Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2004 Dec 7; 101(49): 17056–17060.
Published online 2004 Dec 1. doi:  10.1073/pnas.0408046101
PMCID: PMC535401

Strand invasion promoted by recombination protein β of coliphage λ


Studies of phage λ in vivo have indicated that its own recombination enzymes, β protein and λ exonuclease, are capable of catalyzing two dissimilar pathways of homologous recombination that are widely distributed in nature: single-strand annealing and strand invasion. The former is an enzymatic splicing of overlapping ends of broken homologous DNA molecules, whereas the latter is characterized by the formation of a three-stranded synaptic intermediate and subsequent strand exchange. Previous studies in vitro have shown that β protein has annealing activity, and that λ exonuclease, acting on branched substrates, can produce a perfect splice that requires only ligation for completion. The present study shows that β protein can initiate strand invasion in vitro, as evidenced both by the formation of displacement loops (D-loops) in superhelical DNA and by strand exchange between colinear single-stranded and double-stranded molecules. Thus, β protein can catalyze steps that are central to both strand annealing and strand invasion pathways of recombination. These observations add β protein to a set of diverse proteins that appear to promote recognition of homology by a unitary mechanism governed by the intrinsic dynamic properties of base pairs in DNA.

Keywords: genetic recombination, phage λ

“Single-strand annealing” and “strand invasion” are common terms for two molecular paradigms or pathways of homologous recombination that mediate double-strand break repair and are widely distributed in nature (Fig. 1). Bacteriophage λ encodes its own relatively simple recombination system, called “Red,” which comprises three proteins, including an exonuclease that processively degrades single-stranded DNA in the 5′-to-3′ direction (1) and an associated protein, β, that can anneal complementary single strands (2, 3). Under various circumstances, these enzymes promote recombination through either single-strand annealing or strand invasion. Recombination of phage λ has been studied extensively as a model system (4), and, more recently, the activities of the λ enzymes in vivo have been used as potent tools in bioengineering (5).

Fig. 1.
The two paradigms of homologous recombination. (A) Single-strand annealing as exemplified by the activities in vitro of the exonuclease and β protein of phage λ. Small filled black circles represent 5′ ends. (B) Strand invasion ...

In normal genetic crosses, recombination of phage λ is almost entirely independent of the host's RecA protein and occurs largely by single-strand annealing, a pathway for which replication produces suitable substrates, including overlapping double-strand ends (Fig. 1 A) and possibly single-strand gaps located at replication forks. When replication is restricted, however, recombination mediated by λ Red proteins is dependent on RecA and occurs largely by single-strand invasion (Fig. 1B) unless two nonallelic double-strand breaks are introduced (Fig. 1 A), which alleviates the need for RecA and favors single-strand annealing (4, 6). Under other experimental conditions, when the λ Red system must pair a double-stranded end with intact duplex DNA, RecA is required, and recombination is thought to occur by strand invasion (4, 5, 7, 8).

In bioengineering experiments, the incorporation of single-stranded DNA can be accomplished by β protein without the need for either λ exonuclease or Escherichia coli RecA protein (9). In that circumstance, some genetic alterations are best explained by single-strand annealing, whereas others are more readily explained by strand invasion (5).

In addition to its annealing activity in vitro, β protein has been shown to promote strand exchange in a branched oligonucleotide substrate (10). The activities of the two λ proteins, exonuclease and β, plus DNA ligase, can account for a version of single-strand annealing in which two overlapping double-stranded ends can be perfectly spliced to produce intact DNA (Fig. 1). In the splicing reaction simulated in vitro, λ exonuclease plays two roles: (i) it resects the overlapping double-stranded ends to produce complementary single strands, and (ii) it perfectly trims the product of annealing so that excess 3′ single-stranded branches are assimilated into duplex DNA (11, 12). The β protein also may play two roles in splicing DNA molecules. The annealing activity of β protein can join the complementary single strands made by λ exonuclease, and the strand-exchange activity (10) of β protein may drive a concerted reaction by which 5′ ends are presented to λ exonuclease for degradation, whereas 3′ ends are concomitantly assimilated (see Fig. 1).

As indicated in this brief review, previous biochemical studies on β protein can readily be related to single-strand annealing but not to strand invasion, which, however, also has been implicated in recombination mediated by the λ Red system. In this report, we show that β protein can initiate strand invasion in vitro through the formation of displacement loops (D-loops) in superhelical DNA or by strand exchange between colinear single-stranded and double-stranded molecules.

Materials and Methods

DNA Substrates. The sequences of 83-mer single-stranded oligonucleotides G16(–), G26(–), and G37(–) were described in ref. 13. The numbers in each name refer to the percent G+C content, and the complementary strands are denoted as G26(+), etc. Concentrations of DNA are expressed as moles of nucleotides for single strands or moles of base pairs for duplex DNA. Oligonucleotide EG673 was derived from G16(–) by introducing 10 A/T transversions whose locations are underlined in the following sequence: AAATGAACATAAAGTAAATAAGTATAAGGTAAATACTAAATAAGAAAATGAATAAACTTAGTAAATAAAGAAAAGGTAATAAA.

Plasmids pEG47, pEG61, and pEG177 were made by inserting G16(+)·G16(–), G37(+)·G37(–), and G26(+)·G26(–) duplexes, respectively, into 2.8-kb plasmid pIBI30 (International Biotechnologies). Plasmids pEG47, pEG61, and pEG177 were isolated by using a plasmid purification kit (Qiagen, Chatsworth, CA). Covalently closed supercoiled plasmid pEG47 also was isolated by equilibrium centrifugation in CsCl/ethidium bromide gradients (14). There was no detectable difference in the efficiency of D-loop formation supported by these two preparations of substrates; nonetheless, plasmid pEG47 that was isolated from CsCl/ethidium bromide gradients was used for all experiments except those shown in Fig. 5.

Fig. 5.
Effect of G+C content of substrates on formation of D-loops. Reactions were carried out as described in Materials and Methods. The concentration of 32P-labeled oligonucleotides was 2.5 μM, and of β protein was 3 μM, except in the ...

Plasmid pEG47 has a single recognition site for endonuclease HindIII, which was used to linearize the plasmid.

Purification of β Protein. The whole coding sequence of β protein was amplified by PCR from phage λ DNA. The resulting DNA fragment was inserted into expression vector pQE-80L (Qiagen) in frame with a 5′-end sequence coding six histidine residues that function as a metal-binding domain in the translated fusion protein. The resulting plasmid pEG185 was introduced into E. coli DH10B (BRL). Synthesis of the fusion protein, which is under control of the T5 phage promoter, can be induced by isopropyl β-d-thiogalactoside (IPTG). For isolation of β protein, E. coli DH10B/pEG185 cells were grown at 37°C in tryptone broth medium (14) containing 100 μg/ml ampicillin. When the OD600 of the culture reached 1.5, the production of β protein was induced by addition of 1 mM IPTG. Cells were harvested 2.5 h later and lysed by lysozyme treatment, followed by sonication. β protein was purified by nickel-chelate affinity chromatography on Ni-NTA resin (Qiagen) as recommended by the manufacturer. His-tagged β protein, eluted from the Ni-NTA column by 250 mM imidazole, was passed through gel filtration column PD-10 (Bio-Rad), preequilibrated with buffer B (1 mM EDTA/1 mM 2-mercaptoethanol/10% glycerol/20 mM potassium phosphate buffer, pH 7.2). The protein was further purified by hydroxyapatite chromatography by using a gradient of 20–200 mM potassium phosphate. The purified protein consisted of a single species as judged by gel electrophoresis. No degradation of [32P]G16(–) or [32P]G16(–)·G16(+) was observed by polyacrylamide gel electrophoresis after 30 min of incubation with 3 μM β protein in buffer T (see below). Untagged β protein was isolated as described in ref. 15. In control experiments, we did not observe a difference between the ability of His6-tagged vs. untagged β protein to promote D-loop formation and strand exchange.

Formation of D-Loops. Single-stranded DNA oligonucleotides G16(–), G26(–), or G37(–), 32P-labeled at the 5′ end, were incubated at 37°C with β protein in buffer T (20 mM Tris·HCl, pH 7.4/25 mM NaCl/100 μg/ml BSA/0.5 mM DTT). After 10 min, a plasmid carrying a target sequence homologous to one of the above single-stranded oligonucleotides was added. At intervals, aliquots were taken, and the reaction was stopped by adding 0.3% SDS. Samples were subjected to electrophoresis in a 0.8% agarose gel in TBE buffer (14). The gel was dried, and reaction products were quantitated on a PhosphorImager (Molecular Dynamics).

Strand Exchange. Oligonucleotide G16(–) or oligonucleotide EG673 (see above) was incubated with β protein at 37°C in buffer T (see above) plus 2 mM MgCl2. After 10 min, duplex DNA, [5′-32P]G16(–)·G16(+) was added, and after further incubation at 37°C for 40 min, strand exchange was stopped by adding 0.3% SDS. Samples were subjected to electrophoresis in a nondenaturing 12% polyacrylamide gel.

Strand exchange mediated by recA protein was carried out by incubation of oligonucleotide G16(–) with recA protein at 37°C in a buffer consisting of 25 mM Hepes (pH 7.4), 1 mM MgCl2, 1 mM DTT, 1 mM ATP, and 100 μg/ml BSA. After 10 min, the concentration of MgCl2 was increased to 10 mM, and duplex oligonucleotide [5′-32P]G16(–)·G16(+) was added. After further incubation at 37°C for 40 min, the reaction was stopped and analyzed as described above.


Formation of D-Loops by β Protein. To look for the formation of D-loops, we used an A+T-rich 32P-labeled oligonucleotide, G16(–), and superhelical DNA, pEG47, that contained a homologous sequence (Fig. 2). As seen by gel electrophoresis of deproteinized reaction mixtures, a band that migrated more slowly than the labeled oligonucleotide formed within 1 min and reached a yield of 37% within 60 min. A single cleavage of the superhelical DNA by HindIII before incubation of the complete reaction mixture (lanes 4 and 5) or after the reaction (lane 10) completely eliminated the more slowly migrating band. Both of these observations support the identification of the slowly migrating band with superhelical DNA containing a D-loop. Superhelical DNA, pEG177, which lacked a region homologous to G16(–), did not support the formation of D-loops (lanes 1–3). The maximal yield of D-loops occurred at ≈1 molecule of β protein per nucleotide residue (Fig. 3).

Fig. 2.
β protein forms D-loops in supercoiled DNA. The concentrations of 32P-labeled oligonucleotide and β protein were 2.5 μM and 1.5 μM, respectively. The concentration of heterologous superhelical DNA (pEG177; lanes 1–3), ...
Fig. 3.
Stoichiometry of D-loop formation by β protein. The concentrations of [32P]G16(–) ssDNA and plasmid pEG47 dsDNA were 2.5 μM and 0.15 mM, respectively. Reaction conditions were as described in Materials and Methods. All reactions ...

Because RecT protein, a functional analog of β protein, is inhibited by Mg2+ (19), we examined its effect on the formation of D-loops by β protein. As shown in Fig. 4, concentrations of Mg2+ up to 2 mM did not affect the activity of β protein, but higher concentrations inhibited the reaction.

Fig. 4.
Effect of MgCl2 on the formation of D-loops. [32P]G16(–) oligonucleotide and β protein were present at 2.5 μM. G16(–) was present in 1.7-fold excess over homologous sites in the pEG47 plasmid. MgCl2, at the indicated concentrations, ...

Effect of the Base Composition of Substrates on Formation of D-Loops. As the G+C content of substrates was raised from 16% to 37%, the yield of D-loops decreased (Fig. 5). The yield of D-loops at 30 min was 50% for a substrate with 16% G+C but only 4% for a substrate with 37% G+C (Fig. 5). No D-loops were detected when β protein was omitted or when the single-stranded and duplex substrates were heterologous.

Strand Exchange Promoted by β Protein. We previously observed that β protein promoted strand exchange when the duplex substrate had a single-stranded end that was complementary to the end of a homologous single-stranded substrate. After annealing the two complementary ends, β protein was able to propagate exchange through the rest of the DNA molecule, preferentially in the 5′-to-3′ direction with regard to the single-stranded substrate. Mismatched base pairs were used to block spontaneous branch migration and thus demonstrate the active role of β protein in the exchange.

The newly discovered ability of β protein to form D-loops in superhelical DNA, as described above, led us to examine whether β can promote strand exchange between a fully duplex substrate and a single strand (Fig. 6). The substrates in the experiment shown were A+T-rich 83-mer oligonucleotides. As a heterologous or homeologous control, we used the same oligonucleotide with 10 A/T transversions. At 40 min, 41% of the [32P]G16(–) strands originally present in the parental duplex substrate, [32P]G16(–)·G16(+), were present as single strands (Fig. 6), whereas the yield of single strands from the homeologous control was only 4%. For comparison, at 20 min, the yield of single strands from a reaction catalyzed by RecA protein was 75%.

Fig. 6.
Strand exchange promoted by β protein. The substrates were 83-mer oligonucleotides G16(–) or EG673, the latter a derivative of G16(–) containing 10 A/T transversion mismatches, to provide a heterologous control, and duplex [5′- ...

The result just presented could have resulted from bona fide strand exchange or from a previously undetected melting or helicase activity of β protein. To exclude the latter explanation, we incubated β protein with [32P]G16(–)·G16(+) (1.2 μM) for 20 min under the standard conditions for strand exchange, after which we added a 7-fold excess of unlabeled G16(–) (8.5 μM) together with stop solution. Had β protein melted [32P]G16(–)·G16(+), the excess of unlabeled G16(–) strands added with the stop solution should have competitively annealed with G16(+) and produced single-stranded [32P]G16(–), but no detectable single strands were produced (Fig. 7, lane 3). We have previously shown that renaturation of G16 strands readily occurs under the conditions of this experiment in the presence of stop solution (20). Thus, we conclude that β protein, acting on oligonucleotide substrates, promotes bona fide strand exchange.

Fig. 7.
Strand exchange (S.E.) mediated by β protein is not due to helicase activity. Lanes 1 and 2, strand exchange at 0 and 40 min, respectively, was carried out as described in Materials and Methods. The concentration of dsDNA was 1.2 μM, ssDNA ...

As expected from the observations on the formation of D-loops, the yield of strand exchange reactions was reduced by a substrate of higher G+C content (26% G+C, data not shown).

The time courses of D-loop formation and strand exchange are shown in Fig. 8. Although the conditions and substrates for the two reactions were not directly comparable, it is of interest to note that the maximum yield of D-loops was reached within a few minutes, whereas strand exchange was an order of magnitude slower, which is consistent with the relative rates of formation of synaptic complexes vs. the completion of strand exchange as measured in stopped-flow experiments on RecA protein (16).

Fig. 8.
The kinetics of D-loop formation and strand exchange. In the reaction for D-loop formation, β protein was present at a ratio of one molecule of protein per nucleotide residue of ssDNA. The latter, [32P]G16(–) 83-mer (final concentration, ...


The Red genes of bacteriophage λ have been implicated biologically in the two universal paradigms of homologous recombination, single-strand annealing and strand invasion (4, 5). Single-strand annealing is a molecular splice mediated by the annealing of complementary single strands; the most detailed knowledge of its enzymology comes from studies in vitro of λ exonuclease and β protein (Fig. 1A) (4). The concept of strand invasion is more complex. Even in its simplest form, it involves the recognition of homology in duplex DNA by an “invading” single strand. The prototype for this reaction was supplied by the discovery that E. coli RecA protein catalyzes the formation of D-loops in superhelical DNA (17). The nature of that invasion, however, remained unclear until the serendipitous discovery that base composition powerfully affects both invasion and strand exchange (13, 18). Further exploration led to the conclusion that recognition of homology, involving the invasion of duplex DNA by a single strand, is mediated preferentially by the exchange of A·T base pairs (13, 16). Moreover, a diverse group of proteins, including E. coli proteins RecA and RecT, human proteins Rad51, Dmc1, and Rad52, and yeast Rad52, all promote strand invasion by a related mechanism as indicated by the preferential role of A·T base pairs (13, 1921). A recent study provided direct physical evidence that in reactions catalyzed by E. coli RecA protein, the exchange of A·T base pairs occurs fast enough to account for strand invasion and for the consequent formation of synaptic complexes containing three strands of DNA that remain together until the rest of the original base pairs undergo exchange (Fig. 1B) (16).

The present experiments show that the ATP-independent β protein can also promote strand invasion that depends on A·T base pairs. This result provides a plausible connection between the biochemical properties of β protein and the observed roles of the λ recombination enzymes in strand invasion as variously seen in phage recombination and in bioengineering experiments that make use of those enzymes. In addition, similar biological and biochemical roles are played by β protein and RecT protein (19), which are evolutionarily related (22). Taken together, the cited experiments make the case that strand invasion and at least the initiation of strand exchange are closely related mechanistically, although a large literature demonstrates that propagation of strand exchange over kilobase distances is mechanistically more complex (23).

Finally, we return to the observation that the genetically unrelated and structurally diverse proteins cited above can catalyze strand invasion in vitro. This finding, of course, does not mean that all of these proteins play equivalent biological roles but rather suggests that they have the potential to play wider roles than previously supposed. The diversity of proteins that can catalyze strand invasion further suggests that the structure and intrinsic dynamic properties of DNA govern invasion, for which various unrelated proteins can provide catalytic surfaces. The dominant role of DNA structure in strand invasion has been proposed previously on the basis of various observations: on the structure of DNA within RecA and Rad51 filaments (24), on the topological structure of various nucleoprotein filaments (25), and on the mechanism of strand invasion (16).


We thank Jan Zulkeski for administrative assistance. This work was supported by National Institute of General Medical Sciences Grant R37-GM33504.


Author contributions: N.R., E.I.G., B.B., and C.M.R. designed research; N.R., E.I.G., and B.B. performed research; N.R., E.I.G., B.B., and C.M.R. analyzed data; and C.M.R. wrote the paper.

Abbreviation: D-loop, displacement loop.


1. Little, J. W. (1967) J. Biol. Chem. 242, 679–686. [PubMed]
2. Kmiec, E. & Holloman, W. K. (1981) J. Biol. Chem. 256, 12636–12639. [PubMed]
3. Muniyappa, K. & Radding, C. M. (1986) J. Biol. Chem. 261, 7242–7478. [PubMed]
4. Kuzminov, A. (1999) Microbiol. Mol. Biol. Rev. 63, 751–813. [PMC free article] [PubMed]
5. Court, D. L., Sawitzke, J. A. & Thomason, L. C. (2002) Annu. Rev. Genet. 36, 361–388. [PubMed]
6. Stahl, M. M., Thomason, L., Poteete, A., Tarkowski, T., Kuzminov, A. & Stahl, F. W. (1997) Genetics 147, 961–977. [PMC free article] [PubMed]
7. Clyman, J. & Belfort, M. (1992) Genes Dev. 6, 1269–1279. [PubMed]
8. Poteete, A. R. & Fenton, A. C. (1993) Genetics 134, 1013–1021. [PMC free article] [PubMed]
9. Ellis, H. M., Yu, D., DiTizio, T. & Court, D. L. (2001) Proc. Nat. Acad. Sci. USA 98, 6742–6746. [PMC free article] [PubMed]
10. Li, Z., Karakousis, G., Chiu, S. K. & Radding, C. M. (1998) J. Mol. Biol. 276, 733–744. [PubMed]
11. Cassuto, E. & Radding, C. M. (1971) Nat. New Biol. 229, 13–16. [PubMed]
12. Cassuto, E., Lash, T., Sriprakash, K. S. & Radding, C. M. (1971) Proc. Natl. Acad. Sci. USA 68, 1639–1643. [PMC free article] [PubMed]
13. Gupta, R. C., Folta-Stogniew, E., O'Malley, S. O., Takahashi, M. & Radding, C. M. (1999) Mol. Cell 4, 705–714. [PubMed]
14. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY), 2nd Ed.
15. Karakousis, G., Ye, N., Li, Z., Chiu, S. K., Reddy, G. & Radding, C. M. (1998) J. Mol. Biol. 276, 721–731. [PubMed]
16. Folta-Stogniew, E., O'Malley, S., Gupta, R. C., Anderson, K. S. & Radding, C. M. (2004) Mol. Cell 15, 965–975. [PubMed]
17. Shibata, T., DasGupta, C., Cunningham, R. P. & Radding, C. M. (1979) Proc. Natl. Acad. Sci. USA 76, 1638–1642. [PMC free article] [PubMed]
18. Gupta, R. C., Folta-Stogniew, E. & Radding, C. M. (1999) J. Biol. Chem. 274, 1248–1256. [PubMed]
19. Noirot, P., Gupta, R. C., Radding, C. M. & Kolodner, R. D. (2003) EMBO J. 22, 324–334. [PMC free article] [PubMed]
20. Bi, B., Rybalchenko, N., Golub, E. & Radding, C. (2004) Proc. Nat. Acad. Sci. USA 101, 9568–9572. [PMC free article] [PubMed]
21. Kumar, J. K. & Gupta, R. C. (2004) Proc. Nat. Acad. Sci. USA 101, 9562–9567. [PMC free article] [PubMed]
22. Iyer, L., Koonin, E. & Aravind, L. (2002) BMC Genomics 3, 1471–2164.
23. Cox, M. M. (2003) Annu. Rev. Microbiol. 57, 551–577. [PubMed]
24. Shibata, T., Nishinaka, T., Mikawa, T., Aihara, H., Kurumizaka, H., Yokoyama, S. & Ito, Y. (2001) Proc. Nat. Acad. Sci. USA 98, 8425–8432. [PMC free article] [PubMed]
25. Egelman, E. H. (2001) J. Mol. Biol. 309, 539–542. [PubMed]
26. Symington, L. S. (2002) Microbiol. Mol. Biol. Rev. 66, 630–670. [PMC free article] [PubMed]
27. Paques, F. & Haber, J. E. (1999) Microbiol. Mol. Biol. Rev. 63, 349–404. [PMC free article] [PubMed]
28. Kowalczykowski, S. C. & Eggleston, A. K. (1994) Annu. Rev. Biochem. 63, 991–1043. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.
  • Taxonomy
    Taxonomy records associated with the current articles through taxonomic information on related molecular database records (Nucleotide, Protein, Gene, SNP, Structure).
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...