Best in a crisis, Boger averted one now by breezily refocusing the debate. Forget Schreiber, he counseled the scientists returning from Pittsburgh. All that matters, all that ever matters, is data, data that now was emerging spectacularly, data that was Vertex’s, exquisitely if not uniquely, to exploit.

It was a powerful balm—redemption in the next experiment—and Boger administered it staunchly, without recrimination or doubt. He reminded them that this was the moment they’d worked for. Thanks to Schreiber, who as promised had quickly released the computer coordinates for his and Clardy’s X-ray structure, Vertex finally had sufficient information to make a first substantive pass at designing drugs. It had detailed blueprints both of the native protein and of the protein in bound conformation with FK-506. Meanwhile, Navia, moiling penitently at the bench while the others were in Pittsburgh, had grown earring-sized cocrystals of FKBP-12 and Vertex’s lead semaphor compound, 367, and had passed them to Yamashita, who, revivified and cramming for his medical school boards, was now within days of solving that structure as well. And there was, whatever one thought of it, Schreiber’s calcineurin bonanza. Excusing the source of much of the information and the significant new problems it presented for drug design, Boger couldn’t have been more pleased. Exuberant, he left the scientists few excuses to mope.

He himself overcame the loss to Schreiber as he often had: by minimizing Schreiber’s role, writing it off to a character flaw, damning the academic star system that allowed Schreiber to take credit for it, and then, to ensure his point, ridiculing the discovery. If Schreiber now considered Boger too self-reliant for his own good, Boger dismissed him as a black hole, a view freshly supported by the stream of publications following the Pittsburgh conference.

According to this paper trail, the first evidence of a specific recepter for an immunophilin complex came not from Schreiber’s lab but from that of a Stanford immunologist named Irving Weissman. Weissman was world renowned, a leader in the molecular biology of T cells, far more accomplished—and famous—within cell biology than Schreiber. Five months earlier, one of his graduate students, Jeff Friedman, had discovered that cyclophilin and cyclosporine together bound to an unidentified protein with a molecular weight of about 55,000. Weissman called Schreiber, who he knew was looking for a similar receptor for the FKBP-12/FK-506 complex. “It is clear to all of us that there had to be a common mediator, and this seemed to be a candidate,” Friedman says. Schreiber flew to Palo Alto. He took Friedman to dinner at an Italian restaurant, where Friedman agreed to share his data. The work, obviously of great promise, was the essence of his doctoral thesis.

Back in Cambridge, Jun Liu, a postdoc in Schreiber’s lab, quickly identified Friedman’s protein as calcineurin. Working further, he found that it was a common binder for both complexes and determined that two other small proteins were involved as chaperones. Thus, as the publications now indicated, Schreiber’s lab had broken the story, but not without being given the basic facts by Weissman and Friedman.

Yet in Pittsburgh, Schreiber had credited Liu strenuously but had scarcely acknowledged the contribution of the Stanford group. To the world, calcineurin—like FKBP-12 and its structure—was Stu Schreiber’s discovery. Schreiber had used his prerogative in announcing the work to assume overall credit for it. Said Friedman, “My view of Stuart is that he’s pretty much a megalomaniac. I will never work with the guy again.”

To Boger, who scarcely needed persuading, this was yet another example of Schreiber’s persistent self-aggrandizing. Schreiber had seen the whole picture, had put himself in a position to enable it to emerge. It was a brilliant example of scientific leadership. But Schreiber apparently had not had the initial key insight himself. Typically, Boger dismissed not only the accomplishment but the implications, particularly Schreiber’s assertion that cyclosporine and FK-506 were “molecular glue.” Scoffing at the idea that the effector domains of cyclosporine and FK-506 were strong enough to tat together two much larger proteins on their own, he said: “This is Uri Geller bending spoons. Don’t show me the evidence. It’s impossible.”

Boger’s cleverness and deft moral certainty blunted most of the attacks on his leadership but not the scientists’ despondency over what was now expected of them. Schreiber’s assertion that FK-506 wasn’t a drug at all but an extraordinarily fortuitous dollop of molecular adhesive, in fact, had calamitous implications for everyone in the field, particularly Vertex. Indeed, the impact could hardly have been worse.

Boger notwithstanding, there was little question that Vertex’s rationale for improving on FK-506 had taken a severe, if not mortal, turn. In less than two years, the project’s degree of difficulty had increased exponentially. What had started with a simple enzyme inhibition problem—to insert a tighter binder than FK-506 into the active site of FKBP-12—had been replaced by a quagmire. If Schreiber was right, and what was necessary for immunosuppression was to inhibit not FKBP but calcineurin, everyone in the field now had to change horses. They faced the immeasurably harder task of mimicking a molecular architecture that was largely unknown, that involved as many as five entities that changed shape, perhaps dramatically, on contact, and that would require the bound structure of at least several, if not all, of them to visualize. It was like designing pieces for a sloshing, scissoring three-dimensional jigsaw puzzle, blindfolded, with outmoded templates.

Boger, as ever, was confident such an effort was possible. “The idea that FK-506 is the best it can be is ridiculous,” he said. His faith that drugs discovered through screening are accidents of nature and thus, by definition, flawed was unshaken. He dismissed recent reports that Merck, despite hundreds of man-years, had been unable to make a single change to FK-506 without sharply reducing its biological activity. Typically, and with notably less success than usual, he exhorted the scientists to join him.

There were solid reasons for the scientists’ reluctance. It was, after all, extraordinary. However unintentionally, Schreiber’s tying of FK-506 and cyclosporine to a single partner protein legitimized the view, promoted most assiduously by Fujisawa, that the two drugs were, in fact, not accidents at all. According to this interpretation, the molecules were evolved specifically by nature to do exactly what they seemed to be doing in T cells: tacking FKBP-12 and cyclosporine to calcineurin (all of which are found in all cells and in all organisms) as part of some more universal biochemical interplay. Such a view would help explain the drugs’ almost mystical activity. Sandoz, for instance, with nearly a decade head start on Merck, had reportedly made 1200 cyclosporine look-alikes, virtually all of which were less potent. From the standpoint of simply tatting together relevant proteins, FK-506 and cyclosporine were beginning to look more and more like what Boger and the scientists could only find ruinous, inconceivable: They were beginning to look perfect.

Boger, unsurprisingly, rejected this view. He was especially caustic regarding the implication that because the molecules had defied improvement by conventional medicinal chemistry, they were therefore unbeatable as drugs. Even if they had evolved to perfection inside their respective microbes to glue proteins together, he reasoned compellingly, they had not evolved to become perfectly bioavailable in humans cells, to survive the gut, to be optimized for reducing side effects. In science, Boger knew, God was in the details, and 4 billion years of microbial evolution, however perfect in itself, was still no basis for designing drugs. That the universality of their targets implied increasingly that the drugs’ side effects were inextricably bound to their activity was a point Boger prudently failed to raise.

It was here—their therapeutic profiles—that FK-506 and cyclosporine were still most vulnerable and where Boger was sure Vertex could still win. However, it remained a “religious question,” as he might have noted, whether Vertex could make a better molecule in its present circumstances, that is, before it ran out of money. Aldrich, especially, worried about this. He continued to believe Boger’s predictions but was unwilling to dismiss Merck’s and Sandoz’s failures as mere wrongheadedness. “If the big boys can’t do it,” he said, “it makes you wonder.” As so often before, Boger’s fearlessness, determination, energy, and focus were a rallying point within the company. But ultimately his success depended on his ability to convince those who worked for him that they could do what many of them were beginning to think impossible. “Josh drives us faster than he should,” said Tung, “but not as many people are as ambitious as he is, and not many people get as much done with as little as he does.”

Boger came to the first project council meeting after Pittsburgh wry, supportive, and, as he often was when he thought the scientists were becoming too self-involved, deeply provocative. The councils, which had gone on unevenly without Boger during the IPO and with increasing tension ever since, had become exhibit A in the indolent mutiny of the past few weeks. Privately, some of the scientists hoped that Boger now would simply disband them, put himself in charge, name department heads and project managers, and tell them what to do.

Nothing was less likely. Far from causing him to renounce his social experiment, the moment occasioned for Boger its first real test. To retrofit a project spontaneously in the wake of new information—to move quickly where big companies like Merck couldn’t—this was Boger’s ideal. The councils, like Lenin’s Soviets, were his sword for permanent revolution, for doing things the way he’d determined was best. He was hardly about to lay it down. Besides, he confessed disarmingly, he didn’t know what to do, not specifically. He would learn from them, from the results of their experiments. When he had more information, he’d decide.

In the short term, he was more concerned about resources. He strolled to the whiteboard and began scribbling numbers. Vertex had thirty-five researchers in immunophilins: five in chemistry, nine in biophysics, twenty-one in biology. As an index of the company’s priorities, the numbers implicitly rebuked the biologists, who had complained the loudest in Pittsburgh and who now swallowed their tongues. He then followed each number with a question mark under the heading Future. Insisting that the overall number wouldn’t grow, he told them to prioritize their experiments and return within a few weeks with new staff adjustments.

Though he’d have preferred not to be considering a complete overhaul in the company’s lead project just two months after telling the world it was on the verge of having a drug candidate, Boger loved these moments, loved to rile things up. He compared them to “free climbing,” where climbers scale mountain walls without ropes. Throughout September, he reveled as the scientists struggled to assemble a response. Certain experiments were obvious. Before they reoriented the project toward designing inhibitors of calcineurin—a challenge so daunting that if it were to come up in the New Project Council, Vertex’s long-range planning group, it would have been howled down at once—they had to test Liu’s and Friedman’s data. Schreiber and his allies could, after all, have been wrong. Vertex needed quantities of calcineurin and an assay for testing its inhibition. Meanwhile, they would keep making and testing molecules against FKBP-12. Livingston’s proposal for shutting down chemistry notwithstanding, Vertex had already generated small molecular inhibitors of the protein that were both immunosuppressive and orally available without knowing the ultimate receptor. It would be foolish not to continue.

Attracted by Boger’s rough optimism, goaded by his energy and enthusiasm, most of the scientists affected a speedy convalescence. They sloughed off the last of the postcalcineurin panic and, by the end of September, were moving forward again. They had another incentive: Vertex’s stock. As disconnected as ever from events within the company, it had begun climbing in August and by September 26 was at $15 a share—a 66 percent increase. It was now at the precise level that Vertex had first proposed four months earlier and had been forced to slash during the IPO.

“Two months ago we couldn’t sell it for $10,” Aldrich groaned absurdly on a day that Vertex’s stock jumped 7 percent. He spent the afternoon fending off reporters wanting to know why the company suddenly was so hot. The run-up, Aldrich understood, had little to do with Vertex per se. As if to confirm that the company had gone public at the worst possible moment, Wall Street’s brutal fence-sitting in June and July had resolved itself in a second manic romp almost as euphoric as the first. Again biotech boomed. Again the market hallucinated. A start-up called Medlmmune, years away from making immune system modulators for treating AIDS, had gone out at $9.25 in May; by the end of August it was at $27. Another company, Somatogen, counting on a questionable demand for artificial blood substitutes, jumped from $19 to $36—in a month. Story stocks once again had a kind of herd immunity: They were inured, even against their own shaky internals, by the artificial robustness of those around them.

It was caveat emptor all over again. Nothing—or at least nothing the public knew about—had changed to warrant the increase in Vertex’s price. Indeed, the very moment its paper worth was soaring by two-thirds, some of the scientists were despairing of it ever doing what it said. It was hard, under such irreconcilable circumstances, not to be cynical, and Boger moved swiftly to remind the researchers that Wall Street’s fickleness should not be confused with the reality of their progress or their worth. If the year on Wall Street had proven anything, it was that investors didn’t—couldn’t, perhaps shouldn’t—know what they were buying. Discovering drugs was a hairy, uncertain business, freighted with precipitate moments like this one when everything simply and stunningly, to use Holman’s phrase, “fell out of bed.” Small unprofitable companies were nightmares of disorder and discord. Like sausage making and writing laws in Churchill’s famous apothegm, it was probably best not to know too intimately what went on with them. And yet an industry that lived by the story died by the story. The point again came home chillingly in October when a lackluster company called Anergen saw its stock rocket 400 percent, then crash as a result of widely misinterpreted news accounts about another company promoting a similar technology. Like a look-alike in a mystery, Anergen was hijacked and taken for a white-knuckle ride by mistake. The company stopped answering its phones.

Boger discouraged any competitive rejoicing. Anergen was working in autoimmune diseases. There was no telling what the toll would be for other small companies doing the same when irate investors realized how they’d allowed themselves to be deluded. “It’s like a Sunday school talking point,” Boger said. “Yes, there’ll be a day of reckoning, but that doesn’t mean the damage is going to be inflicted justly. Weren’t there any good people in Sodom and Gomorrah, or were they all wicked? I’m not worried about us. I’m worried about others falling on us.”

The stock run-up was a safety valve for the disheartened, especially Yamashita, who planned to use the proceeds to put himself through medical school—to escape. Others, particularly Murcko, were annoyed. “I wish it would go to $4,” he said unironically. He was thinking about future stock options, where lower prices would be more valuable.

In mid-September, Yamashita completed the complex structure of FKBP-12 and 367, giving Vertex its first detailed look at the bound anatomy of one of its own compounds. Murcko inhaled the data like a drowning skater suddenly discovering a pocket of air under the ice. At last, he would be doing more than modeling; no more bald speculation, no qualifications, no apologies. Within minutes he had the new structure up on the screen, side by side with Schreiber’s and Clardy’s structure of FKBP-12/FK-506. He would show, once and for all, why FK-506 was an immensely potent drug and 367, a weak imitator. He would see, see, at last what he had to do.

The juxtaposition was stark, suggestive. FK-506, as Schreiber had predicted, had a protruding effector domain that reached out from the portion of the molecule that bound to FKBP-12. The right-hand flag of 367’s semaphor, meanwhile, lay crumpled in the active site, also reaching, but not as far. Together the structures looked like before and after pictures in a shaving ad: FK-506 sticking hairlike from the follicle of the binding pocket; 367, same full follicle, but a shorn stump, cut clean. Even without knowing how the molecule fit with calcineurin, Murcko knew Vertex was going to have to put something into the region beyond the protein’s undulant surface, a hook of some kind. Admiringly, he noted that Schreiber’s effector hypothesis was substantially right, though not necessarily his assertions about calcineurin. “It’s one percent of the protein in the brain,” he said. “I don’t think you want to just knock it out.”

And yet Schreiber’s hypothesis didn’t explain everything. Looking at the two structures three dimensionally, Murcko noted the relative insignificance of the effector. It was like a nail jutting from a warped plank: enough, perhaps, to join it to another piece of wood, but not if one wanted them to stick. For that, one wanted more surface area contact, a coat of glue. Whether or not he needed a molecular doorstop—a group of atoms to prop open the flap region—Murcko now doubted more than ever that the effector region alone accounted for the drug’s action. The atomic configuration around the active site, as Boger had speculated, seemed also to be involved.

Murcko pinned Vertex’s computer network with dozens of studies comparing the binding energy of the two complexes. He had always talked about designing drugs as an “iterative process,” a kind of smart person’s trial and error. Not that he considered himself smarter than the chemists, but by modeling different atomic configurations, then calculating their efficiency, he could predict which ones would be most potent. He could see how moving a few atoms an angstrom or two to fill an empty pocket might make a compound bind more tightly. He could juggle electron clouds, substitute charges, make thermodynamic adjustments measured in millionths of calories. The key was to take the information he was generating and wrap it into specific suggestions that Armistead, Saunders, and the other chemists would listen to, suggestions that were not only sound, but easily tested. No chemist wanted to hear about a potentially great molecule that might take a month to synthesize. Who would bother? Murcko’s degree of confidence—a sore point for all modelers, and particularly Murcko, who had grown more and more anxious to prove both the validity of his methods and his own worth and whose frustrations had long been huge—depended not only on the quality of his predictions, but on his ability to persuade the chemists to make what he designed.

Weighing everything—the obvious need to build out into the effector region, the differences between 367 and FK-506, the unchartable but apparently significant changes in the surface—Murcko concentrated on the semaphor. Drawn on paper, its arms lay outstretched in a Y. Three dimensionally, however, he could see that they crumpled on binding, collapsing like a fighter’s arms in the clenches. Here, possibly, was an opportunity. Harding had once compared 367 to a skater falling through a hole in the ice. As long as its arms were loose, it would slide through. Rigidify them somehow, though, and they might catch. The molecule might save itself.

Murcko modeled several new molecules with arrays of atoms inserted to keep the semaphor pointed outward, outward against the flap, outward toward the effector region. Calculating a substantial improvement in binding, he interpolated which ones might most improve the drug’s overall activity while being easist to make, then suggested them to the chemists.

There is a presumption among nonscientists that research breakthroughs are inherently dramatic—great eureka truths, blazing revelations, thunderous insights, shouts of joy, the combined emotional stimulation of a great college basketball rivalry and a soaring aria. More often, the opposite is true. Some small deed is made. An adjustment. A scientist, frustrated, half rises with an idea just different enough to elicit a subtly better result that reinforces the conviction to go on. Always, the seed is the moment when one experiment, one reagent, one method is chosen over another, a small, but critical, branch point.

The chemists welcomed Murcko’s suggestions, but ambivalently, noncommitally. Such communications between modelers and bench chemists are inherently awkward, since a decision by a medicinal chemist to make someone else’s compound, besides requiring a commitment of time and energy, is equally a choice not to pursue some favored idea of one’s own. And yet the chemists were stuck. For months, Armistead, Saunders, and the others had been making derivatives of 367 that bound ever more tightly but with no corresponding gain in immunosuppression. They, too, had concluded that the molecules weren’t extending far enough and had faulted the semaphor. The problem, from a production standpoint, was how to shore it up without making it so hard to construct that it became worthless. Said Armistead, “The bottom line for us was that we knew it took Stuart’s group six to eight months to synthesize 506BD. We didn’t want to do that.”

Working in series, Armistead and Saunders concentrated on one of Murcko’s ideas: inserting a small planar ring of six carbon and six hydrogen atoms into the semaphor’s crotch, as a chock, to pinion its appendages. Six compounds later, Saunders made a molecule—563—that was three times more active in cells than Vertex’s previous best hit. Within weeks, Patsi Nelson noted a commensurate gain in activity in mice; the molecule was immunosuppressive in animals. It was the most potent Vertex had made yet.

The chemists, their confidence rebounding, immediately began scaling up production: for more assays; for toxicological studies; for Navia and Yamashita, who began at once to try to cocrystallize it with FKBP-12 so that Murcko could compare that complex with 367; especially for shipment to Chugai, to whom Boger had again promised just such a gain in molecular activity.

It was a bravura leap, an exceptional piece of science, not a great theoretical breakthrough, but a great practical one. Gains in cellular activity tend to come slowly, unpredictably. Yet another one like this and Vertex would have a drug candidate. Suddenly, Murcko’s “feedback loop” was installed for one brief, exhilarating cycle and had worked perfectly, just as Boger had predicted.

Boger was electric, infused. The process was unambiguous proof of his concept for drug design, a prototype moment, as momentous in its way as Schrieber’s triumph with calcineurin. “It wasn’t the obvious, next medicinal chemistry thing to do,” he told the project council, uncharacteristically understating the achievement. Vertex had deliberately made a more biologically active molecule. It had predicted, on the basis of structural information, an incremental gain in activity, thereby designing a better compound. For an instant, it had turned on a switch illuminating the mystery of molecular binding and introduced choice into a process that for fifty years had swung between random sampling and brute force, between frustration and luck.

Fifty years earlier, Tishler and the other pioneers of scientific drug discovery launched the process of screening systematically for new drugs that placed the drug industry in its modern-day arc. They had found extraordinary drugs in epochal quantities and had learned to reassemble their atoms to make them even better. But they hadn’t been able to improve them at will. Rationalizing the process of discovery, making it fabulously productive and profitable, they’d fallen short of the ultimate prize: to control it, to drive it ahead on the gales of their own brilliant imagining.

That had always been Boger’s goal, the hard, irreducible thing inside that impelled him. He had come to discovering drugs not, as Tishler had, as a moral act, but as an intellectual one, and now it was in its first, primitive fruition at Vertex. He was there by the standard that measured most, his own intellectual satisfaction. He was controlling the process, its architect, its avatar. Boger had hoped that Tishler had understood this, but it seems equally likely that he died feeling hurt by Boger’s defection from Merck and believing that it was meant to be so.

By the norms of science, Vertex’s new information loop was an invisible conquest. It wasn’t publishable. There would be no talks about it. One chemist’s deliberations in choosing to make certain molecules over others was far too slender a data point for making any public assertions on structure-based drug design or anything else.

And yet ironically, for he had gone to great strains to make mountains out of far less, Boger seemed not to care. Schreiber was right about him: Boger’s fundamental connection to the outside world was different from his, different from most scientists’. They thrived on recognition for the sake of their careers if nothing else. Boger, too, needed attention, but he needed the idea of himself more, of his freedom, of being able to accomplish what he set out to on his own because he had figured out what to do and how to do it. And what he had set himself to do was to perfect a system of using proprietary information in the advantageous design of new drugs. Better molecules, he was absolutely certain, would follow. Someday, the drug industry’s pipelines would be choked with them, and the people who made them first would, one hoped, go into teaching, as Boger often said he still wanted someday to do.

His competition with Schreiber again had come down to a siblinglike match over parity. Schreiber had made himself a beacon of synthetic chemistry in biology to illuminate its secrets. He was a discoverer. Boger aimed at the next rung: to control biological activity, to impose himself on it in order to change it. This explained their life choices, the speed at which each flung himself ahead and the difference in their respective views. Science exalts conceptual breakthroughs; business, practical ones. Thus although Schreiber appeared at times, such as in Pittsburgh, miles ahead of Boger, they were still straining neck-and-neck for the larger prize: to make a lasting mark on chemistry, science, the world. Each one’s work informed, completed, made whole, and was mutually essential to the other’s.

The scientists, understandably, regarded the situation more ambiguously. Five-sixty-three was not a drug. It had undergone none of the critical tests that make a molecule an approved therapeutic. It could be blindingly toxic, cause hallucinations, hypertension, strokes. Vertex’s iteration, its one turn of the crank, was just that: a single isolated success. As the technology was brought forward and improved, it would take perhaps dozens of such increments to demonstrate that the concept worked well enough to be used with any degree of confidence. Even then, it might be useless for reducing the overall side effects associated with FK-506 and cyclosporine, for these side effects now, with Schreiber’s and Starzl’s work, appeared more than ever to be linked inextricably with the drug’s action rather than with extraneous atoms that could simply be cut away and discarded. There were enormous problems, chiefly with certain kinds of toxicity, that lay forever beyond the promised land of structure data.

Nor was 563 a strict display of the lock-and-key, we-like-to-be-sure-of-the-biology paradigm long promoted by Boger. The scientists still had no sure understanding of the biological events they were trying to control. Was calcineurin the ultimate target? If so, how did the complexes bind to it? What of the other FKBPs? Murcko had built his better molecule by examining keys only. Of the tumblers he was trying to hit, he knew depressingly little, less than that. He knew nothing at all.

Murcko, curmudgeonly as ever, reproved Boger’s euphoria. “I think all the arguments that we make, Abbott can make the same,” he said, referring to the large drugmaker known for having some of the drug industry’s best biophysicists. “On what basis can we claim to be uniquely qualified to be doing rational drug design?”

The question lingered like an unresolved chord throughout the fall as the scientists shook off the last effects of Schreiber’s victory and dove, one by one, back into their work. It was more than rhetorical. As Boger had warned, the great risk in science wasn’t being wrong, it was overstating one’s data. In truth, it was more than that. It was deciding what to conceive possible in the first place. Scientists like Boger were deeply, immutably contradictory by nature. They were shamans. Either they were right or they weren’t. Eventually, with data, they would know. Everyone would. Choosing to believe them in the meantime could be like stepping off the edge of the world.

Could Vertex design drugs? Could better drug molecules be designed? Boger was as certain as ever.

Who else thought so failed to interest him.

“I know,” he smiled impetuously, serenely. “Everyone who matters already knows.”