Lac operon: Two gems from the early days (1|2)

by Christoph

When I re­cently wrote about noise in the base­ment of the lac operon, I stum­bled across two pa­pers in said base­ment, both of which I – and not only I – con­sider gems of early mol­e­c­u­lar bi­ol­ogy: Iso­la­tion of the lac re­pres­sor (Gilbert & Müller-Hill, 1966), and The nu­cleotide se­quence of the lac op­er­a­tor (Gilbert & Maxam, 1973). Dis­clo­sure: Both pa­pers were sub­jects of my gra­du­a­tion exam (bio­chem­istry) in 1984, which prob­a­bly ex­plains my long-stand­ing fond­ness for them. Here in part 1, I will dust off the lac re­pres­sor story a bit to make this gem sparkle again...

To give you an idea of the vibes in which the hunt for the lac re­pres­sor be­gan 65 years ago, I quote be­low from the Pa­JaMo pa­per by Pardee, Ja­cob & Monod (1959), which Jon Beck­with said was "a flash of clar­ity in the murky sea of vaguely out­lined the­o­ries that had been pro­posed up to this time" for ge­netic reg­u­la­tion. You've been warned, but if you're cu­ri­ous to nav­i­gate this 'murky sea...,' I rec­om­mend as com­pass Who discov­ered mes­sen­ger RNA? by Matthew Cobb (2015) (Open Ac­cess). Now the Pa­JaMo quote: 

(5) … Al­though the wide oc­cur­rence of re­pres­sion ef­fects is cer­tain, the sit­u­a­tion re­vealed with the present sys­tem, namely a ge­netic "com­plex" com­pris­ing, be­sides the "struc­tural" genes (z, y) a re­pres­sor-mak­ing gene (i) whose func­tion is to block or re­gulate the expres­sion of the neigh­bor­ing genes is, so far, unique for en­zyme sys­tems.
...
(6) Im­plicit in the re­pres­sor model are two crit­i­cal ques­tions, ... These ques­tions are:
  (a) What is the chem­i­cal na­ture of the re­pres­sor? Should it be con­sid­ered a pri­mary or a sec­ondary prod­uct of the gene?
  (b) Does the re­pres­sor act at the level of the gene it­self, or at the level of the cyto­plasmic gene-prod­uct (en­zyme-form­ing sys­tem)?

FIG. 4. The bind­ing con­stants of the wild-type and i ^t mu­tant strains. Sam­ples iso­lated from iso­genic wild-type and i ^t diploid strains were di­a­lyzed against many dif­ferent con­cen­tra­tions of ¹⁴C IPTG as de­scribed un­der Fig. 2. The amount of IPTG bound at each ex­ter­nal IPTG concen­tra­tion, nor­mal­ized to a pro­tein con­cen­tra­tion of 10 mg/ml, is plot­ted against the ra­tio of bound to free IPTG. Sin­gle-side bind­ing should obey the re­la­tion: (Amount bound) = (num­ber of sites) – Km (bound/free). Source. Fron­tispiece: Model of lac re­pres­sor tetramer bound to two lacO op­er­a­tors, loop­ing-out of in­ter­ven­ing DNA. Source

(Ad­dress­ing the ques­tions in re­verse or­der) Ques­tion (6b) may seem a bit far-fetched from to­day's per­spec­tive – we now know that the re­pres­sor (LacI) binds to the gene it­self, that is, to the op­er­a­tor (lacO), see part 2 – but based on their ge­netic ex­per­i­ments Pardee et al. (1959) could not rule out the pos­sibility that the re­pres­sor binds to and blocks the 'cyto­plasmic gene-prod­ucts', β‑galacto­sidase (LacZ) and per­me­ase (LacY) (or binds to and blocks George E. Pa­lade's micro­somes (ri­bo­somes), the 'en­zyme-form­ing sys­tem', known since 1955). To raise the ques­tion of the chem­i­cal na­ture of the re­pres­sor (6a) was ob­vi­ous, but it was in­trigu­ing to ask whether it was the pri­mary or sec­ondary gene prod­uct of the repres­sor gene, lacI. The 'pri­ma­ry gene prod­uct' would be an RNA tran­script, An Un­sta­ble Inter­mediate Car­ry­ing In­for­ma­tion from Genes to Ri­bo­somes for Pro­tein Syn­the­sis, whose ex­istence was first demon­strated by Bren­ner et al. (1961) and that we now know as mes­sen­ger RNA (mRNA). The 'sec­ondary gene prod­uct' would be the trans­la­tion prod­uct of the pri­mary tran­script, a pro­tein.

In fact, Ja­cob & Monod (1961) fa­vored in their land­mark pa­per Ge­netic reg­u­la­tory mecha­nisms in the syn­the­sis of pro­teins a model in which the lac re­pres­sor was the pri­mary tran­script of the lacI gene, the mRNA, which, how­ever, could not be sub­stan­ti­ated. Mean­while, Gilbert & Müller-Hill had other ideas...

Iso­la­tion of the lac re­pres­sor

Gilbert & Müller-Hill felt they were on the right track with their as­sump­tion that the lac re­pres­sor is a 'secon­dary gene prod­uct' of the lacI gene, a pro­tein, since they (and oth­ers) had suc­ceeded in iso­lat­ing Sup­pressible reg­u­la­tor con­sti­tu­tive mu­tants of the lac­tose sys­tem in Es­cherichia coli (that is, non­sense mu­ta­tions in the lacI gene with a lacI ^c phe­no­type that can be par­tially sup­pressed by su­pres­sor tRNAs). They set about de­sign­ing an as­say that "de­tects and quan­ti­tates this re­pressor by mea­sur­ing its bind­ing to an in­ducer, as seen in this case by equi­librium dial­y­sis against ra­dioac­tive IPTG (iso­propyl-thio-galac­to­side)." And they con­tinue: "we de­sired some means that would not de­pend on the spe­cific mod­els that might be imag­ined for the ac­tual mech­a­nism of re­pres­sion. The min­i­mal as­sump­tion on which the as­say is based is that there should be some in­ter­ac­tion be­tween the re­pres­sor and the in­ducer."

FIG. 5. Sed­i­men­ta­tion of the re­pres­sor. Sam­ples of 0.2 ml of a con­cen­trated re­pres­sor frac­tion were lay­ered on 5‑ml gra­di­ents, from 5 to 30% glyc­erol, in TMS buffer. The gra­di­ents were spun for 8.5 hr at 50,000 rpm at 40°C. In one tube, cata­lase and β‑galactosidase were run as sed­i­men­ta­tion mark­ers. For the ex­per­i­ment shown in (a) the re­pres­sor frac­tion was brought to 2.5×10^–7 M ¹⁴C IPTG and lay­ered on a gra­di­ent contain­ing 1.2×10–⁷ M ¹⁴C IPTG. Af­ter the cen­trifu­ga­tion, two-drop sam­ples were col­lected di­rectly into scin­til­la­tion vials and counted with Bray's so­lu­tion. The back­ground level in the gra­di­ent is about 1000 cpm; the counts are plot­ted as a per cent of the back­ground level (open cir­cles). For the ex­per­i­ment shown in (b), a par­al­lel gra­di­ent was run with­out IPTG. The op­ti­cal den­sity was mea­sured on each two-drop sam­ple at 260 mμ (=nm) (—-) and 280 mμ (=nm) (——). Three-tube sam­ples were pooled and placed in dial­y­sis sacs. Their vol­umes were re­duced by bury­ing the sacs un­der dry G 200 Sephadlex for 1 hr. Then the sam­ples were di­a­lyzed against ¹⁴C IPTG. The ex­cess bound, cor­rected to 0.100 ml fi­nal vol­ume in­side the dial­y­sis sac, is plot­ted as a bar. super­imposed on both ex­per­i­men­tal curves is the po­si­tion of the mark­ers taken from a par­al­lel gra­di­ent. Source

Con­cep­tu­ally, their ap­proach was sim­i­lar to the clas­si­cal method for iso­lat­ing en­zymes: grad­u­ally re­fined frac­tion­a­tion of to­tal cel­lu­lar pro­tein and in vitro test­ing of the in­di­vid­ual frac­tions by an en­zyme as­say. Since they had a few clues that IPTG is not me­tab­o­lized in the cells – such as lac­tose, which is con­verted to al­lo­lac­tose – but acts di­rectly as an in­ducer, this ap­proach seemed straight­forward (¹⁴C‑labelled IPTG had just be­come com­mer­cially avail­able then).

Ini­tial at­tempts to tag the lac re­pres­sor by affin­ity la­bel­ing with IPTG failed (as Müller-Hill later re­called). In ret­ro­spect, this was ac­tu­ally ad­van­ta­geous, as it brought them to use equi­lib­rium dial­y­sis as a de­tec­tion method for the re­pres­sor (much like an en­zyme as­say, which would have been im­pos­si­ble with affin­ity la­bel­ing). The prob­lem was ob­tain­ing suf­ficient quan­ti­ties of lac re­pres­sor! They es­ti­mated from Ki­netic stud­ies of beta-galac­tosi­dase in­duc­tion that "the wild­type re­pres­sor should have an affin­ity of the or­der of 6×10^–6 M for IPTG, and if any bind­ing is to be de­tectable [by equi­lib­rium dial­y­sis], the re­pres­sor concen­tration must ex­ceed a few times 10^–6 M." They "would have to frac­tion­ate and con­cen­trate the re­pres­sor by at least a fac­tor of 100, work­ing blindly, be­fore de­tect­ing any ef­fect." They over­came this hur­dle in three steps: first, they iso­lated a lacI mu­tant (i ^t) with a ~ten-fold higher Km for IPTG (see FIG. 4.), sec­ond, they made a diploid de­riv­a­tive of the i ^t mu­tant strain (that is, the i ^t al­lele on a F' plas­mid in ad­di­tion to the chro­mo­so­mal copy) to have twice the re­pres­sor con­cen­tra­tion at the start, and third, they frac­tion­ated late‑log cell ex­tracts with sper­mine and am­mo­nium sul­fate to have more lac re­pres­sor per mg pro­tein. Thus they ob­tained "ma­te­r­ial that will draw the in­ducer into a dial­y­sis sac to a con­cen­tra­tion 1.5–2 times that out­side thes ac (50–100% ex­cess bind­ing) at a pro­tein con­cen­tra­tion of 10 mg/ml." To this ma­te­r­ial, "The la­beled IPTG is bound re­versibly; it can be freely di­a­lyzed into, out of, and even into the sac again. This is il­lus­trated in Ta­ble 1."

Among the nu­mer­ous neg­a­tive and pos­i­tive con­trols – all of which gave the ex­pected re­sults – one stands out as, yes... el­e­gant: An iden­ti­cal par­allel pu­rifi­ca­tion was run on 50-gm lots of diploid i ^t con­trol cells and diploid (am­ber-sup­pres­sor-sen­si­tive) i ^– cells that ex­press a trun­cated lac re­pressor (see above). At the end of the pu­rifi­ca­tion, sam­ples of each pro­tein so­lution were di­a­lyzed against sev­eral dif­fer­ent IPTG con­cen­tra­tions: no bind­ing was ob­served for ma­te­r­ial from the i ^– cells.

Gilbert & Müller-Hill de­ter­mined a mol­e­c­u­lar weight of 150–200 kDa for their prepa­ra­tion of the lac re­pres­sor ob­tained by glyc­erol gra­di­ent cen­trifu­ga­tion. From the yield of lac repres­sor iso­lated from the to­tal amount of processed pro­tein, they cal­cu­lated that the cells con­tained no more than about ten re­pres­sor mol­e­cules per lacI gene copy. They phrase the ac­tual Aus­sagekraft of their ex­per­i­ments care­fully and list the prop­er­ties of the lac re­pres­sor:

The abil­ity to bind IPTG is not at­tacked by RNase or DNase. It is de­stroyed by pronase. The bind­ing site can be in­ac­ti­vated by tem­peratures above 50°C. These are prop­er­ties of the part of the mol­e­cule that in­ter­acts with the in­ducer. As yet there is no way of see­ing the rest of the mol­e­cule.

This is the iso­la­tion of the lac re­pres­sor, no ques­tion, but not its pu­rifi­ca­tion ac­cord­ing to the high stan­dards of the pro­tein spe­cial­ists (see Ko­rn­berg (2009) Why pu­rify en­zymes?). The im­pu­­rities in their prepa­ra­tion men­tioned in the cap­tion to Fig. 5. would have been ap­par­ent in Coomassie-stained SDS-PAGE, but this tech­nique was not yet lab stan­dard at the time (popu­larized by Laemmli in 1970). Above all, they would have seen in de­na­tur­ing SDS-PAGE that their (na­tive) 150–200 kDa lac re­pres­sor is in fact a tetramer, which be­came ap­par­ent when Beyreuther et al. (1973) de­ter­mined The amino-acid se­quence of lac repres­sor: 347 amino acids (38.5 kDa). Fried­man et al. (1995) de­ter­mined the Crys­tal Struc­ture of lac Re­pres­sor Core Tetramer and Its Im­pli­ca­tions for DNA Loop­ing (Fig­ure: PDB-Model, Fron­tispiece). The Struc­tural Analy­sis of Lac Re­pres­sor Bound to Al­losteric Ef­fec­tors by Da­ber et al. (2009) re­vealed that bind­ing of IPTG in­duces a con­forma­tional change in the re­gion con­necting the N- and C‑terminal do­mains to the pro­tein, thereby pre­venting ef­fi­cient bind­ing of the dimer to op­er­a­tor DNA (see here a di­a­gram).

Fig­ure: PDB‑Model of a the lac re­pres­sor tetramer (with my added de­scrip­tions)

I hope you agree that it was worth dust­ing off this gem. With­out fur­ther ado, let me fin­ish by quot­ing two sen­tences from an es­say by Benno Mül­ler-Hill (1990) ti­tled, you guessed it, The Iso­la­tion of the lac Re­pres­sor: "One day [in au­tumn 1966], late in the af­ter­noon, I saw Mo­nod stand­ing in the door. 'Hello' he said. 'Hello' I an­swered. And then he sim­ply said 'Benno, af­ter all it was pedes­trian', bowed and dis­ap­peared, leav­ing me baf­fled for a long time. Now, many years later, I like to in­ter­pret his com­ment as 'slow, down to earth, with only a few tech­ni­cal crutches' and think it was, af­ter all, a com­pli­ment."