Biochemistry 811

Lecture on Evolution of Signal Transduction and Comparative Genetics

D. Nelson, last modified Nov. 26, 2002 5PM

Reading: Berg, Tymoczko and Stryer 5th edition, parts of Chapter 15 and literature references

Dr. Fains recent lectures were on signal transduction. This involves a cell responding to a stimulus. Both proteins and small molecule second messengers are key players. This lecture will deal with the evolution of these pathways. Where did they arise, and how far back on the tree of life do they exist.

There are a variety of ways to transmit a signal across a cell membrane and into the interior of the cell, or even into the nucleus. Many of these pathways use the same components, so these may be viewed as modules of the signal transduction process. A list of these components is given below.

1) Seven transmembrane segment receptors (also called serpentine receptors) coupled to heterotrimeric G proteins. G-protein coupled receptors (GPCRs)

2) Single transmembrane segment receptors (receptor tyrosine kinases, receptor guanylyl cyclases, histidine kinases and two component regulators)

3) Nuclear Receptors

4) Ion channels, either voltage gated, ligand gated or mechanosensitive

5) calcium/ calmodulin

6) cyclic nucleotides cAMP, cGMP, adenylate cyclase and phosphodiesterases

7) Nitric oxide synthase and soluble guanylate cyclases

8) phospholipases/ sphinomylinase and products diacylglycerol, phosphoinositides and ceramide

9) protein kinase cascades and associated phosphatases

Lets take the first six and look at the evolutionary history of these systems. On the tree of life, plants, animals and fungi branch from nearly the same point, about 1.2 billion years ago. This tight cluster of branches is called the crown group of eukaryotes. Plants and several groups of protists branched off first followed by Amoebozoa (includes the cellular slime mold Dictyostelium). Fungi and animals diverged soon after. Dictyostelium discoidium has been well studied because it assembles from single cells to form a slug that develops into a fruiting body that makes spores. Of course, this involves signal transduction, so we will look at some signal transduction pathways in Dictyostelium as a representative primitive eukaryote. We will also consider yeast, since the whole genome is sequenced and it has been thoroughly studied as a model organism. In 1998, C. elegans the nematode worm had its genome sequence completed. In March 2000 the Drosophila genome was placed in Genbank. On Dec. 14 2000 the Arabidopsis genome was published in nature. These genomes will offer insights into signal transduction in animals and plants but it may take years to analyze all the gene data.

Seven Transmembrane Receptors

Seven transmembrane segment receptors are eukaryotic proteins. Animals are especially fond of these receptors. The genome of the nematode worm is now done and it contains 1049 G-protein coupled receptors. These may be chemosensory receptors. Almost all of these are orphan receptors without known ligands. In fact, only one ligand has been linked to its chemoreceptor in C. elegans and that is diacetyl. This ligand binds to the odr-10 gene product that normally mediates attraction to diacetyl. However, if the gene is expressed in another neuron that usually mediates repulsion from a substance then the worm is repelled by diacetyl. This shows that the receptor only triggers a pathway and the response is dependent on which neuron the gene is expressed in. The worms genome has 18,424 genes total, so 1049 odor receptors is about 5.7% of the worms genome dedicated to chemosensory receptors. Current estimates suggest that 300 of these genes are pseudogenes and about 200 are not expressed in neurons, so they are not likely to be chemoreceptors, but that still leaves 500 serpentine receptor genes that are probable functioning chemoreceptors. This is the largest family of genes in the worms genome. (see Seven-transmembrane proteins as odorant and chemosensory receptors. Science. 1999 Oct 22;286(5440):707-11, and Chemosensory signaling in C. elegans. Bioessays. 1999 Dec;21(12):1011-1020.)

Even yeast have examples of these receptors, but not nearly as many. In a brief search, I found only 3: STE2, STE3 and GPR1. These are receptors for pheromone (STE2 and STE3) and a receptor that is induced during starvation. The ligand for GPR1 is glucose which causes a cAMP spike in starved yeast cells(Mol. Microbiol. 38, 348-358 2000). The heterotrimeric G protein alpha subunits for these receptors are named GPA1 (for STE2 and STE3) and GPA2 (for GPR1). There only appear to be two G protein alpha subunits in yeast, and two in the fission yeast S. pombe. Based on these results, the fungi have not exploited these useful receptors very much. They seem to be used for mating purposes and for sensing the concentration of glucose in the environment. Dictyostelium also uses them for cell to cell communication in the developmental program of aggregation and sporulation.

Dictyostelium branches on the eukaryotic lineage within the crown group of eukaryotes (animals, fungi, amoebozoa, plants, stramenopiles and alveolates). The signal transduction mechanisms present in Dictyostelium should be present in animals and fungi, but not necessarily in plants. There are genome projects being done now that include Dictyostelium and Giardia, one of the most distant eukaryotic organisms known. These projects should aid greatly in understanding the earliest events in eukaryotic evolution, including development of signal transduction pathways. But the genomes are not done yet, so the story is fragmentary so far.

Searches for seven transmembrane receptors and G-proteins in Dictyostelium show that there are four cAMP receptors called cAR1 to cAR4. These are able to sense cAMP secreted by the Dictyostelium cells as they aggregate to form a stalk. These four receptors are expressed at different times during the developmental program of the organism. However, one report (EMBO J 1998 Sep 1;17(17):5076-84) refers to deletion of the single G-protein beta subunit. There appears to be only one beta subunit that is shared by all four receptors. The situation is not the same for the alpha subunits, since there is sequence data for at least eight distinct alpha subunits. It is not clear if there are eight different seven transmembrane receptors for all these alpha subunits, or if the four cAR receptors use more than one, possibly at different times.

One interesting finding concerning these serpentine receptors in Dictyostelium is that some responses are not dependent on the beta subunit. These seem to be G-protein independent responses that are linked to the receptor. One of these is calcium influx, though the amount of calcium influx is less than with the intact G-proteins. Here we may be seeing a remnant of early evolution of the serpentine receptor/G-protein couple. At one point in the past there had to be a receptor without G-proteins like the bacteriorhodopsin light driven pump discussed below. If the cAMP receptors can carry out some of the same functions, but at a reduced efficiency, then this might be a vestige of the original receptor function without G-proteins. A related issue is activation of receptors without ligand bound. The Nov. 98 TIBS has an article on Agonist Independent Regulation of Consitutively Active G-protein Coupled Receptors. Constitutive signaling without ligand bound has been documented now for several GPCRs.

The seven transmembrane receptors have a remarkable similarity to bacteriorhodopsin and halorhodopsin. These are a light driven proton pump and a light driven chloride pump from the archaebacteria Halobacterium salinarium (previously named halobium). The seven transmembrane segments enclose a retinal attached to a lysine residue of the last helix by a schiffs base. This retinal undergoes an all-trans to 13- cis isomerization when it absorbs a photon. The conformational change is transmitted to the protein and it is used to pump protons or chloride and form a gradient. In vertebrate eyes, the protein rhodopsin is also a seven transmembrane segment protein with retinal attached to a lysine also in the last helix. In this protein the conformational change is 11-cis to all trans and it activates a heterotrimeric G protein that activates a cGMP phosphodiesterase which reduces cGMP concentrations. The result is closing of sodium ion channels and electrical stimulation of a nerve.

The common ancestry between bacteriorhodopsin (BR) and eukaryotic receptors might be considered unconvincing if the retinal attachment and seven transmembrane segments were the only evidence in its favor. However, the relationship is supported based on more detailed evidence. BR has two close relatives in Halobacterium salinarium that are light driven sensory receptors SRI and SRII. These also contain retinal, but they are tightly coupled to two transducer proteins HtrI for SRI and HtrII for SRII. If the sensory rhodopsins are expressed without the transducers they function as proton pumps. The tight association between the sensory rhodopsins and the transducer prevents proton pumping. Instead, the conformational change caused by light is used to stimulate the transducers. These proteins are histidine kinases that control a phosphorylation cascade that regulates the flagellar motor as in E. coli chemotaxis. Different wavelengths of light can either attract or repel the bacteria. In this way, the transducer proteins are similar to two component regulators described below.

The most convincing evidence for common ancestry of bacterial and eukaryotic rhodopsins is in the mechanism of signal transduction. Both bacterial and vertebrate rhodopsins form a protonated schiffs base on the G helix where retinal is attached. Both form a salt bridge between this protonated schiffs base and a conserved carboxyl group on helix C. Absorbance of a photon breaks the salt bridge and shifts the proton to the carboxyl group. In BR the net result is opening of a half channel facing the cytoplasm and transport of 2 H+ across the membrane. In sensory rhodopsin, the signal is transmitted to the Htr transducer protein and on into the cell by phosphorylation. In eukaryotes, the trimeric G proteins separate into alpha and (beta, gamma) subunits.

There is no sequence similarity between bacteriorhodopsin and vertebrate rhodopsin. However, the overall structure and the placement of retinal are two pieces of evidence that suggest a common ancestor. If this is correct, then the archaebacterial ancestor of eukaryotes may have contributed a seven transmembrane precursor of modern seven transmembrane receptor proteins. J.L. Spudich, an expert in these proteins does not believe that archaeal and visual rhodopsins are descended from a common ancestor. He argues that the 22 amino acids that make up the binding pocket for retinal are not conserved in the vertebrate rhodopsins and the retinal structure is significantly different. High resolution crystal structures are available now for bacteriorhodopsin trapped in various stages of the transport cycle. (See Science 286, Oct. 8, 255-260 and 252 commentary 1999); halorhodopsin (Science 288 1390-1396 2000) and bovine rhodopsin (Science 289, 733-734, 739-745 2000) The interpretation of the structures suggests that the seven helices are arranged in different ways. The consensus seems to be that the vertebrate rhodopsins (and therefore all the GPCRs) are not related to the archaeal rhodopsins (Curr. Opin. Struct. Biol. 11, 420-426 2001).

Therefore, the origin of seven transmembrane segment receptors cannot be traced back to light sensitive receptors in archaebacteria. The similarity between these proteins seems to be a spectacular case of convergent evolution. A new development in 1999 is the discovery of a fungal protein called nop-1 (opsin from Neurospora crassa) that has been overexpressed in Pichia pastoris. The protein binds retinal as a schiff base and probably acts as a rhodopsin like molecule in N. crassa (PNAS 96, 8034-8039 1999) This protein has 20 of 22 conserved amino acids in common with archaeal bacteriorhodopsin. There are other related proteins in N. crassa that do not have the lysine residue that makes the schiff base, so these may have lost the ability to use retinal as a stimulus, but this is expected in the evolution of the seven transmembrane receptors that do not contain retinal and are driven by ligand binding. (for reviews search for JL Spudich). At the current time the origin of the seven transmembrane receptors is not clear.

A 1998 report claims to have found the first seven transmembrane receptor in plants (Plakidou-Dymock S, Dymock D, Hooley R. A higher plant seven-transmembrane receptor that influences sensitivity to cytokinins. Curr Biol 1998 Mar 12;8(6):315-24). The Arabidopsis genome published in Dec. 2000 shows 27 GPCR domains present, so these receptors are not common in plants. Additional support for GPCRs in plants comes from the fact that Arabidopsis has a single G-protein alpha subunit GPA1 (Curr. Biol. 11, R869-R874 2001). Knockout of the alpha subunit in rice causes a dwarf phenotype and abnormal seeds (Plant Phisiol. 42, 789-794 2001). These GPCR proteins are then found in four main branches at the top of the eukaryote tree, animals, fungi, amoebozoa and plants. Beyond that no sequence data is available for GPCRs in the lower eukaryote groups.

Single Transmembrane Segment Receptors

 Some receptors with single transmembrane segments unite as dimers and have enzymes associated with them on the cytosolic side of the membrane. These enzymes can be tyrosine kinases, serine/threonine kinases, tyrosine phosphatases, guanylyl cyclases that form cGMP or histidine kinases. The histidine kinases are ancient signal tranduction molecules found in all three domains of life. These are called two component regulators. This system is used in bacterial chemotaxis, where a receptor binding ligand activates a histidine kinase CheA to phosphorylate itself. The PO4 group is rapidly transfered to an Asp residue on another soluble protein CheY called the response regulator. This protein then carries out some regulatory function. In chemotaxis CheY binds to the flagellar motor and influences the direction it turns. Clockwise rotation causes tumbling which is a kind of retreat from a repellant substance. Counterclockwise rotation makes the E. coli swim smoothly ahead. CheY that is phosphorylated causes clockwise rotation. The HtrI and HtrII proteins of Halobacterium are similar to CheA, except they have two transmembrane segments instead of one.

Dictyostelium has a his kinase response regulator system. The receptor DhkA binds a secreted peptide SDF2 that triggers autophosphorylation on histidine. The PO4 is transfered to RegA, a cAMP phosphodiesterase which is inhibited. The increase in cAMP activates protein kinase A (PKA).

Yeast also has one of these response regulatory his kinase receptors for sensing osmotic stress. It feeds into the Hog1 MAP kinase pathway (see below).

Receptor tyrosine kinases are only found in metazoa[animals] (J. Mol. Evol. 44, 242-252 1998). One has been cloned from a sponge, which is as far back as one can go in the metazoa. This gene is interesting because it contains only two introns. More advanced animals have more introns in their tyrosine kinase genes. Since the sponge gene appears to branch on a tree before the other metazoan receptor tyrosine kinases, this suggests that introns were added after this branching occurred, favoring the introns late theory.

Nuclear Receptors

 Nuclear receptors were called the largest family of transcription factors in a 1994 Annual Rev. of Biochem. article (vol. 63, p. 453). In a Nov. 98 article, more than 300 sequences were known (PNAS 95, 13442-13447). These are internal receptors in contrast to the membrane receptors we just discussed. The ligands for these receptors are hydrophobic and diffuse through the membranes of cells without specific transport proteins to bring them in. They include steroids, retinoids, vitamin D and thyroid hormone. Vincent Laudet has been active in the study of the evolution of these receptors. His group has searched for the presence of nuclear receptors in many types of animals, plants and lower eukaryotes. The approach was an exhastive PCR study using highly conserved regions of the nuclear receptors to make degenerate PCR primers that could be used to detect almost any member of the group. The result was that nuclear receptors are strictly metazoan (animal). They were not found in plants, algae, fungi or protists. Among animal species, they were not found in sponges, but they were present in Cnidarians (radial animals like jellyfish, hydras and corals) and all other more advanced animals. There are 270 nuclear hormone receptors in C. elegans alone.

Many nuclear receptors do not have ligands identified. These are called orphan receptors. One surprising finding in this exhaustive search for receptors was that a given ligand did not appear in only one branch of the tree of receptors. Also, the receptors with known ligands were predominantly found in recenty evolved subfamilies. The authors concluded that the ability to bind ligand has only evolved recently and the ancestral nuclear receptors were probably orphan receptors. This also suggests that many orphan receptors do not have any ligand that binds to them. To function without ligand the receptor is proposed to have two conformations. One binds to DNA and regulates gene function. The other does not bind DNA. The equilibrium between the two states determines how much time the gene is turned on. Binding a ligand shifts the equilibrium to the DNA binding conformation. Of course other things could shift this equilibrium too, like phosphorylation, temperature, etc. This idea is new and will have to be debated and tested. In 1999 a nomenclature system was adopted for nuclear receptors (A unified nomenclature system for the nuclear receptor superfamily. Cell. 1999 Apr 16;97(2):161-3). Nomenclature is a serious problem as more genomes are sequenced and it needs to be addressed for most large families of genes.

Ion Channels

 Ion channels include neurotransmitter activated channels like the acetylcholine receptor (ligand gated), voltage gated channels and mechanosensitive channels that respond to changes in membrane properties that could be altered by mechanical forces caused by high or low osmolarity conditions or shear forces. The ligand gated channels like the acetylcholine receptor are more advanced. They are found at synapses and these are nerve specific. They are not found outside animals.

The mechanosensitive channels are found widely in nature in bacteria, archaebacteria and eukaryotes. These are the simplest channels, since thay do not need a ligand binding site or a voltage sensor. They do have to exist in at least two states (open and closed) with possibilities for additional substates. The nature of channels is determined by patch clamp techniques, where the size of the channel is estimated from the current it passes and the duration of open and closed states can be measured. A mechanosensitive channel has been purified from the archaeon Haloferax volcanii. It was only 37kDa, much smaller than vertebrate ion channels that can be about 2000 amino acids and 200kDa.

Voltage gated ion channels include the porins of bacteria, mitochondria and archaea. However, porins are quite different from the nerve based ion channels for Na, K and Ca. These channels have alpha helical transmembrane segments, while porins have a beta barrel structure. The two protein superfamilies must have separate evolutionary origins. Among the K, Na and Ca channels, the K channel appears to be the oldest. It is a tetramer of identical subunits, whereas the Na and Ca channels have four domains, each similar to one K channel. K channels have been found in bacteria, in 1998 a crystal structure was determined of the Streptomyces lividans K channel. It is tetrameric, with each identical subunit contributing two transmembrane segments. The pore is in the center. It made the cover of Science (Vol. 280, 69-77 April 3 issue). Some eukaryotic K channels have six transmembrane segments rather than two, but the first four seem to be extra. The last two transmembrane segments are very similar to the bacterial K channel. The evolution of Na and Ca channels probably started from a K channel. Recently, a Na channel has been cloned from a jellyfish (Cnidarian). These are the most primitive organisms with nerves.

Calcium/ Calmodulin

 Calmodulin is a calcium binding protein that activates many different enzymes such as plasma membrane calcium pumps, phosphodiesterases, kinases(CAM kinase II) and phosphatases (calcineurin). It binds calcium at low micromolar concentrations, changes its conformation and binds to its target enzymes. Calmodulin is found in every eukaryote ever looked at, but it is not found in bacteria. A report has been published that a calmodulin like protein is present in Halobacterium salinarium an archeabacterium. This protein binds calcium, it activates cAMP phosphodiesterase and is inhibited by calmodulin inhibitors, but the sequence is not known yet. The Halobacterium genome was completed in summer 2001, but it is not publicly available yet. A search of the annotation at the genome website looking for calmodulin did not find any hits.

The structure of calmodulin has four EF hands that are motifs that bind calcium. The EF hands are numbered I, II III and IV. I and III are more similar to each other than they are to II and IV. II and IV are more similar to each other than they are to I and III. The interpretation of this is that the protein evolved from a gene with a single calcium binding motif that duplicated and diverged for a time then duplicated again. The original one domain gene is now vanished into the distant past and may not exist any more, or it will probably be impossible to detect the similarity to it. It is unfortunate that many evolutionary questions of origins cannot be solved because the process of evolution has erased the books.

Cyclic Nucleotides

 cAMP is found in all three domains of life: bacteria, archaea and eukarya. In eukaryotes, it activates protein kinase A (PKA). It is secreted as a signal by Dictyostelium and it binds to seven transmembrane receptors in this organism. Bacteria use it to regulate gene expression in catabolite repression, where glucose can repress the synthesis of genes needed to metabolize other carbon sources. The glucose signal activates a cAMP phosphodiesterase that decreases cAMP concentrations. cAMP normally binds to a protein called CRP (cAMP receptor protein) or CAP (Catabolite gene-activator protein) and this protein binds to DNA and activates gene expression of an operon. When the cAMP level drops, the CAP protein can no longer activate the genes and their transcription is reduced or stopped. cAMP is one of the most ancient of regulatory signaling molecules.

Comparative Genetics and Human Evolution, What Makes Us Human

 Humans and chimpanzees have DNA sequences that are 98.74% identical (over 3 million bases). When one considers that our DNA is 3 billion base pairs, that means that only 38 million bases differ between humans and chimps. According to the Feb 15 2001 Nature only 1.1% of our genome is exons (however the gene count estimates have almost doubled since then, so assume 2% is exons) The other 98% is non-coding sometimes called junk DNA, then the number of base pairs that differ in genes is about 756,000. If we have 70,000 genes (this gene count estimate seems to the current best guess) that is about 11 base differences per gene. Most genes are not going to be seriously changed by a handful of base changes. In fact, most of these changes are probably going to be silent and result in the same amino acid sequence, or a very conservative change like arginine for lysine. It is very unlikely that the function of a protein will be changed by these small conservative differences. Hemoglobin will still be hemoglobin.

So what are the differences that make us different from chimps? These will probably be very small changes, perhaps only a single base difference in genes that make a difference in development. An article in Science 281, 1432-1434 1998 (Sept. 4), discusses this question. The suggestion is made that small changes in proteins like transcription factors could have large effects. For example, a transcription factor that controlled the length of time that the brain was able to grow would have a large effect on brain complexity and the number of synapses that could form. One biochemical difference that has been discovered is a defect in a human gene that modifies sialic acid. Humans do not have a functional copy of this gene because it has a 92 bp deletion. The result is a change in the cell surfaces where glycoproteins are expressed. (A structural difference between the cell surfaces of humans and the great apes. Am J Phys Anthropol. 1998 Oct;107(2):187-98) (A mutation in human CMP-sialic acid hydroxylase occurred after the Homo-Pan divergence. Proc Natl Acad Sci U S A. 1998 95(20):11751-6).

One of the FAQs (Frequently Asked Questions) that was asked on the Research Genetics site for the GeneFilters was can you do a cross-species hybridization with the filters. The answer was yes you could and if the genes were more than 70% identical good signals could be detected under selected conditions of hybridization. Since the two genomes of humans and chimps are so similar, it would not even matter that they were different species. DNA microarray analysis of 60,000 human genes (5 Affymetrix chips) between chimps and humans could be done right now to look for expression level differences in specific tissues or developmental stages. It is likely that very few genes would show significant differences in gene expression and these would be the ones that would make the best candidates for those genes that make us human. Svante Paabo reports his group is doing this now "We have furthermore studied the relative levels of expression of 20,000 genes in humans, chimpanzees and macaques. A number of genes that differ significantly in their expression levels between humans and chimpanzees have been identified." From an abstract at the meeting: Evolutionary Genomics New Paradigm of Biology in the 21st Century, November 4-6, 2001 Atami Korakuen Hotel, Atami, Japan. Nature from July 6, 2000 reports a Japanese project to compare humans and chimps. The Silver Project It is called silver for the chemical symbol of silver Ag which is also short for Ape genome.

For more discussion see this link>

Comparative Genetics and Human Races

 Studies of human populations from around the world have shown that 85% of all the variation in our DNA can be found in any small population no matter where it is on earth (Science 286, 451-453 1999). That means that 85% of all the variation there is in humans occurred before the migration from Africa. Another 6% of the total variation is seen from within different groups on the same continent. That only leaves 9% of the observed variation in our DNA that is linked to differences in continents. This means the idea that human races are genetically separate from one another is wrong. So why do races look so different? The case of differences in skin color has been linked to a hormone receptor named MC1R for melanocortin- stimulating hormone receptor 1. Skin color is largely determined by the ratio and concentrations of different melanin pigments made in melaoncytes. This gene MC1R has been shown to have multiple alleles (DNA sequence variants) in humans. These occur in the 1st and 7th transmembrane segments of the protein and the first extracellular domain. Certain alleles are seen in 80% of people with red hair and skin that burns rather than tans. These same alleles are seen in less than 4% of caucasian people who tan rather than burn. These particular alleles are never seen in Africans. Different alleles are found in Asians, some in as high as 70% of the sampled group. (Genetics 151, 1547-57 1999) The red hair alleles show a loss of function when expressed in COS cells and assayed for cAMP production upon addition of hormone (Biochem Biophys Res Commun 260, 488-91 1999) The variation at this site is probably an adaptation to different sun exposure levels in these different populations. It is expected that differences in hair color and texture, differences in facial structure and other features of human races will be controlled by alleles of one or a small number of genes that affect these superficial visible differences. The true differences between human races will be minor sequence variations in receptors or some equivalent genes and these have nothing to do with our humanity.

Pathway Engineering

The Blue Rose Project

Florigene, a company based in Australia is interested in expressing proteins in flowers, not to make the protein, but to engineer in a pathway for flower pigments. For centuries, a blue rose has been the subject of fiction. It was mentioned in the Arabian Nights. There is no such thing as a blue rose however, because the key enzyme in the pathway to blue or purplish pigments is lacking in the rose family. This enzyme is a flavonoid 3'5' hydroxylase. This enzyme acts on anthocyanins that are already hydroxylated at the 3' and 4' positions to add a third hydroxyl at the 5' position of the anthocyanin B ring. This pigment is bluish or purplish in color and its specific absorption properties can be modified by pH, metal ions and copigments.

Excerpt from the public information sheet on PR-35: Planned release of transgenic rose. (This is part of an Australian regulatory agency impact statement) "The flower colour modification gene is the 'blue' gene from petunia. It encodes an enzyme, flavonoid 3'5' hydroxylase, which is required for synthesis of a class of pigments called delphinidins. Only plants containing delphinidin are able to produce blue flowers, and rose plants do not naturally synthesise delphinidin. The expectation is that the transgenic rose will have an altered flower colour, such as blue, violet, purple or lilac."

Florigene has developed methods to transform genes into carnations, chrysanthemums and roses. They have cloned the genes for flavonoid 3'5' hydroxylase from petunia flower petals and expressed these genes in carnations. This has led to purplish colored carnations. They have not got all the additional factors worked out yet to get a true blue color expressed, but they are working on it. In addition, they are transforming the genes into chrysanthemums and roses. The estimated world wide market for a blue rose is in the 3-5 billion dollar a year range, so it is worth the initial trouble to engineer this pathway into roses. Quote from a web page from March of 1998 "Dr. Steve Chandler of Florigene said recently that they expect a prototype of a true blue rose to be developed in 1998 or 1999." Link to page But another site dated May 21, 1999 says: But hopes of a blue rose have been thwarted because, it seems, rose petals are naturally acidic and this prevents the blue pigment being expressed. The company hopes to sidestep the problem - either by finding a conventional rose variety that is less acidic, or genetically-altering the plants to become more alkaline. Florigene.com web site In 2001, the company still has not given up on the blue rose, but it is elusive. Dr. Fred Guengerich, a well known P450 researcher at Vanderbilt and his collaborator from Australia have published on this: Gillam, E.M.J., and Guengerich, F.P. (2002) IUBMB Life 52, 1-7. "Exploiting the Versatility of Human Cytochrome P450 Enzymes: The Promise of Blue Roses from Biotechnology."