The Façade of E. coli K‑12

by Elio

In the early days of re­com­bi­nant DNA re­search, ca. 1973, a thun­der­ous de­bate took place re­gard­ing the safety of ge­netic en­gi­neer­ing. On the one side were those who pro­posed that mess­ing with the genome could lead to the emer­gence of novel and po­ten­tially per­ni­cious pa­tho­gens. On the other side were those who rea­soned that the re­sult­ing or­gan­isms were un­likely to be more vir­u­lent than those al­ready in ex­is­tence. The his­tory of that era has been re­counted many times. (For links to some ac­counts, click here.) Cen­tral to the ar­gu­ments was the con­vic­tion that the most pop­u­lar or­gan­ism used for cloning, E. coli strain K‑12, was un­ques­tion­ably a non-pathogen. Sure enough, hu­man vol­un­teers who drank cul­tures of K‑12 did not be­come col­o­nized and walked away with no un­to­ward ef­fects. Through­out the argu­ments that en­sued, con­fi­dence in the benev­o­lence of K‑12 re­mained un­bro­ken.

Fig­ure 1. In­ter­nal­iza­tion of E. coli K‑12 in host epi­thelial cells. Left: HU mu­tant strain SK3842. Right: wild type. Cells were in­fected with an MOI of 100 bac­te­ria per cell. Source

This long­stand­ing ar­ti­cle of faith has now been shaken. A re­port from the lab­o­ra­to­ries of S. Kar in New Delhi and S. Ad­hya at the NIH alerts us to the fact that the K‑12 strain is not as be­nign as we thought. It has been known for some time that the abun­dant non­in­va­sive strains of E. coli in our mi­cro­biota carry cryp­tic (i.e., un­ex­pressed) genes for vir­u­lence fac­tors. The news here is that these genes are ac­ti­vated by a mu­ta­tion in a pro­tein called HU, a small, ba­sic, DNA-bind­ing pro­tein. HU is the most abun­dant of the so-called nu­cleoid-as­so­ci­ated hi­s­tone-like pro­teins of E. coli that are in­volved in main­tain­ing the struc­ture of the nu­cleoid. By con­trol­ling DNA su­per­he­lic­ity, HU reg­u­lates many processes, in­clud­ing gene ex­pres­sion, DNA re­pair, and the ini­ti­a­tion of chro­mo­some repli­ca­tion. There­fore, mu­ta­tions in HU are, of ne­ces­sity, pleiotropic, af­fect­ing many cel­lu­lar func­tions. In E. coli, these mu­tants usu­ally have a mild phe­no­type due to the pres­ence of other pro­teins with anal­o­gous prop­er­ties. Al­though many HU mu­tants grow at rates com­pa­ra­ble to the wild type, some show de­fects in cell di­vi­sion and re­sponse to tem­per­a­ture or UV shocks.

Fig­ure 2. At­tach­ment and en­try process. In­testi­nal cells (Int-407) in­fected with the HU mu­tant, E. coli SK3842, show­ing dif­fer­ent stages dur­ing in­va­sion. (i) Small mem­brane pro­jec­tions (ar­row) on the host cell sur­face near the point of bac­te­r­ial at­tachment. (ii) Mem- brane pro­tru­sions ex­tending from ei­ther side of the an­chored bac- teria. (iii) Mem­brane pro­trusions fuse to en­gulf the at­tached bac­te­ria in a mem­brane-bound vac­uole. Source

Now, how­ever, those re­searchers have found a mu­tant strain (SK3842) that car­ries two amino acid sub­sti­tu­tions in the HU gene and that ex­presses sev­eral path­o­genic­ity-as­­so­ci­ated genes. This strain is ca­pa­ble of in­vad­ing in­testi­nal cells. All of a sud­den, E. coli K‑12 is no longer the par­a­digm of an in­of­fen­sive do­mes­ti­cated strain. Mamma mia! Is it that easy to con­vert a harm­less strain into a vir­u­lent one? Even if two si­mul­ta­ne­ous amino acid sub­sti­tu­tions do not hap­pen all that of­ten, one would think that given the long his­tory of K‑12 mu­ta­ge­n­e­sis and shot­gun cloning, some­thing of this sort could have arisen by now. Ap­par­ently it did not, or at least the con­struc­tion of a new dan­ger­ous bac­terium hasn't been an­nounced.

But a closer look at the HU mu­tant strain is il­lu­mi­nat­ing. When added to in­testi­nal cells, it in­vades them, repli­cates in them, but does not kill them. Maybe it's not such a dan­gerous pathogen af­ter all. The genes that are turned on in­clude those for curli, ex­tra­cel­lu­lar fibers that are in­volved in ad­he­sion to host cells and sub­se­quent in­va­sion, as well as biofilm for­ma­tion. Also, this strain in­duces an ex­ten­sive re­arrange­ment of the host cell's cy­toskele­ton, lead­ing to the for­ma­tion of sur­face pro­tru­sions that en­gulf the bac­te­ria and fa­cil­i­tate their en­try into the cells. Once in­side, other things also hap­pen (or don't hap­pen.) The mu­tants es­cape from the phago­somes, a step ne­ces­sary for their sur­vival, but they then down­reg­u­late the ex­pres­sion of a cy­to­toxic pro­tein, he­molysin E. As a re­sult, they cause no vis­i­ble dam­age to the host cells. The mu­tant does not in­duce apop­to­sis and it even sub­verts the process via a novel mech­a­nism. Host cell and bac­teria live to­gether more or less in peace.

The fi­nal proof that this mu­tant is not a se­ri­ous pathogen is demon­strated in its power to kill mice. Both mu­tants and wild type have the same LD₅₀ when in­jected intrape­ri­toneally into mice. For both, the LD₅₀ is around 2 × 10⁸ bac­te­ria, which is a lot of bac­te­ria. The au­thors say: The fact that SK3842 en­gen­ders a well-de­fined sur­vival pro­gram in the host cell and shows no in­crease in vir­u­lence in an an­i­mal model sig­ni­fies that SK3842 not only is an un­con­ven­tional de­riv­a­tive of ex­tra­cel­lu­lar E. colibut also fol­lows a non­canon­i­cal in­tra­cel­lu­lar re­la­tion­ship with the host cell.

The take-home lessons? Don't as­sume that a friendly strain, es­pe­cially one that has vir­u­lent cousins, is non-path­o­genic. On the other hand, don't as­sume that en­gi­neer­ing a pathogen is all that easy.

Ref­er­ence

Koli P, Su­dan S, Fitzger­ald D, Ad­hya S, & Kar S. 2011. Con­version of Com­men­sal Es­cherichia coli K‑12 to an In­va­sive Form via Ex­pres­sion of a Mu­tant Hi­s­tone-Like Pro­tein. mBio, 2 (5). PMID 21896677