Protein sequence analysis: protein sequence was obtained from Entrez internet server, http://www.ncbi.nlm.nih.gov. Sequence matrix comparison was done with DNA Strider program ver 1.2 on a Macintosh computer. Sequence analysis was done with the Expasy tools on the internet, http://www.expasy.ch. operated by University of Geneva [Appel 1994]

Protein structure analysis: observations using Insight software (MSI, ver. 95) were directed toward those that affect inhibition and that were not covered in previous publications. Special consideration was given to examining the structure for consistency with the widely held model of 5’ attachment and 5’ overhang of the temporary cuts in DNA.

Homology generation of human topoisomerase: the general steps of building a homology model are: identify a template having sufficient identity, generate sequence alignment & coordinate, optimize to relieve contacts, and conduct quality validation [Peitsch 1997]

The PDB coordinate file of the S. cerevisiae topoisomerase 2, residues 420-1178, was kindly provided by Dr. James Wang (now available online from Brookhaven PDB as 1bgw). Resolution was reported as 2.7 A. Using the homology module of Insight software [MSI, ver 95] on Silicon Graphics Indigo workstation, the PDB file was visualized and the sequence was extracted. The human sequence obtained from NCBI Entrez was imported and aligned pair-wise.

Because of poor match-up of the first few residues and our observation that a section of high identity begins at 428/438, the homology alignment was begun there. Some small gaps in the alignment were treated with the loops software tools. Some substantial gaps in the crystal must be skipped. Because the main interest is in the active site tyrosine, the long helical arms (after 974) are not done in this model.

The resulting model therefore includes residues (yeast/human):

yeast: 428-------639 715—787 875-----974

hum: 438-------648 736---808 896-----997

using 25 residues per space the scale appears as:

……………..---------…---….----…………………

Of these seven discontinuities, three are near the active site tyrosine and the loop routine was used to connect them with the main structure. Bumps routine was used to determine amino acid overlaps >.5 and to resolve them.

SC: 420-RKSRITNY-PKLEDANKAGTKE

HUM: 430-khn.ikgi-...D...D..GRN

SC: 441-GYKCTLVLTEGDSALSLAVAGLAVVGRDYYGCYPLRGKMLNVREASADQI

HUM:451-STE...I.......KT...S..G.....K..VF.....I.......HK..

SC: 491-LKNAEIQAIKKIMGLQHRKKYEDT---KSLRYGHLMIMTDQDHDGSHIKG

HUM:501-ME....NN.I..V...YK.N...EDSL.T....KI.......Q.......

SC: 538-LIINFLESSFLGLLDIQGFLLEFITPIIKVSITKPTKNTIAFYNMPDYEK

HUM:551-.L...IHHNWPS..R-HR..E......V...--.-n.QEM...SL.EF.E

SC: 588-WREEESHKFTWKQKYYKGLGTSLAQEVREYFSNLDRHLKIFHSLQGNDKD

HUM:597-.KSSTPNHKK..V.........TSK.AK...ADMK..RIQ.KYSGPE.DA

SC: 638-AA 715-NLKSELKVAQ

HUM:647-.I 736-.D.R.V....

SC: 725-LAPYVSECTAYHHGEQSLAQTIIGLAQNFVGSNNIYLLLPNGAFGTRATG

HUM:746-..GS.A.MSS.....M..MM...N..........LN..Q.I.Q....LH.

*

SC: 775-GKDAAAARYIYTE

HUM:796-...S.SP...F.M

SC: 875--GTIEEIEPLRYRMYGRIEQIGDNVLEITELPARTWTSTIKEYLLLGLS-

HUM:896-k......A.NQ.VIS.EVAILNSTTI..S...V....Q.Y..QV.EPMLN

SC: 923-GNDKIKPWIKDMEEQHDDN-IKFIITLSPEEMAKTRKIGFYERFKLISPI

HUM:945-.TE.TP.L.T.YR.Y.T.ttv..VVKMTE.KL.EAERV.LHKV...QTSL

SC: 972-SLM

HUM:995-TCN

. = identity

- = gap

For sections of high similarity, the human sequence was assigned coordinates equivalent to the yeast topoisomerase template.. Kabsch-Sander secondary prediction was used to guide alignments for portions having less similarity. For breaks in the alignment the loops tools and bump check were used to join the similarity sections. The ending structure was refined for 200 iterations with backbone fixed and then without constraints.

Because a portion of the second monomer of the topoisomerase homo-dimer is near the active site, a portion (560-639) of the human structure was duplicated and placed relative to the alpha monomer by superimposition over the Saccharomyces topoisomerase crystal coordinates.

Mutant structure analysis: literature was scrutinized for topoisomerase mutations induced by drugs for which the site of change was known and mapped onto the crystal structure.

Trans-esterification semi-empirical calculation: this procedure was based on Spartan software on a SGI computer using three common hamiltonians, AM1, PM3 and MNDO on a tyrosine before and after attachment to the phosphate.