One of our reporters tried to do CRISPR. He failed miserably

I speak biology fluently, but the molecular complexities of the novel genome-editing tool called CRISPR left me as befuddled as when I peruse descriptions of the inflationary universe. So I decided to test what one investigator told me: CRISPR (for "clustered regularly interspaced short palindromic repeats") may sound intimidating, but it is so simple to use that "any idiot" could do it.

I would give it a try.

CRISPR works best at crippling, or knocking out, genes, so that's how I choose to use it. But I aim high: I target an immune gene that, I theorize, could lead to insights into reducing the harm done by Zika virus. (My admittedly wild hypothesis is that the gene, CD32, may help drive Zika virus to copy itself to higher levels if a person was previously infected with dengue and has antibodies to that virus.)

Roland Wagner, a postdoc in the lab of Sumit Chanda at the Sanford Burnham Prebys Medical Discovery Institute in San Diego, California, agreed to serve as my CRISPR sensei. An experienced rock climber originally from Austria, Wagner approaches everything methodically. He pulls up the sequence of the CD32 gene, which has five distinct protein-coding regions. If we cut the DNA in one region, the gene most likely would be knocked out: It would no longer make its protein.

CRISPR uses a guide made of RNA to direct molecular scissors—part of the CRISPR-associated protein, or Cas9—to exact spots in a genome. We could buy the guide RNA (gRNA), but the idea appalls Wagner. "I would assume it's probably $500 to buy the gRNA, but I wouldn't know," he says. "We're making our own and we're spending about $5."

The gRNA sequence must complement a stretch of 20 nucleotides on the segment of the CD32 gene we want to cut. But the same DNA sequence could recur elsewhere in the genome, leading the molecular scissors to cut in the wrong place. Such "off-target" effects can cause mayhem, and eliminating them is a key goal of those honing their CRISPR skills. To make the match more specific, Cas9 requires an additional sequence flanking the targeted 20 nucleotides: N-G-G, in which "N" can be any nucleotide. Where Cas9 finds N-G-G immediately following the 20 nucleotides, it attaches to and opens the double helix, allowing the gRNA to bind. Cas9 then cuts each strand of the DNA.

To homebrew our gRNA, Wagner copies the sequence of the CD32 segment we've identified and pastes it into a freely available database, Optimized CRISPR Design, that looks for a matching set of 20 nucleotides followed by N-G-G. There are 41 options within CD32. The database scans the entire human genome to see whether there are identical matches elsewhere—potential sites of off-target cuts. We select a sequence that appears unique, and then he goes to another website and orders a stretch of DNA—an oligonucleotide—with that sequence.

The oligo arrives, and I lose my modern pipetting virginity. I have not worked in a lab since I was an undergraduate more than 30 years ago. Back then, I learned a pipetting technique that probably was invented by Louis Pasteur: I put a finger in my mouth and then sucked up a chemical into a thin glass tube, capping it with my fingertip when I had drawn up enough.

Now, at Wagner's lab bench, I face a rack of fancy plastic gizmos that look like squirt guns but enable users to suck up precise microliters of liquid with a push of a button. My task is to pipette the oligo from one tiny test tube into another. The second tube holds a plasmid, which is a circular piece of DNA that will act as a Trojan horse. This plasmid, customized for CRISPR experiments, already holds the gene for Cas9. It also contains a 60-nucleotide "hairpin" sequence that ultimately will attach to the 20 nucleotides I add to make the full gRNA.

I use one of the fancy pipettes to move the oligo into the plasmid tube, and I also add buffer, water, and an enzyme. If all goes well, the enzyme will cut open the plasmid, removing a piece of its DNA and allowing the oligo to take its place. All does not go well.

"Oops!" Wagner says as I pipette the enzyme. "You failed a little bit." I apparently hit the pipette button before submerging the tip into the liquid.

In the end, I manage the procedure. After waiting for the chemical reactions to take place, we take my CRISPR plasmid to an electrophoresis machine, a tray that has wires hooked to it. We add a liquid that quickly turns to gelatin, and then I pipette a few drops of my CRISPR plasmid into different lanes on the device. I flip a switch to apply an electric current, which should separate the DNA into bands based on weight. The small piece of DNA I cut out with the enzyme should form a distinct band.

My gel electrophoresis only has one band, from the plasmid.

"It doesn't look like it worked," Wagner says gently. "I didn't want to be all picky, but it could be that you messed up the enzyme with your pipetting." He allows that when he was starting out, his experiments often failed. "I'd go home and I'd say, 'I hate my life,'" he confided. "There are a lot of setbacks in science."

I've already learned that any idiot cannot do CRISPR: It takes, at least, basic laboratory skills.

Wagner conducts the experiment in parallel with me, and his plasmid properly incorporates the oligo that will guide Cas9 to its target. We then coax a cell line made from an embryonic kidney into taking up the Trojan horse plasmid. After a few days, we isolate the DNA from the cells, amplify it with the polymerase chain reaction, and use electrophoresis to show that the CD32 gene has been cut into pieces. Voilà, our knockout worked.

"You did great," Wagner tells me. "Way to go!"

I did not do great. But CRISPR did its job.