mRNA in E. coli : Not Lost With­out Trans­la­tion

Nanos gi­gan­tum humeris in­si­d­entes
—Bernard of Chartre, XI­Ith Cen­tury

by Roberto

Di­rectly trans­lat­ing from the Latin, the quoted phrase reads "dwarfs ride on the shoul­ders of gi­ants." This metaphor and its many de­riv­a­tives have been used over the cen­turies to mean that new knowl­edge is dis­cov­ered by build­ing on pre­vi­ous dis­cov­er­ies. Most of the time, that is the per­ception we have when we think of progress in sci­ence. But every now and then, a new find­ing seems to come out of nowhere, greatly chal­leng­ing prior ideas. Such might be the case on how we came to gain new knowl­edge re­gard­ing the struc­tural or­ga­ni­za­tion of the ba­sic processes of tran­scription and trans­la­tion in Es­cherichia coli.

Fig­ure 1.A por­tion of an E. coli chrom­osome be­ing tran­scribed (left to right) and be­ing simulta­neously trans­lated. The ar­row points to the pu­tative site where RNA poly­merase is first bound to the DNA. The chains of dark bod­ies are poly­somes, that is, sev­eral ri­bo­somes on the same mRNA mol­e­cule. Note that the size of the poly­somes in­creases with the dis­tance from the up­stream site where tran­scrip­tion be­gan. Source

Let me start with a quick re­count of old knowl­edge. For sev­eral decades – go­ing back to the early days of mo­le­cular bi­ol­ogy – the ac­cepted view was that tran­scrip­tion and trans­la­tion in E. coli were cou­pled, mean­ing that as soon as an RNA poly­merase be­gan to pro­duce a nascent mRNA, ri­bo­somes bind it and ini­ti­ate trans­la­tion. Per­haps the most strik­ing early ev­i­dence in sup­port of this coup­ling was pro­vided by the elec­tron mi­cro­graphs that Miller and col­leagues pub­lished in 1970. These ima­ges are re­markable and seemed to with­stand the test of time. So much so that Elio chose one of them for the first "Pic­tures Con­sid­ered" post of STC in 2013. I find the ima­ge so beau­ti­ful in its sim­plic­ity and power that I show it here once again (Fig­ure 1). In­di­rect ev­i­dence for cou­pling had ac­cu­mu­lated through­out the 1960s (in­clud­ing the first de­scrip­tion of polysomes by Elio) but this pic­ture came to ce­ment in people's minds the gen­eral con­cept. It all made per­fectly sense. E. coli be­cause it had no nu­clear mem­brane – like all Bac­te­ria and Ar­chaea – made this cou­pling not only pos­si­ble but al­most un­avoid­able; it was an easy con­cept to ac­cept. The cou­pling came to be seen as a defi­ning fea­ture that dis­tin­guished mem­bers of the do­mains Bac­te­ria and Ar­chaes from the Eu­karya, as very nicely de­scribed in 2014 by Nanne Nan­ninga in an­other STC post. Tran­scrip­tion-trans­la­tion cou­pling was also deemed most appro­priate to achieve fast growth rates and it al­lowed for the evo­lu­tion of mech­a­nisms whereby tran­scription was reg­u­lated by the state of the ri­bo­somes fol­low­ing close by.

Fig­ure 2. Mech­a­nism of tran­scrip­tional at­te­n­u­ation of the trp operon. Source

Many fea­tures of E. coli gene ex­pres­sion stud­ied ex­ten­sively in the 1970s fit per­fectly in the con­text of cou­pled tran­scrip­tion and trans­la­tion. Let me briefly re­mind you of the de­tails of my two per­sonal fa­vorites: at­ten­u­a­tion and po­lar­ity. In at­ten­u­a­tion, be­cause of the cou­pling, the RNA poly­merase gets in­for­ma­tion re­gard­ing the state of the trans­lat­ing ri­bo­some and ei­ther ter­mi­nates tran­scrip­tion or con­tin­ues to tran­scribe genes down­stream (Fig­ure 2). This is be­cause a ri­bo­some loads onto the nascent mRNA and ini­ti­ates trans­la­tion of a leader pep­tide so that it closely trails the RNA poly­merase. In the case of atten­uation con­trol of amino acid biosyn­thetic oper­ons, the leader pep­tide gene is rich in codons for the cog­nate amino acid. When there are plen­ti­ful tR­NAs charged with that amino acid, trans­la­tion of the leader pep­tide pro­ceeds with the ri­bo­some reach­ing the ter­mi­na­tion codon. By do­ing so, the ri­bo­some masks a por­tion of the mRNA and the sub­se­quent bases of the mRNA fold into a stem loop struc­ture that tells the RNA polyme­rase to ter­mi­nate tran­scrip­tion. How­ever, if the amount of charged tR­NAs for the cog­nate amino acid is low, the ri­bo­some stalls at the cor­re­spond­ing codons. This re­sults in an al­ter­na­tive mRNA struc­ture, an an­titer­mi­na­tor, that pre­cludes the for­ma­tion of the ter­mi­na­tor. As a con­se­quence, the RNA poly­merase con­tin­ues tran­scrib­ing and thus the down­stream genes are ex­pressed. It is clear that at­ten­u­a­tion ab­solutely de­pends on the cou­pling of tran­scrip­tion and trans­la­tion.

Tran­scrip­tional po­lar­ity refers to the phe­nom­e­non where non-sense mu­ta­tions (where trans­la­tion termi­nation codons re­placea sense codon) in genes lo­cated in the pro­moter prox­i­mal end of ope­rons cause de­creased ex­pres­sion of genes lo­cated dis­tal to the pro­moter. The mol­e­c­u­lar mecha­nism of po­lar­ity is con­cep­tu­ally sim­i­lar to at­ten­u­a­tion. In wild-type oper­ons, trans­lat­ing ri­bo­somes fol­low the RNA poly­merase very closely and thus naked mRNA is not ex­posed. But, when a non-sense mu­tation causes the early re­lease of the trans­lat­ing ri­bo­some, the mRNA is ex­posed and the ter­mi­nation fac­tor Rho can bind at nor­mally oc­cluded sites, re­sult­ing in early tran­scrip­tion ter­mi­nation. Here again, it is clear that this in­ter­pre­ta­tion of po­lar­ity de­pends on the cou­pling of tran­scription and trans­la­tion.

Fig­ure 3. Sub­cel­lu­lar lo­cal­iza­tion of mRNA tran­scripts cor­re­lates with sub­se­quent loc­alization of their pro­tein prod­ucts. Fluor­escence micro­scopy im­ages of cells ex­pressing MS2-GFP (gre­en) and tran­scripts con­taining or lack­ing bind­ing sites for MS2-GFP. "6×bs" af­ter a gene's na­me de­notes that the MS2 bind­ing sites were in­troduced im­me­di­ately next to this gene: a cat and b lacY. Adapted from Source

So much for old knowl­edge, now comes the sur­prise. It's Feb­ru­ary 2011 and a pa­per, led by Keren Nevo-Dinur from the group of Orna Am­ster-Choder (He­brew Univer­sity of Jerusalem), bears the un­ex­pected ti­tle "Transla­­tion-In­­de­pen­dent Lo­cal­iza­tion of mRNA in E. coli." The ab­stract goes on to state: "In con­trast to the view that tran­scrip­tion and trans­la­tion are cou­pled in bac­te­ria, our re­sults show that, sub­se­quent to their syn­the­sis, cer­tain mR­NAs are ca­pa­ble of mi­grat­ing to par­tic­u­lar do­mains in the cell where their fu­ture pro­tein prod­ucts are requi­red." What had they done to al­low such a bold claim? Their stud­ies were both el­e­gant and tho­rough. At the heart of their ap­proach was the clever tag­ging of the mRNA of two genes whose fi­nal pro­tein pro­ducts had two dif­fer­ent cel­lu­lar ad­dresses: lacY, en­cod­ing the lac­tose per­me­ase – an in­te­gral mem­brane pro­tein – and cat, en­cod­ing the sol­u­ble (that is, cy­to­plas­mic) chlo­ram­pheni­col acetyl trans­ferase. How were the mR­NAs tagged? Us­ing a cute trick. Six re­peats of a se­quence which, when tran­scribed into RNA, be­comes bound by the MS2 phage coat pro­tein are first cloned into the gene of in­ter­est. Then, a fu­sion of the MS2 phage coat pro­tein and the green flu­o­res­cence pro­tein (MS2-GFP) is ex­pressed in the cell. Vi­su­al­iz­ing MS2-GFP flu­o­res­cence thus iden­ti­fies the lo­ca­tion of the mRNA. The re­sults were strik­ing: the lacY mRNA lo­cal­ized to the mem­brane all around the cell pe­riph­ery while the cat mRNA lo­cal­ized in the cy­to­plasm in a he­lix-like pat­tern (Fig­ure 3). The au­thors car­ried out all the re­quired and ex­pected con­trols to en­sure that these re­sults were not ar­ti­fac­tual (in­clud­ing fol­low­ing the fate of other mR­NAs, for ex­am­ple, one en­cod­ing the mem­brane-bound sugar perme­a­se BglF). Then they went a step fur­ther. To de­ter­mine if the mRNA lo­cal­iza­tion was a prop­erty in­her­ent to the RNA and not the pro­tein prod­uct, they un­cou­pled tran­scrip­tion and trans­la­tion through the use of pro­tein syn­the­sis in­hibitors. Sure enough, the spe­cific lo­cal­iza­tion of the diffe­rent mR­NAs was not af­fected by pro­tein syn­the­sis in­hibitors! In ad­di­tion, when they un­cou­pled tran­scription and trans­la­tion by in­tro­duc­ing a stretch of rare leucine codons at the be­gin­ning of the bglF gene (cre­at­ing a "trans­la­tional block­age", they still saw the mRNA lo­cal­iz­ing to the mem­brane. All in all, the au­thors made a very com­pelling case that un­trans­lated mR­NAs can lo­cal­ize to dis­tinct cel­lu­lar ad­dresses in keep­ing with where their pro­tein prod­uct is to func­tion. With this se­ries of ex­per­i­ments they ap­peared to have put a ma­jor dent on a "defin­ing" fea­ture of cells lack­ing a nu­clear mem­brane, the Bac­te­ria and Ar­chaea: the cou­pling of tran­scrip­tion and trans­la­tion.

Fig­ure 4. RNAs in E. coli ex­hibit asym­met­ric dis­tribution on a tran­scrip­tome-wide scale. Source

Re­mem­ber, these re­sults were pub­lished in 2011. Yet, here at STC (and else­where) there was still a strong sense that in Bac­te­ria and Ar­chaea, tran­scrip­tion and transla­tion were cou­pled. Per­haps the slow ac­cep­tance of this new Weltan­schau­ung (world­view) was due both to the strength of the ev­i­dence fa­vor­ing cou­pling and the fact that only a few genes were an­a­lyzed in the 2011 publi­ca­tion. But now, just last week, a pa­per by Kan­na­iah et al., also from the Am­ster-Choder group, presents a global analy­sis of the spa­tiotem­po­ral or­ga­ni­za­tion of the E. coli tran­scrip­tome. To carry out the global analy­sis they de­veloped a tech­nique that frac­tion­ates the cell con­tents and se­quences the mRNA present (they call the ap­proach "Rloc-seq"). The re­sults are im­pres­sive, they de­tect who­le­sale spe­cific lo­cal­iza­tion of mRNA – at least half of it in­de­pen­dent of trans­la­tion – whereby the mRNA pla­ce­ment cor­re­lates with the fi­nal des­ti­na­tion of the pro­tein prod­uct (Fig­ure 4).

In all fair­ness, to some in­ves­ti­ga­tors the lack of cou­pling was not all that un­ex­pected. As far back as 2000, there were re­ports that there was clear seg­re­ga­tion be­tween the ri­bo­somes and the RNA poly­merase within cells. It just seemed that there was no room for most of the ri­bo­somes to be near the nu­cleoid, where the vast ma­jor­ity of the RNA poly­merase could be found.

Do these re­sults com­pletely over­turn the prior no­tion of cou­pled tran­scrip­tion and trans­la­tion in cells with­out a nu­cleus? Per­son­ally, I do not think so. Per­haps that is due to the fact that in this ag­ing mind I carry too much in­tel­lec­tual bag­gage. Still, I do feel that the ev­i­dence for cou­pling is ex­tremely strong; I hope I ar­gued that case strongly enough at the out­set. I am ex­tremely impres­sed by the pa­pers pre­sent­ing the re­sults of mRNA lo­cal­iza­tion. But I was sur­prised that nei­ther of these two key pa­pers of­fers a dis­cus­sion of these new re­sults in light of what is known about such processes as at­ten­u­a­tion and po­lar­ity, which ob­vi­ously de­pend on close cou­pling be­tween tran­scription and trans­la­tion. In my own think­ing, I look at that com­pelling elec­tron mi­cro­graph of Miller et al. (Fig­ure 1) and see the scale bar. Even a sin­gle gene tran­script is likely to be around 0.5 mi­crons in length. That's pretty long to have the RNA poly­merase still tran­scrib­ing while the 5' end of the mRNA, loaded with ri­bo­somes, can be reach­ing a dis­tant cel­lu­lar ad­dress. I think that spe­cific mRNA lo­cal­iza­tion and tran­scrip­tion-trans­la­tion cou­pling are not mu­tu­ally ex­clu­sive. But I would like to hear from the ex­perts in the field. I sense that even though these re­sults ap­pear to shat­ter prior claims, the new knowl­edge still stems – at least in part – from "rid­ing on the shoul­ders of gi­ants."