Two neglected but valuable genetic tools for Escherichia coli and other bacteria: In vivo cosmid packaging and inducible plasmid replication

In physiology and synthetic biology, it can be advantageous to introduce a gene into a naive bacterial host under conditions in which all cells receive the gene and remain fully functional. This cannot be done by the usual chemical transformation and electroporation methods due to low efficiency and cell death, respectively. However, in vivo packaging of plasmids (called cosmids) that contain the 223 bp cos site of phage λ results in phage particles that contain concatemers of the cosmid that can be transduced into all cells of a culture. An historical shortcoming of in vivo packaging of cosmids was inefficient packaging and contamination of the particles containing cosmid DNA with a great excess of infectious λ phage. Manipulation of the packaging phage and the host has eliminated these shortcomings resulting in particles that contain only cosmid DNA. Plasmids have the drawback that they can be difficult to remove from cells. Plasmids with conditional replication provide a means to “cure” plasmids from cells. The prevalent conditional replication plasmids are temperature‐sensitive plasmids, which are cured at high growth temperature. However, inducible replication plasmids are in some cases more useful, especially since this approach has been applied to plasmids having diverse replication and compatibility properties.

plasmid-encoded genes upon entry.In contrast, infection with bacteriophages can be highly efficient.All cells of a culture can be infected.Moreover, to enable phage replication, evolution has optimized the infection process to minimalize metabolic disturbance of the host.This approach allows completely off-to-on transcriptional control.The cosmid-transduced genes enter a naive cytosol, which bypasses concerns of toxic genes and leakiness of regulated promoters.Examples are the λCE6 phage (see below) and dealing with leaky promoters by off-to-on expression in lipoic acid synthesis, which requires only minute enzyme levels (Zhao et al., 2003).In the latter study, the question was whether a protein modified by attachment of octanoic acid could be converted to a lipoic acid modified protein.Was the octanoylated protein the direct precursor to the lipoylated protein?This required accumulation of the octanoylated protein in the absence of lipoic acid synthase.Following octanoylated protein accumulation, lipoic acid synthase would then be restored to test the hypothesis.The straightforward approach would have been to delete the lipoic acid synthase gene and put synthesis of lipoic acid synthase under a tightly controlled promoter.However, none of the available tightly controlled promoters were sufficiently tight to prevent conversion of the octanoylated protein, which precluded a strictly controlled experiment.The solution was to provide lipoic acid synthase by transformation with a λ phage encoding the enzyme into all cells (Zhao et al., 2003).Mass spectroscopy of the purified proteins demonstrated that introduction of lipoic acid synthase gave direct conversion of the octanoylated protein to lipoic acidmodified (lipoyl) protein.These experiments proved that lipoic acid is assembled on the enzyme proteins on which it functions, a previously unknown synthetic mechanism.If this experiment had been done by electroporation or chemical transformation with a lipoic acid synthase plasmid, the background of the dead and untransformed cells would have obscured the results.This is because the octanoylated protein in these cells would not be converted to lipoylated protein but would be present and perhaps dominate the mass spectra.Antibiotic selection of the electroporated cultures would lose the time resolution required for proof of the hypothesis.The surviving cells would make lipoylated protein, but the mechanism would be unclear.
To take advantage of the properties of phage infection, bacterial genes have been introduced into the genomes of several phages, the most popular being phages M13 (and close relative phages f1 and fd) and λ.Recombinant versions of these phages can be used to introduce bacterial and foreign DNA segments into the cells of growing Escherichia coli cultures (Jones et al., 1986;Studier & Moffatt, 1986;Tabor, 2001).However, phages have advantages and disadvantages as DNA introduction vehicles.A major advantage of phage M13 is that the double-stranded replicative form of the phage genome can be treated as a plasmid in cloning procedures and vectors containing a multiple cloning site plus lacZ blue-white screening are available (Jones et al., 1986;Tabor, 2001).Moreover, plasmid derivatives that carry the M13 phage replication origin and DNA packaging determinant (called phagemids) can be replicated and packaged into M13 virions by use of a helper phage (Qi et al., 2012).However, a major disadvantage is that infection of E. coli with M13 is slow and inefficient.
The filamentous phage particles have the task of binding end on to pili encoded by the F factor extending from the cell and there is less than one F pilus per cell (Daehnel et al., 2005).This results in mixed populations of uninfected and asynchronously infected cells.Moreover, M13 DNA replication is of low fidelity and deleted derivatives of the viruses or phagemids often accumulate (Ehrlich et al., 1993;Michel & Ehrlich, 1986;Vilette et al., 1996).
Phage λ has different advantages and disadvantages.Unlike M13, phage λ rapidly and efficiently infects its host.However, phage λ has a large genome (48.5 kb) and, thus, unique restriction sites are few and difficult to introduce.This results in challenging cloning into specialized vectors that have been deleted for nonessential λ genes to make room for introduced DNA.Bacterial genes can also be inserted into the λ genome by homologous recombination, but this requires construction of suitable acceptor phages and donor plasmids plus a screen or selection to identify the recombinant phages.The utility and construction of such λ transducing phages are demonstrated by the λCE6 phage of Studier and Moffatt (1986).The expression of plasmids encoding toxic genes using the powerful T7 RNA polymerase was problematical because leaky expression of the polymerase killed the host.To cope with this problem, Studier and Moffat constructed λCE6, which encodes T7 RNA polymerase.Upon delivery of T7 RNA polymerase by λCE6 transduction, very toxic target genes that could not be expressed previously were rapidly expressed at high levels.
Phage λCE6 is commercially available and provides a useful tool.
However, the construction of the phage required in vitro ligation of the T7 RNA polymerase gene into a minimized λ phage vector followed by in vivo recombination and screening to introduce two mutant λ genes needed for efficient phage production (Studier & Moffatt, 1986).Note that this approach has become more difficult because cloning into λ vectors has fallen into disuse.This is because the main advantage of λ vectors was the ability to clone DNA fragments up to about 20 kb (Loenen & Brammar, 1980).
However, artificial chromosome BAC and PAC plasmid vectors can carry inserts of 150-350 kb (Biradar et al., 2014;Osoegawa et al., 2001).Moreover, BAC and PAC plasmid vectors are available from a central repository (Addge ne.org), whereas λ vectors have no such repository.
In contrast to λ cloning, plasmid cloning is the dominant technique of molecular genetics.In vivo packaging of plasmids containing the cos region of λ (called cosmids) into λ particles has the advantages of λ clones such as λCE6 in expression of toxic genes but has the ease of plasmid cloning.Cosmids were described many years ago (Collins, 1979;Collins & Hohn, 1978, 1992).The 223 bp cos region of λ contains the determinants for processive packaging of the oligomeric phage DNA and includes the site (cos) at which the cohesive ends of λ virion DNA are generated (Catalano et al., 1995;Cue & Feiss, 2001).When the cos sequence is present in a plasmid and the plasmid is of sufficient length (~37 to 52 kb), the plasmid can be packaged into phage particles either in vitro (Collins, 1979;Collins & Hohn, 1978, 1992) or in vivo (Feiss et al., 1982).In vitro packaging is often used for preparation of clone banks, but the titers (~10 8 /mL) of in vitro packaged cosmids are too low to transduce sufficient cell mass for biochemical or physical analyses.In vivo packaging of cosmids upon induction of λ lysogens (or infection with λ) was discovered over 20 years ago (Feiss et al., 1982;Miwa & Matsubara, 1982, 1983) but has only occasionally been utilized in genetic analysis.The major drawbacks of these in vivo cosmid packaging systems were low yields of packaged cosmids and large excesses (10-to 1000-fold) of infectious phage particles over the particles containing cosmid DNA (Feiss et al., 1982;Miwa & Matsubara, 1982, 1983).
In a first attempt to construct a system that would package only cosmid DNA, the strict spacing of the packaging phage cos region was disrupted by insertion of a kanamycin resistance cassette into the cos sequence (Cronan, 2003).The phage was inserted into the E. coli chromosomal λatt site to become a prophage.The prophage retained the λ gam and red genes, required for efficient packaging of small cosmids (Figure 1).The gam gene product is required to inactivate the host recBC nuclease.This allows the rolling circle replication required to synthesize concatemers of small cosmids that are sufficiently large (37 to 52 kb) to be packaged into λ particles (Miwa & Matsubara, 1982, 1983).Induction of λ DNA replication in a host strain shifts cosmid molecules into the rolling circle mode of DNA replication used by λ, although the mechanism is not well understood (Shimada et al., 1979;Umene et al., 1978Umene et al., , 1979)).Expression of the λred genes gives a marked increase in the level of packageable cosmids because λred-catalyzed homologous recombination is extremely efficient (Court et al., 2002;Sharan et al., 2009).λRed generates packageable DNA molecules by recombining DNA molecules that are not otherwise substrates for packaging (e.g., the circles resulting from conventional plasmid θ replication) (Miwa & Matsubara, 1983).Inactivation of the prophage cos site resulted in a very large increase in cosmid packaging probably due to lack of competition with phage DNA packaging.However, a variable background of infectious phage particles was present.The infectious phage resulted from repair of the altered phage cos region by homologous recombination with either the cosmid cos sequence or the cos sequence of one of the cryptic defective prophages in the E. coli K-12 chromosome (Cronan, 2003(Cronan, , 2013;;Fisher & Feiss, 1980) (Redfield & Campbell, 1987).Repair was catalyzed by the very efficient λRed recombination required for rolling circle replication and cosmid packaging.Since λRed recombination could not be eliminated from the packaging phage, the only means to prevent recombinational repair of the altered cos was removal of all cos regions except that of the cosmid.
In later work, replacement of the entire cos region of a λ prophage with a chloramphenicol resistance cassette prevented repair by recombination with the cosmid cos site (Cronan, 2013).However, a background of infectious λ phage remained.Two defective cryptic prophages Qin and DLP12 (also called Qsr) (Redfield & Campbell, 1987) resident in the E. coli K-12 genome contain cos sites plus extensive flanking regions essentially identical to those of λ and thus are excellent substrates for recombinational repair of the deleted cos region of the packaging phage (Redfield & Campbell, 1987).The sequences that bracket the cos region of the packaging phage could not be deleted because these genes encode particle structural proteins.Hence an E. coli K-12 strain lacking the λ-like defective cryptic prophages was required.Fortuitously, in the studies of the effects of cryptic prophages on host physiology, Wang and coworkers deleted all nine cryptic prophages from an E. coli strain called BW25113∆9 (Wang et al., 2010).Strain BW25113∆9 was lysogenized with the ∆cos chloramphenicol-resistant λ phage to give the cosmid packaging strain.In this strain, recombinational repair is prevented because the only cos site present is that of the cosmid to be packaged (Cronan, 2013).
The packaging prophage contains two mutations that facilitate production of cosmid-containing particles (Cronan, 2013).A temperature-sensitive mutation (cI 857 ) in the λcI repressor allows induction of phage replication by temperature shift from 30°C to 42°C and a mutation in the S lysis gene (Sam7) that prevents lysis of the induced cells (Altman et al., 1983;Goldberg & Howe, 1969).
The S gene mutation allows the induced cells to be concentrated by centrifugation and then lysed by treatment with chloroform.Preparations of 10 11 -10 12 cosmid transduction units with no detectable lytic phage were routinely obtained (Cronan, 2013).Addition of 5-10 cosmid packaged particles per cell gave cosmid-encoded β-galactosidase levels that were essentially the same as those of the same strain carrying the resident cosmid indicating that virtually all cells received a cosmid (Cronan, 2013).In other experiments, a toxic protein encoded by a cosmid was expressed at high level (Cronan, 2013).The cosmid titer needed will depend on how many cells are necessary to obtain the required data.For example, production of a protein to be purified would require at least a 100 mL culture at ~5 × 10 8 cells/mL so ~5 × 10 11 cosmid particles would be required.If the assay is radioisotope incorporation, then a large decrease in scale would be indicated.
The receptor on the E. coli outer surface that λ binds to initiate infection is called LamB (Randall-Hazelbauer & Schwartz, 1973).
LamB is a component of the maltose operon where it facilitates diffusion of maltose and maltodextrin across the outer membrane.& Schwartz, 1973).Some close relatives of E. coli (Shigella and Enterobacter) express functional LamB receptors but the LamB protein of Salmonella enterica, another close relative, is not a λ receptor (Palva et al., 1987).However, transformation of S. enterica with plasmids encoding E. coli LamB allows λ to inject cosmids into this bacterium (Palva et al., 1987).This approach has been successfully extended to Vibrio cholerae, Erwinia carotovora, Klebsiella pneumoniae, Yersinia enterocolitica, Photorhabdus luminescens, and Agrobacterium tumefaciens (Brzostek et al., 1995;Palva et al., 1987;Tang et al., 2010).In some gram-negative bacteria, LamB fails to express well but mutations can be selected that allow LamB expression and λ adsorption (Ludwig, 1987).Hence,  2021)).After induction, the phage genome is excised from the host chromosome by Int/Xis action to give a circular molecule containing a cos site.This molecule first undergoes conventional θ bidirectional replication, which switches to rolling circle (σ) replication after sufficient Gam has been made to inactivate the host RecBCD nuclease.Rolling circle replication continues to make long double-stranded concatemers containing several head-to-tail copies of the λ genome until a cos site is bound by λ terminase.Terminase introduces nicks on opposite strands offset by 12 bp within the cos sequence and separates the cos strands to form the first cohesive end, which remains bound to terminase (Catalano, 2000).This complex then binds an empty phage head (called a prohead) after its assembly and begins to translocate the DNA into the prohead (both the translocation ATPase and the concatemer-cutting endonuclease reside in terminase).As the head expands and fills, the bound terminase scans the translocated DNA for a cos that is appropriately spaced to give a full head.When the head is full, terminase cuts the second cos site to generate the other cohesive end of the packaged λ genome (Catalano, 2000).Terminase then dissociates from the filled heads and the λ tails are attached to produce mature infectious viruses containing the linear λ genome (Catalano, 2000).(Panel b) Model of conversion of newly transduced linear cosmid concatemers to multimeric covalently closed plasmids in the recipient bacterial cell.Upon entry of cosmid concatemers into the host cell the cohesive ("sticky") ends generated in packaging anneal to form a circular DNA having two nicks.The nicks are sealed by the host DNA ligase followed by the host gyrase, which converts the circular DNA to negative supercoils.The sizes of these molecules will vary within the λ packaging constraints (37 to 52 kb) and the size of the cosmid.In the cartoon, the cosmid packaged would be 9-10 kb in size.The terminase packaging motor skips over cos sites until the head is full at which time the second cos is recognized and cleaved (Catalano, 2000).Circular concatemer molecules are stable but host RecA recombination will successively decrease the size of the molecules by recombination between the direct repeats (Miwa & Matsubara, 1983).In a formal sense, these are recombination-mediated deletion events but since the deletion products have functional replication origins, they are not lost.

Hazelbauer
pHSG415r replicate at 30°C but not at 42°C.However, these plasmids tend to be large and of low copy number, which complicates cloning procedures.Moreover, these plasmids cannot be used to generate temperature-sensitive (Ts) mutations to study essential genes.The mutation in the pHSG415r-derived plasmids is a point mutation in the rep gene that can revert, which complicates the curing process.Although pHSG415r is reported to replicate well at 30°C (Hashimoto-Gotoh et al., 1981), recombinant derivatives often grow poorly at 30°C, which provides selection for reversion of the rep gene mutation (Hashimoto-Gotoh et al., 2000).An example is the plasmid pCP20 (Cherepanov & Wackernagel, 1995) utilized in some recombineering protocols, which is often lost from 30°C cultures grown without selection.Plasmids requiring induction for replication are an excellent alternative to the pHSG415r derived plasmids.Gil and Bouche (1991) pioneered inducible replication plasmids by constructing plasmids that required addition of isopropyl β-D-1thiogalactopyranoside (IPTG) for replication.These authors constructed plasmid pAM34, a pBR322-derived ampicillin-resistant vector in which transcription of RNA II, the primer of pBR322 DNA replication, is driven by the lacZYA promoter in place of the native pBR322 promoter (Figure 2) (Gil & Bouche, 1991).The lac-ZYA promoter is regulated by the LacI repressor encoded on the plasmid.Although valuable, plasmid pAM34 has several disadvantages.The plasmid contains the most widely used the plasmid replication origin and antibiotic resistance determinant often present in other plasmids and cannot be used in glucose-containing F I G U R E 2 Inducible plasmid replication.Plasmid pAM34 (upper panel) was constructed by Gil and Bouche (1991) by replacing the native pBR322 RNA II promoter with the lacZYA promoter.This placed plasmid replication under control of the LacI repressor encoded on the plasmid (the high expression LacI q allele was used) (Gil & Bouche, 1991).RNA II is the primer for DNA polymerase I catalyzed replication of the plasmid.RNA I transcribed from a promoter on the opposite strand is an antiprimer, which controls copy number by annealing to RNA II and blocking the priming of DNA replication.The small (63 residue) protein Rop (also called Rom) dimerizes, binds, and stabilizes the RNA II-RNA I hybrid providing further copy number control.In the cases where the pBR322 origin was replaced with those of the related plasmids, p15a and RSF1030, RNA I primer regulation is retained although Rop regulation is absent.Rop does not bind these RNA II-RNA I hybrids because they have shapes that differ from that of the pBR322 RNA II-RNA I hybrid and Rop recognizes the shape rather than the sequence of the hybrid (Predki et al., 1995).RNA I is also responsible for the incompatibility seen when two different plasmids having the same origin are present in the same cell (Tomizawa & Itoh, 1981).Plasmids with pBR322, p15a, and RSF1030 origins are all compatible with one another because they encode different RNA I molecules (Davison, 1984;Selzer et al., 1983).The aadA spectinomycin resistance cassette is a "stuffer" included to expedite cloning using sites in the flanking multiple cloning sites (Gil & Bouche, 1991).The gene designated as bla is the β-lactamase of pBR322 responsible for ampicillin resistance.Lower panel: In plasmid pCY1108 the entire lacZYA promoter and lacI gene of pAM34 were replaced with the Tn10 tetracycline regulatory system.This placed plasmid replication under control of the TetR repressor (Cronan, 2016).media (see below).To cope with the first two disadvantages, the repertoire of IPTG-dependent plasmids was recently expanded by construction of IPTG-dependent vectors having different origins of replication, those of p15A and RSF1030 (Srinivas et al., 2019).These plasmids are from the same family as pBR322 and use the same replication mechanism, but the origin sequences differ such that each of these vectors is compatible with pBR322-derived plasmids and with one another (Srinivas et al., 2019).Surprisingly when propagated in the absence of IPTG, these new vectors were cured from the host cells significantly more rapidly and efficiently than pAM34 (Srinivas et al., 2019).In addition, versions of pAM34 were constructed that encode resistance to commonly used antibiotics in place of ampicillin (Srinivas et 2019).

| TE TR AC YCLINE-DEPENDENT REPLI C ATI ON PL A S MIDS
The lacZYA promoter of pAM34 contains the binding site for the complex of the CAP cyclic-AMP activator protein with cyclic-AMP.The low cyclic-AMP levels in cultures grown with glucose (the prevalent carbon source used in minimal media) shut down the lacZYA promoter even in the presence of IPTG (Aggarwal & Narang, 2022).To cope with the need to avoid glucose using the IPTG-dependent vectors, the lac regulatory system of pAM34 was replaced with the well-understood transposon Tn10 tetracycline (Tet) regulatory system (Figure 2) (Cronan, 2016).The Tn10 Tet regulatory system has been used to control gene expression in diverse bacteria as well as in eukaryotes (Schonig et al., 2010).
The Tn10 Tet regulatory system consists of two genes, tetA and tetR.The tetA gene encodes a TetA(B) tetracycline efflux pump, whereas tetR encodes a repressor protein inactivated upon binding of tetracycline or tetracycline analogs.The two genes are adjacent and transcribed bidirectionally.The pAM34 lac promoter and lacI Q repressor were replaced with the tetA promoter and tetR repressor sequences, thereby placing transcription of RNA II under Tet control (Cronan, 2016).The induction of RNA II requires only nontoxic (ng/mL) levels of either tetracycline or anhydrotetracycline.Curing of these plasmids by plating in the absence of inducer produced tiny colonies, which required replating to obtain plasmid free strains (Cronan, 2016).However, if the strain contained a tetracycline efflux pump, withdrawing inducer gave essentially immediate curing.The efflux pump could be encoded either on the chromosome or on a readily cured plasmid (Cronan, 2016).Note that anhydrotetracycline is not a tetracycline pump substrate (Chopra et al., 1982;Rasmussen et al., 1991).Versions of the tetracycline-dependent plasmids encoding resistance to commonly used antibiotics in place of ampicillin were constructed.
Plasmids that use the replication origins of the pBR322 family of plasmids have a narrow host range, probably limited to Salmonella enterica, various Shigella strains, and some Klebsiella and Yersinia strains.However, an inability to transform a plasmid of this family into a bacterial species cannot be taken as a failure to replicate since other factors posit (restriction, modification, Crispr, etc.).The wellstudied pBR322 family of plasmids is unusual in that replication is unidirectional and thus readily regulated via transcription.The unidirectional replication probably facilitates the rolling circle replication of the cosmids triggered by phage λ.Indeed, the bidirectional replication of the F factor may account for lack of packaging of small F plasmid-derived cosmids (Miwa & Matsubara, 1982).Bidirectional replication is found in many broad host range plasmids and would be considerably more difficult to regulate than the pBR322 family plasmids.
In conclusion, both in vivo cosmid packaging and inducible plasmid replication are valuable tools that should now be very useful in physiology and synthetic biology.Indeed, the two tools can be combined.
IPTG-dependent plasmids containing cos are efficiently packaged and transduced (Cronan, 2003).Therefore, a cosmid can be efficiently delivered into a host and then efficiently cured from the host.
cosmid transduction could be broadened to Pseudomonas and Rhizobium strains.REPLI C ATI ON Plasmids are a key tool of bacterial genetics but have the shortcoming that elimination from the host is often difficult.The ability to efficiently eliminate (cure) a plasmid when the product encoded by the plasmid is no longer needed or would become problematical (e.g., Cas9) is limited.A widely utilized temperature-sensitive replication plasmid origin is that of pHSG415r, a derivative of plasmid pSC101 (Hashimoto-Gotoh et al., 1981).Plasmids derived from F I G U R E 1 λ DNA replication and cosmid packaging.(Panel a) Outlines the mechanism of λ DNA replication.For details, see Catalano (2000), Catalano et al. (1995), Catalano and Morais (