Phage Re­sponses to the SOS Re­sponse

by Merry Youle

No mat­ter where a bac­terium lives, its DNA is some­times dam­aged by UV ra­di­a­tion, tox­ins, re­ac­tive oxy­gen species (ROSs), or other nasty things. Faced with such as­saults, as well as prob­lems in­her­ent in DNA repli­ca­tion, bac­te­ria evolved so­phis­ti­cated mech­a­nisms to de­tect and re­pair DNA dam­age. One widely con­served sys­tem, the SOS re­sponse, was first de­scribed in 1974 in E. coli (cited here). When this bac­terium rec­og­nizes that it has a prob­lem, it halts cell di­vi­sion and ac­ti­vates a net­work of genes to re­pair the dam­age (here's a pa­per with a use­ful sum­mary, as well as more de­tails on one player, LexA).

Fig­ure 1. SOS sand. Source

Launch­ing an SOS Re­sponse

Fig­ure 2. SOS re­sponse mech­a­nism. Source

A key agent in re­cov­ery is a re­com­bi­nase, RecA. It re­pairs DNA by swap­ping ho­mol­o­gous re­gions be­tween flawed strands to gen­er­ate one com­plete, un­dam­aged chro­mo­so­me. Ho­mo­logues of RecA have been found in all cells in all three do­mains ex­am­ined so far — it's Rad51 in us hu­mans — as it is es­sen­tial for re­ac­ti­vat­ing stalled re­pli­ca­tion forks as well as for men­ding dam­age. When re­gions of sin­gle-stranded DNA (ss­DNA) re­sult from UV ex­posure, for ex­am­ple, they must be pro­tected from at­tack by cel­lu­lar nu­cle­ases. Such re­gions are quickly coated by many copies of SSB pro­tein (Sin­gle-Strand DNA-Bind­ing pro­tein) that are sub­se­quently dis­placed by RecA. Each RecA monomer pro­tects only three nu­cleotides, but RecA monomers poly­mer­ize into long fil­a­ments that wrap around the DNA.

Fig­ure 3. RecA in Ac­tion. Re­searchers have sol­ved the struc­ture of RecA and cap­tured dif­fe­rent steps in the process of ho­mol­o­gous re­com­bi­na­tion. The RecA/Rad51 pro­tein per­forms the tricky task of pair­ing up a bro­ken strand with its backup copy. The struc­ture shown here, from PDB en­try 3CMX, shows E. coli RecA sur­roun­ding the het­erodu­plex, af­ter the dam­aged strand has been paired with the in­tact copy. The struc­ture was de­ter­mined us­ing an en­gi­neered form of recA that when ex­pressed yield­ed pro­teins com­posed of five tan­dem copies of RecA. The struc­ture re­vealed that the DNA is stretched when it binds to recA, with a dis­tinc­tive pat­tern of three bases to each pro­tein sub­unit (Click to see the full fil­a­ment). Source

The fil­a­ment form of RecA — also known as ac­ti­vated RecA or RecA* — is the ac­tive agent dri­ving RecA's sec­ond con­tri­bu­tion to the SOS re­sponse: in­creased ex­pres­sion of the more than 43 genes scat­tered across the genome that com­prise the SOS reg­u­lon. The in­ter­me­di­ary be­tween the RecA fil­a­ments and the reg­u­lated genes is an­other key cel­lu­lar pro­tein, the re­pres­sor LexA.

Dur­ing nor­mal growth, LexA re­presses ex­pres­sion of all the genes of the SOS reg­u­lon. Struc­turally, LexA con­tains two do­mains joined by a hinge re­gion. Sta­ble LexA di­mers bind to DNA at spe­cific, highly con­served sites (SOS boxes) and pre­vent ex­pres­sion of the down­stream genes. Ex­pres­sion of the genes in the SOS reg­u­lon re­quires de­struction of LexA. The struc­ture of ac­ti­vated RecA al­lows LexA dimers to em­bed deeply into the fil­a­ment groove. This em­bed­ding al­ters LexA's con­for­ma­tion and stim­u­lates its auto-cat­alytic cleav­age, thereby end­ing its re­pressor ac­tiv­ity. When the DNA has been re­paired, the LexA level re­turns to nor­mal and re­pres­sion of the SOS reg­u­lon is re­stored.

An In­vi­ta­tion to a Phage

The host's SOS re­sponse in­vites ex­ploita­tion by tem­per­ate phages, and phages do not of­ten turn down such in­vi­ta­tions. In a lyso­gen, the SOS re­sponse pro­vides a sig­nal that no­ti­fies an as­tute pro­phage that the cell is in trou­ble and may have suf­fered ir­repara­ble DNA dam­age. It might be pru­dent for the prophage to go lytic at this point, asap!, and get a crop of prog­eny as­sem­bled while the host can still sup­port phage repli­ca­tion. How a prophage de­cides to jump ship has been best stud­ied for phage λ, a tem­per­ate phage in­fect­ing E. coli. For a λ prophage to main­tain ly­so­ge­ny it must si­lence many of its genes, specif­i­cally those in­volved in repli­ca­tion, syn­the­sis of struc­tu­ral pro­teins, and host ly­sis. At the same time, it ac­tively ex­presses a few genes, in­clud­ing its re­pres­sor CI whose job it is to si­lence those genes. This re­pres­sor is sim­i­lar to LexA in struc­ture and func­tion. Like­wise, CI is prompted by RecA fil­a­ments to self-cleave, thereby flip­ping the switch from lysogeny to lytic repli­ca­tion.

Fig­ure 4. Pro­posed model of the E. coli LexA re­pres­sor bound as a dimer at the cka gene reg­u­la­tory re­gion. The three α‑helices of the N‑terminal do­main are marked as H1 to H3, re­spec­tively. Source

As of­ten is the case, the λ way of end­ing lysogeny be­came the par­a­digm, but λ's way is not the only way. More proof for that comes from phage GIL01, a tem­per­ate phage that in­fects the Gram-pos­i­tive in­sect pathogen Bacil­lus thuringien­sis. Since GIL01 car­ries ter­mi­nal pro­teins co­va­lently linked to both ends of its chro­mo­some, its prophage can't in­te­grate into the host chro­mo­some. In­stead, dur­ing lysogeny GIL01 main­tains sta­bly as an in­de­pen­dent lin­ear prophage. Nev­er­the­less, like phage λ, GIL01 repli­ca­tion is in­duced when UV or mit­o­mycin C (MMC) treat­ment elic­its the SOS re­sponse in the lyso­gen. As with phage λ, in­duc­tion re­quires RecA. But when re­searchers looked for a phage-en­coded re­pres­sor com­parable to λ's CI, they found no good can­di­dates.

To pur­sue this far­ther, they iso­lated clear plaque (cp) mu­tants, i.e., mu­tants that could not es­tab­lish lysogeny but could repli­cate only lyt­i­cally. These mu­tants mapped to three lo­ca­tions in the ge­nome. Of par­tic­u­lar in­ter­est was the re­gion named dinBox1 (dam­age-in­ducible Box1) up­stream of genes for repli­ca­tion and reg­u­la­tion. That re­gion was sim­i­lar to the 14 bp con­sen­sus se­quence for LexA bind­ing sites (SOS boxes) in B. sub­tilis. Point mu­ta­tions in that re­gion abol­ished lysogeny. Did GIL01 per­haps use the host's LexA as its re­pres­sor? The re­searchers wanted to test this by ge­ne­ra­ting a host strain lack­ing LexA, but LexA is es­sen­tial for B. thuringien­sis' sur­vival. In­stead they ge­ne­ra­ted a host strain with a non-cleav­able LexA. As you would ex­pect, these cells were hy­per­sen­si­tive to DNA dam­age from MMC since cleav­age of the LexA pro­teins is nec­es­sary to mount an SOS re­sponse and boost DNA re­pair. Cleav­age of LexA was also found to be nec­es­sary to in­duce the GIL01 prophage.

Fig­ure 5. Par­tial phys­i­cal map (De­tail ) of phage GIL01. Pre­dicted ORFs are de­picted as open box­es. All tran­scrip­tion is right­ward. Prob­a­ble gene func­tions are in­di­cated above cer­tain ORFs. Pro­mot­ers are shown as filled arrow­heads, and each pro­moter re­gion is en­larged be­low the map. Boxes for –35 and –10 are de­pic­ted as black rec­tan­gles, and an­gled ar­rows rep­re­sent tran­scrip­tion start sites. Gray boxes show the lo­ca­tions of con­served LexA-bind­ing sites in re­la­tion to the three pro­mot­ers. Source

Based on these and many other ex­per­i­ments, the re­sear­chers con­cluded that LexA di­rectly re­presses the lytic path­way for this phage by bind­ing to dinBox1, an SOS box in the prophage genome. Treat­ment of the lyso­gens with MMC trig­gered LexA cleav­age as part of the host's SOS re­sponse, and LexA cleav­age in­duced the GIL01 pro­phage to repli­cate and ul­ti­mately lyse the cell. Us­ing the host's LexA in­stead of mak­ing its own LexA-like re­pres­sor re­duces the num­ber of genes the phage re­quires by one. This might mat­ter to GIL01. Phage λ has 74 genes to spend, com­pared to 30 for GIL01. More­over, a λ pro­phage re­lies on the host for its repli­ca­tion and pas­sage to both daugh­ter cells, while GIL01 has the added bur­den of need­ing sev­eral genes to repli­cate its 'plas­mid-like' lin­ear prophage and par­ti­tion the copies dur­ing cell di­vi­sion.

An Ac­com­plice

But this is not the whole story for GIL01, as two other clues later re­vealed. First, LexA also binds to a re­gion ad­ja­cent to dinBox1, a re­gion dubbed dinBox1b. Mu­ta­tion of con­served nu­cleotides in ei­ther re­gion re­duces LexA bind­ing; mu­ta­tions in both sites nearly elim­i­nate bind­ing al­to­gether. Sec­ond, some GIL01 cp mu­tants mapped to ORF7 in­di­cat­ing that the en­coded pro­tein, Gp7 (gene prod­uct 7) is re­quired for lysogeny. These two ob­ser­va­tions turned out to be in­ti­mately re­lated. By form­ing a sta­ble com­plex with LexA, Gp7 in­hibits its self-cleav­age. LexA now binds more sta­bly to SOS box­es in the prophage chro­mo­some, thereby in­hibit­ing prophage in­duc­tion. It also binds more firmly to host SOS boxes, and thus im­pedes the host's SOS re­sponse. At first glance this is surely puzz­ling. Why would a phage want to com­pro­mise its host's DNA re­pair mech­a­nism and at the same time hin­der its own es­cape from the dam­aged ship?

I emailed that ques­tion to the re­searcher, Na­dine For­ne­los. In her re­sponse she de­scribed a per­ti­nent re­cent ex­per­i­ment: "I have over­ex­pressed gp7 in the phage's nat­ural host, Bacil­lus thu­rin­gien­sis GBJ002 [with­out GIL01], and sub­jected the cells to MMC treat­ment. With this sim­ple ex­pe­ri­ment, I in­tended to demon­strate that cells in which gp7 was ex­pressed were more vul­ner­a­ble to geno­toxic treat­ment. Sur­pris­ingly, this was not the case." In­stead she found that ex­pres­sion of gp7 in­creased cell sur­vival 8 – 10-fold.

She then went look­ing for clues in the cel­lu­lar genome and found a chro­mo­so­mal prophage (CP). More­over, this CP en­coded a LexA-like re­pres­sor, the tac­tic we saw ear­lier used by phage λ. In her ear­lier ex­per­i­ments she had in­duced GIL01 repli­ca­tion with 0.05 μg·ml^–1, a con­cen­tra­tion with mi­ni­mal ef­fect on host sur­vival. In the re­cent gp7 ex­pres­sion stud­ies she used a higher MMC con­cen­tra­tion (0.1 μg·ml^–1) that re­duced host sur­vival 100 – 1000-fold. She ten­ta­tively con­cluded that at the lower MMC con­cen­tra­tion, GIL01 repli­cated with­out in­ter­fer­ence from the re­pressed CP. At the higher MMC con­cen­tra­tion, the CP was in­duced, repli­cated, and lysed the cells. Ex­pres­sion of Gp7 likely in­hib­ited CP in­duc­tion at the high MMC con­cen­tra­tion, thus ex­plain­ing her ob­ser­va­tion of in­creased cell sur­vival.

Gp7 is the first phage fac­tor known to in­ter­act di­rectly with LexA to mod­u­late both prophage in­duc­tion and the host's SOS re­sponse. This story por­trays prophage in­duc­tion dur­ing the SOS re­sponse as a highly-evolved reg­u­la­tory dance sen­si­tive to dif­fer­ences in stress level and re­pres­sor bind­ing affini­ties to var­i­ous SOS boxes. Out­side the lab, it is not a one phage–one bac­terium world. Com­pe­ti­tion for bac­te­r­ial hosts is fierce, and a phage must de­fend against stow­aways.