The integrase family of tyrosine recombinases: evolution of a conserved active site domain

doi:10.1093/nar/25.18.3605

This Article

	Abstract
	Print PDF (976K)
	Alert me when this article is cited
	Alert me if a correction is posted

Services

	Email this article to a friend
	Similar articles in this journal
	Similar articles in ISI Web of Science
	Similar articles in PubMed
	Alert me to new issues of the journal
	Add to My Personal Archive
	Download to citation manager
	Search for citing articles in: ISI Web of Science (174)
	Request Permissions
	Commercial Re-use Guidelines for Open Access NAR Content

Google Scholar

	Articles by Esposito, D.
	Articles by Scocca, J. J.
	Search for Related Content

PubMed

	PubMed Citation
	Articles by Esposito, D.
	Articles by Scocca, J. J.

Social Bookmarking

	What's this?

The integrase family of tyrosine recombinases: evolution of a conserved active site domain

The integrase family of tyrosine recombinases: evolution of a conserved active site domain Dominic Esposito* and John J. Scocca¹

Laboratory of Molecular Biology, NIDDK, National Institutes of Health, 5 Center Drive MSC0560, Bethesda, MD 20782, USA and ¹Department of Biochemistry, The Johns Hopkins University School of Hygiene and Public Health, 615 N. Wolfe Street, Baltimore, MD 21205, USA

Received June 3, 1997; Revised and Accepted August 4, 1997

ABSTRACT

The integrases are a diverse family of tyrosine recombinases which rearrange DNA duplexes by means of conservative site-specific recombination reactions. Members of this family, of which the well-studied lambda Int protein is the prototype, were previously found to share four strongly conserved residues, including an active site tyrosine directly involved in transesterification. However, few additional sequence similarities were found in the original group of 27 proteins. We have now identified a total of 81 members of the integrase family deposited in the databases. Alignment and comparisons of these sequences combined with an evolutionary analysis aided in identifying broader sequence similarities and clarifying the possible functions of these conserved residues. This analysis showed that members of the family aggregate into subfamilies which are consistent with their biological roles; these subfamilies have significant levels of sequence similarity beyond the four residues previously identified. It was also possible to map the location of conserved residues onto the available crystal structures; most of the conserved residues cluster in the predicted active site cleft. In addition, these results offer clues into an apparent discrepancy between the mechanisms of different subfamilies of integrases.

INTRODUCTION

The integrases are a family of proteins that recombine DNA duplexes by executing two consecutive strand breakage and rejoining steps and a topoisomerization of the reactants (1 ,2 ). The first member of this family, the well-studied lambda Int protein, promotes integration and excision of the phage genome from that of the host (3 ); other family members function in the maintenance of plasmid copy number (4 ,5 ), the elimination of dimers from replicated chromosomes (6 ) and in alteration of cell-surface components (7 ), as well as in the life cycle of temperate phages (8 -10 ). All these processes involve the conservative site-specific recombination of two DNA partners. The initial definition of the family was based on comparisons of seven sequences, and three invariant residues were identified: a His-X-X-Arg cluster and a Tyr residue (11 ). Alignment of 28 sequences identified a fourth invariant position, occupied by an Arg residue (12 ). These four conserved residues are located in the C-terminal half of the protein sequences, and occur in the order Arg, His-X-X-Arg and Tyr, with Tyr closest to the C-terminus. Mutations introduced at each of these conserved or invariant positions in several different systems produced proteins inactive in recombination, as would be expected if these positions corresponded to active site residues (12 -15 ). In those systems where the question has been examined, reaction proceeds through a covalent intermediate in which the DNA 3'-phosphoryl group is esterified to the hydroxyl group of the conserved tyrosyl residue (16 ); consequently the group is also known as the tyrosine recombinase family. The emergence of large scale genomic sequencing and the recognition of the role of tyrosine recombinases in the terminal stages of bacterial chromosome replication have increased the number and functional diversity of the known members of this family.

None of the earlier comparisons examined the possible evolutionary relationships among members of the family. Such an analysis has been hampered by the limited degree of similarity among the known integrases. This diversity is due in part to the fact that many of these proteins are bifunctional DNA binding proteins. For instance, both the HP1 and lambda phage integrases bind to two distinct specific DNA sequences (17 ,18 ). In both cases, the N-terminal portion of the protein exhibits specific binding to one sequence, while the C-terminal region, which contains the active site residues, interacts with a different sequence present at the recombination sites. Even those proteins that bind only to the recombination site and have a single recognition specificity (Flp, Xer) are very different in the N-terminal half of their sequences. These considerations suggested that the most informative comparisons and alignments of family members would be obtained if they were confined to the C-terminal portions of the sequences. We have therefore undertaken to align the C-terminal segments of 81 members of the tyrosine recombinase family available in the DNA database, and to explore the evolutionary relationships among these sequences. Sequence similarities outside the four invariant positions suggest the presence of residues potentially important for the reactions promoted by the family members. Evolutionary relationships emerging from the comparison indicate that integrases aggregate into subfamilies based on their biological roles. Recently, the structures of the C-terminal active site domains of two members of the family, HP1 integrase (19 ) and lambda Int (20 ), have been determined, allowing us to place conserved residues within the three-dimensional structure. The majority of these residues occupy positions either within the active site cleft or very near it. This suggests that the prokaryotic members of the family share a common architecture in their active sites. Another clear result of these comparisons was that the yeast-derived recombinases formed a group which was at best remotely related to the prokaryotic group. The extreme divergence in sequence between these two groups may explain the apparent mechanistic discrepancy between them; two protomers of the yeast FLP recombinase together form a single active site (21 ), while each protomer of HP1 integrase contains a complete active site (19 ).

MATERIALS AND METHODS

Computer programs

GCG programs (http://www.gcg.com) were run on the NIH Helix server. BLAST searches were performed on the NCBI web site (http://www.ncbi.nlm.nih.gov). The PHYLIP package for evolutionary analysis (http://evolution.genetics.washington.edu) was run on a Power Macintosh. ClustalW alignments were performed with Oxford Molecular MacVector software (http://www.oxmol. com).

Sequence acquisition

Previously identified tyrosine recombinase protein sequences (12 ,22 ) were obtained from GenBank. Using 30 amino acid segments from the conserved regions of several of the known family members, BLAST searches were done and additional members of the family were identified and retrieved from GenBank. Several sequences were identified as putative recombinases when deposited with GenBank, and were also retrieved for this analysis. A total of 135 sequences were identified in these searches; of these, a number were completely identical. The most commonly duplicated sequence was TnpA of Tn21, which occurs in several species, and in numerous cloning vectors. After all duplicate sequences were removed, 81 unique sequences remained. These sequences, and their GenBank accession numbers, are identified in Table 1 . Two partial sequences (LO L5 and MP int) encoding the C-terminal ends of these presumptive proteins both contained sufficient regions of similarity to warrant their inclusion in the family.

Sequence alignments

Multiple alignment of the sequences was carried out in steps. First, the automatic Higgins alignment of MacDNASIS was combined with manual refinement to produce a crude alignment of all 81 sequences. From this initial alignment an ~200-residue region of similarity was chosen starting 10 amino acids prior to the first conserved Arg, and ending five amino acids after the conserved Tyr. This region contained almost all the similarities shared among the family members, and was used in subsequent alignments and comparisons. At this stage it was obvious that the six eukaryotic sequences were clearly distinct from the rest of the family. They were therefore aligned separately. Four randomly selected groups of 19 sequences were then each subjected to ClustalW alignment to produce four refined alignments. From these alignments, different sets of 19 sequences were randomly chosen and realigned with ClustalW, and this process of random selection and realignment was repeated twice more to produce a final alignment of the 75 sequences. This final alignment was then subjected to a single round of further refinement using GCG PILEUP. The final output of the programs was examined, and minor adjustments to the alignment were made manually. The eukaryotic recombinases were aligned separately with GCG PILEUP, and the aligned sequences were manually aligned with the PILEUP results from the prokaryotic recombinases.

Evolutionary analysis

The Macintosh version of the public domain PHYLIP package was used to construct evolutionary trees. Seventy-nine of the aligned sequences from PILEUP (excluding the partial LO L5 and MP int sequences) were input into the PROTDIST program to identify separation distances. All of the available PROTDIST distance algorithms were attempted; in no case were finite distances for all 79 sequences obtained. Removal of the six eukaryotic sequences produced a set of finite distances using the Kimura algorithm, while additional removal of two prokaryotic sequences (pSE101 and pSE211) allowed all 71 sets of distances to be determined under the Kimura, Dayhoff PAM and Categories algorithms. The Dayhoff PAM calculated distances of the 71 proteins were used as the input file for the KITSCH algorithm. The input order of sequences was randomized three times and the best-fit trees were identified using a power factor of 2 with no subreplicates. Using the KITSCH data, DRAWGRAM was used to construct a phenogram of the most likely tree. Two additional runs of the PROTDIST and KITSCH programs were carried out using the same protein sequences in a scrambled order; in both cases, similar consensus trees were determined. Attempts to map the two prokaryotic and six eukaryotic sequences back onto the final tree failed, suggesting they are beyond the evolutionary distance calculable with any of the algorithms.

RESULTS

Selection of sequences for analysis

Eighty-one different tyrosine recombinase sequences were retrieved using the criteria described in Materials and Methods; these sequences are listed in Table 1 . These were aligned initially to identify the regions of maximal similarity. Not surprisingly, it was almost impossible to include the N-terminal sequences of these proteins in any reasonable alignment. These regions of the proteins seem to have diverged so extensively that their relationships, if any, have been entirely obscured. The exception to this rule was the Fim family of proteins, in which the region of similarity begins near the N-termini of the proteins, suggesting that these proteins may lack the bipartite DNA binding specificity. This divergence in the N-terminal segments of the family is consistent with the proposed function of these regions in binding to DNA sites located away from the recombination points. These arm or organizing sites contribute importantly to the directionality and specificity of integrases in organizing the condensed intasomes, and consequently differ from system to system (23 ,24 ).

Table 1 The 81 tyrosine recombinases analyzed in this study

Name Host Location Gene Accession R1 R2 Y

BS codV Bacillus subtilis chromosome codV U13634 149 249 281

BS ripX B.subtilis chromosome ripX U32685 15 112 144

BS ydcL B.subtilis cryptic prophage ydcL AB1488 201 316 351

CB tnpA Clostridium butyricum chromosome tnpA Z29084 194 302 335

Col1D Escherichia coli miniF plasmid D X04967 82 205 237

CP4 E.coli cryptic prophage CP4-57 int U03737 248 340 373

Cre E.coli phage P1 int X03453 173 292 324

D29 Mycobacterium smegmatis phage D29 int X70352 166 278 310

DLP12 E.coli phage DLP12 int M31074 215 317 349

DN int Dichelobacter nodosus chromosome orf X98546 122 212 245

EC FimB E.coli chromosome fimB X03923 47 144 176

EC FimE E.coli chromosome fimE X03923 41 139 171

EC orf E.coli chromosome orf U73857 306 399 432

EC xerC E.coli chromosome xerC M38257 148 243 275

EC xerD E.coli chromosome xerD M54884 148 247 279

[Phi]11 Staphylococcus aureus phage phi11 int M34832 195 299 332

[Phi]13 S.aureus phage phi13 int U01875 197 290 323

[Phi]80 E.coli phage phage phi80 int X04051 256 355 387

[Phi]adh Lactobacillus gasseri phage phi-adh int M62697 218 333 366

[Phi]CTX Pseudomonas aeruginosa phage phiCTX int S75107 209 333 367

[Phi]LC3 Lactococcis lactis phage phiLC3 int X57797 203 319 352

FLP Saccharomyces cerevisiae 2[mu] plasmid FLP J01347 191 308 343

[Phi]R73 E.coli retronphage R73 int M64113 241 333 366

HI orf Haemophilus influenzae chromosome orf1572 U32831 73 161 190

HI rci H.influenzae chromosome rci U32821 174 259 291

HI xerC H.influenzae chromosome xerC U32750 145 240 272

HI xerD H.influenzae chromosome xerD U32716 147 246 278

HK22 E.coli phage HK022 int X51962 212 311 342

HP1 H.influenzae phage HP1 int U24159 207 283 315

L2 Mycoplasma sp. phage L2 int L13696 144 236 268

L5 Mycobacterium tuberculosis phage L5 int P22884 205 317 349

L54 S.aureus phage L54 int M14371 182 301 334

[lambda] E.coli phage lambda int J02459 212 311 342

LL orf Lactobacillus leichmannii chromosome orf X78999 153 252 283

LL xerC L.leichmannii chromosome xerC X84261 145 244 276

LO L5 Leuconostoc oenos phage L5 int L06183 * 43 79

MJ orf Methanococcus jannaschi chromosome orf U67489 177 278 310

ML orf Mycobacterium leprae chromosome orf U00021 162 264 296

MP int Mycobacterium paratuberculosis chromosome int L39071 * 78 114

MT int Mycobacterium tuberculosis chromosome int Z80225 204 284 316

MT orf M.tuberculosis chromosome orf Z74024 166 264 296

MV4 Lactobacillus delbrueckii phage MV4 int U15564 266 372 407

P186 E.coli phage 186 int X04449 203 280 312

P2 E.coli phage P2 int M27836 194 272 304

P21 E.coli phage P21 int M61865 228 335 364

P22 Salmonella typhimurium phage P22 int X04052 216 317 349

P4 E.coli phage P4 int X05947 245 351 385

P434 E.coli phage 434 int M60848 212 311 342

PA sss Pseudomonas aeruginosa chromosome sss X78478 146 240 272

PM fimB Proteus mirabilis chromosome fimB Z32686 58 155 187

pAE1 Alcaligenes eutrophus plasmid pAE1 orf L34580 257 356 388

pCL1 Chlorobium limicola plasmid pCL1 fim U77780 41 137 169

pKD1 Kluyveromyces lactis 2[mu] plasmid FLP P13783 187 301 338

pMEA Amycolatopsis methanolica plasmid pMEA300 orf L36679 217 333 366

pSAM2 Streptomyces ambofaciens plasmid pSAM2 orf X14899 208 330 363

pSB2 Zygosaccharomyces bailii 2[mu] plasmid FLP M18274 190 304 342

pSB3 Zygosaccharomyces bisporus 2[mu] plasmid FLP P13784 187 302 339

pSDL2 Salmonella dublin plasmid pSDL2 resV A38114 74 197 229

pSE101 Saccharopolyspora erythraea plasmid pSE101 orf L11597 217 392 425

Table 1. continued

Name Host Location Gene Accession R1 R2 Y

pSE211 Saccharopolyspora erythraea plasmid pSE211 orf M35138 214 382 414

pSM1 Zygosaccharomyces fermentati 2[mu] plasmid FLP P13770 207 320 358

pSR1 Zygosaccharomyces rouxii 2[mu] plasmid FLP P13785 140 254 292

pWS58 L.delbrueckii plasmid pWS58 orf Z50864 172 280 312

R721 E.coli plasmid IncI2 (R721) rcb X62169 156 239 272

Rci E.coli plasmid IncI1 (R64) rci X12577 155 238 271

SF6 Shigella flexneri phage Sf6 int X59553 234 327 360

SLP1 Streptomyces coelicolor plasmid SLP1 orf X71358 274 400 432

SM orf Serratia marcescens chromosome orf D50438 155 289 321

SsrA Methanosarcina acetivorans plasmid pC2A ssrA U78295 158 259 291

SSV1 Sulfolobus sp. virus 1 int X07234 211 281 314

T12 Streptococcus pyogenes phage T12/T270 int U40453 201 304 337

Tn21 E.coli transposon Tn21 int M33633 146 280 312

Tn4430 Bacillus thuringiensis transposon Tn4430 int X07651 145 237 269

Tn554a S.aureus transposon Tn554 tnpA X03216 198 305 338

Tn554b S.aureus transposon Tn554 tnpB K02987 363 468 500

Tn7 E.coli transposon Tn7 int L10818 135 269 301

Tn916 Entercoccus faecalis transposon Tn916 int M37184 225 346 379

Tuc Lactobacillus lactis phage Tuc2009 int L31348 203 319 352

WZ int Weeksella zoohelcum chromosome orf U14952 102 202 234

XisA Anabaena sp. nifD locus xisA P08862 287 384 416

XisC Anabaena sp. hupL locus xisC U08014 306 401 433

Listed are the name used to identify the protein in the alignments, the host organism the protein was identified in, the location of the protein (chromosomal, phage- encoded, plasmid, etc.), the name of the gene encoding the protein, the GenBank accession number from which the sequence was retrieved, and the amino acid positions of the three completely conserved residues found in all tyrosine recombinases. R1 is the arginine found in Box A, R2 is the arginine found in Box B and Y is the catalytic tyrosine found in Box C. The two sequences which contain asterisks in the R1 column come from incomplete DNA sequence data which are missing N-terminal portions of the proteins. In addition, the sequence of BS ripX is only partial, but contains the entire C-terminal region studied in this paper. All other entries consist of complete protein sequences.

Because of these divergent N-terminal segments, the analysis concentrated on the C-terminal portion of each sequence. The similarities in sequence began 10 amino acid residues prior to the initial conserved arginine residue; no significant similarity among a majority of family members was identified upstream of this point. The region analyzed ended five amino acids after the active site tyrosine; again, after this region, no similarity was observed. The segments being compared were between 123 and 208 amino acid residues in length depending on the individual protein.

All of the 81 sequences retrieved initially contained at least three of the four `invariant' residues identified previously (12 ). An additional sequence, the integrase of phage AAU2, was identified by its depositors as being a member of the family (25 ). However, inspection of this sequence showed that it had no significant homology to other family members, and was lacking several of the `invariant' residues, suggesting that it had been misclassified. The remaining 81 sequences were compared using both automated and manual methods to produce a best-fit alignment; this alignment required a span of 223 amino acids to accommodate gaps. Table 1 lists the sequences, their origins, and the positions of the two invariant arginine residues and the invariant tyrosine residue in the complete protein sequences. As will be discussed below when the evolutionary relationships of these sequences are examined, the six sequences derived from yeast plasmids form a group marked by extensive divergence from the prokaryotic representatives. Consequently the 75 prokaryotic and six fungal sequences were aligned separately. Initial alignments revealed significant similarities among certain family members, suggesting the presence of subfamilies of presumably homologous proteins. Clear global similarities among nearly all of the proteins were apparent as well. These global similarities occurred in three major clusters, designated Box A, Box B and Box C. Each cluster surrounded one or more of the conserved residues identified in earlier work. The spacing between Box A and Box B varied considerably, while the spacing between Boxes B and C was less variable. The locations of these boxes in representative members of the family are compared in Figure 1 .

Nucleic Acids Research