Lasker Awards

Stories

An Amazing Turn of Events

Michael N. Hall

In 1965, a group of scientists from Montreal arrived on Easter Island, also known as Rapa Nui, to
collect soil samples. This was the inauspicious beginning of an extraordinary and unpredictable
series of events that make up a wonderful biomedical story. Distinct achievements and discov-
eries that shaped this story, in which I had the good fortune to be a protagonist, were recognized
with a Lasker Award in 2012 and now in 2017.
‘‘I liked the name TOR
The purpose of the scientists’ journey to a remote island in the South Pacific was
to prospect for exotic microbes that might produce novel metabolites that could
because it was easy to
be developed into drugs—in this case, antifungal drugs. This is how drugs were say and remember. Joe
developed in those days. In 1975, they did indeed isolate such a metabolite, which
they named rapamycin, after Rapa Nui. However, while developing rapamycin as
had a more romantic
an antifungal, they found that it had the undesirable side effect of suppressing the reason for favoring it.’’
immune system. Rapamycin was therefore abandoned and largely forgotten until
it caught the attention of two bold transplant surgeons. In the 1980s and 1990s,
Roy Calne and Thomas E. Starzl exploited the formerly undesirable immunosuppressive prop-
erties of rapamycin and two other natural compounds (cyclosporin A and FK506) to advance
transplantation from an experimental procedure to an accepted treatment for organ failure. For
establishing the clinical utility of cyclosporin A, FK506, and rapamycin as immunosuppressive
drugs, Calne and Starzl shared the 2012 Lasker Clinical Medical Research Award. The citation
reads, ‘‘For the development of liver transplantation, which has restored normal life to thousands
of patients with end-stage liver disease.’’

I had the pleasure of seeing Tom Starzl again a couple years ago, after a long hiatus, and we were
soon reminiscing about the old days. At 89 years of age, his enthusiasm for science was undi-
minished. Shortly thereafter, I received a letter from Tom with reprints of some of his rapamycin
papers from the early ‘90s. He ended the letter by writing, ‘‘Have these 2 papers passed or flunked
the test of time?’’ Needless-to-say, they had passed.

Calne and Starzl’s achievements were heroic and profoundly important, but the underlying
mode(s) of action of the three drugs was unknown other than that they inhibited signaling in T cells.
In 1989, Joe Heitman, an outstanding postdoc, joined my laboratory at the Biozentrum of the
University of Basel after his M.D. and Ph.D. studies at Cornell and The Rockefeller University.
Given the medical breakthrough made possible by the new immunosuppressive drugs, and thus
the heightened sense of excitement surrounding them, he decided to study the drugs’ molecular
mechanism of action. This was a departure from our ongoing research, not least because we were
primarily a yeast genetics laboratory, and Joe wanted to study drugs used, or being developed for
use, in humans. Joe and Rao Movva, a collaborator and group leader at the local pharmaceutical
company Sandoz (now Novartis), had the brilliant albeit risky idea to use yeast genetics to study
the drugs. They were, of course, making the assumption that the drugs’ molecular targets were
conserved from yeast to human. I could justify this departure from our normal research by viewing
the compounds as probes to identify novel signaling pathways. Nevertheless, it seemed to many
that our experiments were physiologically irrelevant—giving human drugs to yeast cells?!
Colleagues were rolling their eyes, and rightfully so.

Joe’s initial experiments were with cyclosporin A and FK506. Although these two drugs had little to
no effect on growth of our laboratory yeast strain, Joe characterized the yeast FK506-binding
protein (FKBP) encoded by the gene he named FPR1. Rao later gave us some rapamycin, which
Joe found arrested yeast cells in the G1 phase of the cell cycle—the same effect the drug had on
T cells! This gave us confidence that our risky idea might have merit. Moreover, Joe found that
FK506 antagonized rapamycin toxicity (structurally similar, FK506 and rapamycin competitively
bind FKBP) and that strains in which he had deleted the FPR1 gene were fully resistant to
rapamycin. This suggested that an FKBP-rapamycin complex was the active, toxic agent in the
cell, rather than rapamycin alone. To identify the target of the FKBP-rapamycin complex, Joe
Michael N. Hall in his lab

Cell 171, September 21, 2017 ª 2017 Elsevier Inc. 3

 Lasker Awards
Stories

(L to R) Jeannette Kunz, early 1990s. Joe Heitman, 1990. Robbie Loewith, early 2000s

isolated several spontaneous rapamycin-resistant mutants. At this stage, Joe was furiously
working around the clock to finish the characterization of his mutants before returning to New York
City to complete clinical rotations. By early December 1990, he knew that the mutants were
defective in the FPR1 gene, as expected based on his earlier experiments, or that they harbored a
dominant mutation in at least one other unknown gene. The December 2 entry from Joe’s lab
notebook shows his progress in characterizing two resistant mutants (R1 and R17) defective in the
unknown gene(s). As indicated in his notes, he wondered whether the mutations were in the same
gene or in different genes. Knowing there was at least one new gene (‘‘gene X’’), he began listing
candidate gene names (FAP, FIP, RAT, TFR, TOR, PAF, PIF, RAR) that fit the yeast three-letter-
acronym nomenclature convention. As is now well known, we chose TOR, for target of rapamycin.
I liked the name TOR because it was easy to say and remember. Joe had a more romantic reason
for favoring it. In German, TOR (das Tor) means ‘‘gate,’’ and at the time, we viewed TOR as the
entry to the cell cycle, much like the gates in the wall that formerly encircled Basel were the entry
into the city. Joe was not deterred by the fact that the masculine form of the German noun Tor
(der Tor) means ‘‘fool.’’ To resolve the issue of one versus two new genes, Joe put the diploid from
an R1  R17 cross on sporulation medium (‘‘spo plates’’) to eventually dissect apart the spores
and determine whether the mutations were linked or segregated independently. He managed to
complete this experiment, among many others, before leaving Basel in late December. He showed
that the rapamycin resistance-conferring mutations in R1 and R17 were unlinked and thus in two
distinct genes: TOR1 and TOR2. This was the discovery of TOR, at least as genetic loci in the yeast
genome. We published Joe’s findings in Science on August 23, 1991. Coincidentally, on that same
day, Stuart Schreiber published a paper in Cell, showing that the phosphatase calcineurin is a
common target of cyclosporin A and FK506.
‘‘To the students of
What did the two TOR genes encode? To answer this question, two very talented
students in the lab, Jeannette Kunz and later Stephen Helliwell, cloned and
today, it probably seems
sequenced the TOR genes. This was much more difficult than anticipated because odd that we once
the TOR genes are among the largest in yeast and were thus under-represented
in the size-fractioned genomic libraries we had generated for cloning purposes
thought cell growth is a
(Hall, 2016, Mol. Biol. Cell 27, 2804–2806). The sequencing revealed that the two spontaneous, passively
TORs were novel, highly similar kinases and also that Joe’s rapamycin-resistance-
conferring mutations were missense mutations that, as shown later, prevented
controlled process that
FKBP-rapamycin binding to TOR without otherwise affecting TOR activity. Rapa- just happens when
mycin acts by forming a complex with FKBP/FPR1, which in turn directly binds and
inhibits TOR. The mammalian TOR (mTOR) gene was cloned shortly thereafter by building blocks are
several other groups, including Stuart Schreiber’s, David Sabatini in Solomon available.’’
Snyder’s, and Robert Abraham’s groups, and was shown to encode a kinase
similar to yeast TOR, hence the name mTOR. Schreiber’s laboratory subsequently
demonstrated that a reconstituted mTOR variant containing a missense mutation analogous to one
of our yeast TOR mutations conferred rapamycin resistance in mammalian cells. Thus, the mode of
action of rapamycin was conserved, as presumed by Joe and Rao when they initiated the project.

What is the cellular role of the TOR kinases other than being the target of rapamycin? Answering
this question took several years and led to our most gratifying and possibly most important
discovery (Hall, 2016, Mol. Biol. Cell 27, 2804–2806). We originally thought that the role of TOR

4 Cell 171, September 21, 2017

Lasker Awards
Stories

Section of December 2, 1990, entry from Joe Heitman’s lab notebook, the first appearance of the name TOR.

was to control the cell cycle. This model was based on the cell cycle arrest we observed upon TOR
inhibition. However, during a fateful seminar I gave in Vienna in late 1993 to members of the
Institute of Molecular Pathology, including the cell cycle expert Kim Nasmyth and his group, I was
disabused of this notion. As I learned in Vienna, our kinase did not fit neatly into what was then
known about cell cycle control. This view was reinforced as we began to discover that TOR
controlled processes related to macromolecular synthesis and unrelated to the cell division cycle.
We then began to consider that the true role of TOR was to control cell growth (increase in cell size/
mass), rather than cell division (increase in cell number), in response to nutrients. This was a
particularly important conceptual advance because it cast our hitherto confusing results in a new
light. We suddenly understood our results. It was also a paradigm shift because cell growth was
not thought to be actively controlled. To the students of today, it probably seems odd that we once
thought cell growth is a spontaneous, passively controlled process that just happens when
building blocks are available. Over the subsequent years, we and others showed that TOR acti-
vates several anabolic processes and inhibits catabolic processes, confirming that TOR controls
mass accumulation and thereby cell growth. In 2000, we proposed that TOR is a central controller
of cell growth. The misleading cell cycle arrest we had observed was an indirect consequence of a
cell growth defect.

The next major problem was to elucidate the mechanism by which TOR controls cell growth.
Jeannette Kunz had earlier knocked out the two TOR genes, finding that the two TORs are not
functionally equivalent. More precisely, she had found that loss of TOR1 had no effect on cell
viability, whereas knockout of TOR2 was lethal. When she knocked out both TOR1 and TOR2,
cells were not viable, but the cells displayed a G1 arrest not observed upon loss of TOR2 alone.
This suggested that TOR1 has one function, while TOR2 has two functions, one of which is
redundant with the single TOR1 function. Further analyses, by several excellent students and
postdocs, including Stephen Helliwell, Nic Barbet, Anja Schmidt, Marc Bickle, Thomas Beck,
Estela Jacinto, José Luis Crespo, Tobias Schmelzle, Martin Dietmar, Vittoria Zinzalla, Alexandre
Soulard, and Mitsugu Shimobayashi, revealed that the two TOR ‘‘functions’’ are two distinct
signaling pathways. TOR1 signals via an effector pathway that activates protein synthesis and
other processes. TOR2 can substitute for TOR1 in this pathway, but also has an independent
function, signaling via a different effector pathway to control the actin cytoskeleton. These
observations provided a logical conceptual framework—TOR signals via two pathways to inte-
grate temporal and spatial control of cell growth—but we still did not understand the molecular
basis of this signaling complexity. To investigate the complexity of TOR signaling, Robbie Loewith,
an undaunted cold-tolerant Canadian postdoc, and my long-term technician Wolfgang Oppliger

Cell 171, September 21, 2017 5

 Lasker Awards
Stories

entered the cold room to purify the TOR proteins from yeast. The idea was to develop a purification
scheme that would be sufficiently gentle to allow co-purification and thereby identification of TOR
interacting proteins. This was difficult because we did not know whether TOR even had interacting
proteins and, thus, what the appropriate purification scheme might be. Robbie’s biochemical
characterization, also aided by Stephan Wullschleger, revealed that the TORs are part of two
structurally and functionally distinct complexes, which we named TORC1 and TORC2. Either
TOR1 or TOR2 assembles into TORC1, but only TOR2 assembles into TORC2. Robbie then
showed that the two TORCs correspond to the two previously identified TOR signaling pathways,
thereby providing a molecular basis for the complexity of TOR signaling. Estela Jacinto, Pazit
Polak, Nadine Cybulski, Asami Hagiwara, Aaron Robitaille, Raúl Durán, Marion Cornu, and Verena
Albert in our laboratory and others showed that the two complexes are structurally and func-
tionally conserved in mammals, where they are known as mTORC1 and mTORC2. David Sabatini,
in particular, played a key role in elucidating the mammalian TOR complexes.

Over the course of the last 15 years, it has become apparent that mTOR is implicated in a wide
variety of diseases in addition to allograft rejection, including cancer and diabetes, and in aging.
This, in turn, has spawned a great deal of activity in the pharmaceutical industry to develop new
drugs that target the mTOR signaling network. Such drugs may restore normal life to even more
patients. Thus, due to an amazing turn of events, rapamycin has had impact in medicine,
fundamental biology, and the pharmaceutical industry. It all started with an unknown scientist
bending over to pick up a handful of dirt in a faraway land and time.

6 Cell 171, September 21, 2017