The algae in the ocean are a diverse bunch, one type of which is the silica-based diatom genus Pseudo-nitzschia. Pseudo-nitzschia species are pennate eukaryotes: long, glass microscopic pens in waters along the California coast. When algae grow in large numbers, changing the color of the seawater, they are said to bloom. Pseudo-nitzschia blooms every year, and sometimes, when it does, it produces domoic acid.
Domoic acid (hereafter DA) is a neurotoxin poisonous to marine vertebrates. When the toxin is produced, it moves up the food chain and into filter feeders like mussels and crabs. This is a major part of the diet of marine mammals like sea lions, who are killed off by the neurotoxin. If humans are exposed via ingestion of infected shellfish, it can result in dizziness, vomiting, abdominal cramps, and amnesia. It can damage the kidneys and can be fatal if taken in high doses, or if the person has kidney problems. DA is also teratogen causing fetal developmental problems if ingested by pregnant women. As a result, when these harmful algal blooms (HABs) occur, shellfish fisheries are temporarily shut down.
We don’t know why Pseudo-nitzschia produced DA. We don’t know it’s physiological function. We don’t know why some blooms become HABs, and some remain DA-free. And until recently, we didn’t know how they made it.
A group of oceanographers tackled this problem last year by identifying the biosynthetic pathway Pseudo-nitzschia uses to synthesize DA. They published it in a Science paper, Biosynthesis of the neurotoxin domoic acid in a bloom-forming diatom. Not only does this increase their understanding of DA production, helping us identify key risks and protect human and environmental safety better, it will help us understand tools in biotechnology and how scientists think about biosynthetic pathways.
What is a biosynthetic pathway? Let’s start with one many of us are familiar with: photosynthesis. We all know it’s not actually quite as simple as:
6CO2 + 6H2O + energy → C6H12O6 + 12O2
Oh no. This is a long, complicated process of light reactions plus carbon fixation, coming in three different varieties, and involving the shuffling of electrons in an unnecessary amount of steps. Electrons get shuffled through the light reactions when enzymes in Photosystems II and I break apart water molecules, and the cytochrome B6-F complex pumps the resulting hydrogen ions to get some potential energy built up. ATP and NADH are used in carbon fixation, using this energy to build not glucose but some random molecule called glyceraldehyde-3-phosphate, which can then be converted into sugar or starch as needed.
This is a biosynthetic pathway. There are many enzymes and proteins needed to get each step of photosynthesis to work, and each enzyme is encoded in the DNA of a cell. When the sun rises and the chloroplasts in the cells wake up, those genes turn on and start producing enzymes to get the job done.
It’s the same with other molecular builds. Photosynthesis produces Gal3P, but there are pathways in the human body that produce dopamine and hemoglobin. There are gene clusters in the human genome that produce enzymes that synthesize dopamine out of phenylalanine. And the Pseudo-nitzschia genome has genes that produce enzymes for the synthesis of domoic acid. Well, toxic species of Pseudo-nitzschia do, at least.
The authors of Biosynthesis of neurotoxin started with a few premises. Other publications had guessed that DA was likely synthesized out of two other molecules: GPP and L-Glu. Other sources showed that Pseudo-nitzschia is more likely to produce DA under conditions like increased CO2 partial pressure (pCO2) and limited phosphate availability. So they looked at the molecular structures of GPP, L-Glu, and DA and asked: What kind of chemical reaction is needed to get from these two molecules, to the molecule DA?
Because that’s what cells do. They are remarkable chemical laboratories. Their genes code for enzymes that help chemical reactions along (they catalyze chemical reactions). Cells can synthesize large and complicated molecules with incredible chemical and physical properties. And they do it a lot better than we can. Insulin, for example, has two disulfide bonds that make it difficult for us to produce in a chemistry lab. We’ve engineered microbes, however, that can synthesize it for us, which we then harvest for them to purify and administer to diabetics.
Where did they begin in tackling the problem?
The authors began with chemistry as a window into formulating their first steps. Previous studies have suggested that domoic acid comes from two molecules, geranyl pyrophosphate (GPP) and L-glutamic acid (L-Glu). Their molecule structures can be seen in Fig. 1A.
So you can imagine, if you cut off that phosphate group from GPP (see all the P’s and O’s? that’s the phosphate group on geranyl pyrophosphate) and attach it to the N sticking out on L-Glu, you’ve got most of the domoic acid. You just have to change a few functional groups (hello, organic chemistry) and you’re there. How does that work?
The authors of the paper have helpfully color-coded the domoic acid. There are red, blue and black molecules, and there are red, blue, and black parts of the DA. The red oxygen atoms are converted from an oxygen molecule into a fancy oxygen-containing functional group attached to the blue (GPP) segment of DA. In organic chemistry terms, we have changed an alkane into a carboxylic acid. In redox chemistry terms, we have oxidized a reduced molecule. Let’s briefly consider both.
Organic chemistry talks a lot about functional groups. The basics are related to each other: alkanes contain only carbon-hydrogen bonds. (Each notch in the line is a carbon atom, and the hydrogens are so common we don’t bother to draw them in.) When we replace a C-H bond with a C-O bond, we are really taking away an H+ and adding an O-2. That changes the redox state of the molecule; it has become oxidized. Every time you do this – break an -H bond, add an -O bond, you are oxidizing the molecule.
Seeing and noting this, the authors knew that a redox reaction would be required to get DA out of GPP, L-Glu, and that O2 molecule. There are enzymes encoded in DNA that can do redox reactions. One such enzyme common across living things is cytochrome P450 (CYP450). If they could find a CYP450 associated with DA production, they postulated, they could find a key to the whole DA pathway.
What did they do in lab to answer their questions?
The authors decided that a transcriptome sequencing approach would be a good start to locating this mysterious redox enzyme. Transcriptome sequencing is a method that allows scientists to identify not just what genes are present in the genome, but which are turned on under given conditions.
We already know DA-inducing conditions: increase the pCO2, and decrease the phosphate availability. When this happens, the genes Pseudo-nitzschia uses to produce DA are turned on. That means they are read and transcripted in the nucleus by RNA, and translated into proteins outside of the nucleus. (This is nothing new – the Central Dogma of biology is this transcription/translation process, of DNA to RNA to proteins.) As another human biology analogy, if I were to sequence the genome of a neuron, I’d find genes for making insulin, hemoglobin, and dopamine. But if I were to sequence the transcriptome – which genes are actually being used? – I’d never find hemoglobin or insulin genes! I may not find dopamine, either, though. Who knows if those genes are being used at present? But if I feed my patient chocolate whilst somehow reading the transcriptome of their neurons, I might find that a condition like ingesting chocolate induces dopamine production, and I start seeing dopamine genes show up in the transcriptome.
That’s what they did with Pseudo-nitzschia. They exposed it to known DA-inducing conditions, and they looked at the transcriptome. But they didn’t just care which genes were expressed under DA-inducing conditions; they wanted to know which genes turned on in between regular conditions and DA-inducing ones. When they looked at the change in expression, they found that only 2.5% of the RNA transcripts were upregulated with limited phosphate, and that this small subgroup of transcripts shrunk even more when they added the other condition of increased pCO2.
And guess what? Of that handful of genes leftover, one of them was a CYP450. You can see the data in Fig. 1B from the paper.
Let’s spend a moment reading this graph. The colors denote expression levels. Green means the gene is being upregulated; it has been turned on and is being expressed in the cell. Purple means it is downregulated, with black the color denoting the gradient. Changing phosphate conditions from high (there’s plenty to go around) and low (limited availability causes the largest switch in gene regulatory activity. All of these genes are upregulated when phosphate is limited; they become much greener. Some experience a greater change in upregulation than others; for example, the glycoside hydrolase RNA transcript jumps from black to green, but the dioxygenase jumps from a deep purple to green. According to the key, that’s a increase in upregulation of ~2 logFC for glycoside hydrolase, and ~4 for dioxygenase (don’t worry about what “logFC” means for now; just recognize it as the unit by which upregulation is measured here).
Then there is the condition of increased pCO2. Looking at the 40S ribosomal protein RNA transcript, increasing pCO2 actually downregulates the gene (it moves from lighter green to darker green). In the terpene cyclase, though, you can see an increase of upregulation under increased pCO2, denoted by the change to a lighter shade of green.
The authors chose CYP450 as their favorite from this list – they had been looking for it already as a redox enzyme. Then they located it in the Pseudo-nitzschia genome and found it nestled in a gene cluster alongside, lo and behold, three other genes that followed the same pattern from their RNA transcripts, shown in Fig. 1C. This was actually a serendipitous discovery; the authors weren’t certain if the genes would be clustered together or scattered throughout the genome. So finding four genes all upregulated under DA-inducing conditions clustered together was a convenient surprise.
All four of these are upregulated under limited phosphate and increased pCO2. And as part of a gene cluster, they likely function together as well. Scientists know the chemical activity of terpene cyclase and dioxygenase, so they used that known chemical activity to piece together a pathway for the biosynthesis of DA. (Also note the names “dabA,” “dabB,” etc. assigned to each gene. These gene names were assigned by the authors and merely stand for “domoic acid biosynthesis A,” etc.)
In the next post we’ll look at what that biosynthetic pathway actually is, and how they tested to see if these genes did, in fact, make domoic acid.
Also stay tuned for a deep-dive post on the methods used in lab.
Brunson, J. K., McKinnie, S. M. K., Chekan, J. R., McCrow, J. P., Miles, Z. D., Bertrand, E. M., … Moore, B. S. (2018). Biosynthesis of the neurotoxin domoic acid in a bloom-forming diatom. Science. https://doi.org/10.1126/science.aau0382
Genetic Engineering of Insulin Production
Really great visuals on genetic engineering for insulin production specifically but also genetic engineering in general.
Khan Academy, everyone’s favorite online reference source, has a video “Introduction to genetic engineering” that includes talk about synthetic insulin production.
Great visual & explanation of redox reactions, applied to organic chemistry applications.
& don’t forget to check Khan Academy.