Portal de Periódicos da CAPES

One, Two, Infinity

2002; Elsevier BV; Volume: 9; Issue: 12 Linguagem: Inglês

10.1016/s1074-5521(02)00286-7

1879-1301

Larry Gold, Ed Brody, Joe Heilig, Britta Swebilius Singer,

CRISPR and Genetic Engineering

Individual cells limit production of unnecessary metabolic capacities by monitoring their immediate environment and regulating gene expression appropriately. This presumably results in greater efficiency and energy utilization while allowing rapid response to environmental changes. Cells respond to changes in individual metabolites such as carbon and nitrogen sources or vitamins by regulating expression of genes necessary for synthesis or utilization of the specific metabolite. Curiously, although the ultimate chemical signal to which cells respond is the metabolite itself, until now the regulatory factotum usually has been a protein. Ron Breaker and his colleagues [1Nahvi A. Sudarsan N. Ebert M.S. Zou X. Brown K.L. Breaker R.R. Chem. Biol. 2002; 9: 1043-1049Abstract Full Text Full Text PDF PubMed Scopus (586) Google Scholar, 2Winkler W. Nahvi A. Breaker R.R. Nature. 2002; 419: 952-956Crossref PubMed Scopus (964) Google Scholar] have published two papers that demonstrate regulation of translation through direct interaction of a small molecule metabolite with mRNAs encoding proteins necessary for their uptake or utilization. In the first paper, the btuB leader sequence, 5′ to the translational initiation domain for btuB, interacts directly with coenzyme B12 in such a way as to cause structural changes to the RNA that sequester the translational initiation domain from ribosome inspection. The btuB gene encodes the receptor by which B12 is transported into the cell. Hence, the impact of this regulatory loop is to reduce synthesis of the transporter when intracellular B12 levels are sufficient. In the second paper, the 5′ leaders of thiM and thiC mRNAs are shown to bind directly to thiamine (vitamin B1), again leading to an mRNA structural alteration such that the biosynthetic enzymes are depressed (translationally) when the intracellular thiamine levels are sufficient. The mechanisms for the translational repression are not known with certainty, but one can deduce the most likely scenarios. In E. coli, translation is initiated by a ribosomal interaction with the Shine and Dalgarno region/initiation codon of an mRNA as well as initiator tRNA and initiation factors [3Gold L. Pribnow D. Schneider T. Shinedling S. Singer B.S. Stormo G. Annu. Rev. Microbiol. 1981; 35: 365-403Crossref PubMed Scopus (484) Google Scholar, 4de Smit M.H. van Duin J. Prog. Nucleic Acid Res. Mol. Biol. 1990; 38: 1-35Crossref PubMed Scopus (163) Google Scholar]. The easiest way to currently understand the two B-vitamin regulatory mechanisms described by Breaker's group is to imagine an mRNA with two alternative structures, open and closed, in equilibrium. The open structure allows translational initiation, while the closed hides the Shine and Dalgarno region from the ribosome (using different intramolecular base pairing). If the metabolite is preferentially recognized by the closed structure, thus altering the equilibrium, translation initiation will be lowered (Figure 1). Exactly the same principles can alter other mRNA (and pre-mRNA) functions (to be discussed later; Figures 1 and 4). These two systems described by Breaker represent a novel appreciation of the power of mRNA functional capacities, leading us to wonder (once again [5Gold L. Brown D. He Y. Shtatland T. Singer B.S. Wu Y. Proc. Natl. Acad. Sci. USA. 1997; 94: 59-64Crossref PubMed Scopus (119) Google Scholar, 6Gold L. Singer B. He Y.Y. Brody E. Curr. Opin. Genet. Dev. 1997; 7: 848-851Crossref PubMed Scopus (22) Google Scholar]) how far such RNA magic may go. The authors interpret their data in the glow of work by Harold B. White III [7White III., H.B. J. Mol. Evol. 1976; 7: 101-104Crossref PubMed Scopus (377) Google Scholar]. White observed more than 20 years ago that many of the present nucleotide cofactors might have been among the nucleotides present in ancient RNA or DNA, thus providing oligonucleotides with both coding and catalytic capacities. He proposed, explicitly, that as the protein world emerged and became the modern basis of catalysis, a contemporaneous partitioning of function could have left cofactors free of oligonucleotides (to function as aids to the proteins that needed them for catalysis) and left oligonucleotides free to encode proteins. White predicted an earlier world in which "shapes and tapes" were found in the same ancient molecules, and that division of labor led to a simpler code (only four simple nucleotides) and better catalysts (proteins with cofactors, which were free to evolve or not, but were no longer bound by the rules for oligonucleotides). White had a substantial vision into what we now call the "RNA world," although he did not claim the idea as conclusively as did Gilbert [8Gilbert W. Nature. 1986; 319: 618Crossref Scopus (1961) Google Scholar]. Because these two examples in E. coli of aptamer activity within an mRNA are in the context of expression systems that participate in the synthesis of cofactors, Breaker interprets his findings as a reflection of the world that was. We love Harold B. White's and Ron Breaker's work, yet we wonder if remnants are less important to this discussion than continuous selection for coordinated regulation over the last several billion years, well after Harold B. White's cofactors had left their previous oligonucleotide homes. A more dramatic extrapolation of the findings would be to wonder if these two examples in E. coli reflect a rich world of regulatory phenomena involving RNAs expressing their "shape" coincidentally with their "tape"—that is, can we wonder if genomes encode large numbers of aptamers to alter metabolic pathways? Explicitly, we wish to consider the evolutionary pressures for a variety of aptamers that do exactly what Breaker has found, and to consider as well the organisms in which we are likely to find the phenomenon most frequently. We will conclude that these regulatory phenomena are even more important in the eukaryotes. Let us first spend a few moments reviewing the SELEX protocol [9Tuerk C. Gold L. Science. 1990; 249: 505-510Crossref PubMed Scopus (8014) Google Scholar, 10Ellington A.D. Szostak J.W. Nature. 1990; 346: 818-822Crossref PubMed Scopus (7593) Google Scholar] through which nonnatural aptamers are identified (Figure 2). Double-stranded DNA with a randomized sequence region is prepared chemically. One strand is synthesized with flanking fixed regions (for PCR or other amplification), and the complementary stand is prepared enzymatically. If the synthetic DNA is prepared on a 10 μmol scale, and 1 μmol of final product is obtained, the starting library may contain about 6 × 1017 unique sequences (when the random region is longer than about 30 nucleotides). The double-stranded sequence library (tape) is converted enzymatically to a single-stranded (DNA, RNA, or modified oligonucleotide) shape library from which shapes with desired properties are selected. Because the flanking sequences allow amplification of culled "winners," one can perform multiple rounds of enrichment, and eventually most targets yield an aptamer. In many ways, the SELEX protocol is merely the single-stranded oligonucleotide version of phage, ribosome, or mRNA display for peptides and larger proteins [11Wilson D.S. Keefe A.D. Szostak J.W. Proc. Natl. Acad. Sci. USA. 2001; 98: 3750-3755Crossref PubMed Scopus (332) Google Scholar, 12Jermutus L. Honegger A. Schwesinger F. Hanes J. Pluckthun A. Proc. Natl. Acad. Sci. USA. 2001; 98: 75-80Crossref PubMed Scopus (227) Google Scholar, 13Jermutus L. Ryabova L.A. Pluckthun A. Curr. Opin. Biotechnol. 1998; 9: 534-548Crossref PubMed Scopus (129) Google Scholar, 14Rodi D.J. Makowski L. Curr. Opin. Biotechnol. 1999; 10: 87-93Crossref PubMed Scopus (201) Google Scholar]. The primary difference is that the identified aptamer is an oligonucleotide rather than a protein. Bruce Eaton has said repeatedly in his seminars that RNA/aptamer scientists must suffer from "side group envy" (B. Eaton, personal communication). The length of the random regions in classical SELEX experiments is small [15Cerchia L. Hamm J. Libri D. Tavitian B. de Franciscis V. FEBS Lett. 2002; 528: 12-16Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar, 16Hesselberth J. Robertson M.P. Jhaveri S. Ellington A.D. J. Biotechnol. 2000; 74: 15-25PubMed Google Scholar, 17Brody E.N. Gold L. J. Biotechnol. 2000; 74: 5-13PubMed Google Scholar, 18Hicke B.J. Stephens A.W. J. Clin. Invest. 2000; 106: 923-928Crossref PubMed Google Scholar]. At NeXstar, we focused on finding aptamers for therapeutics that could be synthesized at reasonable cost, and tended to use 30 to 40 random positions in a starting SELEX library. This was a compromise: longer single-stranded oligonucleotides must provide more shape opportunities than shorter oligonucleotides, but we knew that the cost of aptamer drugs would be high and a function of their length. Some SELEX experiments were performed with libraries containing up to 200 randomized positions, even though no aptamer derived from a long randomized region ever requires all or even most of the nucleotides of that aptamer to do the binding. In fact, no strong published data show that smaller randomized regions would have failed to yield an equivalent aptamer. The best short aptamer ever identified for a protein target may be the vascular endothelial growth factor (VEGF) aptamer that is presently in its pivotal clinical trial for macular degeneration; that molecule, now called Macugen by EyeTech Pharmaceuticals (http://www.eyetk.com), is but 28 nucleotides long, has an approximately 50 pM Kd, binds to VEGF with high specificity, and prevents VEGF from binding to its receptors [19Green L.S. Jellinek D. Bell C. Beebe L.A. Feistner B.D. Gill S.C. Jucker F.M. Janjic N. Chem. Biol. 1995; 2: 683-695Abstract Full Text PDF PubMed Scopus (245) Google Scholar]. It is true that most aptamers have been selected from very large starting libraries of oligonucleotide sequences, substantially larger than the sequence space available in modern genomes (SELEX starting libraries can be as large as 1015 unique molecules, about 105 times more than the human genome.). However, SELEX experiments are done in a week with 5 to 15 rounds of selection, while evolution and selection in biology had a billion years with perhaps a "round" of selection every day; the point is that 1011 divisions during unicellular growth can more than make up, via mutation, for a smaller starting library (the genome). We know a great deal today about aptamers that have been selected against protein targets and also against small molecule (metabolite) targets [20Gold L. Polisky B. Uhlenbeck O. Yarus M. Annu. Rev. Biochem. 1995; 64: 763-797Crossref PubMed Scopus (747) Google Scholar, 21Hermann T. Patel D.J. Science. 2000; 287: 820-825Crossref PubMed Scopus (1333) Google Scholar]. RNA aptamers have Kds between 10 nM and 10 pM for proteins (which are present usually in cells at concentrations between μM and nM), while aptamer Kds are between 10 μM and 10 nM for small molecules (which are present usually in cells at concentrations far higher than protein concentrations, perhaps mM to μM). Aptamers that have been selected in vitro seem ideally suited for interacting with proteins or small molecules at their physiological concentrations. This leads to wondering whether aptamer selection occurs frequently in vivo, leading to flagrant use of aptamers in all creatures. The first natural aptamer discovered in biology probably was the operator, although that "aptamer" consisted of double-stranded DNA. However, the large number of DNA sites to which specific proteins bind does reflect our first principle: evolution can demand and get tight and specific binding between genomic nucleic acid sequences and proteins. The second class of natural single-stranded RNA aptamers (and the first sites that really were aptamers) were discovered in the small E. coli RNA phages MS2, R17, and Qβ [22Romaniuk P.J. Lowary P. Wu H.N. Stormo G. Uhlenbeck O.C. Biochemistry. 1987; 26: 1563-1568Crossref PubMed Scopus (156) Google Scholar, 23Carey J. Cameron V. de Haseth P.L. Uhlenbeck O.C. Biochemistry. 1983; 22: 2601-2610Crossref PubMed Scopus (203) Google Scholar, 24Witherell G.W. Uhlenbeck O.C. Biochemistry. 1989; 28: 71-76Crossref PubMed Scopus (91) Google Scholar]. These phage face serious regulatory problems (such as clearing the input RNA genome of ribosomes before replication of the plus strand [to make a minus strand] results in a lethal collision between two immovable objects), and the phage all evolved wonderful binding sites comprising folded RNAs that allowed tight and specific binding of proteins; these sites are natural aptamers. Oddly, these fascinating viral RNA structures often were seen as the best these poor creatures could do because they had no chance to use transcriptional regulation. (People actually said in the 1970s and 1980s [in print] that translational regulation would not occur in an organism that had a chance to regulate gene expression transcriptionally.) Translational regulation was discovered in the bacteriophage T4 (which has a real DNA genome): T4 gene 32 protein regulates its own translation, T4 gene 43 protein regulates its own translation, and T4 regA protein regulates the translation of twenty or so mRNAs [25Russel M. Gold L. Morrissett H. O'Farrell P.Z. J. Biol. Chem. 1976; 251: 7263-7270Abstract Full Text PDF PubMed Google Scholar, 26Gold L. O'Farrell P.Z. Russel M. J. Biol. Chem. 1976; 251: 7251-7262Abstract Full Text PDF PubMed Google Scholar, 27Andrake M. Guild N. Hsu T. Gold L. Tuerk C. Karam J. Proc. Natl. Acad. Sci. USA. 1988; 85: 7942-7946Crossref PubMed Scopus (40) Google Scholar, 28Winter R.B. Morrissey L. Gauss P. Gold L. Hsu T. Karam J. Proc. Natl. Acad. Sci. USA. 1987; 84: 7822-7826Crossref PubMed Scopus (61) Google Scholar]. The sites responsible for translational regulation were called "translational operators" (a phrase chosen from transcription envy), but these sites are natural aptamers. A similar set of natural aptamers has been identified in E. coli: several ribosomal proteins were found to regulate their own translation by binding to the ribosome binding sites on their cognate mRNAs. Usually the proteins that regulated their own synthesis were proteins with binding sites for specific sites in ribosomal RNA, and thus the mRNA aptamers were thought to be mere mimics of that functional binding site [29Nomura M. Gourse R. Baughman G. Annu. Rev. Biochem. 1984; 53: 75-117Crossref PubMed Scopus (585) Google Scholar, 30Nomura M. Yates J.L. Dean D. Post L.E. Proc. Natl. Acad. Sci. USA. 1980; 77: 7084-7088Crossref PubMed Scopus (168) Google Scholar]. Nevertheless, one saw creeping into the literature mRNA sequences that (usually) were secondary sites that helped prevent excessive synthesis of specific proteins [31Springer, M., Lesage, P., Graffe, M., Dondon, J., Grunberg-Manago, M., Moine, H., Romby, P., Ebel, J.-P., Ehresmann, C., and Ehresmann, B. (1990). In Post-Translational Control of Gene Expression, J.E.G. McCarthy and M.F. Tuite, eds. (Berlin: Springer-Verlag), pp. 309–323.Google Scholar]. This type of autogenous translational repression is now quite common. We think of transcriptional attenuation in bacteria as an example of regulatory events caused by a natural aptamer; the aptamer comprises specific codons within an mRNA [32Yanofsky C. Nature. 1981; 289: 751-758Crossref PubMed Scopus (458) Google Scholar]. When a ribosome is translating the 5′ end of a polycistronic mRNA that encodes the enzymes needed for the biosynthesis of an amino acid, codons in the leader peptide are recognized by the cognate aminoacylated tRNA (during translation) if the amino acid is present in sufficient quantities; those codons are not recognized if that amino acid concentration is too low. Translation of the leader peptide causes the remaining polycistronic mRNA to not be transcribed, thus signaling to the operon that additional biosynthetic enzymes are not needed. This example of feedback repression operates through an mRNA-amino acid recognition (thus an aptamer for the amino acid, sort of), but the actual mechanics of feedback involve the tRNA, the aminoacyl-tRNA synthetase, and the entire ribosome. There exist many other (far simpler) ways to have high amino acid concentration direct lower synthesis of the requisite enzymes for amino acid synthesis, but evolution takes what it can get (We will return to this point repeatedly, yet never make this point more exquisitely than Jacob made it more than 25 years ago [33Jacob F. Science. 1977; 196: 1161-1166Crossref PubMed Scopus (1621) Google Scholar].). These observations are not restricted to bacteria. Years ago, Klausner and his colleagues described natural aptamers at the 5′ or the 3′ ends of mRNAs that could impact translation or mRNA stability [34Harford J.B. Klausner R.D. Enzyme. 1990; 44: 28-41Crossref PubMed Scopus (74) Google Scholar, 35Binder R. Horowitz J.A. Basilion J.P. Koeller D.M. Klausner R.D. Harford J.B. EMBO J. 1994; 13: 1969-1980Crossref PubMed Scopus (234) Google Scholar, 36Gray N.K. Hentze M.W. EMBO J. 1994; 13: 3882-3891Crossref PubMed Scopus (222) Google Scholar], in either case altering the protein output by a feedback mechanism dependent on iron regulatory protein (IRP), which binds IREs (iron-responsive elements) in mRNA only in iron-depleted cells. Similarly, the number of proteins thought to enhance or repress splicing is growing; aptamer sequences near exon-intron or intron-exon junctions give splicing a remarkable flexibility and coordination with metabolism. In Saccharomyces cerevisiae, ribosomal protein L30 binds to a 36 nucleotide bulged stem-loop in rRNA. Excess L30 binds to a similar stem-loop structure containing the 5′ splice site in L30 pre-mRNA and inhibits utilization of this splice site. A further twist on aptamer use in this system is the finding that the translation initiation AUG of L30 mRNA is also wrapped up in a similar structure, so that excess L30 protein autogenously regulates both splicing and translation during gene expression [37Vilardell J. Warner J.R. Mol. Cell. Biol. 1997; 17: 1959-1965Crossref PubMed Scopus (126) Google Scholar, 38Vilardell J. Yu S.J. Warner J.R. Mol. Cell. 2000; 5: 761-766Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar]. Splicing control in metazoans is even richer in aptamer like modulating sequences. Intronic enhancers [39Inoue K. Hoshijima K. Sakamoto H. Shimura Y. Nature. 1990; 344: 461-463Crossref PubMed Scopus (235) Google Scholar] and exonic enhancers [40Hedley M.L. Maniatis T. Cell. 1991; 65: 579-586Abstract Full Text PDF PubMed Scopus (176) Google Scholar] participate in a cascade of sex-dependent splicing events in the development of Drosophila melanogaster. In mammals, splicing regulation seems to be the norm rather than the exception. It is now clear that SR proteins, a group that had been previously lumped together as one class because of their serine-arginine repeats, actually bind to individual aptamer sequences in exons to give exquisitely fine-tuned splicing patterns [41Liu H.X. Zhang M. Krainer A.R. Genes Dev. 1998; 12: 1998-2012Crossref PubMed Scopus (420) Google Scholar]. We note that when aptamers are selected against proteins that bind to natural aptamers, the resulting artificial aptamer (which is always short compared to the genomic window of opportunity) may not (and usually does not) resemble the natural aptamer. This is really not more than a statement that complex structures can utilize long-range RNA interactions that cannot happen in a randomized sequence only 30 or 40 nucleotides in length, although it must also be remembered that nature selects for optimal sequences (with respect to physiological needs), whereas artificial SELEX chooses maximal binding affinity. The presence of many natural aptamers that serve to control protein expression is the intellectual space within which the Breaker papers reside. The startling element in the Breaker papers derives from the fact that the feedback inhibition occurs in response to the concentration of a small molecule rather than a protein, even though the specificity and affinity data for synthetic aptamers aimed at small molecules is so robust (see [21Hermann T. Patel D.J. Science. 2000; 287: 820-825Crossref PubMed Scopus (1333) Google Scholar] for a wonderful review). RNA binding activities (as aptamers) can be generalized to include "open" and "closed" forms of attenuator sequences [32Yanofsky C. Nature. 1981; 289: 751-758Crossref PubMed Scopus (458) Google Scholar], splicing elements [37Vilardell J. Warner J.R. Mol. Cell. Biol. 1997; 17: 1959-1965Crossref PubMed Scopus (126) Google Scholar, 38Vilardell J. Yu S.J. Warner J.R. Mol. Cell. 2000; 5: 761-766Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar], transcriptional termination sites [42Platt T. Annu. Rev. Biochem. 1986; 55: 339-372Crossref PubMed Google Scholar], frame-shifting sequences [43Weiss, R.B. (1990). In Post-Translational Control of Gene Expression, J.E.G. McCarthy and M.F. Tuite, eds. (Berlin: Springer-Verlag), pp. 579–602Google Scholar], nuclear exit determinants [44Pasquinelli A.E. Ernst R.K. Lund E. Grimm C. Zapp M.L. Rekosh D. Hammarskjold M.L. Dahlberg J.E. EMBO J. 1997; 16: 7500-7510Crossref PubMed Scopus (197) Google Scholar, 45Grimm C. Lund E. Dahlberg J.E. Proc. Natl. Acad. Sci. USA. 1997; 94: 10122-10127Crossref PubMed Scopus (20) Google Scholar], RNA decay elements [46Ross J. Microbiol. Rev. 1995; 59: 423-450Crossref PubMed Google Scholar], etc.; that is, any RNA might fold into two states in equilibrium, such that the perturbation of that equilibrium by an aptamer-specific ligand can have a regulatory impact (Figure 1) which is manifested through either more or less RNA, more or less activity, more or less transport to an appropriate intracellular location, and so on. Francois Jacob said it best in "Evolution and Tinkering" [33Jacob F. Science. 1977; 196: 1161-1166Crossref PubMed Scopus (1621) Google Scholar]. Evolution has a path, but that path is hard to predict and derives from historical accidents. While there is a perhaps illogical history to evolution, time (at least abundant time) can be the driver of metabolic coordination. Michael Savageau has spent a lifetime wondering about the likelihood that evolution selects, over abundant time, optimal solutions to serious problems [47Savageau M.A. Proc. Natl. Acad. Sci. USA. 1977; 74: 5647-5651Crossref PubMed Scopus (87) Google Scholar]. At some important level, time allows playfulness (okay, call it randomness), and playfulness, when continued over abundant time, leads to optimal solutions. Jacob's analysis deals with the mechanism of evolution but does not deny the value of abundant time. Savageau argues, persuasively, that time is so abundant that we can forget sub-optimal solutions if they are not robust. The Savageau work suggests that evolution cannot hold a "position" unless it is sufficiently robust to survive playfulness. Jacob and Savageau represent the extremes of thought, even though for sure they would agree about almost everything in detail; each has made points that live comfortably within the other's worlds, once time is allowed into the dialog [48Savageau M.A. Biochemical Systems Analysis. Addison-Wesley Pub. Co, Reading, MA1976Google Scholar]. Why are we speaking about Jacob and Savageau? The regulation of gene expression is an area in which many quasi-equivalent/quasi-different solutions appear quasi-robust. This explains the extraordinarily different mechanisms by which bacterial phages control transcription during their developmental programs [49Brody E.N. Kassavetis G.A. Ouhammouch M. Sanders G.M. Tinker R.L. Geiduschek E.P. FEMS Microbiol.Lett. 1995; 128: 1-8Google Scholar]; that is, evolutionary pressures cannot drive bacteriophages to all use the same mechanism because several biochemical mechanisms are equivalently robust (at least approximately equivalent). Thus, we need to understand the pressures that might allow Breaker's observations to be commonplace: once again can we extrapolate from two to infinity. Organisms struggle to utilize their biochemistry efficiently by coordinating vast numbers of different molecules and reactions within a cell or within an organism. What a struggle this must be. The world of biology has but a few robust mechanisms that ought to be selected and utilized whenever possible—one robust mechanism for coordination must be the use of polycistronic transcription. Clearly, nothing is more directly aimed at biochemical coordination than polycistronic transcription. If a cell requires enzymes 1 through 10 to accomplish some biochemical pathway, what could be more robust than to place those enzymes into an operon, express them coordinately from a single promoter, and allow the product (of the translational yield of each gene times the activity of each enzyme) to provide biochemical coordination? Nothing is more straightforward than operons to effect coordinated biochemistry, and both eubacteria and archea use operons to accomplish coordinate biochemistry. This leads to a hypothesis (really based on the presumption of robustness in the Savageau sense, without recourse to mathematical modeling—Michael would hate this) that the evolution of polycistronic transcription is an enormous event that vastly simplifies intracellular biochemical coordination. The recent literature gives us reason to rejoice. In some important papers by Tom Blumenthal and his colleagues [50Blumenthal T. Evans D. Link C.D. Guffanti A. Lawson D. Thierry-Mieg J. Thierry-Mieg D. Chiu W.L. Duke K. Kiraly M. et al.Nature. 2002; 417: 851-854Crossref PubMed Scopus (271) Google Scholar, 51Liu Y. Huang T. MacMorris M. Blumenthal T. RNA. 2001; 7: 176-181Crossref PubMed Scopus (18) Google Scholar, 52Evans D. Perez I. MacMorris M. Leake D. Wilusz C.J. Blumenthal T. Genes Dev. 2001; 15: 2562-2571Crossref PubMed Scopus (31) Google Scholar], we learn that C. elegans uses operons. Furthermore, we learn that C. elegans uses operons less frequently than E. coli but far more frequently than almost any other eukaryote. This leads to a tiny extension of the "tree" of life (Figure 3) and an idea that extends Breaker's work from remnant to novel coordination mechanisms. We believe that the division of life into the three kingdoms occurred prior to selection of polycistronic life, and that further (isolated within kingdoms) evolution had the opportunity to independently evolve polycistronic life as a solution to a big biochemical problem: coordination. Because both the eubacteria and the archea have had many more organismic replication events than have had eukaryotes, only those kingdoms have had the opportunity to optimize coordination through polycistronic life. C. elegans has had the good fortune, stochastically, to discover and use robustly that polycistronic life, but even C. elegans uses polycistronic life infrequently (about 15% of the genes are transcribed polycistronically). Interestingly, the coordination achieved in C. elegans is not followed by polycistronic translation; rather, trans-splicing events create individual mRNAs for translation. Polycistronic life is a robust solution to the coordination problem. From within a world of polycistronic life, relatively minor "tinkering" probably allows pathways of all kinds to provide buffered responses to the external world those organisms face. But the world we care most about, mammalian life, has not evolved polycistronic expression. Life without polycistronic regulation must include robust alternatives. Mammals with large genomes and slow replication cycles require strategies that work at this moment, even though future mammals may well "discover" polycistronic transcription. Perhaps splicing flexibility (introduced into evolutionary mechanisms a la Jacob and refined a la Savageau) is sufficiently optimized to preclude polycistronic expression. Nevertheless, we are finally back to Breaker and his discovery.