Saccharomyces cerevisiae Red1 protein exhibits nonhomologous DNA end–joining activity and potentiates Hop1-promoted pairing of double-stranded DNA

Elucidation of the function of synaptonemal complex (SC) in Saccharomyces cerevisiae has mainly focused on in vivo analysis of recombination-defective meiotic mutants. Consequently, significant gaps remain in the mechanistic understanding of the activities of various SC proteins and the functional relationships among them. S. cerevisiae Hop1 and Red1 are essential structural components of the SC axial/lateral elements. Previous studies have demonstrated that Hop1 is a structure-selective DNA-binding protein exhibiting high affinity for the Holliday junction and promoting DNA bridging, condensation, and pairing between double-stranded DNA molecules. However, the exact mode of action of Red1 remains unclear, although it is known to interact with Hop1 and to suppress the spore viability defects of hop1 mutant alleles. Here, we report the purification and functional characterization of the full-length Red1 protein. Our results revealed that Red1 forms a stable complex with Hop1 in vitro and provided quantitative insights into their physical interactions. Mechanistically, Red1 preferentially associated with the Holliday junction and 3-way junction rather than with single- or double-stranded DNA with overhangs. Although Hop1 and Red1 exhibited similar binding affinities toward several DNA substrates, the two proteins displayed some significant differences. Notably, Red1, by itself, lacked DNA-pairing ability; however, it potentiated Hop1-promoted intermolecular pairing between double-stranded DNA molecules. Moreover, Red1 exhibited nonhomologous DNA end–joining activity, thus revealing an unexpected role for Red1 in recombination-based DNA repair. Collectively, this study presents the first direct insights into Red1's mode of action and into the mechanism underlying its role in chromosome synapsis and recombination.

Meiosis is a specialized form of cell division during which the diploid cells undergo a single round of DNA replication followed by two rounds of chromosome segregation to produce haploid gametes. During meiosis I, homologous chromosomes, one from each parent, pair up and exchange genetic material by crossing over (1,2). In most eukaryotes, a meiosis-specific, dense, proteinaceous structure, called the synaptonemal com-plex (SC), 2 is necessary for crossing over (1,2). A mature SC consists of two parallel structures, a pair of lateral elements that run along the entire length of the paired chromosomes and a centrally located structure (the central element), which are interconnected by transverse filaments (1,2). Although the ultrastructure of the SC is evolutionarily conserved, the protein components share a very low level of amino acid sequence homology (2)(3)(4).
S. cerevisiae Hop1, a structural component of the axial/lateral elements of SC, contains a Cys2/Cys2 zinc finger motif that is essential for its function (7,8,31). Previous studies have

Overexpression and purification of the full-length S. cerevisiae Red1 protein
Full-length untagged S. cerevisiae Red1 was expressed in and purified from Escherichia coli BL21 (DE3) CodonPlus-RIL cells. An analysis of whole-cell lysates from uninduced and induced cell cultures by SDS-PAGE followed by Coomassie Blue staining showed the presence of Red1 of 95.5 kDa in the pellet and soluble fraction (Fig. 1A). The band corresponding to 95.5 kDa was undetectable in the whole-cell lysates of cells grown in the absence of isopropyl ␤-D-1-thiogalactopyranoside. Using conventional chromatographic methods, untagged Red1 protein was purified to apparent homogeneity (Fig. 1A, lane 5). In addition to the 95.5 kDa band, a proteolytic breakdown product of Red1 with a lower molecular weight was seen on an SDS-polyacrylamide gel. This premise was corroborated by Western blot analysis using anti-Red1 antibodies (Fig. 1B). The amount of full-length untagged Red1 in the protein preparation(s) used for the elucidation of its functional activities, which are described below, was Ͼ90%.

Affinity and kinetic measurements of the Red1 and Hop1 interaction
Genetic, co-immunoprecipitation, and co-localization studies have demonstrated the physical interaction between Red1 and Hop1 proteins (12,43,46). A systematic mutant screen has identified a mutation in RED1 (red1-K348E) that inhibited its interaction with Hop1, but it did not affect its homo-oligomerization (49). The yeast two-hybrid assay with lexA-RED1 and Yeast Red1 is a structure-selective DNA-binding protein  variant forms exhibited a strong positive signal, indicating that the Red1 homo-oligomerization is mediated via its C-terminal region (12,49). However, the affinity and kinetics of their interaction have not been investigated.
To confirm that the purified untagged protein was Red1, its ability to participate in protein-protein interactions with Hop1 was tested. For this purpose, two independent assays were used: far Western blotting and surface plasmon resonance (SPR). The interaction of Red1 with Hop1 was determined by a far Western blot assay using Red1 as a probe. The S. cerevisiae DNA damage checkpoint control protein Rad17 was used as a negative control. Red1 was observed to specifically interact with Hop1 in the assay, whereas Rad17 failed to do so ( Fig. 2A). For an equilibrium analysis of affinity, increasing amounts of Red1 were added over immobilized Hop1 on a BIAcore CM5 sensor chip surface. The SPR response increased with increasing analyte concentration and reached equilibrium rapidly, reflecting fast kinetics. The representative sensorgrams are shown in Fig. 2B. The SPR response data were corrected for bulk refractive index effects and fitted using a 1:1 Langmuir binding model (A ϩ B 7 AB). An analysis of the affinity and kinetic parameters indicated that Red1 binds quantitatively to Hop1 with a higher on-rate (2.44 ϫ 10 4 M Ϫ1 s Ϫ1 ) and slower off-rate (5.06 ϫ 10 Ϫ3 s Ϫ1 ). The apparent dissociation constant (K d ) value derived from these data for the binding of Red1 to Hop1 was ϳ0.18 M. These findings support the conclusion that the purified Red1 can interact in vitro with Hop1 and could also have biological activities.

Red1 exhibits high affinity for DNA recombination intermediates
Previous studies have revealed preferred DNA-binding sites for Red1 along the meiotic chromosomes (41,42). For instance, the frequency and distribution analysis of Red1 foci in DNase I-digested chromosome spreads, after immunostaining with anti-Red1 antibodies and antibodies to histone 2B, has revealed a tight association of Red1, but not histone 2B, along the axes of both homologs (42). Another study has shown that Red1 localizes preferentially to the DSB-active domains along chromosome III (41). In addition to being a component of the lateral element, Red1 also serves as a downstream signaling factor in the DNA damage surveillance pathway during meiosis (45). Furthermore, Red1 associates with Ddc1 and Mec3, which are the subunits of the 9-1-1 complex involved in meiotic DNA damage surveillance (45). On the basis of these findings, it was hypothesized that Red1 can function directly via its interaction with genomic DNA sequences.
To gain an insight into the mechanistic aspects of Red1 function, a set of experiments were carried out to determine whether it has the ability to bind a variety of DNA substrates that are believed to be post-DSB recombination intermediates (Table 1). Six different substrates were designed by annealing combinations of synthetic oligonucleotides (ODNs) ( Table 2). The DNA substrates included a branch-immobile Holliday junction (HJ) containing four 20-bp arms, a 3-way junction with three 20-bp arms, a flap structure containing 40-bp duplex DNA with a 20-nucleotide 5Ј-flap single strand, a 40-bp bluntended duplex DNA, a Y-shaped junction containing a 20-bp duplex region (dsDNA) with two 20-nucleotide heterologous single strands, and a 20-bp DNA duplex with a 20-nucleotide 3Ј single-strand overhang and ssDNA. The electrophoretic mobility shift assays were performed by incubating a constant amount (0.5 nM) of the specified 32 P-labeled substrate with increasing concentrations of Red1 ranging from 50 to 500 nM. The products of binding reactions were resolved by 6% PAGE under nondenaturing conditions and visualized as described under "Experimental procedures." It was found that Red1  Yeast Red1 is a structure-selective DNA-binding protein showed significant differences in the relative binding affinity for different DNA substrates. Red1 preferentially bound to branched DNA structures, namely the HJ and 3-way junction, over the other substrates tested; the extent of its association increased with the addition of increasing Red1 concentrations (Fig. 3, A and B). In the assays performed with the 5Ј-flap, splayed DNA, and blunt-ended duplex DNA, a similar pattern of interaction was noticed, but the binding was relatively weak (Fig. 3, C-E). In sharp contrast to the binding with HJ and blunt-ended DNA duplex, very little or no binding was seen with ssDNA or partial DNA duplex containing a 3Ј-overhang (Fig. 3, F and G).
To ascertain the DNA binding preferences, the amount of DNA bound was quantified and plotted against a range of input concentrations of Red1. It was found that the Red1 binding exhibits a hyperbolic dependence on its concentration and a hierarchical substrate preference (Fig. 3H). Specifically, Red1 shows a higher binding affinity toward HJ, followed by replication fork (RF) Ͼ dsDNA Ͼ 5Ј-flap Ͼ Y-shaped structure Ͼ dsDNA with a 3Ј-overhang Ͼ ssDNA. These results support a model in which ssDNA or ssDNA tethered to a duplex is not energetically favorable to the DNA-binding activity of Red1. Next, the equilibrium dissociation constants for the interaction of Red1 with different DNA substrates were derived. The relative DNA affinity of Red1 was estimated from slopes of linear regression curves fitted to the data points. As seen in Table 3, the affinity of Red1 to the HJ is significantly greater, yielding an apparent K d value of 66 nM, which is 2-fold lower than the value

Red1 forms highly stable complexes with the HJ and 3-way junction
To further characterize the specificity of Red1-DNA interactions, the effect of ionic strength on the stability of preformed Red1 nucleoprotein complexes was tested. The assays were set up with a fixed amount of Red1 that was sufficient to bind all of the DNA present in the reaction mixtures (Fig. 4, lane 2). The reactions were quenched, and the products were resolved on EMSA gels. As shown in Fig. 4, marked differences were observed in the stability of complexes formed by Red1 with various substrates. In good agreement with the data presented in Fig. 3, Red1 formed highly stable complexes with the HJ and 3-way junction, and they were not affected even by the presence of 500 mM NaCl. However, the stability of complexes formed with blunt-ended dsDNA was affected by changes in the ionic strength. To determine quantitative changes, the amount of free and bound DNA in each lane was quantified, and the data are expressed relative to salt concentration. It was found that ϳ80% of the complexes formed with blunt-ended dsDNA dissociated in the presence of 1 M NaCl, in contrast to a modest dissociation (ϳ20%) of protein-DNA complexes formed with either the HJ or RF at the same concentration (Fig. 4D). The progressive retardation of free DNA (Fig. 4C) could be due to the corresponding increase of salt in the reaction mixture. These results support the idea that Red1 forms thermodynamically stable complexes with the HJ and RF than with the bluntended dsDNA and that hydrophobic interactions also contribute to the stability of the complex.

Red1 potentiates Hop1-promoted pairing of DNA double helices containing G/C-rich regions
The mechanisms underlying the processes of homology search and synapsis of homologous chromosomes are not fully understood (1,2,4). It has been hypothesized that G4 DNAforming sequences along the entire length of a chromosome may be used to bring together the four chromatids to establish synapsis during meiosis (50 -54). Our previous work has shown that Hop1 promotes intra-and intermolecular pairing of double-stranded DNA molecules (33,34,37). The pairing was further augmented by a centrally positioned G/C-rich region (33,34). In light of these studies, we asked whether Red1, by itself, can promote intermolecular pairing between two DNA duplex molecules. To verify this, an in vitro assay using a 48-bp synthetic duplex DNA with a single centrally positioned 8-bp G/C sequence was employed (33,34). First, the ability of Red1 to bind to the 48-bp DNA duplex containing a centrally positioned 8-bp G/C sequence was tested in comparison with a mixed-sequence DNA duplex of identical length (Fig. 5, A and B). The assay was performed with a constant concentration (2 nM) of DNA duplex and increasing concentrations of Red1 ranging from 0 to 2 M. The reaction products were resolved on nondenaturing polyacrylamide gels. As shown in Fig. 5 (C and E), incubation with Red1 shifted the free DNA to a slowly migrating protein-DNA complex, which increased with increasing Red1 concentration. However, the relative binding efficiency of Red1 for mixed-sequence DNA duplex was higher compared with the DNA duplex containing the G/C region. To determine whether Red1 promotes pairing between two DNA duplex molecules, experiments were conducted in parallel. In these tests, the reaction mixtures were deproteinized by proteinase K and SDS, and the products were separated on 10% nondenaturing polyacrylamide gels. In both cases, no evidence was found to support a role for Red1 in the pairing of two duplex DNA molecules (Fig. 5

, D and F).
Previous studies have shown genetic and physical interaction between Red1 and Hop1. The overexpression of HOP1 sup-

Yeast Red1 is a structure-selective DNA-binding protein
presses certain red1 alleles, and vice versa (12,42,46). In line with these findings, we asked whether Red1 can modulate the DNA-pairing activity of Hop1. We performed EMSA with the same 32 P-labeled DNA duplex containing a centrally embedded 8-bp G/C region. In one experiment, a constant amount of DNA duplex was incubated with increasing concentrations of Hop1. In a second experiment, the same reaction was performed with a constant amount of Hop1 in the presence of increasing concentrations of Red1. The reaction mixtures were deproteinized, and the products were analyzed as described above. It was found that Hop1 can promote intermolecular pairing between two duplex DNA molecules that are highly reminiscent of those seen before (33, 34) (Fig. 6B). Importantly, however, the findings of the second experiment pointed out that the DNA-pairing activity of Hop1 is potentiated by Red1 in a concentration-dependent manner (Fig. 6C). Further, quantitative analysis of the product formed with a PhosphorImager revealed that Red1 is able to stimulate the DNA-pairing activity of Hop1 by 2-fold (Fig. 6D). We next investigated the effect of sequence-specific and -nonspecific competitor DNAs on the pairing activity. The assay was performed as described above, except that the unlabeled competitor DNA duplex was added in various concentrations. The reaction products were resolved on nondenaturing polyacrylamide gels. Fig. 7 (A and B) shows gel images illustrating significant qualitative differences in the pattern of inhibition by specific (DNA duplex containing a single centrally embedded 8-bp G/C region) and nonspecific (mixed-sequence DNA duplex) DNAs on the pairing activity of Hop1 and Red1. The addition of a specific competitor strongly decreased product formation in a concentration-dependent manner compared with a nonspecific competitor; in both cases, no intermediate species were generated. In fact, differences in the pattern of suppression between specific and nonspecific DNAs were evident at all of the concentrations that were examined. A quantitative analysis of the amount of product formed revealed that the suppression of DNA pairing was more pronounced in the presence of specific sequence compared with nonspecific DNA (Fig. 7C). The difference is most obvious at a 100-fold molar excess of each of the competitors; here, the specific competitor DNA suppressed product formation to Ͼ95%, whereas the nonspecific competitor suppressed it to 5%.

Red1 promotes intramolecular bridging of non-contiguous segments in a circular duplex DNA
To gain further insights into the DNA binding mode of Red1, atomic force microscopy (AFM) was used to reveal the architecture of Red1-DNA complexes. Following a published protocol (37,38), the binding of Red1 to pUC19 circular duplex DNA was first examined using the conditions developed for its binding to DNA. After incubation, 5-l aliquots were placed on freshly cleaved mica; they were allowed 2 min for absorption and then visualized using tapping-mode AFM. At low concentrations of Red1, the association of Red1 with DNAs was observed at various positions, leading to intramolecular bridging between non-contiguous segments of DNA (Fig. 8A, ii-iv). Although a similar pattern of binding was observed at higher concentrations of Red1, the nucleoprotein filament frequency and length increased at this concentration (Fig. 8B, i-v). Quantification of the frequency of various types of Red1-bound DNA structures revealed at least four different types of DNA binding events: 1) binding of Red1 at random positions; 2) Red1-bound intramolecular nodes; 3) intramolecular stem-loop structures, and 4) Red1 nucleoprotein filaments. The contour length of the Red1-DNA complexes gradually decreased ϳ3-fold concomitant with the increase in Red1 concentration, pointing out that Red1 induces DNA compaction. Interestingly, in contrast to Hop1 (37), no intermolecular alignment of circular plasmid DNA molecules was detected in the presence of Red1 under these conditions. A number of AFM images were subjected to statistical analysis to quantify the Red1-DNA complexes. The analysis revealed that Ͼ80% of the plasmid DNA molecules are associated with Red1 to generate various structures (Fig. 8, C  and D).  2-11), respectively. In D and F, the assay was carried out with a 2 nM concentration of the specified 32 P-labeled duplex DNA in the absence (lane 1) or presence of increasing Red1 concentration. After a 30-min incubation, the reaction mixtures in C and E were deproteinized and analyzed as described under "Experimental procedures" (D and F). The positions of free DNA and the product are indicated on the left side of the gel image.

Red1 promotes nonhomologous DNA end joining
These foregoing results prompted us to further explore (using AFM) the mode of binding of Red1 to linear plasmid DNA. To this end, the pUC19 linear duplex DNA at two concentrations of Red1 was incubated, and the protein-DNA complexes were visualized as described above. The results were striking. The formation of concatemeric DNA was observed; this is due to the end-to-end joining of unit lengths of linear plasmid DNA (Fig. 9, A and B). It was also noticed that Red1 localizes mostly at the ends and interface between two linear DNA duplex molecules, in addition to being bound at many different sites on the DNA molecule. However, in contrast to Hop1 (37), crossover structures resulting from the intermolecular alignment of two linear duplex DNA molecules were not observed in the presence of Red1. Nonetheless, these results are consistent with the idea that Red1 binding may lead to higherorder assembly of DNA through the end joining of linear duplex DNA helices. Further, a quantification analysis suggested that the linear multimerization promoted by Red1 increases in a concentration-dependent manner (Fig. 9C). Despite the differ-  Yeast Red1 is a structure-selective DNA-binding protein ent nature of the assays, the AFM evidence that Red1 was unable to promote the pairing of DNA molecules is in good agreement with data from the gel assays (Fig. 6).

Hop1 stimulates the intermolecular ligation of linear dsDNA molecules promoted by Red1
In light of the fact that Red1 and Hop1 proteins physically interact with each other (12,39,43,46), we tested the effect of Hop1 on nonhomologous DNA end-joining activity of Red1. The assay was carried out by incubating blunt-ended linear dsDNA with Red1 in the presence of various concentrations of Hop1. The DNA products of this reaction were separated by agarose gel electrophoresis (see "Experimental procedures"). Whereas Red1 alone displayed a significant amount of DNA end-joining activity (Fig. 10, lane 5), Hop1 greatly potentiated the ligation of linear dsDNA molecules to form dimers, trimers, etc. (Fig. 10, lanes 6 -9). The products were found to be sensitive to exonuclease III treatment, thus confirming the formation of linear multimers of the blunt-ended 2.6-kb plasmid DNA substrate (Fig. 10, lanes 10 -12). We note that Hop1 by itself promotes bridging between linear dsDNA molecules to form multimers (37), which is not apparent at the concentration tested here (Fig. 10, lane 4). Therefore, the question of whether Red1 can stimulate Hop1-promoted DNA endjoining activity needs further study. Collectively, these in vitro results support the genetic evidence and further underscore the importance of physical interaction between Hop1 and Red1.

Discussion
Despite significant progress in defining the protein components of S. cerevisiae SC and their localization to specific substructures of the tripartite complex, the function of these different proteins is not fully understood. Toward this goal, a biochemical approach was adopted to elucidate the mode of action of Red1. In this study, we provide the first direct evidence that Red1 is a sequence-nonspecific but structure-selective DNA-binding protein, associates preferentially with DNA substrates that are considered as obligatory meiotic recombination intermediates, promotes DNA bridging and nonhomologous DNA end joining, and potentiates Hop1-promoted pairing of dsDNA molecules. Moreover, Hop1 greatly stimulates the DNA end-joining activity of Red1.
Previous genetic and biochemical studies on S. cerevisiae have established the molecular mechanism(s) of homologous  Yeast Red1 is a structure-selective DNA-binding protein recombination (HR) in meiosis. Briefly, HR between paired homologous chromosomes begins with the incision of a DSB at the recombination hot spot by a topoisomerase-like protein, Spo11 (55,56). Following the generation of 3Ј single-stranded tails on either side of the break by the coordinated action of helicase/nuclease activities encoded by MRE11/XRS2/RAD50/ SAE2 genes, the Rad51 and Dmc1 proteins polymerize on ssDNA thus generated to form a helical filament (1, 2, 4). The resulting nucleoprotein filament invades homologous dsDNA to produce D-loops. Subsequently, the D-loops are extended by repair synthesis to form heteroduplex DNA. A number of accessory factors either facilitate the activity of Rad51/Dmc1 or promote strand transfer by themselves. The strand exchange yields a 3-way junction. When the heteroduplex DNA encounters the ssDNA-dsDNA junction on the invading strand, the 3-stranded joint molecule is converted into a 4-stranded Holliday junction. Multiple pathways exist in eukaryotic cells for processing and resolution of the HJ (57). In somatic cells, four proteins comprising BLM, topoisomerase IIIa, RMI1, and RMI2, termed the BTR complex (in S. cerevisiae, orthologs of the BTR complex are Sgs1-Top3-Rmi1 (STR complex)), promote the convergent migration of two HJs to produce a hemi-catenane that can be processed by topoisomerase action. In S. cerevisiae sgs1 mutants, the joint molecules are processed by Mus81-Mms4 (the ortholog of MUS81-EME1) and Yen1 (the ortholog of GEN1), generating both crossover and non-crossover products (58 -61). The second and third pathways utilize structure-selective endonucleases that cleave the HJs to generate crossover and non-crossover products (57).
Understanding of the mechanisms of homologous DNA pairing, strand exchange, and HJ resolution in eukaryotes has advanced considerably since the isolation of Rad51 and Dmc1 proteins (62,63). However, knowledge of the functions of SC components is scanty. Genetic studies have shown that red1 and hop1 mutations perturb the same subset of events during meiotic recombination (39). Accordingly, an overexpression of RED1 suppresses or escalates certain non-null hop1 alleles (11,64). The mutation of RED1 in a DMC1 diploid cells leads to the disappearance of DSBs, thereby giving rise to the display of meiotic division abnormalities and non-viable spores (48,66,67). These genetic findings are most consistent with our observations that Red1 possesses DSB-binding and DNA endjoining activities. Further, red1 mutants are pleiotropic; consequently, they show a number of mutant phenotypes, including unusual defects in chromosome structure and synapsis (39,67). Genetic data from S. cerevisiae indicate that mutations in hop1, red1, and mek1 specifically reduce interhomolog recombination and generate inviable spores, indicating that HOP1 acts as a barrier to sister chromatid repair (68). Besides Hop1, Red1 also interacts with Mek1 and subunits of the 9-1-1 checkpoint module (Rad17-Mec1-Ddc1) through the Ddc1-Mec3 subunit-specific motifs (45,69). Consistent with the genetic data, immunoprecipitation experiments have shown that Red1/ Hop1 and Red1/Mek1 form complexes in meiotic cells (46,69,70,71). Whereas the role of Red1 in acting as a barrier to sister chromatid repair in meiotic cells has been appreciated for over 20 years, the exact mechanism of RED1 function in establishing interhomolog bias has not been elucidated. To interpret the genetic results, it is important to determine the biochemical properties of Red1 and its functional relationship to Hop1.
Several experiments in the current investigation demonstrate that purified Red1 associates preferentially with the duplex DNA and 3-way junction as well as the Holliday junction compared with ssDNA and dsDNA with single-stranded overhangs. Furthermore, the broad substrate specificity of Red1 indicates that it can engage in a wide range of DNA binding events and induce condensation and topological changes. It also suggests that these functional outcomes may vary according to the specific substrate and the availability of interacting partners. The finding of the present study that Red1 induces DNA bridging, condensation, and formation of stem-loop structures implies a strong probability of its involvement in the folding of prophase I chromosomes into radial loops; this, in turn, may facilitate tethering of the loops to the SC axial/lateral elements (72). Nevertheless, what does it mean for Red1, a lateral element protein of SC formed during early meiosis, to have a high affinity for a late DNA recombination intermediate, such as the HJ? Although Red1 is thought to be a crucial regulator of early meiosis, it physically associates with the ring-shaped 9-1-1 complex, a downstream effector essential for the meiotic DNA Yeast Red1 is a structure-selective DNA-binding protein damage surveillance pathway (45). Association of 9-1-1 with Red1 is required for meiotic checkpoint activation (45,69), and Red1 by itself binds to DSBs and promotes DNA end joining (this study), but its function at different phases of meiotic prophase I remains unclear. Given these findings, we propose that Red1 also functions at mid-and late pachytene phases by monitoring the DNA integrity and progression of meiotic recombination intermediates, including the DSBs and Holliday junction.
Previous studies have implicated Red1 in several aspects of meiotic recombination (5,6,39,48). Evidently, R/G-band regions modulate the frequency of homologous recombination; Red1 promotes the formation of DSBs at both R-and G-bands and then assists the loading of Dmc1 on ssDNA, specifically by counteracting the adverse effects of R-bands (41). Significant insights into the role of G-quadruplexes in meiosis emerged from genetic studies in S. cerevisiae; deletion of KEM1/SEP1, which encodes a G-quadruplex DNA-specific nuclease, arrests the cells at meiotic prophase I (73). Genome-wide studies on S. cerevisiae have revealed the existence of Ͼ1500 sequence motifs with the potential to form G-quadruplex structures; these motifs overlap with the meiotic DSBs (74,75). Additionally, a strong correlation has been established between meiotic DSBs, recombination rate, and G/C content in S. cerevisiae (76,77). The occurrence of high concentrations of potential G-quadruplex forming sequences in the S. cerevisiae genome strongly implicates a role for these sequences in meiotic recombination (78,79). Furthermore, the MRX complex, composed of Mre11, Rad50, and Xrs2, which is involved in the resection of DSBs, has a high affinity for G4 structures in vitro (80,81). There is good evidence that intragenomic DNA recombination in the pilE locus is controlled by the G-quadruplex DNA (82). Altogether, these findings highlight an important role for the G-quadruplex DNA-based mechanism, possible involvement in chromosome pairing and recombination.
Although both Hop1 and Red1 associate independently with chromosomes at the leptotene stage to form a series of foci, the association of Hop1 is dependent on Red1 (42). Several studies have demonstrated that the incision of programmed meiotic DSBs and the loading of Red1 coincide with GC-rich regions (40,41,76). Intriguingly, we found that Red1 and Hop1 proteins exhibit similar affinities for dsDNA containing the G/C region in vitro, but the functional consequences of the two proteins appear to be different at least in one important way: Red1 lacks the ability to promote the pairing of two linear dsDNA molecules. However, these results do not exclude the possibility that Red1 contains low levels of DNA-pairing activity that is not apparent under these conditions. In addition, the failure of Red1 to promote pairing between duplex molecules is not due to its inability to bind G-quadruplex DNA, because it binds to G4 DNA at nearly the same affinity as does Hop1 (data not shown). On the other hand, the positive interaction between Hop1 and Red1 in potentiating the pairing of two DNA molecules and DNA end joining probably provides a huge advantage in terms of selectivity and specificity during meiosis.
The formation of synapsis product promoted by Hop1 or Hop1 plus Red1 does not reach 100% equilibrium. Although we do not know the precise reason(s), one plausible explanation is that Hop1/Red1 protein(s) also promote pairing of duplex DNA molecules at regions of "low G/C content." Upon deproteinization of the reaction mixtures, the synapsis product(s) formed at low G/C content perish, resulting in 30 -40% product formation. Recently, two independent studies conducted in mice and S. cerevisiae have discovered a key role for a chromosomally tethered proteasome in the regulation of chromosome synapsis and recombination during prophase of meiosis I (83,84). This hints that inappropriate synapsis and protein degradation are important for the interrelated molecular pathways that execute synapsis and recombination of meiotic chromosomes.
What are the consequences of the direct interaction between Red1 and Hop1? Whereas Hop1 and Red1 can independently bind various meiotic recombination intermediates, Red1 and Hop1 unite to perform an important function: potentiate the pairing of two dsDNA molecules and promote nonhomologous DNA end joining. The results reveal that Hop1 acts earlier than Red1, at least in DNA-pairing events. In agreement with these results, genetic evidence has shown the existence of Hop1/Red1 hetero-oligomers (11,12,85). Thus, the functional interaction between Hop1 and Red1 in the pairing of dsDNA molecules is a functional validation of their genetic/physical interaction. Notwithstanding, what specialized function can Red1 perform in the cell if it acts independently? We assume that DNA end joining, DNA bridging, and looping activities of Red1 fit well with the condensation of chromosomes during the meiotic prophase I. In summary, this study, for the first time, has revealed a biochemical function for Red1. In particular, our results highlight a crucial role for Red1 during prophase of meiotic division I. In addition to its structural role in the SC, Red1 potentiates pairing of DNA molecules by Hop1 and also promotes the joining of broken DNA ends, which, in turn, might protect meiotic recombination intermediates and prevent the loss of genetic material. The fact that Hop1 and Red1 act independently of each other and also together could allow them to perform different functions at separate steps during chromosome synapsis and recombination.

Source of biochemicals, bacterial strains, proteins, and DNA oligonucleotides
Fine chemicals were purchased from GE Biosciences and Sigma. Restriction endonucleases, T4 DNA ligase, and T4 polynucleotide kinase were purchased from New England Biolabs and Thermo Scientific. All of the reagents used were of analytical grade. The DNA oligonucleotides were synthesized by Sigma-Genosys. [␥-32 P]ATP was purchased from Bhabha Atomic Research Centre (Mumbai, India). E. coli expression strain RIL and the expression vectors pET-22b and pET-28a were purchased from Novagen. Protein purification resins Ni 2ϩ -nitrilotriacetic acid-agarose, SP Sepharose, and Q Sepharose were purchased from Qiagen and Sigma. Fast performance liquid chromatography columns were purchased from GE Biosciences. The Red1 antibodies were custom-made against Red1 protein in rabbits by Imgenex India (Bhubaneswar, India). The full-length Hop1 was purified as described previously (86).

Molecular cloning of the S. cerevisiae RED1 gene
The nucleotide sequence corresponding to S. cerevisiae RED1 (2484 bp) was PCR-amplified from the plasmid plb-1 (11) using specific primers ( Table 2). The PCR product was purified from the agarose gel using Qiagen's gel extraction kit. The product was digested by NdeI and HindIII and directionally cloned into expression pET-22b vector (Novagen). The resulting recombinant plasmid was designated pRED1.

Overexpression and purification of the Red1 protein
S. cerevisiae Red1 was expressed and purified from E. coli BL21 (DE3) CodonPlus-RIL cells, which were transformed with the pRED1 plasmid. A single colony from a freshly grown transformation plate was inoculated into 30 ml of LB broth supplemented with 100 g/ml ampicillin and was grown overnight at 37°C. This overnight culture was then inoculated into 2 liters of LB broth supplemented with 100 g/ml ampicillin. The cells were grown at 37°C with vigorous shaking. The expression of Red1 was induced by the addition of 0.5 mM isopropyl ␤-D-1thiogalactopyranoside after the culture reached an A 600 nm of 0.5; the incubation was extended for 12 h. The cells were collected by centrifugation, washed with STE buffer (10 mM Tris-HCl (pH 8), 100 mM NaCl, and 1 mM EDTA), and resuspended in buffer A (20 mM Tris-HCl (pH 8), 10% glycerol, 50 mM NaCl, and 5 mM 2-mercaptoethanol) and stored at Ϫ80°C. These cells were thawed and lysed by sonication (model GEX-750 ultrasonic processor, Cole-Parmer India Pvt. Ltd, Mumbai, India.) on ice at 51% duty cycles in a pulse mode. The sonicated suspension was centrifuged in a Beckman type 45 Ti rotor at 30,000 rpm for 1 h at 4°C. The supernatant was subjected to (NH 4 ) 2 SO 4 precipitation at 55% saturation followed by centrifugation at 18,000 rpm at 4°C for 30 min. The pellet was resuspended in buffer A and dialyzed against the same buffer. The dialysate was chromatographed on a Q Sepharose column equilibrated with buffer A. The Red1 protein was eluted from the column in the flow-through fractions. The flow-through was subjected to (NH 4 ) 2 SO 4 precipitation at 50% saturation. The precipitated protein was collected by centrifugation at 18,000 rpm for 30 min at 4°C. The pellet was resuspended in buffer B (20 mM HEPES-HCl (pH 8), 1 M NaCl, 10% glycerol, and 5 mM 2-mercaptoethanol) and dialyzed against the same buffer. The dialysate was passed through a Sephadex 200 (S-200) gel filtration column equilibrated with buffer B and eluted using buffer B. The fractions containing Red1 were pooled and dialyzed against buffer C (20 mM HEPES-HCl (pH 8), 250 mM NaCl, 50% glycerol, and 1 mM DTT). The purity of Red1 was assessed by SDS-PAGE, followed by Coomassie Blue staining. Aliquots of the Red1 protein were stored at Ϫ80°C. The concentration of Red1 was determined by the dye-binding method using bovine serum albumin as a standard (87).

Far Western blotting assay
Increasing amounts of purified Hop1 or Rad17 (1-5 g) diluted in buffer A (10 mM HEPES-HCl (pH 7.5), 100 mM NaCl (pH 7.5), and 1 mM EDTA) were spotted onto nitrocellulose membranes. Thereafter, blocking was carried out at 4°C for 4 h with buffer A containing 5% (w/v) nonfat milk and then incubated with Red1 for 9 h at 4°C. The membranes were washed once with buffer A prior to incubation with anti-Red1 antibody for 12 h at 4°C. The membranes were washed six times for 10 min with PBS containing 0.1% Tween 20 (PBST) and incubated with a 1:40,000 dilution of horseradish peroxidase-conjugated anti-rabbit antibody for 2 h at 4°C. Finally, the membranes were washed three times with PBST, and the antibody signals were visualized using a chemiluminescence imaging device.

Preparation of DNA substrates
The ODNs used for construction of the DNA substrates are listed in Table 2. The ODNs were labeled at the 5Ј-end using [␥-32 P]ATP and T4 polynucleotide kinase (88). The DNA substrates (depicted in Table 1) were constructed using different combinations of ODNs (Table 2). Briefly, stoichiometric amounts of the purified ODNs were annealed by incubation in 0.3 M sodium citrate buffer (pH 7) in a 100-l reaction mixture containing 3 M NaCl at 95°C followed by slow cooling to 4°C over a period of 2 h. The Holliday junction, Y-shaped substrates, and other intermediates of DNA replication/repair were prepared and characterized as described (65). In the case of substrates used for the DNA-pairing assay, the top strand of the duplex DNA was labeled at the 5Ј-end using [␥-32 P]ATP and T4 polynucleotide kinase. The labeled strand was annealed to an equimolar amount of unlabeled complementary strand. To purify the substrates, the reaction mixtures were subjected to electrophoresis on 8% (w/v) polyacrylamide gel in 45 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA and 120 mM KCl at 10 V/cm for 5-12 h. The bands corresponding to specific substrates were excised from the gel and eluted in TE buffer (10 mM Tris-HCl (pH 7.5), 1 mM EDTA).

Surface plasmon resonance measurements
All of the measurements were performed with a Biacore 2000 optical biosensor (GE Healthcare). Protein immobilization, binding experiments, and data analysis were performed according to the manufacturer's protocol and software supplied with the instrument. Hop1 was immobilized on the surface of a CM5 sensor chip (ϳ1000 response units/flow cell) using the aminecoupling method. The unbound protein was removed by passing the buffer with a flow rate of 100 l/min. The flow cell 1 was used as a control. The binding experiments were carried out at 30°C using a continuous flow of the running buffer (10 mM HEPES buffer (pH 7.5) containing 150 mM NaCl, 50 mM EDTA and 0.005% P-20 surfactant) at a flow rate of 50 l/min. Increasing concentrations of Red1 (40 -320 nM) were added over the CM5 sensor chip. Regeneration was carried out by 4 M MgCl 2 . The data obtained for the interaction of Red1 with Hop1 were corrected for nonspecific binding by the automatic subtraction of blank data from in-line reference flow cells. The data were analyzed using a 1:1 Langmuir binding model. A global fit of the data was used to determine the kinetic constants using BIAcore evaluation software version 3.0.

DNA-binding assay
The reaction mixtures (20 l) contained 10 mM HEPES-HCl (pH 7.5) or 10 mM Tris-HCl (pH 7.5), a 0.5 nM concentration of the specified 32 P-labeled DNA substrate, and increasing concentrations of the specified protein. The mixtures were incu-Yeast Red1 is a structure-selective DNA-binding protein bated at 30°C for 30 min. The reaction was terminated by the addition of 2 l of loading dye (0.25% xylene cyanol, 0.25% bromophenol blue in 20% glycerol). The reaction mixtures were electrophoresed on a 6% native polyacrylamide gel in 44.5 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA at 10 V/cm at 4°C for 3-6 h. The gels were dried and exposed to a PhosphorImager screen (FLA-9000); the bands were visualized using the software supplied by the manufacturer. The data were quantified using the UVI band map software package and plotted in GraphPad Prism software.
After incubation at 30°C for 30 min, the mixtures were deproteinized by the addition of proteinase K (0.2 mg/ml) and SDS (0.2%) followed by incubation at 37°C for 20 min. The reactions were terminated by the addition of 2 l of loading dye (0.25% xylene cyanol, 0.25% bromophenol blue in 20% glycerol). The reaction mixtures were electrophoresed on a 10% native polyacrylamide gel in 44.5 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA at 10 V/cm at 4°C for 6 h. The gels were dried and visualized by the FLA-9000 PhosphorImager. The data were quantified using the UVI band map software package and plotted in GraphPad Prism software.

AFM imaging of DNA and Red1-DNA complexes
The plasmid pUC19 was isolated from E. coli DH5␣ cultures and used for imaging the Red1-DNA complexes. The plasmid DNA was purified using the Thermo Scientific Maxi-prep kit and resuspended in water. Subsequently, the plasmid was treated with chloroform/isoamyl alcohol to remove the intrinsically bound proteins. Blunt-ended linear pUC19 DNA was generated by digestion of circular pUC19 DNA with SmaI. Red1 (100 and 300 nM) was incubated with 5 ng of circular or linear pUC19 DNA in 10 l of reaction mixture containing 20 mM Tris-HCl (pH 8) and 2 mM MgCl 2 . After incubation at 30°C for 30 min, 5 l of aliquot was applied to the surface of freshly cleaved mica. The mica surface was rapidly rinsed with nanopure deionized water and air-dried. The images were acquired in tapping mode using an SNL (silicon tip on nitride lever) probe (Agilent Technologies, spring constant 21-98 newtons/m).

DNA ligation assay
One hundred ng of blunt-ended 2.6-kb pUC19 DNA (generated by digestion of circular pUC19 DNA with SmaI) was incubated in the absence or presence of specified concentrations of Red1 and Hop1 in a 20-l reaction mixture containing 20 mM Tris-HCl (pH 8). After incubation at 30°C for 30 min, 2 l of T4 DNA ligase buffer and 10 units of T4 DNA ligase (Thermo Scientific) were added, and the incubation was extended at 30°C for 20 min. For reactions involving exonuclease III (Exo III), samples were further incubated at 37°C for 30 min in the presence of 5 units of Exo III (New England Biolabs). The reactions were terminated by the addition of 1 l of 20% SDS and 1 l of 10 mg/ml proteinase K and incubated at 37°C for 30 min.
The samples were electrophoresed through 1.1% agarose in 45 mM Tris borate buffer (pH 8.3) containing 1 mM EDTA at 80 V for 7 h and visualized by staining with ethidium bromide.
Author contributions-K. M. conceived and designed the project and wrote the manuscript. R. K. performed experiments and contributed to the writing of the manuscript. I. G. performed the AFM measurements. K. M., R. K., and I. G. analyzed the data and interpreted the results. All authors read and approved the final manuscript.