with several mutator alleles of different specificity showed that the failure to find spontaneous Ebg mutations other than asp92→asn and trp977→cys was not the result of a mutational bias imposed by the spontaneous mutagenesis machinery of E. coli (Hall 1995), but that study could not determine whether other substitutions, such as asp92(D)→gln (G), for instance, might have produced equally improved enzyme. Figure 4 shows the amino acid replacements that could result from each of the possible single-base substitutions at positions 92 and 977. The experimental evidence shows that none of those replacements allows effective lactose or lactulose hydrolysis. The phylogenetic evidence indicates either that asn92 and cys/trp977 are the only acceptable amino acids at those positions, or that all of the single base-substi-
FIG. 4.—Evolutionary potential at two active sites of the ebgA gene. A, Residue 92 of EbgA. B, Residue 977 of EbgA. The darkly shaded boxes represent amino acids found at those positions in all bgalactosidases. The lighter shading indicates single-base substitutions which are never observed in experimental studies of Ebg evolution and thus must be deleterious to b-galactosidase activity. The unshaded residues are codon changes which can be reached only after traversing the deleterious single-site mutations.
tution replacements are so deleterious that they constitute a deep selective valley that has not been traversed in over 2 billion years. In either case, the natural evolutionary potential of Ebg enzyme with respect to lactose and lactulose hydrolysis is limited to those two replacements.
Ebg arose from a functional clade A b-galactosidase. The ‘‘natural’’ function of Ebg enzyme is unknown, but, given the observed degree of sequence and open reading frame conservation, it is unlikely that Ebg performs no function and is therefore not under selection. The only two active-site residues that differentiate Ebg from the other Clade A enzymes are asp92 and trp977. Neither the asn92→asp nor the cys977→trp substitution would have eliminated effective b-galactosidase activity of Ebg’s ancestor, but those two substitutions together certainly did so. Not only is the wild-type Ebg virtually inactive toward lactose and lactulose, but its activity toward eight other b-galactosides is so low that it is not, in any meaningful sense, a functional b-galactosidase. All of the b-galactosidases in this study have been identified on the basis of their lactase activity. Since it is clear that b-galactosidases arose long before lactose was being produced by mammals, it might be the case that some lactose-negative organisms retain homologs of these b-galactosidases. It will be interesting to see if, as more sequences accumulate in the databases, non-lactase proteins that belong to clade A begin to appear. Despite the strong conservation of proteins in the ancient clade A, it is clear that those proteins do not provide a function that is universally required, because no clade A homologs, as detected by BLAST, are present in any of the five other genomes that have been completely sequenced.
Acknowledgments
This study was supported by grant #NP-932 from the American Cancer Society. We are grateful to Simon Easteal for an essential suggestion about the fitness of possible multiple base-substitution mutants.
LITERATURE CITED
ALTSCHUL, S. F., W. GISH, W. MILLER, E. W. MYERS, and D. J. LIPMAN. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403–410.
BURTON, J., and M. L. SINNOTT. 1983. Catalytic consequences of experimental evolution. Part 1. Catalysis by the wildtype second b-galactosidase (ebgo) of Escherichia coli: a comparison with the lacZ enzyme. J. Chem. Soc. Perkin Trans. II 359–364.
CALUGARU, S. V., B. G. HALL, and M. L. SINNOTT. 1995. Catalysis by the large subunit of the second b-galactosidase of Escherichia coli in the absence of the small subunit. Biochem. J. 312:281–286.
CALUGARU, S. V., S. KRISHNAN, C. J. CHANY II, B. G. HALL, and M. L. SINNOTT. 1997. Larger increases in sensitivity to paracatalytic inactivation than in catalytic competence during experimental evolution of the second b galactosidase of Escherichia coli. Biochem. J. 325:117–121.
ELLIOTT, A. C., S. K. M. L. SINNOTT, P. J. SMITH, J. BOMMUSWAMY, Z. GUO, B. G. HALL, and Y. ZHANG. 1992. The catalytic consequences of experimental evolution. Studies on the subunit structure of second (ebg) b-galactosidase of Escherichia coli, and on catalysis by ebgab, an experimental evolvant containing two amino acid substitutions. Biochem. J. 282:155–164.
FENG, D.-F., G. CHO, and R. F. DOOLITTLE. 1997. Determining the divergence times with a protein clock: update and reevaluation. Proc. Natl. Acad. Sci. USA 94:13028–13033.
HALL, B. G. 1978a. Experimental evolution of a new enzymatic function: II. Evolution of multiple functions for EBG enzyme in E. coli. Genetics 89:453–465.
. 1978b. Regulation of newly evolved enzymes. IV. Directed evolution of the ebg repressor. Genetics 90:673–691.
. 1981. Changes in the substrate specificities of an enzyme during directed evolution of new functions. Biochemistry 20:4042–4049.
. 1982a. Evolution of a regulated operon in the laboratory. Genetics 101:335–344.
. 1982b. Transgalactosylation activity of ebg b-galactosidase synthesizes allolactose from lactose. J. Bacteriol. 150:132–140.
. 1990. Directed evolution of a bacterial operon. BioEssays 12:551–558.
. 1995. Evolutionary potential of the ebgA gene. Mol. Biol. Evol. 12:514–517.
HALL, B. G., P. W. BETTS, and J. C. WOOTTON. 1989. DNA sequence analysis of artificially evolved ebg enzyme and ebg repressor genes. Genetics 123:635–648.
HALL, B. G., and N. D. CLARKE. 1977. Regulation of newly evolved enzymes. III. Evolution of the ebg repressor during selection for enhanced lactase activity. Genetics 85:193– 201.
HALL, B. G., and D. L. HARTL. 1975. Regulation of newly evolved enzymes. II. The ebg repressor. Genetics 81:427– 435.
HALL, B. G., M. MURRAY, S. OSBORNE, and M. L. SINNOTT. 1983. The catalytic consequences of experimental evolution. Part III. Construction of reaction profiles for hydrolysis of lactose by ebgo, ebga, and ebgb enzymes via measurements of the enzyme-catalyzed exchange of galactose1-18O by 13C NMR spectroscopy. J. Chem. Soc. Perkin Trans. II 1595–1598.
HALL, B. G., and T. ZUZEL. 1980. The ebg operon consists of at least two genes. J. Bacteriol. 144:1208–1211.
HUANG, W., J. PETROSINO, M. SIRSCH, P. S. SHENKIN, and T. PALZKILL. 1996. Amino acid sequence determinants of blactamase structure and activity. J. Mol. Biol. 258:688–703.
JACOBSON, R. H., X. J. ZHANG, R. F. DUBOSE, and B. W. MATTHEWS. 1994. Three-dimensional structure of beta-galactosidase from E. coli. Nature 369:761–766.
LI, B. F. L., S. OSBORNE, and M. L. SINNOTT. 1983. Catalytic consequences of experimental evolution. Part 2. Rate-limiting degalactosylation in the hydrolysis of Aryl b-D-galactopyranosides by the experimental evolvants ebga and ebgb. J. Chem. Soc. Perkin Trans. II 365–369.
ROLSETH, S. J., V. A. FRIED, and B. G. HALL. 1980. A mutant ebg enzyme that converts lactose into an inducer of the lac operon. J. Bacteriol. 142:1036–1039.
SAITOU, N., and M. NEI. 1987. The neighbor-joining method: a new method for reconstructring phylogenetic trees. Mol. Biol. Evol. 4:406–425.
SRINIVASAN, K., B. G. HALL, and M. L. SINNOTT. 1995. The catalytic consequences of experimental evolution. Catalysis by a ‘‘third generation’’ evolvant of the second b-galactosidase of Escherichia coli, Ebgabcde and Ebgabcd, a ‘‘second generation’’ evolvant containing two supposedly ‘‘kinetically silent’’ mutations. Biochem. J. 312:971–977.
SRINIVASAN, K., A. KONSTANTINDIS, M. L. SINNOTT, and B. G. HALL. 1993. Large changes of transition state structure during experimental evolution of an enzyme. Biochem. J. 291:15–17.
SWOFFORD, D. L. 1993. PAUP: phylogenetic analysis using parsimony. Illinois Natural History Survey, Champaign.
THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673–4680.
SIMON EASTEAL, reviewing editor
Accepted May 11, 1998