Evo­lu­tion of a Su­per­pathogen

by S. Mar­vin Fried­man

A bit of his­tory   The en­ter­obac­terium Yersinia pestis, the causative agent of the plague, was re­spon­si­ble for many ter­ri­fy­ing out­breaks over the course of recorded his­tory, but the most dev­as­tat­ing pan­demic was known as the Black Death. Oc­cur­ring be­tween the years 1347 and 1351, it is es­ti­mated that it claimed about one-third of Europe's pop­u­la­tion, half of that of China, and one-eighth of Africa's. This dreaded dis­ease can de­velop in three dis­tinct forms, de­pend­ing upon the ma­jor site of in­fec­tion: bubonic plague (lymph nodes), sep­ticemic plague (blood ves­sels) and pneu­monic plague (lungs). Plague is pri­mar­ily trans­mit­ted by the bite of fleas that have fed on in­fected an­i­mals, es­pe­cially ro­dents.

Fig­ure 1. The Great Plague. Source

Y. pestis was first iso­lated in 1894 by the French-Swiss bac­te­ri­ol­o­gist Alexan­dre Yersin, who humbly named it 'Pas­teurella pestis'. It was re­as­signed to its cur­rent Yersinia no­ta­tion in 1967. In the genus Yersinia, there are two ad­di­tional species that are path­o­genic in hu­mans: Yersinia pseudo­tu­ber­cu­lo­sis and Yersinia en­te­ro­co­l­it­ica. Both cause rel­a­tively mild food- and wa­ter-borne en­teric dis­eases that are trans­mit­ted via the fe­cal-oral route. Y. en­te­ro­co­l­it­ica and Y. pseudo­tu­ber­cu­lo­sis di­verged be­tween 41 and 186 mil­lion years ago, whereas Y. pestis evolved from Y. pseudo­tu­ber­cu­lo­sis only within the last 1,500 to 6,400 years, mak­ing it a rel­a­tive new­comer on the evo­lu­tion­ary stage. For a pa­le­omi­cro­bi­o­log­i­cal dis­cus­sion of the plague in this blog, see here. So what changed in that rel­a­tively short times­pan to al­low Y. pestis to be­come the in­fa­mous, flea-ex­ploit­ing su­per­pathogen we've come to know and dread?

Fleas and Yersinia pestis   Con­sider its mech­a­nism of spread. When a flea dines on Y. pestis-tainted blood, the bac­te­ria form a biofilm in the gut of the flea, block­ing di­ges­tion and leav­ing the flea con­sti­pated and hun­gry. When the flea then bites an­other host, bac­te­ria break loose from the biofilm and en­ter the new host. By con­trast, when fleas in­gest blood tainted with Y. pseudo­tu­ber­cu­lo­sis, the bac­te­ria pro­duce a pro­tein that is toxic to them and kills ~40% of the un­lucky par­tak­ers. The loss of this toxic pro­tein in Y. pestis al­lowed it to ex­ploit a vec­tor-borne route of trans­mis­sion that re­mains un­avail­able to Y. pseudo­tu­ber­cu­lo­sis. How­ever, lit­tle is known about the toxic agent other than it is a non-se­creted pro­tein that is ac­tive in the flea di­ges­tive tract. Now Chouikha and Hin­neb­usch iden­ti­fied this pro­tein, fill­ing in an im­por­tant piece of the plague puz­zle.

Fig­ure 2. Flea (X. cheopis) col­o­nized by Y. pestis. Source

How come Y. pestis is so flea-friendly?   In or­der to iden­tify the toxic pro­tein, they car­ried out a lit­tle taste test with fleas known to be sen­si­tive to the toxin. They di­vided the fleas into groups and fed them cel­lu­lar frac­tions from Y. pseudo­tu­ber­cu­lo­sis en­riched for mem­brane, cy­to­plas­mic, or periplas­mic pro­teins. Fleas fed on blood meals con­tain­ing the periplas­mic or cy­to­plas­mic frac­tions were un­af­fected, while those fed on the mem­brane frac­tion showed more signs of tox­i­c­ity than un­lucky guests at a cruise­ship buf­fet. They fur­ther sep­a­rated the mem­brane frac­tions by low-speed cen­trifu­ga­tion into su­per­natant (MS) and pel­let (MP) frac­tions.  Blood meals con­tain­ing the MS frac­tion were very toxic to fleas, whereas those con­tain­ing the MP frac­tion were non­toxic, in­di­cat­ing that their mys­tery toxin is a par­tially sol­u­ble, mem­brane-as­so­ci­ated pro­tein.

The au­thors then used two-di­men­sional gel elec­trophore­sis to sep­a­rate the mixes of pro­teins in MS frac­tions, com­par­ing those de­rived from Y. pseudo­tu­ber­cu­lo­sis with those from Y. pestis. Of nine­teen pro­teins spe­cific to the toxic Y. pseudo­tu­ber­cu­lo­sis MS frac­tion, four are com­po­nents of a mul­ti­meric ure­ase com­plex en­coded by the ure­ABCE­FGD operon. In­trigu­ingly, the ureD ho­molog in Y. pestis has a frameshift mu­ta­tion that in­tro­duces a pre­ma­ture stop codon, thus ren­der­ing Y. pestis strains in­ca­pable of de­grad­ing ure­ase, whereas Y. pseudo­tu­ber­cu­lo­sis is ure­ase-pos­i­tive and thus able to de­grade urea. Could si­lenc­ing ure­ase have been the modus operandi that en­abled fleas to be­come highly suc­cess­ful vec­tors for trans­mit­ting the plague?

Fig­ure 3. Yersinia ure­ase is re­spon­si­ble for tox­i­c­ity to X. cheopis and O. mon­tana fleas. A Y. pseudo­tu­ber­cu­lo­sis and Y. en­te­ro­co­l­it­ica ure­ase mu­tants are not toxic to fleas. Mor­tal­ity of fleas 24 h af­ter feed­ing on blood con­tain­ing wild-type, ΔureD, ΔURE, or com­ple­mented ΔURE (pWKS-URE­p­stb) strains of Y. pseudo­tu­ber­cu­lo­sis IP32953 (Y. pstb) or Y. en­te­ro­co­l­it­ica 8081 (Y. ent). B Re­ac­ti­va­tion of the si­lenced ure­ase ac­tiv­ity in Y. pestis and ex­pres­sion of Y. pseudo­tu­ber­cu­lo­sis ure­ase in Y. pestis and E. coli lead to flea tox­i­c­ity. Mor­tal­ity of fleas 24 h af­ter feed­ing on blood con­tain­ing wild-type Y. pestis KIM6+ with or with­out the empty cloning vec­tor), KIM6+ ex­press­ing the Y. pstb ure­ase clus­ter or the Y. pstb ureG‑D genes KIM6+ in which the ureD pseudo­gene was re­paired or E. coli con­tain­ing the empty cloning vec­tor (pWKS130) or ex­press­ing the Y. pstb IP32953 ure­ase operon (pWKS-URE­p­stb) Source

To test this no­tion, these re­searchers con­structed Y. pseudo­tu­ber­cu­lo­sis strains deleted of ureD or the en­tire ure­ase gene clus­ter. Both mu­ta­tions com­pletely abol­ished the abil­ity of Y. pseudo­tu­ber­cu­lo­sis to hy­drolyze urea, and fleas fed on blood con­tain­ing the mu­tant strains did not ex­hibit any signs of tox­i­c­ity.  Like­wise, Y. en­te­ro­co­l­it­ica was also ren­dered non­toxic by dele­tion of ureD. Com­ple­ment­ing the Y. pseudo­tu­ber­cu­lo­sis URE dele­tion mu­tant with a plas­mid ex­press­ing the wild-type ure­ase lo­cus, re­stored flea tox­i­c­ity. These re­sults clearly show that ure­ase is the pro­tein used by Y. pseudo­tu­ber­cu­lo­sis to make fleas deathly ill, ren­der­ing them poor vec­tors.

Ap­proach­ing the sub­ject from an­other an­gle, if just the mu­ta­tion that pre­vents Y. pestis from hy­drolyz­ing urea were fixed, would Y. pestis then be toxic to the fleas, as well?  In­ser­tion of a sin­gle gua­nine residue in the ureD gene is re­spon­si­ble for shift­ing the read­ing frame and re­moves the of­fend­ing residue. As ex­pected, the "fixed" bug was able to hy­drolyze urea, and caused greater mor­tal­ity in fleas (more than 40%) than wild-type Y. pestis. Thus, the re­ac­ti­va­tion of ure­ase in Y. pestis re­stores its an­ces­tral tox­i­c­ity to fleas. It turns out that the ure­ase doesn't even have to be from Yersinia. Adding pu­ri­fied jack bean ure­ase works as well and treat­ment with an ure­ase in­hibitor re­moved the tox­i­c­ity. Which demon­strates that, what­ever the source, the ure­ase must be en­zy­mat­i­cally ac­tive in or­der to kill the fleas.

Fig­ure 4. Tox­i­c­ity to fleas is de­pen­dent on ure­ase en­zy­matic ac­tiv­ity. A Mor­tal­ity of fleas that fed on blood con­tain­ing Y. pseudo­tu­ber­cu­lo­sis cell lysate (CL) or JBU with or with­out the ure­ase in­hibitor p‑Bq. Con­trol fleas were fed on blood con­tain­ing PBS buffer and p‑Bq. Data are means ± SD of two in­de­pen­dent ex­per­i­ments. B In­hi­bi­tion of ure­ase en­zy­matic ac­tiv­ity by p‑Bq in the sam­ples added to the flea blood meals was ver­i­fied by us­ing a quan­ti­ta­tive ure­ase as­say. Source

Loss of Func­tion, Gain of Vir­u­lence   The loss of ure­ase ac­tiv­ity al­lowed Y. pestis to ac­cess a very ef­fi­cient vec­tor, but did this ad­van­tage come with a fit­ness trade­off? Are there neg­a­tive ef­fects of ure­ase loss in terms of the abil­ity of Y. pestis to sur­vive in the flea gut and spread ef­fi­ciently? Ap­par­ently not. In fleas fed on ei­ther the ure­ase-neg­a­tive wild-type Y. pestis or an ure­ase-pos­i­tive mu­tant, the av­er­age in­fec­tion rate and bac­te­r­ial load were sim­i­lar for both strains, in­di­cat­ing that the ad­van­tage gained due to loss of ure­ase in Y. pestis did not re­quire a trade­off. A sweet deal for Y. pestis, in­deed.

The elim­i­na­tion of ure­ase ac­tiv­ity pre­vented the poi­son­ing — likely by am­mo­nia pro­duc­tion — of fleas, al­low­ing Y. pestis, in con­trast to its prog­en­i­tor Y. pseudo­tu­ber­cu­lo­sis, to use fleas as vec­tors for trans­mis­sion of the plague. This joint ven­ture was highly suc­cess­ful and prob­a­bly un­der­went Dar­win­ian pos­i­tive se­lec­tion.  The ex­tent of the ad­van­tage gained is high­lighted by the fact that the mu­tated ureD al­lele is found in all Y. pestis strains tested world­wide. This is note­wor­thy since a mu­ta­tion by the in­ser­tion of a sin­gle nu­cleotide in a ho­mopoly­meric run via slipped strand mi­s­pair­ing has a high re­ver­sion rate. Such Y. pestis re­ver­tants have ap­par­ently been elim­i­nated from the pop­u­la­tion due to a fit­ness dis­ad­van­tage. It is no won­der, there­fore, that in the pre-an­tibi­otic era Y. pestis evolved as a su­per­pathogen claim­ing more lives than any other in­fec­tious agent in the an­nals of mi­cro­bial patho­gen­e­sis. And all this be­cause of the loss of one sin­gle gene func­tion!

Ref­er­ence

Chouikha I, Hin­neb­usch BJ. 2014. Si­lenc­ing ure­ase: a key evo­lu­tion­ary step that fa­cil­i­tated the adap­ta­tion of Yersinia pestis to the flea-borne trans­mis­sion route. Proc Natl Acad Sci USA, 111, 18709−18714. PMID 25453069