www.elsevier.com/locate/yexpr

Experimental Parasitology 109 (2005) 143–149

Strand asymmetry patterns in trypanosomatid parasites

Daniel Nilsson, Bjorn Andersson*

Center for Genomics and Bioinformatics, Karolinska Institutet, Berzeliusv. 35, SE-171 77 Stockholm, Sweden

Received 12 August 2002; received in revised form 1 December 2004; accepted 1 December 2004

Abstract

The genome organization of kinetoplastid parasites is unusual, with chromosomes containing several long regions of polycistron-ically transcribed genes. The regions where the direction of transcription switches have been hypothesized to contain origins of rep-lication and possibly also centromers and promoters. We report that overall strand asymmetry patterns can be observed inTrypanosoma cruzi and Trypanosoma brucei with optima on strand-switch regions. The base skews of T. cruzi and T. brucei diver-gent strand-switches show patterns analogous to those for bacterial origins of replication, but they differ from those of Leishmania

major. Bias in codon usage and the trypanosomatid unidirectional gene clusters predict most of this skew, but fail to properlyexplain the same trend in intergenic regions, as does the current knowledge of regulatory sequences.� 2005 Elsevier Inc. All rights reserved.

Index Descriptors and Abbreviations: Parasite genomics; Parity rule two deviation; Strand asymmetry; Codon usage; Trypanosomatids; Lm,Leishmania major Friedlin; Tb, Trypanosoma brucei; Tc, Trypanosoma cruzi; Ori, Origin of replication; Ter, Termini of replication

1. Introduction

Studies of trypanosomatid parasites have revealedsurprising phenomena in the past, such as trans-splicingand RNA editing. The first longer contiguous genomicsequences from the genome projects in Leishmania major

Friedlin (Lm), Trypanosoma brucei (Tb), and Trypano-

soma cruzi (Tc) (reviewed in Degrave et al., 2001) havebegun to reveal the genome organization of these para-sites. Trypanosomatid genes are almost exclusively en-coded in unidirectional clusters, with interveningstrand-switches (El-Sayed et al., 2003; Hall et al., 2003;Worthey et al., 2003). The genes in a cluster are tran-scribed polycistronically and trans-spliced to formmature mRNA molecules from each gene. The strand-switches have attracted considerable interest, and havebeen suggested to contain regulatory elements, tran-scriptional promoters (Wong et al., 1994), origins of rep-

0014-4894/$ - see front matter � 2005 Elsevier Inc. All rights reserved.

doi:10.1016/j.exppara.2004.12.004

* Corresponding author. Fax: +46 8 311620.E-mail address: bjorn.andersson@cgb.ki.se (B. Andersson).URL: http://cruzi.cgb.ki.se/ (B. Andersson).

lication, as well as to constitute centromers (Myler et al.,1999) of the trypanosomatid chromosomes, but no con-clusive evidence for any of these elements has as yet beenpresented.

Analyses of genomic sequences have revealed distinctpatterns of deviation from Chargaff�s second rule of par-ity (PR2) (Lobry, 1996; Wu and Maeda, 1987). PR2states that the intra-strand nucleotide concentration[C] = [G], as well as [A] = [T], under equilibriumassumptions given Watson–Crick basepairing and nobias in selection or mutation between the strands(Lobry, 1995; Sueoka, 1995). Many eubacterial genomesshow a strong correlation between coding excess (gene-orientation) and the degree of deviation from PR2,quantified as GC-skew and AT-skew, with optima inthe cumulative measures precisely at the origin (ori)and terminus (ter) of replication (Grigoriev, 1998). TheGC-skew has proven to be more predictive than theAT-skew in this regard. Also, plots of purine excess havebeen demonstrated to correlate with ori and ter in some,but not all, cases (Freeman and Plasterer, 1998). The sit-uation in eukaryotes is less clear, with many local skew

144 D. Nilsson, B. Andersson / Experimental Parasitology 109 (2005) 143–149

optima (minima and maxima), that have been postu-lated to coincide with multiple ori and ter (Shioiri andTakahata, 2001). Recent findings support this (Niuet al., 2003).

Many alternative explanations for the origin of thesepatterns have been proposed (see, e.g., Frank andLobry, 1999, for a review). Several cellular processesare asymmetric with respect to DNA strands and thusconstitute potential causes of asymmetric basedistributions.

Errors introduced in transcription coupled repair ofthe coding strand have been identified as causative ofbase skew in certain organisms (Francino and Ochman,1997; Green et al., 2003). In transcription, the codingstrand (non-template) is more unprotected, whereasthe non-coding (template) strand is not only more pro-tected, but also undergoes correcting excision repair toa greater extent (Mellon and Hanawalt, 1989). Also,during replication, the strands are copied by differentsystems of enzymes, and face different risks of DNAdamage, as well as different repair mechanisms (Fijalk-owska et al., 1998; Izuta et al., 1995; Maliszewska-Tkac-zyk et al., 2000). This can lead to different fidelity ofreplication for the leading and lagging strands (Reyeset al., 1998; Rocha and Danchin, 2001). At least ineubacteria, the direction of replication and transcriptionis often correlated (Brewer, 1988; McLean et al., 1998)so that a skew introduced by either process can be addedto that of the other.

Regular patterns of purine excess and GC- andAT-skew have been found in Lm chromosome 1(McDonagh et al., 2000), surprisingly showing a neg-ative correlation between coding excess and GC-skew,which is the opposite of the pattern found in eubacte-ria. The analysis of Lm led McDonagh et al. to sug-gest that the effects of transcription coupled repairare balanced by ubiquitous and strand un-specifictranscription, which has been found to occur in kine-toplastids (Clayton, 2002).

This interpretation now seems unlikely given resultsfrom studies on tentative promoter activity in trypano-somes that suggest more strand-specific transcription(Martinez-Calvillo et al., 2003), from the non-coding(template) strand only, as in many other organisms. Thisview is complicated by results that show a backgroundlevel of transcription initiation independent of promot-ers (Clayton, 2002, and references therein). It is still un-clear how extensive this background transcription is,and if all regions of the genome behave equally in thisregard.

We find that the skew resulting from unequal codonusage explains much of the correlation between codingexcess and skew seen in the three parasites investigated,but that it is clearly present also in the third codon posi-tion, and furthermore, that this pattern is present inintergenic regions as well. Intriguingly, the chromo-

somes of Tb and Tc show a positive correlation betweencoding excess and GC-skew—the opposite of that of theclosely related Lm.

2. Methods

The sequences used were Tb chromosome I (Gen-Bank Accession No. AL359782), Tc chromosome 3(AF052831, AF052832, and AF052833), and Lm chro-mosome 1 (NC_001905).

The sequence was subdivided into coding and non-coding according to the coordinates given in GenBankCDS entries. All calculations were performed on the5 0–3 0 strand exclusively, compensating for the codon po-sition of any ‘‘reverse’’-strand coding regions accord-ingly. The strand-switches were set to be at nucleotideposition 24,500 for Tc, 78,000 for Lm, and 243,000 forTb.

The cumulative GC (or AT) base skew sRY was calcu-lated as

sjRY ¼Xj

n¼0

Xl

i¼1

dnþiR � dnþi

Y

dnþiR þ dnþi

Y

;

where N is the sequence length, l is the window length,and j 6 N � l is the sequence position. Arbitrary over-lapping window lengths of 400 and 10,000 bp have beenused. dxR is 1 for position x only if there is a purine R atthat position and zero otherwise; dxY analogously forpyrimidine Y.

Coding excess was calculated as

sce ¼XN

n¼1

dnF � dnR;

where dnF is one if and only if base n is in a gene encodedon the forward strand, and 0 otherwise. dnR, analogously,takes the value 1 only for bases in genes encoded on thereverse strand, and 0 otherwise. Thus sce is a cumulativemeasure of which strand the genes are encoded on, withoptima at strand-switches.

The skew contributions for each of the 64 codonswere calculated, and weighted with the codon usage(as estimated by Nakamura et al., 1999) for each ofthe three parasites. This is the expected average skewper nucleotide given uniform amino-acid and stop co-don usage. We note that the codon usage estimationsrely in part on data from the sequences subsequentlyanalysed in this work. The sequences analysed contrib-ute only a small fraction of the CDS entries used inthe codon usage estimation for Tc (39/420) and Lm

(79/1426), but for Tb the fraction is relatively large(325/422).

The tandem repeat analyses were executed as inDuhagon et al. (2001), locating exact repeats using a perl(Wall and Schwartz, 1991) program and with the

D. Nilsson, B. Andersson / Experimental Parasitology 109 (2005) 143–149 145

EMBOSS repeat analysis program etandem (Rice et al.,2000) for location of inexact tandem repeats. A mini-mum and maximum repeat size of 2 was given, uniformrepeats were enabled and the scoring threshold set to 6.The R package (Ihaka and Gentleman, 1996) was usedfor statistical analyses.

Images were produced using R R Development CoreTeam (2004). Sequence handling and skew processingwas implemented in a set of perl programs, availablefrom the author upon request.

All computations were performed on a GNU/LinuxIntel Pentium M 1.6 GHz laptop.

3. Results

We have investigated the cumulative GC and ATbase skews in overlapping windows of 400 bp over aTc chromosome 3 (Andersson et al., 1998) strand-switch, as well as in overlapping windows of 10,000 bpon Tb chromosome I (Hall et al., 2003) and Lm chromo-some 1 (Myler et al., 1999), with qualitatively similar re-sults. Notably, the converging strand-switches analysedfollowed the same pattern of correlation between codingexcess and GC-skew as divergent switches. For brevity,we present the results for one (divergent) strand-switchfrom each organism.

Fig. 1. From the top left and clockwise for the entire region: deviations frTrypanosoma cruzi chromosome 3 (GenBank Accession Nos. AF052831, AFsequence, and the third codon position of the coding sequence.

The results show that overall strand asymmetry pat-terns can indeed be found in Tc and Tb (Figs. 1 and2). As in eubacteria, but contrary to what has beenfound in the closely related Lm (McDonagh et al.,2000) (Fig. 3), a positive correlation between coding ex-cess and GC-skew was found. This pattern was also seenfor the third codon position and in intergenic regions,which indicated that direct amino acid selection is notthe only cause of this pattern.

The weighted sum of codon skews (Table 1) correctlypredicted the sign of the correlation between skew andcoding excess for all three parasites. Codon usage couldnot, however, explain the similar patterns observed fornon-coding regions.

4. Discussion

It is important to note that most of the correlationbetween the skew of coding regions and coding excessin the three parasites can be explained by codon usage,both for the positive correlation in Tb and Tc and thenegative in Lm. Codon usage cannot, however, explainthe correlations observed for non-coding regions. In-stead, it seems clear that mechanisms introducing skewaffect both intergenic and coding regions.

om Chargaff�s second rule of parity and coding excess over a part of052832, and AF052833), the coding sequence, non-coding (intergenic)

Fig. 2. From the top left and clockwise: deviations from Chargaff�s second rule of parity and coding excess over a part of Trypanosoma brucei

chromosome I (GenBank Accession No. AL359782) for the entire region, the coding sequence, non-coding (intergenic) sequence, and the third codonposition of the coding sequence. The dashed vertical line indicates the strand-switch position.

Fig. 3. From the top left and clockwise: deviations from Chargaff�s second rule of parity and coding excess over Leishmania major Friedlinchromosome 1 (GenBank Accession No. NC_001905) for the entire region, the coding sequence, non-coding (intergenic) sequence, and the thirdcodon position of the coding sequence. The dashed vertical line indicates the strand-switch position.

146 D. Nilsson, B. Andersson / Experimental Parasitology 109 (2005) 143–149

Table 1Weighted sum of skew contributions from each codon with codonusage frequency

GC skew AT skew

Leishmania major �0.014 0.039Trypanosoma brucei 0.064 0.060Trypanosoma cruzi 0.073 0.035

D. Nilsson, B. Andersson / Experimental Parasitology 109 (2005) 143–149 147

In the following, four possible explanations will bediscussed: the tentative nature of primary annotations,functional intergenic elements and the related questionof intergenic repeats, strand asymmetric DNA catabo-lism, and possible trypanosomatid ori/ter.

A part of the correlation could be due to the tentativenature of the primary annotation of the genome se-quence. The primary annotations carried out in our lab-oratory and the other centres involved in the threegenome projects are likely to err both by calling non-functional sequences as coding and by missing actualcoding sequences, especially for short genes and func-tional non-messenger RNA transcripts. Systematic biasin the gene calling procedures, and particularly manygenes called on the wrong strand, could affect the corre-lation between coding direction and the skew present onthe genomic sequence.

A second possibility would be the existence of strand-specific regulatory elements, co-oriented with the pro-tein coding sequences that give rise to a base skew.The current understanding of regulatory elements in try-panosomatids is limited. Of the known elements, themost abundant (Hotz et al., 1997) is a polypyrimidine(CT) stretch in the 3 0-untranslated regions, implictedin splicing and polyadenylation, that is present in a largeportion of the predicted genes from all three parasites.This could be a part of the explanation for the skew ob-served in Lm, but cannot explain the pattern seen in Tb

and Tc where the GC-skew is instead positively corre-lated with coding excess. Even though the intergenic re-gions are relatively short, novel strand asymmetricalregulatory elements are likely to be too few and tooshort to affect the overall skew in a significant way.

Related to this is the occurrence of repeated, inter-genic elements with strand asymmetry. Both Tb andTc have highly repetitive genomes, whereas the Lm gen-ome is known to have a lower repeat content (Anders-son et al., 1998; El-Sayed et al., 2000; Myler et al.,1999). The longer, more complex and usually inter-spersed elements, transposons or transposon like se-quences are not likely to be the cause of the baseskew, given the relatively continuous nature of thecurves observed. Even in the highly repeated parasitegenomes, these elements are only present in some inter-genic regions. Unexpectedly, large amounts of short, di-and oligonucleotide repeats are present in the intergenicregions of Tc, and strand asymmetric poly[dT–dG] re-peats have been demonstrated (Duhagon et al., 2001).

When such repeat analyses were performed for these re-gions of Tb and Lm (data not shown), Tb and Tc werefound to be similar in this respect. In the case of dinucle-otide repeats, strong strand asymmetry was detectedonly for poly[dT–dG] Æ poly[dC–dA] in Tb and Tc. InLm both poly[dA–dG] Æ poly[dC–dT] and poly[dC–dC] Æ poly[dG–dG] was found to have strong strandasymmetry in Lm. Thus, we could not rule out that apart of the difference in intergenic strand skew couldbe caused by simple repeats.

As a third possibility, the difference in intergenic skewpattern could be caused by a difference between Lm andthe other two investigated kinetoplastids in a strandasymmetric activity, such as replication or transcriptioncoupled repair. Such mechanistic differences could be ex-plained by a difference in lifestyle between the parasites.One would perhaps expect the extracellular Tb to differin repair or susceptibility to mutation from Lm or Tc

who both invade mammalian host cells. This seemsnot to be the case. But the parasites also have differentvector organisms. Any of the Lm host environmentsmay have forced or allowed a modification in DNAdamage tolerance or repair acting asymmetrically onthe DNA strands.

A fourth possible interpretation would be that thestrand-switch regions constitute origins of replication,at least in Tc and Tb. Skew optima often coincide withori and ter in eubacteria (Brewer, 1988; McLean et al.,1998). The negative correlation between coding excessand GC-skew in Lm would indicate replication in theopposite direction compared to Tb and Tc. This seemsunlikely, due to the strong conservation of synteny be-tween the three parasites (Bringaud et al., 1998; Ghedinet al., 2004). By analogy with the eubacterial skew pat-terns, Lm would require additional functions for resolv-ing head-on DNA and RNA polymerase collisions(Brewer, 1988). No nuclear genome ori or ter have sofar been reported for these trypanosomatids.

In summary, we present base skew patterns fromthree parasites. Most notably, the two trypanosomatidsnot previously subject to this kind of study showed skewpatterns different to that of the closely related Lm. Thepatterns in Tb and Tc were similar to those reportedfor eubacteria. The contribution of the coding regionsto the base skews is demonstrated to be qualitativelypredicted by a skew inherent in the synonymous codonusage. The intergenic regions were also found to exhibitbase skew patterns qualitatively similar to the globalpattern. Four potential explanations for these findingswere discussed.

Acknowledgments

The authors acknowledge Martti T. Tammi for valu-able discussions. We also acknowledge the Swedish

148 D. Nilsson, B. Andersson / Experimental Parasitology 109 (2005) 143–149

Technology Research Council, the Beijer Foundation,and NIH/NIAID U01AI45061 for financial support.

References

Andersson, B., Aslund, L., Tammi, M., Tran, A.-N., Hoheisel, J.D.,Pettersson, U., 1998. Complete sequence of a 93.4-kb contig fromchromosome 3 of Trypanosoma cruzi containing a strand-switchregion. Genome Research 8, 809–816.

Brewer, B.J., 1988. When polymerases collide: replication and thetranscriptional organization of the E. coli chromosome. Cell 53,679–686.

Bringaud, F., Vedrenne, C., Cuvillier, A., Parzy, D., Baltz, D., Tetaud,E., Pays, E., Venegas, J., Merlin, G., Baltz, T., 1998. Conservedorganization of genes in trypanosomatids. Molecular and Bio-chemical Parasitology 94 (2), 249–264.

Clayton, C., 2002. Life without transcriptional control? From fly toman and back again. The EMBO Journal 21 (8), 1881–1888.

Degrave, W.M., Melville, S., Ivens, A., Aslett, M., 2001. Parasitegenome initiatives. International Journal for Parasitology 31, 532–536.

Duhagon, M.A., Dallagiovanna, B., Garat, B., 2001. Unusual featuresof poly[dT–dG][dC–dA] stretches in CDS-flanking regions ofTrypanosoma cruzi genome. Biochemical and Biophysical ResearchCommunications 287 (1), 98–103.

El-Sayed, N., Ghedin, E., Song, J., MacLeod, A., Bringaud, F.,Larkin, C., Wanless, D., Peterson, J., Hou, L., Taylor, S., Tweedie,A., Biteau, N., Khalak, H., Lin, X., Mason, T., Hannick, L., Caler,E., Blandin, G., Bartholomeu, D., Simpson, A., Kaul, S., Zhao, H.,Pai, G., Aken, V.S., Utterback, T., Haas, B., Koo, H., Umayam,L., Suh, B., Gerrard, C., Leech, V., Qi, R., Zhou, S., Schwartz, D.,Feldblyum, T., Salzberg, S., Tait, A., Turner, C., Ullu, E., White,O., Melville, S., Adams, M., Fraser, C., Donelson, J., 2003. Thesequence and analysis of Trypanosoma brucei chromosome II.Nucleic Acids Research 31 (16), 4856–4863.

El-Sayed, N.M., Hegde, P., Quackenbush, J., Melville, S.E., Donelson,J.E., 2000. The African trypanosome genome. InternationalJournal for Parasitology 30, 329–345.

Fijalkowska, I., Jonczyk, P., Tkaczyk, M., Bialoskorska, M., Scha-aper, R., 1998. Unequal fidelity of leading strand and laggingstrand DNA replication on the Escherichia coli chromosome.Proceedings of the National Academy of Sciences of the UnitedStates of America 95 (17), 10020–10025.

Francino, M.P., Ochman, H., 1997. Strand asymmetries in DNAevolution. Trends in Genetics 13 (6), 240–245.

Frank, A., Lobry, J., 1999. Asymmetrical substitution patterns: areview of possible underlying mutational or selective mechanisms.Gene 238, 65–77.

Freeman, J.M., Plasterer, T.N., 1998. Patterns of genome organizationin bacteria. Science 279, 1827.

Ghedin, E., Bringaud, F., Peterson, J., Myler, P., Berriman, M.,Ivens, A., Andersson, B., Bontempi, E., Eisen, J., Angiuoli, S.,Wanless, D., Arx, A.V., Murphy, L., Lennard, N., Salzberg, S.,Adams, M., White, O., Hall, N., Stuart, K., Fraser, C., El-Sayed,N., 2004. Gene synteny and evolution of genome architecture intrypanosomatids. Molecular and Biochemical Parasitology 134(2), 183–191.

Green, P., Ewing, B., Miller, W., Thomas, P., NISC ComparativeSequencing Program, Green, E., 2003. Transcription-associatedmutational asymmetry in mammalian evolution. Nature Genetics33 (4), 514–517.

Grigoriev, A., 1998. Analyzing genomes with cumulative skewdiagrams. Nucleic Acids Research 26 (10), 2286–2290.

Hall, N., Berriman, M., Lennard, N., Harris, B., Hertz-Fowler, C.,Bart-Delabesse, E., Gerrard, C., Atkin, R., Barron, A., Bowman,

S., Bray-Allen, S., Bringaud, F., Clark, L., Corton, C., Cronin, A.,Davies, R., Doggett, J., Fraser, A., Gruter, E., Hall, S., Harper, A.,Kay, M., Leech, V., Mayes, R., Price, C., Quail, M., Rabbin-owitsch, E., Reitter, C., Rutherford, K., Sasse, J., Sharp, S.,Shownkeen, R., MacLeod, A., Taylor, S., Tweedie, A., Turner, C.,Tait, A., Gull, K., Barrell, B., Melville, S.E., 2003. The DNAsequence of chromosome I of an African trypanosome: genecontent, chromosome organisation, recombination and polymor-phism. Nucleic Acids Research 31 (16), 4864–4873.

Hotz, H.-R., Hartmann, C., Huober, K., Hug, M., Clayton, C., 1997.Mechanisms of developmental regulation in Trypanosoma brucei: apolypyrimidine tract in the 3 0-untranslated region of a surfaceprotein mRNA affects RNA abundance and translation. NucleicAcids Research 25 (15), 3017–3025.

Ihaka, R., Gentleman, R., 1996. R: A language for data analysis andgraphics. Journal of Computational and Graphical Statistics 5 (3),299–314.

Izuta, S., Roberts, J.D., Kunkel, T.A., 1995. Replication error rates forG dGTP, T dGTP, and A dGTP mispairs and evidence fordifferential proofreading by leading and lagging strand DNAreplication complexes in human cells. The Journal of BiologicalChemistry 270 (6), 2595–2600.

Lobry, J., 1995. Properties of a general model of DNA evolution underno-strand-bias conditions. Journal of Molecular Evolution 40,326–330.

Lobry, J., 1996. Asymmetric substitution patterns in the two DNAstrands of bacteria. Molecular Biology and Evolution 13 (5), 660–665.

Maliszewska-Tkaczyk, M., Jonczyk, P., Bialoskorska, M., Schaaper,R., Fijalkowska, I., 2000. Sos mutator activity: unequal mutagen-esis on leading and lagging strands. Proceedings of the NationalAcademy of Sciences of the United States of America 97 (23),12678–12683.

Martinez-Calvillo, S., Yan, S., Nguyen, D., Fox, M., Stuart, K.,Myler, P.J., 2003. Transcription of Leishmania major Friedlinchromosome 1 initiates in both directions within a single region.Molecular Cell 11, 1291–1299.

McDonagh, P.D., Myler, P.J., Stuart, K., 2000. The unusual geneorganization of Leishmania major chromosome 1 may reflect noveltranscription processes. Nucleic Acids Research 28 (14), 2800–2803.

McLean, M.J., Wolfe, K.H., Devine, K.M., 1998. Base compositionskews, replication orientation, and gene orientation in 12prokaryote genomes. Journal of Molecular Evolution 47, 691–696.

Mellon, I., Hanawalt, P.C., 1989. Induction of the Escherichia coli

lactose operon selectively increases repair of its transcribed DNAstrand. Nature 342, 95–98.

Myler, P.J., Audleman, L., deVos, T., Hixson, G., Kiser, P., Lemley,C., Magness, C., Rickel, E., Sisk, E., Sunkin, S., Swartzell, S.,Westlake, T., Bastien, P., Fu, G., Ivens, A., Stuart, K., 1999.Leishmania major Friedlin chromosome 1 has an unusual distri-bution of protein-coding genes. Proceedings of the NationalAcademy of Sciences of the United States of America 96 (6),2902–2906.

Nakamura, Y., Gojobori, T., Ikemura, T., 1999. Codon usagetabulated from the international DNA sequence databases; itsstatus 1999. Nucleic Acids Research 27 (1), 292.

Niu, D., Lin, K., Zhang, D., 2003. Strand compositional asymmetriesof nuclear DNA in eukaryotes. Journal of Molecular Evolution 57(3), 325–334.

R Development Core Team, 2004. R: A language and environment forstatistical computing. R Foundation for Statistical Computing,Vienna, Austria, 3-900051-07-0. Available from: <http://www.R-project.org>.

Reyes, A., Gissi, C., Pesole, G., Saccone, C., 1998. Asym-metrical directional mutation pressure in the mitochondrial

D. Nilsson, B. Andersson / Experimental Parasitology 109 (2005) 143–149 149

genome of mammals. Molecular Biology and Evolution 15(8), 957–966.

Rice, P., Longden, I., Bleasby, A., 2000. EMBOSS: The EuropeanMolecular Biology Open Software Suite. Trends in Genetics 16 (6),276–277.

Rocha, E., Danchin, A., 2001. Ongoing evolution of strand compo-sition in bacterial genomes. Molecular Biology and Evolution 18(9), 1789–1799.

Shioiri, C., Takahata, N., 2001. Skew of mononucleotide frequencies,relative abundance of dinucleotides, and DNA strand asymmetry.Journal of Molecular Evolution 53, 364–376.

Sueoka, N., 1995. Intrastrand parity rules of DNA base compositionand usage biases of synonymous codons. Journal of MolecularEvolution 40, 318–325.

Wall, L., Schwartz, R., 1991. Programming Perl. O�Reilly andAssociates, Sebastopol, CA.

Wong, E., de Lafaille, M.C., Wirth, D., 1994. Identification of a cis-acting gene regulatory element from the lemdr1 locus of Leishmania

enriettii. Journal of Biological Chemistry 269 (42), 26497–26502.Worthey, E., Martinez-Calvillo, S., Schnaufer, A., Aggarwal, G.,

Cawthra, J., Fazelinia, G., Fong, C., Fu, G., Hassebrock, M.,Hixson, G., Ivens, A., Kiser, P., Marsolini, F., Rickell, E., Salavati,R., Sisk, E., Sunkin, S., Stuart, K., Myler, P., 2003. Leishmania

major chromosome 3 contains two long convergent polycistronicgene clusters separated by a tRNA gene. Nucleic Acids Research 31(14), 4201–4210.

Wu, C.-I., Maeda, N., 1987. Inequality in mutation rates of the twostrands of DNA. Nature 327, 169–170.