Professor Emeritus of Pharmacology, Stanford University
735 Dolores, Stanford, CA 94305-8427
650-324-9251 (FAX 650-328-841 2)
e-mail: avram.goldstein@stanford.edu

[For mailing: Use street address only, not department or university]

November 7, 1998

Joshua:

Ali | can tell you about my 1952 letter to you -- 46 years ago!!! -- is that it obviously
refers to the mutator strain of E. Coli that | eventually reported in J Bacteriol
70:588,1955. It has been a long journey from bacteria to neuropeptides and drug
addiction!

About my views on medical school teaching, which were implemented in the Stanford
curriculum:

J Med Education 31:365(1956); 33:193(1958) and especially J Med Education
36:686(1961) ("The basic medical sciences in the Stanford Plan’).

You ask when | met you and Esther prior to that 1952 letter. Since | was working on
bacteria and antibiotic resistance then, maybe a Gordon Conference? | don’t recall it.

Finally, yes, Alex is dead serious about his foundation, which has been transformed into
an actual Institute for the study of the genetics and neurobiology of addiction.

By the way, the article from PNAS that you sent him (by Bond et al, including Mary
Jeanne and others) has an extremely misleading title, and in a strange way mixes two
SNPs that they find. The most common SNP (A118G, which produces N40D) affects a
glycosylation site on the N-terminal extracellular tail of the mu opioid receptor. It
increases the affinity and potency of beta-endorphin by a factor of about 3. But it shows
no difference in prevalence between heroin addicts and non-addicts.

On the other hand, C17T, which produces A6V in the same N-terminal domain, does
differ between heroin addicts and non-addicts, but the statistical significance, although
P<0.05 (I compute P=0.036 by Fisher’s exact method) must still be considered
marginal. In any case, all the data on binding and signal transduction concern the other
SNP, not this one. Moreover, the actual discovery of these two SNPs is not original; they
were reported previously by Berreitini et al (ref 12) and also by Bergen et al (ref 13).

In summary, the work is good but prematurely published inasmuch as the SNP with a
physiologic consequence (slightly tighter binding and slightly greater potency of
beta-endorphin) does not have a different frequency in heroin addicts, whereas the
paper tells us nothing about binding or signal transduction for the SNP that is more
common among heroin addicts.

With best wishes,

an