THE POWER OF PROMISCUITY
Enzymes' ability to catalyze alternative reactions may provide a springboard for evolution
AMANDA YARNELL, C&EN WASHINGTON
We often think of enzymes as being exquisitely efficient and specific catalysts. And they are--particularly in comparison with other chemical catalysts. But in fact, many enzymes have been shown to catalyze not only their "natural" reaction but also one or more alternative reactions, albeit at low levels. And in a growing number of cases, the enzyme's "promiscuous" activity is similar to the natural activity of an evolutionarily related enzyme.
Such observations have led some scientists to suggest that nature tends to recruit promiscuous activities of existing enzymes to provide the jumping-off point for the evolution of new enzymes. And it should be easy for nature to introduce promiscuous activity, according to recent work by a team led by biochemistry professor John A. Gerlt of the University of Illinois, Urbana-Champaign, and Jeremy Minshull, formerly of Redwood City, Calif.-based Maxygen.
Many enzymes can accept a range of structurally similar substrates. For example, bacteria use an enzyme called alkaline phosphatase to generate the free phosphate that's necessary for growth. This enzyme can hydrolyze the phosphate monoesters of a wide variety of substrates.
Other enzymes go even further, not only accepting a range of substrates, but also performing seemingly disparate kinds of chemistry on each. One of the earliest demonstrations of such "promiscuity" was published in the 1960s by chemist Yeshayau Pocker, then at the University of Washington. Pocker showed that carbonic anhydrase II--a zinc metalloenzyme whose physiological role is to catalyze the addition of water to carbon dioxide--also has a weak esterase activity.
Chymotrypsin, a well-studied enzyme that digests proteins in the stomach and small intestine, is also promiscuous. It's capable of catalyzing not only its normal amidase reaction but also a mechanistically distinct phosphotriesterase reaction.
That these and many other enzymes can catalyze a host of mechanistically distinct side reactions may seem counterintuitive. Enzymes' remarkable catalytic prowess, after all, is normally chalked up to their ability to perfectly stabilize a particular reaction's transition state. So it may seem surprising that such highly specialized catalysts could be so versatile.
But in fact, most enzyme active sites are littered with potential catalytic groups, notes Daniel Herschlag, a professor of biochemistry at Stanford University. Such functional groups include metal ions and redox cofactors, as well as amino acid side chains that can act as general acids and bases and nucleophiles. "Most enzyme active sites have great chemical potential," Herschlag tells C&EN.
PRIME TARGET Enzyme active sites provide a constellation of potential catalytic residues. Here, the active sites of three evolutionarily related enzymes--MLE, OSBS, and AEE--are nearly identical. Even though they use different substrates and catalyze different reactions, each forms an enolate intermediate. (Oxygens and nitrogens are shown in red and blue, respectively; MLE's carbons are shown in gray, OSBS's in yellow, and AEE's in bluish green.)
COURTESY OF J. GERLT
HARNESSING THAT POTENTIAL can give rise to promiscuous reactions, Herschlag adds. As evidence, he points to the catalytic antibody 38C2, which was originally designed to do aldol reactions. A lysine residue in this antibody's active site catalyzes aldol condensations and other reactions by acting as a nucleophile during the formation of a reactive eneamine intermediate. But Dan S. Tawfik, a senior scientist at Israel's Weizmann Institute, has shown that a lysine in this catalytic antibody's active site can also act as a general base in the Kemp elimination [Protein Sci., 10, 2600 (2001)].
Herschlag suggests that the potential for promiscuity in enzyme active sites may be the key to the evolution of new enzymes, giving nature a head start toward evolving an enzyme that's perfectly suited to catalyzing the promiscuous reaction.
That idea--first formulated in the 1970s by microbiologist Roy A. Jensen, then at the State University of New York, Binghamton--now is being bolstered by the growing appreciation for the evolutionary history of enzymes. Aided by rapidly accumulating genomic sequences, the burgeoning number of three-dimensional structures of enzymes, and the development of powerful bioinformatics tools, enzymologists are beginning to piece together enzyme family trees. As a consequence, in a rapidly growing number of examples, an enzyme that weakly catalyzes a given alternative reaction is evolutionarily related to the enzyme whose biological job is to catalyze that reaction.
Take the vanadium-dependent chloroperoxidase from the fungus Curvularia inaequalis. This enzyme's sequence and structure--particularly in the active site--are very similar to those of a family of phosphatase enzymes, Herschlag notes. "But the phosphatases catalyze phosphoryl group transfer and chloroperoxidase catalyzes the transformation of hydrogen peroxide to chloroperoxide, presumably via a radical reaction," he points out. "Nature has realized that she can do completely different chemistry in the same active site."
A number of research groups have shown that, although the primary catalytic function of chloroperoxidase is to convert Cl into ClOH using a vanadate cofactor, the enzyme also has weak phosphatase activity in the absence of this cofactor. Furthermore, a member of this phosphatase family can bind vanadate and catalyze the haloperoxidation reaction. These observations raise the possibility that a phosphatase may have provided the starting point for evolution of chloroperoxidase--or vice versa.
Then there is atrazine chlorohydrolase, an enzyme produced by soil bacteria that can dechlorinate the herbicide atrazine. This enzyme's amino acid sequence is 98% identical to that of melamine deaminase, a soil bacteria enzyme that deaminates melamine (2,4,6-triamino-1,3,5-triazine). But University of Minnesota, St. Paul, biochemist Lawrence P. Wackett has shown that, whereas atrazine chlorohydrolase has no detectable deaminase activity, melamine deaminase weakly catalyzes dechlorination. This finding has led him to suggest that atrazine chlorohydrolase may have evolved from melamine deaminase after atrazine was introduced into the environment in the 1950s [Biochemistry, 40, 12747 (2001)].
PROMISCUITY MADE EASY A single amino acid change in the active site of MLE or AEE gives both enzymes the ability to weakly catalyze the reaction normally performed by OSBS.
IT'S COMMONLY THOUGHT that in cases like these, divergent evolution began with duplication of the gene encoding the parent enzyme. This would have allowed the organism to tweak the extra gene to evolve a gene encoding a new enzyme with the desired activity--without losing the old (presumably useful) enzymatic activity.
But how does the organism get from here to there? In one scenario, the parent enzyme doesn't catalyze the new reaction. But by making enough random mutations to the parent enzyme's gene, the organism can produce an enzyme that does. Herschlag finds this scenario highly improbable.
Herschlag and his former graduate student Patrick J. O'Brien have suggested that it's far more likely that nature prefers to start with a promiscuous enzyme--that is, one that already has the ability to catalyze the new reaction, albeit relatively poorly [Chem. Biol., 6, R91-R105 (1999)]. Optimizing this existing activity is likely to require many fewer mutations than would be needed to start from scratch. Therefore, in terms of evolutionary potential, promiscuous enzymes have a selective advantage over more discriminating enzymes, Herschlag and O'Brien have suggested.
Gerlt and Minshull's team recently showed how uncomplicated it can be to induce promiscuity. They studied three structurally homologous members of the enolase superfamily of enzymes: L-Ala-D/L-Glu epimerase (AEE, which catalyzes a 1,1 proton-transfer reaction), muconate-lactonizing enzyme II (MLE, which catalyzes a cycloisomerization reaction), and o-succinylbenzoate synthase (OSBS, which catalyzes a b-elimination/dehydration reaction).
Each enzyme forms an enolate intermediate, though none of them displays any detectable promiscuity, Gerlt says. But by changing a single amino acid in AEE and MLE, the team managed to coax these enzymes into weakly catalyzing the OSBS reaction [Biochemistry, 42, 8387 (2003)]. The only difference: The Illinois team members chose the promiscuous mutation in AEE by comparing the enzyme's structure with that of OSBS, whereas the Maxygen team members independently found the homologous mutation in MLE by directed-evolution experiments.
"We started with two enzymes that, as best as we can tell, were not promiscuous and demonstrated that with a single substitution we could generate promiscuity," Gerlt tells C&EN. Others have introduced promiscuous activities into enzymes, but never with a single mutation, he says. That such a small change could have such an effect "simplifies nature's problem of evolving new catalysts," he suggests. And it implies that both promiscuous and genetically near-promiscuous enzymes can be a springboard for evolution.
Catalytic promiscuity's potential impact on the evolution of enzymes aside, enzymes' ability to catalyze alternative reactions has practical implications, too. Scientists hoping to exploit enzymes for organic synthesis may be able to glean something from the lessons learned about catalytic promiscuity, Tawfik suggests. "If you need to create an enzyme with a given new function, you'd better start with an enzyme that can do this reaction--even if it's at a very slow rate--rather than start from scratch," he says. He admits, however, that many scientists attempting to create enzymes for their own purposes have already learned that lesson, in many cases the hard way.
SO PERHAPS more interesting is how the promiscuous activities of enzymes might be systematically characterized. So far, enzymes' promiscuous activities have been discovered either by chance or by looking for a specific reaction based on an enzyme's close relatives. But "no one has tried to systematically screen for promiscuity," Tawfik says. He suspects that the promiscuity of enzymes might be tightly linked to the conformational plasticity of proteins.
His own group is screening small libraries of metalloenzymes for the ability to catalyze promiscuous hydrolytic reactions. Still, a truly random, broad screen for promiscuous activities remains a daunting technical challenge, he says. But when this challenge is overcome, Tawfik predicts, "I think we'll find that enzymatic promiscuity is both widespread and highly useful."