Looking for the Globin Switch

Nancy Marie Brown
September 01, 1996
yellow balls and one large orange ball

I'd heard of the firefly trick—taking the gene that makes the bug blink and sticking it into a heretofore unblinking cell, say, a human red blood cell. I was hoping to see it done when I scheduled an afternoon last April in Penn State biochemist Ross Hardison's lab. But my tourguide, undergraduate researcher Monette Aujay, merely shrugged.

That work was years old.

But did she use it?

She tilted her head, eyed me calmly. Well, yes.

The lab was a large, bright room, divided into aisles by black countertops, each surmounted by head-high shelves. Jars of colored liquids caught the light. A mobile of paper skiiers hung from the high ceiling. A catfish swam in an algae-darkened tank. The Norfolk Island pine by the window still held six candy canes.

Aujay took me from the bench, where she'd been explaining how to do a restriction digest, around the corner to her tiny cubicle. She opened her lab notebook to two diagrams: rings marked out like roulette wheels, each depicting a vector, a small circle of DNA that can be easily slipped into a cell. One quadrant was marked luciferase, the technical name for the blinking gene.

"I'm working with a piece of HS-5 that's 2.5 kilobases long," Aujay said, meaning a snippet of one of the genes that regulate hemoglobin production. "I cut it out of the vector containing luciferase and splice it into a yeast vector." She paused and coolly met my eye. "That is, since last summer I've been trying to put the pieces together."

The lab's version of the firefly technique—putting the pieces together to make a red blood cell blink—was devised by a graduate student, John Jackson, and two undergrads. One of them, Brent Grotch, won a Pfizer fellowship for his work; he's now finishing medical school at Cornell. The other, Aaron Goldstrohm, garnered the Hauser Award for Undergraduate Research in 1994; he recently started grad school at Duke.

Why make a red blood cell light up?

"We're broadly interested in the molecular basis of gene regulation," Hardison explained, "particularly in how the hemoglobin genes are regulated in humans.

"Do you want pictures? Let me get a handful." He gathered together a stack of mounted diagrams, the bits and pieces of a poster he'd presented at a scientific conference. Some were quite large and unwieldy, and when I commented on that, Hardison laughed. The only way he could think of to carry them to the meeting was in the box for his toddler's "doodle table." That raised some scholarly eyebrows.

He pulled a picture from the stack. It showed a looping, squiggling, necklace-like thing all knotted around bars labeled "Matrix" and boxes called "Boundary." He looked at the diagram, looked at me, and decided to back up.

"Hemoglobin is a tetramer," he began, "which means it's made of four proteins. Two are alpha-globins and two are beta-globins." (The proteins are named globins, I learned later, reading Goldstrohm's thesis, because they're globular in shape.)

"It's called hemoglobin," Hardison continued, "because a molecule called a heme, which has iron in it, snuggles right in between the helices of globin proteins. It's a very reversible way of binding oxygen." (Red blood cells, notes Goldstrohm, are "essentially a bag full of hemoglobin"; each cell can hold "some 280 million molecules of hemoglobin.")

"Hemoglobin," said Hardison, "is one of the most well-studied proteins. It's extremely well-studied. It was crystallized in the 1860's—while we were fighting the Civil War. It was the first protein whose crystal structure was solved, in the mid-1950s."

Yet it has its secrets, if not in its structure then in its genes.

"At first blush, these globin proteins look like each other," Hardison said. "They're similar, but they're different. They derive from the same ancestral gene, but they split roughly 450 million years ago."

According to Goldstrohm's thesis, globins go way back—"well over 1 billion years, based on their existence in bacteria and lower eucaryotes," simple cells, like yeast, with well-defined nuclei. Leghemoglobin, which carries oxygen out of plants' root nodules, diverged from the animal varieties a billion years ago. At 600 million years ago, the globins in animals with jaws began to differ from that found in jawless ones. Hemoglobin comes in at about 500 million years ago, and 50 million years later, its globin component split into alpha and beta subunits.

"If you look at contemporary humans," Hardison added, "we have several alpha-globin genes and several beta-globin genes. The alpha-globin genes are on chromosome 16, and the beta-globin genes are on chromosome 11. And even among the beta-globin gene family, we have embryonic, fetal, and adult forms."

Altogether, clarifies Goldstrohm, humans can make six different hemoglobins—three during the embryo stage, one as a fetus, and two as adults—using the different forms of globins. How does the developing fetus know to make one hemoglobin and not another? Where is the switch that turns off, say, the fetal globin gene? And, like a light switch, can this developmental switch be turned back on once it's off?

"It turns out," said Hardison, "that the regulation of these genes is under the control of some DNA sequences close to them, but also of some that are rather far away, what we call Locus Control Regions."

The switch, as it were, was jigsawed into a number of pieces and scattered about the chromosome. To help collect them all you need the firefly gene: You know the switch is working—that you've found the right bits—if the cell lights up.

Nirmal Veeramachaneni, who won the Hauser Award in 1995, giving Hardison's lab back-to-back victories, quoted Kipling on his poster for the Undergraduate Research Fair.

We be of one blood, ye and I...

But not of one hemoglobin.

"The point is," said Hardison, "you've got this basic tetrameric structure that can bind oxygen. But there are many variants. Some we can use. Some can be highly detrimental—such as the one that causes sickle cell anemia.

"When red blood cells with this sickle-cell hemoglobin release their oxygen, they form abnormal sickle-shaped structures. The hemoglobin forms long polymeric arrays. The red blood cells get rigid and sticky, and they can clog up capillaries."

The sickle-cell mutation is one of at least 600 known mutations in human beta-globin proteins. It's fairly common—especially in people of African or Middle Eastern, since inheriting a sickle-cell gene from one parent provides protection against malaria. (The parasite that causes malaria spends part of its life cycle inside red blood cells; when it invades one with the sickle-cell form of beta-globin, the cell sickles and is destroyed—along with its deadly hitchhiker.) Having two sickle-cell genes, however, being homozygous for the disease, can lead to organ failure, paralysis, and death: An estimated 1 in 400 African-Americans are homozygous; 8 to 10 percent carry one copy of the gene.

"That's why we're so interested in these developmental switches. Fetal hemoglobin has the property of counteracting or ameliorating the homozygous sickle-cell state," said Hardison. Surprisingly, the fetal form of hemogloblin made by people with sickle cell anemia doesn't polymerize nearly as badly as the adult form. And, Hardison noted, "everyone who has a bad' beta-globin gene also has a good fetal globin gene.

"Can we, by pharmacological or genetic means, keep the fetal hemoglobin gene switched on? This is really pretty exciting. If we could figure out what it is around the promoter site of fetal globin genes that turns them off in adults, we could design drugs to mimic that—to block a silencer or to turn on an activator."

Checking out potential silencers and activators is what Hardison calls "a nice project for undergraduates or early graduate students. It can be done with fairly crude materials."

And lots of time.

Goldstrohm began working in Hardison's lab as a freshman. "I knew I wanted to do research," he said when I interviewed him at Duke University, where he'd begun graduate school. "I wanted to be a scientist—I wasn't sure of what.

"Ross's lab direction appealed to me. Human genetics was one of my main interests."

He smiled. "Actually, for the first couple of years it didn't matter what I was working on. I was learning the techniques. Just finding stuff was a major task."

By the time he was a senior, though, he was searching a likely strip of DNA for transcription factor binding sites.

He looked at me, looked at my notepad, smiled as if remembering a lesson, and backed up. "Basically, Ross and Webb Miller, a computer scientist at Penn State, have developed a computer program to align DNA sequences from different mammals to get an idea of which areas have been conserved over hundreds of millions of years. These areas may have some important function."

Or, as Hardison had explained it to me earlier, "One approach we're pushing is to use the lessons of evolution. Which short segments tend not to change very much over the course of evolution? The idea is that the most highly conserved sequences in the gene are the important ones. There's a lot of sequence data available," he added, "from humans, bushbabies, rabbits, goats, cows, mice. Which segments are found over all these species?" The enormity of the data—the DNA sequences that code for the beta-like globins in these animals are each 45,000 to 75,000 base pairs long—made the computing difficult, "but once you get them aligned, you can easily pick out conserved regions."

That, however, is just the starting point.

Both Nirmal Veeramachaneni and Aaron Goldstrohm (who were "extremely good, really amazingly good students," according to Hardison) spent a year or more dissecting the chemical workings of one of these conserved regions of DNA.

Goldstrohm took out a photocopy of the looping, squiggling, necklace-like diagram Hardison had shown me briefly before, and pointed to the dot labeled "2." "My site was in Number 2—that's about 1,000 base pairs long—and I was looking at several regions within it. I had about six complexes"—sites on the gene where proteins seemed to bind—"and good evidence that they were real, not just some artifact of the experiment. Other people in the lab had shown that if you deleted some of these regions, it would cause globin gene expression to decrease. But nobody knew what was binding to those sites to activate transcription. I tried for my last eight months to find out."

What he wanted was the protein, the "transcription factor." Before any copying of the gene could start—way before that RNA copy could itself be turned into a globin protein—this transcription factor would have to arrive, settle into its favorite chair, and create the proper environment. Set the mood, as it were. Maybe pull over a congenial buddy --another transcription factor, bound to a different spot of DNA—to create a loop or a squiggle or a knot.

"Which one was it? Where was it binding? I knew something was binding here—what was it doing? Was it turning something on? Or off?"

Goldstrohm smiled. "I ran out of time. That summer I stayed on --" It was quite a sacrifice. A farmboy from western Pennsylvania, he had earlier confided that "the hardest thing about science is you don't build anything. On the farm, you could work all day—and you could see the pile of hay in the barn. In science, you can work for months and have nothing." (Aujay, Hardison's current student, could relate: The afternoon I visited, she had a large teardrop of liquid to show for her day's work, a bit of DNA to test and see if the pieces, this time, had gone together right.)

Goldstrohm looked out the window of Duke University's majestic pink conference hotel: a trim golfcourse, a fringe of trees; no farms in sight. "If I hadn't been in the lab doing real science, I wouldn't have made it," he remarked. "You can read and read and read, but you won't learn until you go into the lab and do the work."

Hardison would agree. At the 1996 Undergraduate Research Fair (which none of his students entered, spoiling his chance at a hat-trick), he was asked to give a speech on "The Value of Undergraduate Research." "The most important thing," he said, "is that you have taken on the difficult and challenging task of research, and have done well in its pursuit. That's where the long-lasting rewards are in research: in learning to formulate a testable hypothesis, a question to pursue; in designing an experiment or project that will inform us about that hypothesis; in learning and executing the techniques; and, when it all works, in learning something important that no one knew before. That's a truly exhilarating and joyous event!

"Yes, many of you received academic credit for your independent studies," he went on, "but I'm sure that is not the primary motivation for taking on this pursuit. And I'm pretty sure that none of the supervising faculty got much recognition from those who calculate student credit hours as the primary measure of productivity of the University."

"At first we're kind of a pain," Goldstrohm acknowledged, putting himself in Hardison's shoes. "But once he gets us going . . .

"I'd come in and we'd argue about the data, and he'd egg me on. He'd smirk and say, Is that really what you think?' He taught me how to design experiments. He'd act like I was submitting a paper and he was reviewing it. And he would get really excited when you'd bring in your data to show him. He'd jump out of his chair . . .

"It must be a really rewarding thing to turn a fledgling scientist into a critical thinker."

A graduate student, Martin Sigg, took over Goldstrohm's project last September.

"Goldstrohm's area," Hardison said, pointing to the knot labeled 1-2-3-4 on his diagram, "is intensively studied. It was a good candidate as early as 1989. The race was on to find out what proteins are active there."

From the work at other labs, he noted, "It looks like there's a binding site here for AP-1—that's activator protein-1." Goldstrohm looked next door on the gene, finding two more possibilities. "One is a very well known protein," Hardison said. "What excites me is we can't figure out what the other protein is. "We may pull out a unique protein there. It may be half of a switch."

Ross C. Hardison, Ph.D., is professor of biochemistry in the Eberly College of Science, 206 Althouse Lab, University Park, PA 16802; 814-863-0113. He is collaborating with Webb Miller, Ph.D., professor of computer science and engineering in the College of Engineering; 865-4551 and graduate students Laura Elnitski, John Jackson, Brian Shewchuk, and Martin Sigg. Their work is supported by the National Institutes of Health.

Last Updated September 01, 1996