Goodbye PAM: Phage λ’s Red recombination system cripples PAMs and helps dodge CRISPR attacks

Abstract

Ever wondered how the phage λ Red recombination system resembles the Red Queen? Hossain et al. (2021) provide an answer in this issue of Cell Host & Microbe. They show that Red debilitates PAM sequences by mutagenic repair of CRISPR-targeted DNA breaks in infecting λ, thus shaping the phage-CRISPR arms race.

Inspired by the Red Queen telling Alice in Lewis Carroll’s Through the Looking-Glass, “Now, here, you see, it takes all the running you can do, to keep in the same place,” Leigh Van Valen introduced the Red Queen as a metaphor to describe the status quo in the ceaseless struggle between predator and prey in an ever-changing environment. On our planet, phages outnumber bacteria 10-fold. Their evolutionary struggle is evident in the diversity of offense-defense strategies found in both. The paper by Hossain et al. (2021) puts a spotlight on one such battle between phage λ and the CRISPR-Cas defense system of its E. coli host.

In bacteria, the CRISPR (clustered regularly interspaced short palindromic repeats) locus with its associated Cas proteins provides a barrier against mobile genetic elements (MGE), including phages, plasmids, and transposons (Faure et al., 2019). In contrast to bacterial innate genome defense such as restriction-modification, CRISPR behaves as an adaptive immune system by storing in its memory box snippets of DNA (“spacers”) from past invaders (Figure 1A, top). Spacers are assimilated into the CRISPR array by dedicated Cas enzymes that recognize a 2- to 4-bp signature PAM element (protospacer adjacent motif) upstream of a seed sequence in a 10-bp “protospacer”, processing and integrating into the CRISPR locus ~30 bp of MGE DNA without PAM. The spacer array produces CRISPR RNAs (crRNA), which complex with Cas nucleases and hunt for PAM-adjacent DNA matches in their MGE targets, cutting and destroying them as they attempt to break and enter again. This process is called interference. CRISPR systems are categorized into two classes, depending on the number of effector proteins involved in interference, with each class type harboring many subtypes (Koonin et al., 2017). In the class I system, multiple effector proteins (Cascade; Cas3/6) are required for CRISPR targeting, while class II systems only have one (e.g., Cas9) (Figure 1A top). Because of their ability to selectively target DNA-containing sequences complementary to CRISPR spacers, CRISPR-Cas systems have been extensively studied both for their biology and for gene editing applications (Pickar-Oliver and Gersbach, 2019).

Figure 1. Simplified mechanisms of CRISPR and λ Red: How λ escapes CRISPR targeting.

(A) Top: on the initial encounter with a phage (1), Cas1/Cas2 integrate PAM-adjacent bits of the phage DNA as spacers (multicolored rectangles) into the CRISPR array. The array is transcribed and processed into crRNAs. On the second encounter with the same variety of phage (2), class I/II nuclease-bound crRNAs search for complementary matches and cut up the invading phage genome. Bottom: λ Red and phage replication. The Red operon encodes three proteins. Gam binds and inactivates the host RecBCD nuclease to protect dsDNA ends during λ rolling-circle replication. Exo and Beta promote recombinational repair of broken DNA (or of two λ concatamers) by HR (homologous recombination).

(B) Houssain et al. (2021) report that λ escapes Cas nuclease breaks next to PAM by promoting mutagenic repair of the break. In the case of type II CRISPR, error-prone Pol IV was required for generating escapers, but the mutation pattern depended solely on Exo-Beta. (This figure was generated with Biorender).

Phages are known to overcome CRISPR defense by encoding Anti-CRISPR (Acr) proteins directed at frustrating every possible step in the mounted bacterial attack, inhibiting the Cas proteins from either binding RNA or binding and cleaving DNA (Davidson et al., 2020)—a case of the Red Queen hard at work. The impetus of the current study by Hossain et al. (2021) came from prior observations that phages that escape CRISPR targeting contain point mutations in protospacers, apparently a different adaptive strategy. To understand how these mutations are generated, the authors engineered both type I and type II CRISPR defenses in E. coli and challenged each with phage λ. Phages that escaped carried mutations in PAM, for which λ’s Red homologous recombination (HR) system was overwhelmingly responsible.

Phages often carry recombination functions that augment those of their host and are important for their development and survival as they move from one host to another. λ Red is one such function and has been well studied (Poteete, 2001). The Red operon encodes 3 proteins: Gam, Exo, and Beta, the latter two constituting the recombination apparatus. On encountering linear dsDNA (double-stranded DNA), Exo degrades the 5′-ending strand, while the 3′-ending strand is preserved, generating a 3′-ssDNA (single-stranded DNA) overhang on which Beta assembles to form a filament, invading homologous dsDNA to restore the break by replication-dependent HR (Figure 1A bottom). The efficiency of this system has revolutionized bacterial gene manipulation by recombineering with Beta alone (Thomason et al., 2014). In E. coli, the main HR system is RecABCD, where, analogous to Exo, the RecBCD nuclease degrades linear dsDNA ends. Unlike Exo, RecBCD chomps both ends until it encounters an 8-bp Chi site, interaction with which effectively changes its behavior such that, like Exo, RecBCD now preferentially degrades the 5′ strand, leaving ssDNA with 3′ overhangs. Like Beta, RecA coats these ends and promotes HR. The role of the third Red protein, Gam, is to bind RecBCD and inhibit its nuclease function. This is critical because after infection, the circular λ genome amplifies by rolling replication. During this phase, the linear λ end of the concatamer is vulnerable to RecBCD (Figure 1A, bottom). Despite this apparently important role, the Red system is not essential for λ development. In the absence of Red, acquisition of a Chi sequence by point mutations is sufficient to replace the Red function by RecABCD (Myers and Stahl, 1994). This knowledge was used by Hossain et al. (2021) to follow virulent (vir) λ (which only goes lytic) without or with engineered Chi sites to test the role of the Red versus RecABCD HR systems in surviving CRISPR-Cas defense.

Their experimental plan consisted of programming type I and II Cas nucleases with crRNA to target six different potential protospacers in infecting λ. Not all spacer matches were equally potent: some matches were strong, in that infecting λ was clobbered; some were moderately effective; and others were weak, in that λ titers were 2 to 3 orders of magnitude higher than expected. A majority of the escapee phage genomes had mutations in the PAM and seed sites of the targeted protospacers, although a few appeared to have acquired either cryptic prophage sequences by HR or small deletions (MMEJ; microhomology-mediated end joining). Given that λ inactivates the host HR via Gam, the authors asked whether Red recombination might be responsible for their escape. Infection with λ vir variants deleted for exo or beta genes showed that with every category of spacer tested, both of these genes were essential for generating escapers. Co-infection with Red⁺/Red⁻ phages in a 1:1 ratio at very high MOI yielded Red⁺ winners every time, hinting at a hitherto unappreciated cis-functioning element in Red action. Even though in a normal λ infection RecABCD repair is not possible, the authors wondered whether, given a chance, the RecA system can also function similarly. They tested this by disabling Red and introducing multiple Chi sites in the λ genome to allow RecA-mediated HR. Escapers emerged, but these were at least an order of magnitude lower than those generated with Red recombination. λ is therefore self-reliant for breaking through CRISPR defense. Note the two different Red Queen strategies at play here: Acrs preempt DNA cleavage, whereas λ Red acts post-cleavage, turning adversity into advantage. For Red, any break should do, giving it more leverage with the Queen.

Why is Red-mediated repair mutagenic? Hossain et al. (2021) tested the possibility that translesion (TL) polymerases induced during the SOS response to dsDNA breaks in E. coli (Fujii and Fuchs, 2020) are responsible for filling ssDNA gaps across the Cas-cleaved sequence during HR. Of the several TL polymerases tested, they found Pol IV to be required for generation of 99% of the phage escapers but only in the type II system. Curiously, while Pol IV was required, the mutation pattern was wholly dependent on Red, suggesting perhaps an unrecognized role for Exo-Beta in influencing error-prone gap-filling. The difference in requirement for Pol IV in the two systems might lie in how the two Cas nucleases handle the cut DNA—a blunt dsDNA cleavage for Cas9 (type II) and ssDNA cleavage followed by unwinding and processive degradation for Cas3 (type I), likely generating different DNA substrates for repair and thus different outcomes of this process.

Red-Beta recombinases constitute a major superfamily of phage recombinases that carry a Rad52-like fold, and they are widely distributed in both temperate and lytic phages (Lopes et al., 2010). Given that Red-like systems are absent in half the lambdoid phages and therefore not absolutely required for viral replication, the authors argue that these systems must have essential functions that maintain them in the other half of the phage genomes. They propose that the role of Red recombinases in CRISPR evasion could be a significant selection pressure that has fixed these systems and allowed their spread. Just when we thought Red was done and dusted, the exciting findings of Hossain et al. (2021) open fresh avenues of inquiry into the evolutionary edge the mutagenic Red system gives MGEs (and their hosts) and into the molecular biology of how Red may co-opt error-prone polymerases as a means to its ends. A Red revival could be coming.