How To Turn An E. coli  Into An Ar­chaeon (Sort Of)

by Elio

Things pop up in mi­cro­bi­ol­ogy that are truly as­tound­ing. The claim that one can par­tially turn a bac­terium into an ar­chaeon rates high on this list. Ar­chaea and Bac­te­ria be­long to dif­fer­ent main branches of the tree of life and have many dis­tinc­tive bio­chem­i­cal and struc­tural dif­fe­rences. Among the most telling of these are their mem­brane lipids, which in Bac­te­ria (and Eu­kary­otes) have straight-chain fatty acids linked to the 3‑carbon of gly­ce­rol through an es­ter link. In con­trast, Ar­chaea have iso­prenoids linked to the glycerol's 1‑carbon via an ether link. In ad­di­tion, some Ar­chaea go fur­ther along their own way and dis­pose of the lipid bi­layer by fus­ing the tails of two phos­pho­lipid mol­e­cules to make a sin­gle layer, os­ten­si­bly for pro­tec­tion. One can turn to other salient dif­fer­ences be­tween Ar­chaea and Bac­te­ria, such as their macro­mol­e­cule-syn­the­siz­ing ma­chinery, the struc­ture of their cell walls, or to en­vi­ron­men­tal their pref­er­ences. But the lipids cer­tainly have a strong hand defin­ing the cel­lu­lar dif­fer­ences be­tween these two do­mains of life. 

Fig­ure 1. Ar­chaeal and bac­te­r­ial phos­pho­lipids. Source

Be­liev­ing as I do that lipids don't al­ways get their proper re­spect and place in the bi­o­log­i­cal lime­light, I wel­come such em­pha­sis. It fol­lows that if dif­fer­ences in lipid com­position mat­ter, that these should be hard to ma­nip­u­late. Imag­ine my sur­prise when I read a pa­per that re­ports the suc­cess­ful cloning into E coli of genes for ar­chaeal enzy­mes in­volved in lipid me­tab­o­lism, with the re­sult that a sub­stan­tial amount of its lipids is now ar­chaeal. This should shake your faith in some­thing. Is noth­ing sa­cred?

Fig­ure 2. Schematic re­pre­sen­tation of the en­gi­neer­ing of E. coli JM109DE3 show­ing the in­te­gra­tion of the MEP-DOXP ope­ron or the IDI gene into the chro­mo­some and the three vec­tors har­bor­ing the genes en­cod­ing ether li­pids en­zymes. For an ex­pla­na­tion of what these ge­netic el­e­ments are see the text of the pa­per

Re­searchers from var­i­ous labs in the Nether­lands engin­eered E. coli to make a mem­brane that has a hy­brid (he­terochiral) com­po­si­tion, con­sist­ing of both ether-linked and es­ter-linked fatty acids, on ei­ther a glyc­erol-1-phos­phate or glyc­erol-3-phos­phate back­bone. The trick, in brief, was achieved by cloning a num­ber of 'ether lipid' genes and oth­ers from var­i­ous Ar­chaea and Bac­te­ria to syn­the­size the ar­chaeal equi­valents of phos­phatidyl­glyc­erol and phos­phatidylethanolamine. To max­i­mize the yield, these ge­nes were placed un­der a lac pro­moter at a strongly tran­scrib­ing do­main of the chro­mo­some, near oriC. Up to 30% of the lipids in this strain had the ar­chaeal con­fig­u­ra­tion. Note that the en­zymes G3PDH and G1PDH, that are in­volved in the syn­the­sis of glyc­erol-1-phos­phate and glyc­erol-3-phos­phate re­spec­tively, do not share any struc­tural or func­tional ho­mol­ogy, be­ing mem­bers of evo­lu­tionarily un­re­lated pro­tein fam­i­lies. The au­thors also tam­pered with the po­lar head group to al­low it to rec­og­nize the ar­chaeal con­stituents. They car­ried out con­sid­er­able work to demon­strate that the lipids made had the proper stere­o­chem­istry. This re­quired de­ter­min­ing the speci­ficity of vari­ous en­zymes in­volved in lipid syn­the­sis to­wards the ar­chaeal and bac­te­r­ial sub­strates, as well as syn­the­siz­ing a num­ber of these com­pounds to com­pare them to the nat­ural ones. By and large, the ex­pected schemes checked out.

Fig­ure 3. Thin-layer chro­matograms of li­pid ex­tracts from wild-type E. coli [JM109 (DE3)] hete­ro­chiral mixed mem­brane E. coli (MEP/DOX­P⁺EL⁺) in­duced early dur­ing growth (OD600 = 0.0) with dif­fer­ent IPTG con­cen­tra­tions and in­cubated un­til sta­tio­nary phase, and the E. co­li strain har­boring the en­tire ether lipid path­way but lack­ing the araM gene (MEP/DOXP⁺EL⁺ AraM^–) treated sim­i­larly.  AG, archae­ti­dyl­gly­cerol (the ar­chaeal equiv­a­lent of phos­phati­dyl­glycerol); CL, car­di­olipin; PE, pho­s­­­phatidyl­etha­nolamine; PG, phos­pha­ti­­dyl-glyc­erol. Source

What is it like to be an E. coli with such an un-bac­te­r­ial load of lipids? When in­duced to a mod­er­ate level, the cells grow well, at about the same rate as the par­ent strain (al­though what they show is a lin­ear plot, which doesn't al­low one to read­ily 'read' what the ex­po­nen­tial growth rate is. A semi-log plot, plus work­ing at lower cell den­si­ties, would have helped here). They found, how­ever, an ap­par­ent lag in growth, sug­gest­ing that the engin­eer­ed strain must un­dergo some ad­just­ments (but ap­par­ent lags are also seen if the in­ocu­lum con­tains few vi­able cells). In­duc­ing the cells to ex­press their 'lipid genes' to a high level, how­ever, causes prob­lems. They now grow more slowly and be­come elon­gated and slightly thin­ner. Growth was slowed down and the cells made 'lob­u­lar ap­pendages' that later be­came de­tached. The likely rea­son is that, when lipid pro­duc­tion is in­creased, other cel­lular processes do not keep pace. Wisely, the au­thors con­cluded that for many of their ex­per­i­ments it was bet­ter to stick to con­di­tions of lower lev­els of in­duc­tion.

Next, how do the en­gi­neered strains re­spond to stresses? Keep in mind the propen­sity of Ar­chaea to be extremo­philes. In­deed, ex­po­sure of these strains to 58°C reveal­ed an en­hance­ment of re­sis­tance as com­pared to the con­trol strains, which did not sur­vive 50°C. The chimeras were also bet­ter able to with­stand freez­ing and thaw­ing, as well as ex­po­sure to bu­tanol. One un­ex­pected and as yet un­ex­plained re­sult was that these cells be­came en­rich­ed for car­di­olipin at the ex­pense of phosphatidyl­gly­ce­rol.

Fig­ure 4. Growth of the het­e­rochi­ral mixed mem­brane strains. (A) Growth of the E. coli MEP/DOXP⁺EL⁺ strain with all of the ether lipid en­zymes [not in­duced (or­ange)], in­duced with 10 μM (red), and in­duced with 100 μM (black) of IPTG added early dur­ing growth (OD600 = 0.0) com­pared with two neg­a­tive con­trol strains: E. coli JM109(DE3) wild-type (blue) and E. coli MEP/DOXP⁺ strain with the in­te­grated MEP-DOXP operon (green). Source

How to rec­on­cile these find­ings with the no­tion that the rea­son Ar­chaea and Bac­te­ria parted ways is that they made lipids that do not work in com­bi­na­tion. This 'dis­cor­dant hy­poth­e­sis' pro­poses that each kind of  homoge­ne­ous mem­brane is more sta­ble than when com­bined to­gether. The au­thors say: "How­ever, it is in­her­ently diffi­cult to test such a hy­poth­e­sis in vivo, as the con­di­tions of early evo­lu­tion would need to be repli­cated and one would need a mi­croor­gan­ism with a mixed mem­brane." So, the present ge­netic ma­nip­u­la­tions yielded just the sys­tem needed for such in vivo test­ing. Clearly, ar­chaeal and bac­te­r­ial lipids can co­ex­ist in the same mem­brane, which gain­says the no­tion that they can't.

Ar­chaeal lipids are thought to con­fer greater re­sis­tance and bet­ter sur­vival un­der harsh con­di­tions. Overtly, this goes for the E. coli con­tain­ing such lipids, whose fit­ness is some­what im­proved upon some stresses. As the au­thors point out, greater ro­bust­ness may be of use in biotech­no­log­i­cal ap­plications. We have en­coun­tered stun­ning sur­prises be­fore about what E. coli can and can­not do. Clone into it a sin­gle gene for the eu­kary­otic pro­tein cave­olin, and the cells be­come re­plete with mem­brane vesi­cles. Is there a gen­eral con­clu­sion to be de­rived from such re­sults? Prokary­otic cells, bac­te­r­ial in the in­stance of the pa­per dis­cussed here, can ex­ist with sig­nif­i­cant de­vi­a­tions from their 'nor­mal' bio­chem­istry and ap­pear to be much more mal­leable than one would have as­sumed. Yes, noth­ing is sa­cred!