"We Are What We Are"

by Christoph

If I were a stage di­rec­tor and con­tracted for an­other re­vi­val of the Broad­way show 'La Cage aux Folles' (in­clud­ing the cho­rus line's "We Are What We Are" ) I would not wa­ver a minute to choose for all lead­ing roles mem­bers of the Pseudomonas fam­ily. All are true pro­fes­sion­als (me­thod act­ing, Meis­ner tech­nique ), ver­sa­tile in act­ing solo, or as a team, in danc­ing and in singing. Singing? In fact they're mul­ti­lin­gual and out­right lo­qua­cious: they em­ploy, of­ten si­mul­ta­ne­ously, sev­eral Quo­rum Sens­ing sys­tems (QS) (Fig. 1). Some of them are the per­fect choice for the roles of vil­lains, and all are well trained in self de­fense: they bring the nec­es­sary CRISPR tools along. But since they're pro­fes­sion­als they don't ad­here to 'open carry', which they con­sider a waste of en­ergy and prone to self-harm­ing. As Bon­nie Bassler and her cowork­ers showed in a re­cent study, Pseudomonas aerug­i­nosa mounts its de­fense, CRISPR, upon at­tack or, more pre­cisely, when it re­alizes that it moves in a crowd where pick­pock­ets (= pha­ges ) might lurk.

Fig­ure 1. The four in­ter­con­nected QS sys­tems of P. aeru­ginosa. Au­toin­ducer syn­thases, LasI, RhlI, Pqs­ABCDH and AmbB­CDE, pro­duce the auto­in­ducers, 3‑oxo-C12-HSL, C4-HSL, PQS and IQS, re­spectively. 3‑oxo-C12-HSL, C4-HSL and PQS, are rec­og­nized by cy­to­plas­mic tran­scrip­tion fac­tors. The re­cep­tor for IQS is cur­rently un­known. The pro­duc­tion of the IQS sig­nal is in­duced un­der phos­phate star­va­tion. The in­di­vid­ual cir­cuits are highly in­ter­con­nected and in­volve autoinduct­ion (red ar­rows). Source. Front­page: P. aerugi­nosa biofilms on glass wool af­ter 24 h as obser­ved by scan­ning elec­tron mi­croscopy. Bar 5 µm. Source

These re­searchers found in pre­lim­i­nary ex­per­i­ments that the QS in­hibitor baicalein pre­vents pro­duc­tion of the vi­ru­lence fac­tor py­ocyanin, whose syn­the­sis is trig­gered by the LasIR and RhIL QS sys­tems. Baicalein also blocked the pos­i­tive ef­fect of an au­toin­ducer (AI) cock­tail on cas3 ex­pres­sion (Cas3 = the scis­sors ) and in­hib­ited the AI‑mediated en­hance­ment of CRISPR adap­ta­tion in wild-type P. aerug­i­nosa PA14 and in a ΔlasI ΔrhlI mu­tant sup­ple­mented with AI.

Fig­ure 2.  QS reg­u­lates CRISPR-Cas ac­tiv­ity. Re­tent­ion of A the con­trol plas­mid pHERD30T, and B the CRISPR-tar­geted plas­mid pCR2SP1 in PA14 (WT) and in the ΔlasI ΔrhlI mu­tant dur­ing growth (100% de­notes no plas­mid loss). AI in­di cates 2 μM 3OC12-HSL + 10 μM C4-HSL. %CFU = per­cent­age of plas­mid-con­tain­ing colony-for­ming units (CFU). Er­ror bars de­note SD from n = 3 repli­cates. Source

They then quan­ti­fied the re­ten­tion of a con­trol plas­mid and a CRISPR-tar­geted plas­mid – that is, a plas­mid which con­tains a pro­to­spacer "known" by the chro­mo­so­mal CRISPR ar­ray and flanked by a pro­to­spacer-ad­ja­cent mo­tif (PAM) that is re­quired for CRISPR in­ter­fer­ence – over time in wild-type PA14 and in a ΔlasI ΔrhlI dou­ble au­to­in­­ducer-syn­thase mu­tant (Fig. 2). No loss of the con­trol plas­mid oc­curred over the course of the ex­per­i­ment in ei­ther strain. With re­spect to the CRISPR-tar­geted plas­mid, no loss oc­curred in ei­ther strain dur­ing growth at low cell den­si­ties, and ad­di­tion of an au­toin­ducer cock­tail (AI) had no ef­fect for up to 3 h. Af­ter 5 h of growth, con­di­tions un­der which QS has ini­ti­ated, plas­mid loss oc­curred in WT cells. In con­trast, at 5 h, min­i­mal loss oc­curred in the ΔlasI ΔrhlI dou­ble mu­tant but ad­di­tion of AI re­stored plas­mid loss to WT lev­els. At 6.5 h, when QS is highly in­duced, mod­est plas­mid loss oc­curred in the ΔlasI ΔrhlI mu­tant, whereas over 20-fold more of the ΔlasI ΔrhlI mu­tant cells re­tained the plas­mid than did wild-type cells or ΔlasI ΔrhlI mu­tant cells sup­ple­men­ted with AI. This re­sult shows that QS is re­quired to po­tently in­duce CRISPR-Cas ac­tiv­ity in PA14. The resid­ual CRISPR-Cas ac­tiv­ity in the ΔlasI ΔrhlI mu­tant sug­gests that an­other, less ef­fi­cient LasI RhlI-in­de­pen­dent CRISPR-Cas ac­ti­va­tion mechanism(s) ex­ists in PA14.

Fig­ure 3. QS reg­u­lates CRISPR-Cas ac­tiv­ity. Effi­ciency of trans­formation (EOT) of PA14 (wild-ty­pe) and des­ig­nated mu­tants at high cell den­sity (OD₆₀₀ = 1) quan­ti­fied as the per­cent­age trans­formation by the CRISPR-tar­geted plas­mid pCR­2SP1 com­pared with that of the con­trol plas­mid pHERD30T lack­ing the tar­geted se­quence. 100% de­notes an EOT ra­tio of 1 for the two plas­mids. In all pan­els, AI in­di­cates 2 μM 3OC12-HSL + 10 μM C4-HSL. Er­ror bars de­note SD from n = 3 repli­cates. Source

The above ex­per­i­ments in­di­cated that QS en­hances plas­mid loss dur­ing cell growth, that is, when the plas­mid has al­ready gen­er­ated copies of it­self. Høy­land-Kroghsbo et al. went on ex­am­in­ing the in­flu­ence of QS on the abil­ity of CRISPR-Cas to elim­i­nate a sin­gle  in­com­ing ge­netic ele­ment, a sin­gle phage for ex­am­ple. To this end, they mea­sured the ef­fi­ciency of trans­for­ma­tion (EOT) of cells grown to high pop­u­la­tion den­si­ties by the CRISPR-tar­geted plas­mid (cal­i­brated to that of the con­trol plas­mid ) in the wild-type strain PA14, in a mu­tant strain lack­ing both CRISPR ar­rays as well as the cas and csy genes, and in the ΔlasI ΔrhlI strain with and with­out the ad­di­tion of au­toin­ducer cock­tail (note that their CRISPR-tar­geted plas­mid mim­icked a sit­u­a­tion where the cells had al­ready encoun­ter­ed an in­fect­ing phage be­fore and were 'primed' for de­fense ).

As ex­pected, the EOT was 100% for the ΔCRISPR Δcas mu­tant be­cause this mu­tant is in­ca­pable of tar­get­ing ei­ther plas­mid (Fig. 3). In con­trast, the EOT was 2% in wild-type PA14 be­cause CRISPR-Cas is fully in­duced and ef­fi­ciently cleaves the tar­geted plas­mid. The EOT was 14% in the ΔlasI ΔrhlI dou­ble mu­tant, show­ing that the CRISPR-Cas im­mune sys­tem is seven-fold less ef­fec­tive than in the wild type when unin­duced. Ad­di­tion of au­toin­duc­ers to the ΔlasI ΔrhlI dou­ble mu­tant re­stored CRISPR-Cas ac­tiv­ity, re­duc­ing the EOT to 4%. Thus, QS reg­u­la­tion of CRISPR-Cas ac­tiv­ity in PA14 is the ma­jor trig­ger for high-level CRISPR-Cas-de­pen­dent im­mu­nity against in­fect­ing el­e­ments.

Fig­ure 4. Model of the mecha­nisms of ac­tion of three anti-CRISPR pro­teins. The type I‑F CRISPR loci con­sist of di­rect re­peat se­quences (black di­a­monds) sep­a­rated by uni­que spacer se­quences (col­ored rec­tan­gles) that were de­rived from for­eign mo­bile ge­netic ele­ments. Dur­ing the ac­qui­si­tion stage, a piece of DNA from an in­vad­ing phage or other mo­bile el­e­ment is cap­tured and in­serted as a spacer in­to a CRISPR ar­ray (yel­low rec­tan­gle). Next, the CRISPR-as­so­ci­ated (Cas) pro­teins are pro­duced, and the CRISPR loci are tran­scribed into long pre-CRISPR-RNAs. The Csy4 en­donu­cle­ase then cleaves the re­peat se­quences to yield ma­ture CRISPR-RNAs (cr­RNAs), each con­tain­ing a sin­gle spacer se­quence (col­ored lines). The cr­RNA-Csy4 com­plex then in­ter­acts with Csy1,Csy2, and Csy3 to form the surveil­lance com­plex. AcrF1 and AcrF2 di­rectly in­ter­act with this com­plex and pre­vent DNA bind­ing. By con­trast, AcrF3 binds to the Cas3 he­li­case-nu­cle­ase pro­tein and blocks its re­cruit­ment. Source

Høyland‑Kroghsbo et al. ar­gue that their "...dis­cov­ery of QS-me­di­ated ac­ti­va­tion of CRISPR-Cas aligns well with pre­vi­ous stud­ies show­ing QS-me­di­ated phage de­fense via down-reg­u­la­tion of phage re­cep­tors re­duc­ing infect­ion rates (see here in STC). QS re­pres­sion of phage sur­face re­cep­tors at high cell den­sity could be the first line of de­fense, ef­fec­tively en­abling bac­te­ria to pre­vent in­fection. If this ini­tial strat­egy fails, the sec­ond line of de­fense be­comes cru­cial: QS ac­ti­va­tion of CRISPR-Cas im­mune de­fense en­ables pur­suit of phages that make it into the cy­to­plasm." It thus makes sense to de­velop, for ex­am­ple, phage ther­a­pies com­bined with au­toin­ducer an­tag­o­nists for cys­tic fi­bro­sis (CF) pa­tients with chronic, an­tibi­otic-re­sis­tant Pseudomonas aerug­i­nosa in­fec­tions (see Mar­vin Friedman's STC post on phage ther­apy here ). Sup­port­ing this per­spec­tive, D'Argenio et al. found a pre­valence of lasR, and less fre­quently rhlR, mu­tants among P. aerug­i­nosa iso­lates from CF pa­tients dur­ing in­fec­tion for rea­sons that are not com­pletely un­der­stood. Al­though such mu­tants would be un­af­fected by QS‑inhibiting the­rapies tar­get­ing LasR ac­tiv­ity, they would al­ready be im­paired in ac­ti­vat­ing the CRISPR-Cas im­mune de­fense and maybe, in ad­di­tion, un­able to down-reg­u­late phage re­cep­tors. The au­toin­ducer an­tag­o­nist in the phage cock­tail for ther­apy would, in ad­di­tion, re­duce the ex­pres­sion of such vir­u­lence fac­tors that de­pend on ac­ti­va­tion via one of the QS sys­tems. And there is one last point: it is known by now that many phages en­code anti-CRISPR pro­teins (Fig. 4). This makes the choice of the most effi­cient phage cock­tail less of an is­sue.