A unique four-stranded model of a homologous recombination intermediate

J. theor. Biol. (1986) 120, 215-222

A Unique Four-stranded Model of a Homologous Recombination Intermediatet ROBERT C. HOPKINS

Division of Sciences, University of Houston--Clear Lake, Houston, Texas 77058, U.S.A. (Received 3 September 1985, and in revised form 10 January 1986) This paper proposes a model of four-stranded DNA synapsis during recombination between homologous segments of two DNA duplexes. The proposed intermediate is one of only two known models having relative chain orientations about the synaptic junction that are consistent with recent topological results on the integrative recombination of bacteriophage lambda. This model has the advantage of providing a mechanism for recognition of sequence homology between duplexes through specific hydrogen-bond formation; other models are discussed in comparison. The new model is based on an alternative family of DNA structures having chain directions opposite to those of the Watson-Crick family of structures. Idealized coordinates for generating both right- and left-handed forms of these alternative structures are presented for further study.

Introduction A particularly well-studied example of site-specific recombination is the integration (and subsequent excision) of the circular form of bacteriophage lambda D N A into the host genome of Escherechia coli (for reviews see Nash, 1981 and Weisberg & Landy, 1983). Within the attachment site (attP) of the double-stranded phage D N A is a specific 15 base pair (bp) core sequence that is identical to a core region within an attachment site (attB) in the D N A of the bacterium. From studies both of phage h and its host and of D N A containing the cloned attachment site regions (Mizuuchi et al., 1981; Nash et al., 1981; Kitts et al., 1984; and references cited therein), several important aspects of phage D N A integration into the E. coli genome can be summarized: (a) A synapsed recombination intermediate forms initially through the association of portions of the two homologous core regions (four D N A strands). It is within the core regions that crossover occurs. (b) At both boundaries of the homologous junction formed, two strands (one from each parental duplex extending in the 5'-direction away from the junction) are cleaved and rotate about their complementary uncleaved strands. Thus, in some temporal sequence, all four strands are cut, enabling strand exchange to occur and resulting in the creation of a 7 bp overlap region of heteroduplex DNA. (c) Recombination is completed with resealing of the exchanged strands and disengagement of the recombined attachment sites, leaving the prophage fully t A preliminary report of this work was presented at the Fourth Conversation in Biomolecular Stereodynamics, State University of New York at Albany, June 4-7, 1985. 215 0022-5193/86/100215 + 08 $03.00/0

(~ 1986 Academic Press Inc. (London) Ltd

216

R.C. HOPKINS

integrated. The resulting prophage attachment sites (attL and attR) contain core regions that are homologous to the core sequences of the phage and bacterium. These integration steps are mediated by a host protein, integration host factor (IHF), and a phage-encoded polypeptide, Int, which has properties of a topoisomerase. In addition to these proteins, excision of the prophage also requires the product of the viral xis gene, Xis.

DNA Topology During Phage Lambda Integration Of particular interest are the topological aspects of the phage integration process since they help to define the microscopic details of site-specific recombination. The initial steps in elucidating this topology have come from studies of supercoiled, partially relaxed, and nicked circular DNA substrates specially created to include both the attP and attB attachment sites (Pollock & Nash, 1983; Nash & Pollock, 1983). Using such substrates, with the cores oriented in an inverted repeat, Pollock & Nash (1983) concluded that a negative solenoidal supercoil, a nucleosome-like structure, is formed in the attP site as part of the A integration process. Thus this type of substrate, even when fully relaxed or nicked, should contain at least one negative interdomainal writhe node in the I n t / I H F recombination system. Further evidence (Nash & Pollock, 1983) suggested that synapsis within the homologous core sequences must occur early in the process and involve tight binding, based on the observation of a unique change in linking number accompanying recombination. From this data it was proposed that two different models were consistent with the observed results: Model 1:. Association of two homologous Watson-Crick duplexes across their major grooves (Fig. l(c)), as first proposed by McGavin (1971), plus a 270 ° clockwise (CW) rotation of each severed strand around its complementary strand, brings each cleaved end into apposition with a cleaved end from the other parent. Model 2: Synapsis of two homologous Watson-Crick duplexes across their minor grooves (Wilson, 1979), followed by melting, reannealling with the complementary strand from the other parent (Fig. l(d)), and finally 90° CW rotation of each cut strand around the complementary uncut strand, brings each broken end into apposition with a broken end from the other parent. In the topological notation of Cozzarelli et al., (1984), the first model produces fraMe=-3 and t e r M e = - l , whereas the second model produces traMe=+l and terMe = +1. Additional evidence from integrative recombination in relaxed forms of these same circular substrates indicates that products consist of about 50% simple circles and 50% knots (Pollock & Nash, 1983). In an impressive recent development, Griflith & Nash (1985) show conclusively by electron microscopy that the knotted forms are trefoil knots uniquely containing three positive writhe nodes. From this result it can be concluded (Griflith & Nash, 1985) that the McGavin-type synapsis (model 1 above, Fig. l(c)) does not account for the observations. In fact, only models of synapsis which lead to terMe = +1 appear to be acceptable, regardless of the direction of rotation of the severed strands.

HOMOLOGOUS

RECOMBINATION

(o)

INTERMEDIATE

)--17

(b)

3' ,3'

,3' 5

5

3

'

5'

Idl

~

(e)

5

5

,

~5, ~'

FIG. 1. Three models for synapsis of homologous DNA duplexes during recombination (Helix twists and base designations have been omitted for clarity. Arrowheads indicate strand cutting sites, wherever visible, for the case of A recombination). (a) Symbol for a Watson-Crick base pair (e.g. Cyd-Guo). (h) Symbol for a homologous base tetrad (see Fig. 2). (c) Model 1: two Watson-Crick duplexes associated across their major grooves (McGavin, 1971). (d) Model 2: two Watson-Crick duplexes associated across their minor grooves followed by melting and formation of a heteroduplex junction (Wilson, 1979). (e) Model 3: two Configuration II duplexes associated across their major grooves, as described here.

An Alternative Model for DNA Synapsis in HomologousRecombination Although model 2 above is consistent with the electron microscope results, there is another model of four-stranded DNA synapsis (Hopkins, 1984) which is also entirely compatible with this topological evidence: Model 3: Association of two homologous Configuration II D N A duplexes (Hopkins, 1981) across their major grooves, shown schematically in Fig. l(e), would result in terMe = +1, in accord with the topological requirement of A integrative recombination (Griffith & Nash, 1985). Rotation of each severed strand about its complementary strand by either 90° CW or 270 ° counterclockwise (CCW) brings a cleaved end into apposition with a cleaved end from the other parent. However, the same arguments (Nash & Pollock, 1983; Griflith & Nash, 1985) applying to model 2 apply to this model; thus 90° CW rotation would seem to be preferred. Structural details of this third candidate for synapsis are discussed in the Appendix.

Discussion All three models of synapsis described above employ the specific hydrogenbonding scheme first proposed by LSwdin (1963) for pairing two identical base

218

a. C. HOPKINS

c,'

"'O C1' HXN-H"

,---< G

,...

a N ~x.r.,iH'" /T "

/o'

X',~.......~=/~) / r~

9--.H-N,

.-H..-O

~C/N." .H--N~ CI'

.N.

N ,c,,

J •

c('

"H~.tq ~./.N

N

FIG. 2. Unique hydrogen-bonding schemes for specifically pairing two identical base pairs (L6wdin, 1963) in four-stranded DNA.

pairs into base tetrads (Fig. 2) compatible with the conjunction of two homologous DNA duplexes. Major differences in the models, then, lie in the relative locations or orientations of the four individual DNA strands in each case. This fact becomes evident upon comparing the spatial arrangement of the 5'-* 3' chain directions in Figs 1(c), l(d) and l(e). In fact, it is the arrangement of individual strands, coupled with the criterion that effective recombination joins similarly-oriented strands of opposite parentage, that determines the value of terMe. Thus, the requirement that terMe=+l for h recombination (Griffith & Nash, 1985) restricts the acceptable strand orientations to those of models 2 and 3 (Figs l(d) and l(e)). This striking result also may have implications for other models that involve synapsis of homologous DNA sequences, such as in general genetic recombination (reviewed -by Howard-Flanders et al., 1984 and Dressier & Potter, 1982). The details of these latter two models are sufficiently different that they should be experimentally distinguishable. For example, the mechanisms by which homology between duplexes is initially sensed must be quite different in the two cases. An electrostatic attraction across the major grooves of homologous duplexes, as a result of the complementary distribution of hydrogen-bond donor and acceptor sites (Fig. 2), is implicit in model 3, whereas no similar interaction for providing specificity between duplexes approaching via their minor grooves is evident. Presumably, Int and/or IHF would be directly implicated in this process, for the latter case. The additional steps of melting and reannealling in model 2 suggest that a junction of two hydrogen-bonded heteroduplexes should form prior to any strand-cutting event (Wilson, 1979). On the other hand, model 3 implies that prior to strand cleavage the two intact parental duplexes should be associated in a quite different non-covalent junction. Since it is difficult to conceive of cooperative, reversible transitions between DNA chain configurations I and II, if has been proposed that Configuration II may be the one which commonly exists in living systems (Hopkins, 1981, 1983). The synaptic recombination intermediate proposed here (model 3) is a direct extension of this hypothesis and is consistent with available evidence (Griffith & Nash, 1985). Alterna-

HOMOLOGOUS RECOMBINATION I N T E R M E D I A T E

219

tively, other e x p e r i m e n t a l results o n the structure o f n a t u r a l D N A in vivo (in contrast to d a t a o n synthetic D N A in vitro) are too meagre to e x c l u d e the possibility that t r a n s i t i o n s b e t w e e n C o n f i g u r a t i o n s I a n d II o c c u r d u r i n g physiological c h a n g e s w i t h i n a cell. I f such t r a n s i t i o n s can occur, they i m m e d i a t e l y suggest a m e c h a n i s m b y which r e c o m b i n a t i o n a n d other cellular processes such as gene e x p r e s s i o n m i g h t be c o n t r o l l e d , p e r h a p s with the aid o f specialized proteins. I thank Star Hopkins for critical review of this manuscript, and Johann Deisenhofer for generously providing his version of the program EREF. This work was supported by grants from the Robert A. Welch Foundation (E-889) and the UH-CL University Available Fund. REFERENCES COZZARELLI, N. R., KRASNOW, M. A., GERRARD, S. P. & WHITE, J. H. (1984). Cold Spring Harbor Syrup. Quant. BioL 49, 383. CRICK, F. H. C. & WATSON,J. D, (1954). Proc R. Soc. Lond. A223, 80. DE SANTIS,P., MOROSE'r'rI,S., PALLESCHI, A. & SAVINO, M. (1981). Biopolymers 20, 1707. DRESSLER, D. & POTTER, H. (1982). Ann. Reo. Biochem. 51, 727. FRANKLIN, R. E. & GOSLING, R. G. (1953). Acta Cryst. 6, 673. GRIFFITH,J. D. & NASH, H. A. (1985). Proc. natn. Acad. Sci. U.S.A. 82, 3124. HOPKINS, R. C. (t981). Science 211, 289. HOPKINS, R. C. (1983). Cold Spring Harbor Syrup. Quant. Biol. 47, 129. HOPKINS, R. C. (1984). Comments MoL Cell. Biophys. 2, 153. HOPKINS, R. C. (1985). In: The Molecular Basis of Cancer, Part A: Maeromolecular Structure, Carcinogens and Oncogenes. (Rein, R. ed.). p. 299. New York: Alan R. Liss. HOWARD-FLANDERS,P., WEST, S. C. & STAZIAK,A. (1984). Nature 3119,215. KIT't'S, P., RICHET, E. & NASH, H. A. (1984). Cold Spring Harbor Syrup. Quant. Biol. 49, 735. LEvl'l'r, M. (1974). Z tool. Biol. 82, 393. LrWDIN, P.-O. (1963). In: Electronic Aspects of Biochemistry. (Pullman, B. ed.). p. 167. New York: Academic Press. McGAVlN, S. (1971). J. moL Biol, 55, 293. McGAVlN, S. (1980). J. theor. Biol. 85, 665. MIEUUCHI, K., WEISBERG,R., ENQUIST,L., MIZUUCHI,M., BURACZYNSKA,M., FULLER,C., HSU, P.-L., ROSS, W. & LANDY, A. (1981). Cold Spring Harbor Syrup. Quant. BioL 45, 429. NASH, H. A. (1981). Ann. Rev. Genet. 15, 143. NASH, H. A., MIZUUCHI,K., ENQUIST, L. W. & WEISBERG, R. A. (1981). Cold Spring Harbor Syrup. Quant. Biol. 45, 417. NASH, H. A. & POLLOCK,T. J. (1983). J. tool. Biol. 170, 19. NORDHEIM, A., PARDUE, M. L., LAFER, E. M., MOLLER, A., STOLLAR, B. D. & RICH, A. (1981). Nature 294, 417. POLLOCK, T. J. & NASH, H. A. (1983). Jr- mot. Biol. 170, 1. RICH, A., NORDHEIM, A. & WANG, A. H.4. (1984). Ann. Rev. Biochem. 53, 791. SAENGER, W. (1984). Principles of Nucleic Acid Structure. pp. 52-86. New York: Springer-Vedag. WEISBERG, R. A. & LANDY, A. (1983). In: Lambda II. (Hendrix, R. W., Roberts, J. W., Stahl, F. W. & Weisberg, R. A. eds). p. 211. New York: Cold Spring Harbor Laboratory. WILSON, J. H. (1979). Proc. natn. Acad. Sci. U.S.A. 76, 3641. ZHURKIN, V. B., LYSOV, YU. P. & IVANOV,V. I. (1978). Biopolymers 17, 377.

APPENDIX There are three i n d e p e n d e n t chiral features associated with a m o l e c u l e o f d o u b l e s t r a n d e d D N A : (1) the e n a n t i o m e r i c forms of the d e o x y r i b o s e residues (D or L), (2) the h a n d e d n e s s o f the helix (right or left), a n d (3) a less w e l l - k n o w n e l e m e n t associated with the sense o f rotation o f the 5 ' ~ 3' c h a i n d i r e c t i o n s a b o u t each

220

R.C. HOPKINS

T.J

1,1

(o)

(b)

FIG. A1. Possible chain chiralities about each pseudo-diad axis (I) in the minor groove of doublestranded DNA (Arrows indicate the 5'--)3' chain directions). (a) CW or Configuration I (Watson-Crick family). (b) CCW or ConfigurationII (an alternative DNA family; Hopkins, 1981). pseudo-diad axis in the minor groove (CW or CCW). As indicated in Fig. A1, this latter chiral symmetry is the basis on which an alternative family of double-helical models for the structure of DNA has been proposed (Hopkins, 1981). The new model family, Configuration II, preserves the original base-pairing scheme in the model family of Crick & Watson (1954), Configuration I, yet differs in having antiparallel chains of the opposite sense. Because of this subtle difference in structure, the Configuration II family of DNA models is topologically distinct from the Watson-Crick family (Hopkins, 1983). In fact, the possibility of such alternative models also has been considered by others in theoretical studies (Zhurkin et al., 1978; McGavin, 1980; De Santis et al., 1981). Although the left-handed Z-forms of D N A (reviewed by Rich et al., 1984) have a dinucleotide repeat unit, they are examples of Configuration II, not Configuration I (Hopkins, 1983). As a result, studies using Z-DNA-specific antibodies (Nordheim et al., 1981; Rich et aL, 1984) suggest that the alternative chain configuration, as Z-DNA, occurs in some eukaryotic chromosomes. Because of the possible involvement of four-stranded (tetraplex) Configuration II D N A is site-specific recombination as discussed above, and because of potential roles in gene regulation and carcinogenesis (Hopkins, 1984, 1985), an idealized set of atomic coordinates is presented here. These data allow both right- and left-handed Configuration II duplexes and tetraplexes to be built for further study or refinement.

Minor groove

Ci

• ( I'

N1,N9~9"04 A ~

N(:J

JI

x

i

8!31~

L

i--Y

N ,N1

N1,N9 1o-85

Minor groove

FIG. A2. Geometryof atoms definingthe glycosylbonds in base tetrads of the idealized four-stranded being described.

models

HOMOLOGOUS RECOMBINATION I N T E R M E D I A T E

221

I n i t i a l c o o r d i n a t e s were o b t a i n e d using K e n d r e w wire models. T h e structures were idealized b y c o n s t r a i n i n g each h o m o l o g o u s l y - p a i r e d tetrad (Fig. 2) to b e p e r p e n d i c u l a r to the helical axis, with the glycosylic base n i t r o g e n s (N1 for p y r i m i d i n e s , N 9 for p u r i n e s ) p o s i t i o n e d at the c o m e r s o f a rectangle, as s h o w n in Fig. A2. Likewise, the C I ' a t o m o f each d e o x y r i b o s e r e s i d u e was p o s i t i o n e d at o n e c o r n e r o f a larger parallel rectangle (Fig. A2), as was d o n e i n a n idealized Configura t i o n I tetraplex ( M c G a v i n , 1971). I n o r d e r to p r o d u c e the final c o o r d i n a t e s , the energy o f each m o d e l was m i n i m i z e d , w i t h o u t c o n s i d e r i n g charges or solvent, u s i n g the m o l e c u l a r m e c h a n i c s p r o g r a m E R E F (Levitt, 1974) with average b o n d a n d angle p a r a m e t e r s t a b u l a t e d b y S a e n g e r (1984).

TABLE 1

Cylindrical atomic coordinates for constructing idealized right- and left-handed Configuration II models Atom

r (A)


Z (A)

Atom

r (A)

~ (°)

NI C2 02 N3 C4 N4 C5 C6

6.14 5.39 6.12 4.04 3"55 2"30 4.64 5"85

Cytosine 47.4 36.2 26.4 36-3 55"6 64"7 67-4 60.2

0-00 0.00 0.00 0-00 0"00 0"00 0-00 0"00

N1 C2 02 N3 C4 04 C5 C5M C6

6.14 5.24 5.81 3"93 3"61 2.57 4-86 5.40 5"98

Thymine 47-4 37.2 25.9 40-3 60"3 72.4 68"4 84"3 60"1

N9 C8 N7 C5 C6 06 N1 C2 N2 N3

6.14 5,64 4.31 4.00 2.99 1.76 3.87 5-22 6.10 5.87

Guanine 47.4 59-9 61.9 43.7 27,6 27.1 9.8 12.7 2.6 24.8

0.00 0-00 0.00 0-00 0.00 0.00 0-00 0.00 0.00 0.00

N9 C8 N7 C5 C6 N6 N1 C2 N3 C4

6.14 5-53 4.19 4.01

C4

5.27

37-I

0-00

C5' C4' 01' CI' C2'

Left-handed chain 10,29 53.9 9.57 50.4 8.20 53.1 7.59 45.7 8.26 37.4

-2.08 -0-89 -0.78 0.00 -0.74

C5' C4' 01' CI' C2'

CY 03' P OL OR 05'

9.64 10.62 11.08 11.12 12.38 10.03

-1-00 -0-06 0.15 -1.16 0.87 1.12

C3' 03' P OL OR 05'

41.3 38-9 30.9 27.4 31,3 27.1

3.16 1.83

4.08 5.35 6.00 5.27

Adenine 47.4 59-4 60.0 41.3 23-3 19.6 7.9 12.6 24.4 37.1

Right-handed chain 9.66 26-7 9-24 35.6 8.12 36.2 7.59 45.7 8.19 49-1 8.69 40.4 9.67 42.7 10.71 49.7 11.63 48.9 11-36 49-4 10.07 57.9

Z (A) 0-00 0.00 0.00 0.00 0"00 0.00 0.00

0"00 0.00 0.00 0.00 0.00 0.00 0-00 0.00 0.00

0.00 0.00 0.00

0,86 0.71 -0.20 0.00 1.30 1.94 2.93 2.86 4.01 1.53 3.01

222

ft. c. HOPKINS

The final atomic cylindrical coordinates are listed in Table 1 for both right- and left-handed structures. In each case, the rise/bp=3.40/~ and twist/bp=34.6 ° (10.4 bp/turn). Torsion angles in the deoxyribophosphate chains are listed in Table 2. By using appropriate diad operations about the X and Z axes, both duplex and tetraplex forms can be readily generated. TABLE 2 Torsion angles in right- and left-handed chains of idealized Configuration H D N A models Symbol a fl 7 8 e ~" X

Defining atoms O3'-P-O5'-C5' P-O5'-C5'-C4' O5'-C5'-C4'-C3' C5'-C4'-C3'-O3' C4'-C3'-O3'-P C3'-O3'-P-O5' O1'-C1'-N9-C4 (Pur.) or OI'-CI'-N1-C2 (Pyr.)

Torsion angle (*) Left-handed chain Right-handed chain 169.3 165.5 154.3 108.0 173.8 -80-4

75.5 -172.7 168.8 83-8 84.8 66.0

145.3

9.0

In the Configuration I tetraplex model of McGavin (1971), the phosphate groups pack closely into the interduplex boundary regions, while the minor groove regions remain open. On the other hand, it is in the minor groove regions of the Configuration II models where a similar closeness of phosphate residues occurs; the open interduplex boundary regions contain only hydrophobic groups, plus the guanine N7 sites. As a result, these latter four-stranded structures are suggestive of unique functions (Hopkins, 1984, 1985). Divalent cations such as Mg ++ are expected to stabilize the Configuration II duplexes and tetraplexes because of the nearness of the backbone chains on complementary D N A strands. Thus, cations might form bridge bonds between neighboring phosphate groups as postulated in an early structural study of DNA (Franklin & Gosling, 1953). The geometry of the models presented here suggests that phosphate oxygen atoms in the minor grooves could provide several nearly octahedral coordination sites for each charge-neutralizing Mg ++.