BioSystems 125 (2014) 22–31

Contents lists available at ScienceDirect

BioSystems journal homepage: www.elsevier.com/locate/biosystems

Mitochondrial swinger replication: DNA replication systematically exchanging nucleotides and short 16S ribosomal DNA swinger inserts Hervé Seligmann * Unité de Recherche sur les Maladies Infectieuses et Tropicales Émergentes, Faculté de Médecine, URMITE CNRS-IRD 198 UMER 6236, Université de la Méditerranée, Marseille, France

A R T I C L E I N F O

A B S T R A C T

Article history: Received 28 July 2014 Received in revised form 8 September 2014 Accepted 29 September 2014 Available online 3 October 2014

Assuming systematic exchanges between nucleotides (swinger RNAs) resolves genomic ‘parenthood’ of some orphan mitochondrial transcripts. Twenty-three different systematic nucleotide exchanges (bijective transformations) exist. Similarities between transcription and replication suggest occurrence of swinger DNA. GenBank searches for swinger DNA matching the 23 swinger versions of human and mouse mitogenomes detect only vertebrate mitochondrial swinger DNA for swinger type AT + CG (from five different studies, 149 sequences) matching three human and mouse mitochondrial genes: 12S and 16S ribosomal RNAs, and cytochrome oxidase subunit I. Exchange AT + CG conserves selfhybridization properties, putatively explaining swinger biases for rDNA, against protein coding genes. Twenty percent of the regular human mitochondrial 16S rDNA consists of short swinger repeats (from 13 exchanges). Swinger repeats could originate from recombinations between regular and swinger DNA: duplicated mitochondrial genes of the parthenogenetic gecko Heteronotia binoei include fewer short AT + CG swinger repeats than non-duplicated mitochondrial genomes of that species. Presumably, rare recombinations between female and male mitochondrial genes (and in parthenogenetic situations between duplicated genes), favors reverse-mutations of swinger repeat insertions, probably because most inserts affect negatively ribosomal function. Results show that swinger DNA exists, and indicate that swinger polymerization contributes to the genesis of genetic material and polymorphism. ã 2014 Elsevier Ireland Ltd. All rights reserved.

Keywords: Mitochondrial replication Swinger DNA polymerization Invertase 30 -to-50 Polymerization Swinger repeat Mitochondrial recombination Gene duplication Asexual vertebrate reproduction

1. Introduction Modern sequencing techniques of total RNA cell contents reveal that some RNA sequences do not match genomic DNA sequences. Some orphan transcripts are chimeric fusions of two or more regular RNAs (Yang et al., 2013). Other mechanisms such as transcription that systematically exchanges between nucleotides (Seligmann, 2013a,b) and 30 -to-50 ‘inverted’ RNA polymerization (Seligmann, 2012a,c,) elucidate the genomic ‘parenthood’ of some other, apparently orphan RNA sequences, called here swinger RNAs. Swinger RNAs probably existed at the earliest stages of evolution of the biomolecular machinery, when left-handed amino acid enantiomers were selected. This is because swinger frequencies and independently estimated swinger polymerization rates are proportional to preferences for tRNA anticodon/acceptor stem aminoacylation by left- rather than right-handed amino acids

(Michel and Seligmann, 2014). Swinger RNAs increase manifold the coding potential of genes, in addition to other mechanisms such as translation of stops (Faure et al., 2011; Seligmann, 2010a, 2011a, 2012b,c,d, 2013d) and coding by tetracodons (Seligmann, 2012e, 2013e, 2014a; Seligmann and Labra, 2013). These considerations, as well as similarities between RNA and DNA polymerizations (Little et al., 1993; Lee and Clayton, 1997; Prado and Aguilera, 2005; Seligmann, 2011b), suggest that swinger DNA polymerization might also exist. Analyses of mitochondrial DNA sequences presented here detect mitochondrial swinger DNA (mainly swinger rDNA), confirming previous results for mitochondrial swinger RNA (Seligmann 2012a; 2013a,c) and especially nuclear swinger rDNA (Seligmann, 2014b). These consist of explorations of GenBank’s nucleotide database for DNA sequences matching mitochondrial DNA sequences assuming systematic nucleotide exchanges. 1.1. Searching for swinger DNA

* Tel.: + 33 (0)4 91385517. E-mail addresses: [email protected], [email protected] (H. Seligmann). http://dx.doi.org/10.1016/j.biosystems.2014.09.012 0303-2647/ ã 2014 Elsevier Ireland Ltd. All rights reserved.

In total 23 types of nucleotide exchanges exist, nine of which are symmetric exchanges (XY, i.e., AC, Seligmann, 2013a) and 14 asymmetric exchanges (X->Y->Z->X, i.e., A->C->G->A,

H. Seligmann / BioSystems 125 (2014) 22–31

Seligmann, 2013b). The wording ‘nucleotide exchange’ and annotationsusing ‘->’ to indicate exchanges between nucleotides reflect the chemical nature of the process, which is adequate in the context of searches for actual DNA sequences produced by swinger replication. However, note as commented by two anonymous reviewers that ‘systematic nucleotide exchanges’ are called bijective transformations in mathematical terms, symmetric and asymmetric exchanges are ‘bijective transformations of order two’ and ‘bijective transformations of higher order (three or four). These terms are used in studies focusing on symmetry properties of the natural coding system (i.e., Gonzalez et al., 2011; Fimmel et al., 2013). The exchange AT + CG produces a sequence that is the inverse complement of the original sequence, and which conserves selfhybridization properties of the original, regular sequence: as all symmetric exchanges, it is self-inverse by definition. Considering that mitochondrial swinger RNA has been previously described, the relative conservation of mitochondrial genomes, and high sampling densities of vertebrate mitochondria, I produced the 23 swinger versions of the human and mouse mitochondrial genomes (NC_012920, NC_05089) and blasted each version, searching matches in GenBank’s non-redundant nucleotide collections. Swinger RNA exists in the human mitochondrion, but even if GenBank includes no human swinger DNA, BLAST (Zhang et al., 2000) will probably detect some of the homologous swinger mitochondrial DNA from some other metazoans, if such mitochondrial swinger DNA exists. 2. Results 2.1. GenBank swinger DNA BLAST analyses (using the software’s interface most stringent alignment criterion ‘megablast’) were done for each of the 23 swinger versions of the mitogenomes of each Homo sapiens (NC_012920) and Mus musculus (NC_005089). These alignment criteria detect only GenBank DNA sequences matching swinger DNA following the bijective transformation AT + CG, and no other swinger type. All detected GenBank swinger DNA originates from vertebrate mitogenomes, but BLAST analyzed all nucleotide sequences in the database, most of which are not mitochondrial. Table 1 presents all 60 sequences detected using each human and mouse swinger mitogenomes. Three and eleven additional sequences are detected only according to the AT + CG bijective transformation of the human, respectively mouse mitogenome. In total, BLAST analyses of these two swinger mitogenomes yield 74 sequences, from 12S rDNA, 16S rDNA, and cytochrome oxidase subunit I (the latter is detected using the human, but not the mouse swinger mitogenome and consists of a single sequence). These results converge with results on eukaryotic rDNA from the nucleus, for which only swinger rDNA from bijective transformation AT + CG was detected (Seligmann, 2014b). 2.2. Statistical considerations regarding alignment results One may question whether the putative swinger sequences detected by BLAST are genuine swinger sequences, or result from the fact that a very large database was explored, hence that alignments are due to chance. This can be addressed by considering the very identities of the detected sequences: all are mitochondrial DNA sequences, are from five independent studies, and these were detected by searching GenBank for sequences matching swingertransformed mitochondrial genomes.

23

2.2.1. All swinger DNAs are mitochondrial sequences GenBank includes 102 million sequences (after excluding 20 million mRNAs that are irrelevant to a search for swinger DNA), among which 3 million (less than 3%) have in their annotation the word ‘mitochondrial’. This means that obtaining by chance hits with mitochondrial sequences (while searching mitochondrial DNA) has P < 0.03 for each of these hits. Fisher’s method for combining P values that sums -2xln(Pi) over all k (here five) independent tests (i ranges from 1 to k) yields a chi-square statistic with 2xk degrees of freedom, which yields a combined P = 0.00012. Taking into account that the input to BLAST was restricted to mitochondrial swinger DNA of vertebrate origins would decrease even more the P value, as vertebrate sequences with ‘mitochondrial’ in their annotation amount to 1.3 million in GenBank (combined P = 0.0000042). Hence, obtaining only alignments with mitochondrial sequences can not be the result of chance due to the size of the explored database. 2.2.2. All swinger DNAs match the AT + CG bijective transformation The same point accounts for the fact that detected swinger DNA sequences (from five independent projects) are all aligning with the same bijective transformation of the mitochondrial genomes. As there are 23 transformations, this has P = 1/23 = 0.0425. Considering this is the case for swinger sequences from all five independent projects, Fisher’s method for combining P values yields a combined P = 0.0005. Hence this other property of the results, that all alignments are with the same transformation of mitochondrial genomes, is highly unlikely to occur randomly. Taking into account both properties of detected swinger DNAs combined (all match a single swinger transformation and have mitochondrial origin) would yield much lower P values. Remember that above P value calculations take into account the size of GenBank’s database. 2.2.3. Statistics for single alignments Above sections examine the statistical significance of the results, considering these as a group. Here the significance of single alignments is considered. The swinger DNA sequences form five groups, according to the study that produced them, according to taxon and gene: three amphibian 12S rDNA sequences (species name followed by alignment length and lowest e value when several sequences are in that group: Bufo bankorensis, 470, 3  1028; Rana sauteri, 460, 3  1024; Hyla chinensis, 445, 7  1039), three groups of swinger 16S rDNA sequences (Mesocricetus brandti, 481, 2  1094; various Mus species, 110, 3  1047; and various Anguilla species, 127, 1 1036; and a fifth group matching the mitochondrial gene of cytochrome oxidase subunit I from another amphibian, the salamander Desmognathus carolinensis (476, 4  1056). The e value is a statistic indicating the number of random matches expected for an alignment of that length and similarity, and for a search of a database of that size. The lower the e value, the more significant is the match. Typically, an e value of 1 104 is considered a good match. Even if one considers that for each human and mouse mitogenomes, 23 swinger versions were blasted, and hence the obtained e values are multiplied by 46, e values of all swinger DNA sequences detected are several orders of magnitude below (hence unlikely to be obtained by chance) than the common e value threshold. Hence, considering all alignments together, as a group, or each alignment alone, alignments detected by BLAST are highly unlikely to be due to chance. Table 1 presents a single e value for each alignment, even when the sequence was detected by blasting different input seed sequences (i.e., human and mouse transformed mitogenomes). In these cases, only the e value obtained with the human

24

H. Seligmann / BioSystems 125 (2014) 22–31

Table 1 A-T + CG swinger DNA sequences detected by megablast in GenBank (July 2014) using swinger versions of complete mitogenomes of Homo sapiens and Mus musculus. * next to species names (column 3) indicate sequences from other species used to detect other swinger sequences, indicated by the same asterisk number in column 1. H, M, HM indicate sequences detected using swinger human, and/or mouse mitogenomes. eV: e value, number of alignments of that size and similarity with the input sequence, randomly expected when searching a sample of GenBank’s size. Indicated e values are with the transformed human sequence if the alignment is obtained with both human and mouse sequences. 993 is short for 9  1093. S

Gene

Species

Start-end

eV

GenBank

Ref

HM HM HM * M M M M M M M M M M HM HM HM HM HM HM HM HM HM HM HM HM HM H M HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM H HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM HM

12 S rRNA

Bufo bankorensis Rana sauteri Hyla chinensis* Rhacophorus moltrechti Mus musculus musculus

409–724 347–662 380–831 380–831 110–1 110–1 110–1 109–1 110–1 109–1 110–1 110–3 111–3 110–3 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 1–473 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1 127–1

328 824 839 1139 347 347 347 343 739 635 230 234 229 132 993 993 294 993 294 289 289 491 289 988 388 186 289 491 630 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 136 635 635 635 635 136

JN117718 JN117722 JN117721 JN117720 M55046 M55044 M55041 M55023 M55051 M55062 M55063 M55052 M55061 M55057 AM904673 AM904669 AM904676 AM904674 AM904677 AM904671 AM904667 AM904666 AM904665 AM904664 AM904675 AM904670 AM904672 AM904668 AM904681 AB188459 AB188458 AB188457 AB188456 AB188454 AB188453 AB188452 AB188451 AB188450 AB188449 AB188448 AB188479 AB188478 AB188477 AB188476 AB188475 AB188474 AB188473 AB188472 AB188471 AB188470 AB188469 AB188467 AB188466 AB188465 AB188464 AB188463 AB188462 AB188461 AB188488 AB188486 AB188469 AB188485 AB188484 AB188483 AB188482 AB188481 AB188480 AB188460 AB188455 AB188447 AB188468 AB188472

$ “ “ “ Fort et al., 1984 “ “ “ “ “ “ “ “ “ # “ “ “ “ “ “ “ “ “ “ “ “ “ “ & “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “

16 S rRNA

Mus musculus domesticus Mus musculus castaneus Mus abbotti Mus spretus

Mesocricetus brandti

Anguilla bicolor

“ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “

H. Seligmann / BioSystems 125 (2014) 22–31

25

Table 1 (Continued) S HM HM H ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** **

Gene

COX I

Species

Start-end

eV

GenBank

Ref

Anguilla interioris Anguilla marmorata Desmognathus carolinensis** Desmognathus fuscus

127–1 127–1 474–1

136 136 456 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2180 2180 6180 6180 2179 1177 2175 4172 1171 6165 1161 8164 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2180 6180 2179 5176 1172 1171 4167 6165 0 0 2175 8169 8164 1132 1171

AB188487 AB188446 AF437469 AF437484 AF437483 AF437479 AF437478 AF437477 AF437476 AF437481 AF437482 AF437480 AF437488 AF437486 AF437494 AF437487 AF437499 AF437485 AF437490 AF437493 AF437501 AF437474 AF437500 AF437489 AF437491 AF437475 AF437496 AF437492 AF437495 AF437498 AF437497 AF437502 AF437503 AF437504 AF437454 AF437453 AF437452 AF437451 AF437437 AF437435 AF437436 AF437455 AF437449 AF437440 AF437448 AF437465 AF437450 AF437438 AF437439 AF437456 AF437457 AF437461 AF437445 AF437444 AF437443 AF437442 AF437460 AF437447 AF437434 AF437468 AF437462 AF437459 AF437458 AF437463 AF437464 AF437466 AF437414 AF437467 AF437441 AF437446 AF437472 AF437470 AF437473 AF437507 AF437471

“ “ Rissler & Taylor, 2003 “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “

Desmognathus imitator Desmognathus monticola

Desmognathus ochrophaeus

Desmognathus orestes

26

H. Seligmann / BioSystems 125 (2014) 22–31

Table 1 (Continued) S ** ** **

Gene

Species Desmognathus quadramaculatus Desmognathus wrighti

Start-end

eV

GenBank

Ref

4177 5161 2130

AF437505 AF437428 AF437413

“ “ “

$Li, K.-W., Chen, L.-H., Sheu, P.Y., Tang, Z.-J., Weng, S.-H., Yu, T.-L., Weng, C.F., 2011, unpublished. #Neuma, K, Yigit, N., Fritzsche, P., Colak, E., 2007. Evidence for a species-wide bottleneck in the golden hamster Mesocricetus auratus–contrasting population histories in two eastern Mediterranean hamster species. Unpublished. &Kuroki, M., Aoyama, J., 2007. Anguillid leptocephali. Unpublished.

mitogenome is given. None of the e values for the mouse mitogenome and not indicated in Table 1 was >1 1010. 2.2.3. Simulations for random sequences as negative control Analyses above indicate that the positive results described in Table 1 are statistically significant, as a coherent group (all results are for the same swinger transformation, and are mitochondrial sequences matching the mitochondrial test sequence used as query in BLAST), and as single alignments (low e values for each and every alignment obtained). A further method to test whether these results are due to chance is by creating a negative control. For that purpose, 102 randomized sequences of the human mitochondrial genome were created, as well as their AT + CG swinger transformation, and these 204 randomized sequences (102 randomized sequences and their 102 transformed sequences) were analyzed by BLAST, as was done for the swinger transformed natural sequence. For 97 of the 102 randomized sequences, BLAST does not detect any alignment with sequences existing in GenBank, which estimates the statistical significance of obtaining any alignment at P = 0.0294. For 97 among 102 swinger transformed randomized sequences, BLAST does not detect any alignment with sequences existing in GenBank, setting the statistical significance of obtaining any alignment at P = 0.0294. The examination of the six alignments obtained by BLAST analyses of these six different randomized sequences strengthens the interpretation of the results in Table 1 as positive results. Indeed, all alignments show 100% identity between random and GenBank sequences, five have 28, and one 29 nucleotides, with e value 0.13 and 0.037, which is much more than the threshold 1 104 considered for good matches. The aligning sequences do not originate from mitochondrial genomes. In other words, because alignments presented in Table 1 are much longer than the few ones obtained for randomized sequences, and because their e values are much lower, BLAST analyses of randomized sequences do not produce any result comparable with those obtained for the transformation of natural sequences. This means that chance has very low probabilities of producing results such as those presented in Table 1. 2.3. Swinger DNA and regular DNA One swinger DNA sequence representative of each of the taxongene combinations in Table 1 was ‘reverse-exchanged’ so as to reflect the ‘regular’ sequence from which it was putatively replicated by AT + CG swinger DNA polymerization (excluding the mouse 16S rDNA sequences, redundant with the initial mouse mitogenome search). In this case as for other symmetric exchanges (but not for asymmetric ones), reverse exchanging the sequences consists in applying to these sequences the same AT + CG exchange as applied to the original human and mouse sequences used for primary detection in BLAST (all bijective transformations of order two are self-inverse by definition, Fimmel et al., 2014). These reverse-exchanged sequences were then blasted in GenBank: reverse-swinger 12S rRNA of Hyla chinensis is 98% identical to regular Hyla chinensis 12S rRNA;

reverse-swinger 16S rRNA of Mesocricetus brandti is 95% identical to regular Mesocricetus auratus 16S rRNA (Genbank does not include any regular 16S rDNA from Mesocricetus brandti); reverseswinger 16S rRNA of Anguilla bicolor has 100% identity with regular 16S rRNA from the same species; and reverse-swinger COX1 of Desmognathus carolinensis is 99% identical to regular Desmognathus carolinensis COX1. 2.3.1. Evolutionary distances and e values Note in this context that e values indicate the quality of the alignment. This means that they depend on the similarity between the query and the detected sequence, the latter depending on phyletic distance. For example, the best alignment between the swinger human sequence and the hamster Mesocricetus brandti sequence has the e value 9  1093, which is very low, but still much higher than the e value between the swinger transformed mouse mitogenome and that same hamster sequence, which is 4  10161 (Table 1 shows only the e value in relation to the human mitogenome for sequences detected by blasting both human and mouse transformed mitogenomes). This difference reflects the trivial fact that mouse sequences are evolutionarily more similar to hamster sequences (which are also rodents), than human sequences. This stresses that results are to some extent dependent on the input sequence used. 2.4. More undetected swinger DNA in GenBank by a second BLAST analysis Differences between swinger sequences detected using the relatively closely related human and mouse mitogenomes indicate that GenBank probably includes other swinger sequences that would be detected by using as ‘search swinger sequences’ mitogenomes from other species, more closely related to the unknown, yet undetected swinger sequences. This point is shown by BLAST analyses using swinger sequences detected and presented in Table 1. BLAST analyses of the swinger 12S rDNA sequence of Hyla chinensis detects a fourth amphibian swinger 12S rDNA sequence from the same study, Rhacophorus moltrechti, indicated by an asterisk (*) in the first column of Table 1. Similar analyses of swinger sequences detected by the first round of BLAST analyses detect no further swinger sequences for rodents (mouse and hamster) and Anguilla. For the salamander Desmognathus, this procedure detects 75 additional homologous swinger sequences from seven other Desmognathus species, indicated by two asterisks (**) in the first column of Table 1. Table 1 includes 149 swinger sequences from four regions in three genes, four amphibian species for 12S rDNA, six murine taxa and three Anguilla species for 16S rDNA. Results indicate that swinger DNA is biased towards rRNA genes, and taxonomically towards amphibians. Using as search seed swinger DNA from species not belonging to primates and rodents would probably detect numerous additional swinger DNA sequences, as indicated by the detection of 75 additional salamander sequences in the phase two search when using as search seed the original salamander sequence detected in the phase 1 search.

H. Seligmann / BioSystems 125 (2014) 22–31

It is important to stress here that BLAST analyses search in each case the complete GenBank nucleotide database, for alignments with the input sequence. In a first phase, I used as input swinger transformations of two sequences, the human and mouse mitogenomes. In a second phase, reported in this very section, the input sequences analyzed by BLAST were a total of 4 swinger sequences detected in GenBank in this first phase and presented in Table 1. For that second phase, one sequence representative of each of the five sequence groups in Table 1 was used, only two (from Hyla and Desmognathus), as reported above, detect additional swinger sequences. 2.5. Short swinger repeats in regular ribosomal RNA The existence of swinger DNA suggests that rare recombinations might occur between regular and swinger sequences. These could leave traces in the regular sequence in the form of short swinger repeats, which would be detected by BLAST analyses between paired sequences, one the regular, and the other, one among 23 swinger versions of the regular sequence. Table 2 presents the short swinger repeats (lengths between 11 and 19 nucleotides) detected by these analyses, between the regular human 16S rDNA sequence and 13 of its 23 swinger versions (no results for the 10 remaining swinger versions). These alignments (example in Fig. 1) are between a specific sequence at location 1 and a sequence of its swinger version at location 2 in that gene, suggesting that the regular sequence at location 2 was produced by integration of the swinger version of the sequence at location 2. Three alignments in Table 2 are detected twice (indicated by asterisks *): an alignment between the regular sequence at location 2 is also detected with the swinger sequence at location 1. This symmetry indicates that the sequence at location 1 could be

27

Fig. 1. Alignments between regular human mitochondrial 16S rDNA and it's AC + GT swinger version. Sequence 3a is the putative AC + GT swinger repeat of regular sequence 1a, detected by aligning swinger sequence 2a with regular sequence 1a. The regular sequence at the location of sequence 2a is actually sequence 3a, which is transformed into sequence 2a by the swinger exchange AC + GT. This transformation reveals the possibility that sequence 2a is a swinger repeat of sequence 1a. Sequence 3a is identical to sequence 1b, whose presumed swinger AC + GT repeat is sequence 3b, which is identical to sequence 1a. Hence scenario A where 3a/1b is the swinger repeat of 1a, and scenario B where 3b/1a is the swinger repeat of 1b are not distinguishable. BLAST detects alignments corresponding to both scenarios.

the swinger version of location 2, or vice versa. At this point, analyses cannot distinguish between these two possibilities. This symmetry between different bijective transformations is described by Fimmel et al. (2014). Taking all these symmetries into account, Table 1 describes 22 different swinger repeats, which cover 22.9% of the total length of the regular 16S rDNA (see Fig. 2 that plots numbers of presumed swinger repeats detected for a nucleotide as a function of the nucleotide position in the human mitochondrial 16S rDNA gene). Note that three short swinger repeats in Table 2 are palindromic (P in last column, Table 2). Palindromes are letter strings that are identical when read from left to right as from right to left. In

Table 2 Swinger repeats detected by BLAST (‘somewhat similar sequences, blastn) in human mitochondrial 16S rDNA. Column 1: swinger type. Columns 2 and 4: start and end positions (in human 16S rDNA) of alignment, for regular and swinger sequences, respectively. Column 3: regular sequence in alignment. Column 5: numbers of nucleotides identical in the alignment/alignment length. Column 6: symmetries in alignment detections: * indicates that an alignment exists also between the regular sequence corresponding to the location of the swinger sequence indicated in column 4 and the swinger sequence corresponding to the regular sequence indicated in columns 2–3. A similar symmetry occurs also between alignments detected for different swinger types, when this occurs, they are indicated in the last column. “P” indicates that the repeat is palindromic. “I” indicates that the swinger repeat location is exactly at the inverted location of the repeated regular sequence, suggesting some kind of inverted palindrome after swinger transformation of the regular sequence. Swinger

Regular start-end

AC AG AT

965–975 742–752 186–197 376–386 78–88 227–241 426–436 554–565 696–707 1533–1543 14–31 757–767 169–186 774–784 119–137 211–221 519–529 1038–1048 1162–1172 1315–1332 139–151 644–654 188–200 286–296 628–638 936–946 470–481 1154–1164

CT AC+GT AT+CG

A->C->G->A A->C->T->A A->G->C->A

A->G->T->A C->G->T->C C->T->G->C

A->C->G->T->A

GGCTCCACGAG ACCCTCACTGT ATGAATTAACTA CTTTAAATTTG GCGATAGAAAT AAGACCCCCGAAACC TAGTCCAAAGA CCTAAAAAATCC GACAATTAACAG ACCCACACCCA CCAAACCCACT-CCACCTT CCAACACAGGC CCTATACCTTCTGCATAA ATAAGGAAAGG AAGGGAAAGATGAAAAATT AAGGAGAGCCA CCAATTAAGAA CACAGCAAGAC AACCCAACCTC CCTCGATGTTGGATCAGG TAACCAAGCATAA CATGAAAACAT GAATTAACTAGAA TAGCAAAATAG ATGTTAGTATA TTGTTCCTTAA GTAGAGAGAGTA CGGAGCAGAAC

Swinger start-end

Id

Symmetry

1263–1273 88–78 197–186 386–376 752–742 554–568 983–973 565–554 707–696 1543–1533 119–137 211–221 1332–1315 1172–1162 14–31 757–767 946–936 638–628 784–774 186–169 188–200 286–296 139–151 644–654 1048–1038 529–519 481–470 1164–1154

11/11 11/11 12/12 11/11 11/11 14/15 11/11 12/12 12/12 11/11 17/19 11/11 16/18 11/11 17/19 11/11 11/11 11/11 11/11 16/18 13/13 11/11 13/13 11/11 11/11 11/11 12/12 11/11

* CT I I AG * * P P P A->G->C->A A->G->C->A A->G->T->A A->G->T->A A->C->G->A A->C->G->A C->T->G->C C->T->G->C A->C->T->A A->C->T->A C->T->G->C C->T->G->C C->G->T->C C->G->T->C A->G->C->A A->G->C->A I I

28

H. Seligmann / BioSystems 125 (2014) 22–31

Fig. 2. Number of swinger repeats detected by BLAST for the human mitochondrial 16S rDNA gene cumulated over all 23 types of swinger versions of that gene, as a function of the nucleotide position for that gene. Lines indicate the running average of this number over a running window of 10 nucleotides. About 23% of the sequence length is covered by at least one swinger repeat.

addition, four alignments are between the regular sequence and the inverse complement of its swinger sequence, at exactly the same location (I in last column, Table 2), apparently another form of palindrome-like phenomenon associated with swinger repeats. The frequent palindromic nature of swinger repeats stresses their probable origin from recombination (or recombination-like) events (i.e., Akgün et al., 1997; Cromie et al., 2000; Zhou et al., 2001; Hallast et al., 2013). In this case recombination would have occurred between regular DNA and swinger DNA, or between regular DNA and swinger RNA (considering that the vertebrate DNA polymerase gamma has a reverse transcriptase activity that discriminates ribonucleotides and corrects such misinsertions to deoxyribonucleotides (Kasiviswanathan and Copeland 2011). 2.6. Recombination and mitochondrial gene duplication Results on putative natural swinger DNA versions of mitochondrial genes can be examined from the point of mitochondrial recombination, considered for a long time as controversial, but now as quite assessed through direct evidence (Ladoukakis and Zouros 2001a; Kraytsberg et al., 2004; Ciborowski et al., 2007) and broad meta-analyses (Ladoukakis and Zouros 2001b; Piganeau et al., 2004). The ubiquitous swinger repeats (Table 2,Figs. 1–2) in regular, functional 16S rDNA indicate that insertions of short swinger sequences are relatively frequent, especially if one considers that Table 2 describes only swinger repeats that were conserved and hence probably do not impair the gene’s function. I suggest that the rare regular recombinations that occur between mitochondrial genomes from female and male gametes limit numbers of swinger repeats, a correction mechanism avoiding the cumulation of swinger repeats impairing gene function(s), in which palindromes are frequently involved (Leach et al., 1997). This can be examined in the rare example of mitogenomes of the parthenogenetic gecko Heteronotia binoei, where the mitogenome of several independent parthenogenetic lineages includes a partial duplication of the mitogenome covering

the control region, the 12S and 16S rRNA genes, cytochrome B, NADH1–2 and NADH5–6 protein coding genes (Fujita et al., 2007). Genbank includes for regular non-parthenogenetic Heteronotia species mitochondrial sequences of the complete NADH2 gene (H. fasciolatus (Genbank KF289034, Pepper et al., 2013); H. planiceps (Genbank HQ840064, Pepper et al., 2011); and H. spelea (GenBank: HG132830, Boussau et al., 2011). Blasting the AT + CG swinger version of these sequences with their regular sequence yields 3, 1 and 3 swinger repeats, respectively, of total lengths 38, 14 and 35 nucleotides (means 2.33 and 29, respectively). These swinger repeats resemble those described for 16S rDNA (Table 2, Fig. 1). These swinger repeats are as predicted fewer, and shorter than obtained for similar analyses of that gene for parthenogenetic H. binoei with no mitochondrial gene duplication (for NADH2 from the five complete (non-duplicated) mitochondrial genomes (EF62807–11), mean number and total length of swinger repeats are 5.6 and 67). For the same gene from H. binoei mitogenomes with two copies of that gene (EF626813–17, Fujita et al., 2007), the mean number of swinger repeats is 1.5, with mean total length 17.5 nucleotides. Hence these partial results indicate that recombination limits the insertion of swinger repeats. The specific results suggest that intragenomic recombination between duplicated versions of the gene (in the same genome) is more frequent than inter-genomic (sexual) recombination. The latter point (intravs. intergenomic recombination) cannot be confirmed at this point, because sequences for additional mitochondrial genes for nonparthenogenetic Heteronotia are still lacking. The issue that the presence of intragenomic gene duplicates associates with a low number of swinger repeats can be tested separately for each gene and each independent lineage where mitogenome duplication occurred in parthenogenetic H. binoei (Fujita et al., 2007), by comparing, for each gene and each lineage, the number and total length of swinger repeats in the two gene copies, with those from regular mitogenomes (no gene duplications) from that parthenogenetic lineage. This yields 27 comparisons, where 10 and seven yield identical numbers and lengths,

H. Seligmann / BioSystems 125 (2014) 22–31

respectively, for regular and duplicated mitogenomes. Duplicated genes from duplicated mitogenomes have fewer swinger repeats in 12 among the 17 remaining cases (one tailed sign test P = 0.036), with shorter total lengths in 15 among 22 comparisons (one tailed sign test P = 0.033). These analyses require confirmation from additional independent data, but they suggest that recombinations are involved in the insertion of swinger repeats, and especially in their limitation. More importantly, they confirm the existence of swinger repeats in regular genes. 3. Discussion The main results presented in Sections 2.1.–2.4. consist of high similarity alignments between mitochondrial swinger sequences produced in silico and actual swinger DNA of type AT + CG in GenBank. Results on swinger DNA confirm previous results on swinger RNA in vertebrate mitochondria, and on swinger DNA from eukaryote nuclei (Seligmann, 2014b). Their interpretation is nevertheless complex, with potentially far reaching conclusions that differ from those previously discussed for swinger RNA. The second body of results (Sections 2.5.–2.6.) indicates that regular mitochondrial genes include numerous short swinger repeats. These might originate from recombination between regular and swinger sequences. Some very limited data comparing swinger repeats in mitochondrial genomes from geckos with regular vs. parthenogenetic reproduction suggest that sexual recombination of mitogenomes modulates the conservation of swinger repeats. Parthenogenetic gekkos with duplicated mitochondrial gene have fewer and shorter swinger repeats than parthenogenetic geckos possessing single mitochondrial genomic gene copies. This indicates that gene duplications increase frequencies of intragenomic recombinations, enabling to replace swinger repeats by sequences from the other gene copy that does not include the swinger repeat. It is probable that most swinger repeats affect negatively gene function, and hence these ‘reverse’ mutations towards sequences not including swinger repeats are probably subsequently favored by natural selection. The very partial data available suggest that intragenomic recombinations are more frequent than sexual (inter-genomic) recombinations, as the former seem to decrease more swinger repeats than the latter. The latter points are based on analyses of very few sequence data and should await further confirmation. 3.1. Why AT + CG? Nucleotide exchange AT + CG is the only type of exchange for which mitochondrial swinger DNA is detected. This exchange was among the more frequent exchanges detected for swinger RNAs, but unlike for swinger DNA, swinger RNAs were detected also for several other types of exchanges (five other symmetric and four asymmetric nucleotide exchanges, Seligmann, 2013a,b). Exchange AT + CG is peculiar first because it is one of the three exchanges (the other two are AT, and CG) that conserve, after systematic exchange, the sequence’s secondary structure formation potential by self-hybridization. Hence, for example, swinger sequences of rRNA genes probably maintain the capacity to form (approximately) regular ribosomal secondary structure. The conservation of secondary structure formation is crucial for rRNA, but much less for mRNA (but see Krishnan et al., 2004a,b, 2008). Considering conservation of self-hybridization as a criterion for viability of swinger rDNA, at this point there is no explanation why no swinger AT DNA, nor swinger CG DNA was detected. This is notable because it implies that the swinger genes could retain some level of ribosomal functionality: if the swinger DNA

29

was not functional, there would be no need for conserving secondary structure and other types of swinger DNA would be detected. Hence mitochondria with AT + CG versions of rRNA genes might be viable. The issue of mitochondrial viability is also suggested by the number of swinger DNA sequences detected: in most cases the swinger sequence is detected from different isolates. Hence these were not sequenced by chance among regular sequences of the same gene: for example, all fifteen 16S rRNAs from Mesocricetus brandti available in GenBank are swinger sequences, so that the closest sequence found in GenBank is from another closely related species, the golden hamster Mesocricetus auratus. For the 16S rRNA of Hyla chinensis, the swinger version is one among two sequences from that gene and species. Nucleotide exchange AT + CG is also peculiar because it produces the 30 -to-50 version of the antisense strand of the replicated 50 -to-30 sequence. It is not possible at this point to know whether a sequence is produced by AT + CG swinger replication in the 50 -to-30 direction, or by 30 -to-50 replication of the inverse complement of the input sequence. 3.2. Primers and swinger DNA Primer-based sequencing functions only if the targeted sequence hybridizes with the designed primer. Therefore usually only swinger DNA containing the primer sequence designed on the premises of regular DNA will be sequenced. The description in Table 2 of swinger repeats within the regular 16S rDNA sequence, presumably originating from rare recombinations between homologous regular and swinger DNA, might be the answer to the conundrum of primer-based amplification of sequences that differ from the regular target sequence. Indeed, if a sequence at location 2 in the regular 16S rDNA is the AT + CG swinger repeat of a sequence at location 1 in the regular 16S rDNA, and location 1 is the target of the regular primer, the same primer will hybridize with the swinger sequence of location 2 and initiate its amplification. Hence swinger DNA sequencing, and the detection of swinger DNA might depend on random matches between swinger repeats and the adequate primer. This would also depend on variation between species in the targeted sequences and the presence of the corresponding swinger repeat. It is this author’s opinion that this explains that swinger DNA is detected for some taxons (i.e., amphibian, rodents), but not for many others. The design of primers for swinger DNA will help elucidate this phenomenon. 3.3. Swinger RNA versus swinger DNA Blast's analyses of the AT + CG swinger version of the human mitochondrial genome detects two mRNAs that had not been detected by previous analyses (Seligmann, 2012a, 2013a) as these were focused on GenBank's EST database. These are 251 (Y16713) and 184 (Y16712) nucleotides long and 93% and 90% identical with the input mitochondrial sequence (after AT + CG exchange). Both mRNAs are extracted from HIVassociated non-Hodgkin's lymphoma (Tarantul et al., 1998, unpublished). This means that for the AT + CG exchange alone, 13 swinger RNAs have been detected in GenBank, and 103 in total after pooling all 23 exchanges for Homo sapiens mitochondria alone. This stresses that swinger RNA is more frequent than swinger DNA, at least in mitochondria. Many transcript copies occur in an organism, and their functional halftimes are short. Hence not each and every transcript must be functional. Therefore occasional swinger RNAs of any type have little consequences. However, swinger replication is lethal unless it conserves the major functional properties of the genes that were

30

H. Seligmann / BioSystems 125 (2014) 22–31

swinger-replicated. This could explain why swinger DNA is rarer than swinger RNA, and is mainly restricted to rRNA and a nucleotide exchange conserving sequence self-hybridization. 3.4. Proteins coded by swinger sequences Previous analyses of swinger versions of mitochondrial genomes (Seligmann, 2012a, 2013a,b,c), as well as analyses based on the circular code theory (Michel and Seligmann, 2014) suggest that swinger RNA may cryptically code for undetected proteins. The detected swinger DNA for salamander COX1 strengthens the hypothesis of swinger coding based until now solely on material evidence from RNA. 3.5. Fast and ultrafast speciation The location of DNA replication origins determines the positions of mutation gradients along mitochondrial genomes (Krishnan et al., 2004a,b). Hence, any SNP in these locations potentially causes abrupt concurrent multiple genome-wide single nucleotide mutations (Krishnan and Seligmann, 2006 Seligmann et al., 2006a,b; Seligmann, 2008, 2010b, 2011b), potentially leading to fast species divergence. Swinger replication has even more abrupt and more profound effects. The occurrence of both regular and AT + CG versions of a gene in the same species (i.e., Hyla chinensis: 12S rRNA; mouse and Anguilla bicolor 16S rRNA; and Desmognathus carolinensis: COX1) suggests that swinger replication does not necessarily imply instantaneous speciation. 3.6. Evolutionary coding strategy? The possibility that swinger replication causes rapid speciations is of course of general interest, but it should not deflect from considering the more fundamental point that the swinger DNA evidence presented here suggests that some genes (at least rRNAs) may be functional in different ‘swinger’ versions. It is probable that the modular structure and evolution of tRNAs (Di Giulio, 2008, 2013) and rRNA genes (Bokov and Steinberg, 2009; Di Giulio, 2010) enables function also after transformations by swinger replication/transcription, especially if these swinger types conserve self-hybridization properties. This implies that rRNAs might have several functional swinger forms, probably increasing adaptability and evolutionary versatility. It is hence probable that gene sequences adapted so as to inherently include several functional swinger versions cryptically embedded in their main sequence, as indicated by the swinger versions of vertebrate mitochondrial 12S and 16S rRNAs described here. These results on mitochondrial swinger DNA confirm similar results for eukaryotic nuclear rDNA (Seligmann, 2014b). This would be in line with considerations on compact coding (Eigen and Schuster, 1977; Rodin and Rodin, 2007), and in general, with cost minimization principles. The fact that regular gene sequences (i.e., 16S rDNA, see Table 2) integrate numerous short swinger repeats apparently produced by recombination between homologous regular and swinger DNA indicates that swinger polymerization is at the origin of a significant part of the genesis of genetic novelty. References Akgün, E., Zahn, J., Baumes, S., Brown, G., Liang, F., Romanieinko, P.J., Lewis, S., Jasin, M., 1997. Palindrome resolution and recombination in the mammalian germ line. Mol. Cell. Biol. 14, 5559–5570. Bokov, K., Steinberg, S.V., 2009. A hierarchical model for evolution of the 23s ribosomal RNA. Nature 457, 977–980.

Boussau, B., Brown, J.M., Fujita, M.K., 2011. Nonadaptive evolution of mitochondrial genome size. Evolution 65, 2706–2711. Ciborowski, K.L., Consuegra, S., de Leaniz, C.G., Beaumont, M.A., Wang, J., Jordan, W. C., 2007. Rare and fleeting: an example of interspecific recombination in animal mitochondrial DNA. Biol. Letters 3, 554–557. Cromie, G.A., Millar, C.B., Schmidt, K.H., Leach, D.R., 2000. Palindromes as substrates for multiple pathways of recombination in Escherichia coli. Genetics 154, 514–522. Di Giulio, M., 2008. Permuted tRNA genes of Cyanidioschyzon merolae, the origin of the tRNA molecule and the root of the Eukarya domain. J. Theor. Biol. 253, 587–592. Di Giulio, M., 2010. A model of the origin of the 5s ribosomal RNA molecule. J. Mol. Evol. 71, 1–2. Di Giulio, M., 2013. A polyphyletic model for the origin of tRNAs has more support than a monophyletic model. J. Theor. Biol. 318, 124–126. Eigen, M., Schuster, P., 1977. The hypercycle. A principle of natural self-organization. Part A: emergence of the hypercycle. Naturwissenschaften 64, 541–565. Faure, E., Delaye, L., Tribolo, S., Levasseur, A., Seligmann, H., Barthélémy, R.-M., 2011. Probable presence of an ubiquitous cryptic mitochondrial gene on the antisense strand of the cytochrome oxidase I gene. Biol. Direct 6, 56. Fimmel, E., Danielli, A., Struengmann, L., 2013. On dichotomic classes and bijections of the genetic code. J. Theor. Biol. 336, 221–230. Fimmel, E., Giannerini, S., Gonzalez, D.L., Struengmann, L., 2014. Circular codes, symmetries and transformations. J. Math. Biol., in press. Fort, P., Bonhomme, F., Darlu, P., Piechaczyk, M., Jeanteur, P., Thaler, L., 1984. Clonal divergence of mitochondrial DNA versus populational evolution of nuclear genome. Evol. Theory 7, 81–90. Fujita, M.K., Boore, J.L., Moritz, C., 2007. Multiple origins and rapid evolution of duplicated mitochondrial genes in parthenogenetic geckos (Heteronotia binoei; Squamata, Geckonidae). Mol. Biol. Evol. 24, 2775–2786. Gonzalez, D.L., Giannerini, S., Rosa, R., 2011. Circular codes revisited: a statistical approach. J. Theor. Biol. 275, 21–28. Hallast, P., Balaresque, P., Bowden, G.R., Ballereau, S., Jobling, M.A., 2013. Recombination dynamics of a human Y-chromosomal palindrome: rapid GC-biased gene conversion, multi-kilobase conversion tracts, and rare inversions. PLoS Genet. 9, e1003666. Kasiviswanathan, R., Copeland, W.C., 2011. Ribonucleotid discrimination and reverse transcription by the human mitochondrial DNA polymerase. J. Biol. Chem. 286, 31490–31500. Kraytsberg, Y., Schwartz, M., Brown, T.A., Ebralidse, K., Kunz, W.S., Clayton, D.A., Vissing, J., Khrapko, K., 2004. Recombination of human mitochondrial DNA. Science 304, 981. Krishnan, N.M., Seligmann, H., Raina, S.Z., Pollock, D.D., 2004a. Detecting gradients of asymmetry in site-specific substitutions in mitochondrial genomes. DNA Cell Biol. 23, 707–714. Krishnan, N.M., Seligmann, H., Raina, S.Z., Pollock, D.D., 2004b. Phylogenetic analyses detect site-specific perturbations in asymmetric mutation gradients. Curr. Comput. Mol. Biol. 2004, 266–267. Krishnan, N.M., Seligmann, H., Rao, B.J., 2008. Relationship between mRNA secondary structure and sequence variability in chloroplast genes: possible life history implications. BMC Genomics 9, 48. Ladoukakis, E.D., Zouros, E., 2001a. Direct evidence for homologous recombination in mussel (Mytilus galloprovincialis) mitochondrial DNA. Mol. Biol. Evol. 18, 1168–1175. Ladoukakis, E.D., Zouros, E., 2001b. Recombination in animal mitochondrial DNA: evidence from published sequences. Mol. Biol. Evol. 18, 2127–2131. Leach, D.R., Okely, E.A., Pinder, D.J., 1997. Repair by recombination of DNA containing a palindromic sequence. Mol. Mirobiol. 26, 597–606. Lee, D.Y., Clayton, D.A., 1997. RNase mitochondrial RNA processing correctly cleaves a novel R loop at the mitochondrial DNA leading-strand origin of replication. Genes Dev. 11, 582–592. Little, R.D., Platt, T.H.K., Schildkraut, C.L., 1993. Initiation and termination of DNA replications in human ribosomal-RNA genes. Mol. Cell. Biol. 13, 660–6613. Michel, C., Seligmann, H., 2014. Bijective transformation circular codes and nucleotide exchanging RNA transcription. Biosystems 118, 39–50. Pepper, M.R., Doughty, P., Fujita, M.K., Moritz, C., Keogh, S., 2013. Speciation on the rocks: integrated systematics of the Heteronotia spelea species complex (Gekkota; Reptilia) from Western and Central Australia. PLoS One 8, E78110. Pepper, M.R., Fujita, M.K., Moritz, C., Keogh, S., 2011. Paleoclimate change drove diversification among isolated mountain refugia in the Australian arid zone. Mol. Ecol. 20, 1529–1545. Piganeau, G., Gardner, M., Eyre-Walker, A., 2004. A broad survey of recombination in animal mitochondria. Mol. Biol. Evol. 21, 2319–2325. Prado, F., Aguilera, A., 2005. Impairment of replication fork progression mediates RNA polII transcription-associated recombination. Embo. J. 24, 1267–1276. Rissler, L.J., Taylor, D.R., 2003. The phylogenetics of Desmognathine salamander populations across the southern Appalachians. Mol. Phylogenet. Evol. 27, 197–211. Rodin, A.S., Rodin, S.N., 2007. Translation of both complementary strands might govern early evolution of the genetic code. In Silico Biol. 7, 309–318. Seligmann, H., 2008. Hybridization between mitochondrial heavy strand tDNA and expressed light strand tRNA modulates the function of heavy strand tDNA as light strand replication origin. J. Mol. Biol. 379, 188–199. Seligmann, H., 2010a. Avoidance of antisense antiterminator tRNA anticodons in vertebrate mitochondria. Biosystems 101, 42–50.

H. Seligmann / BioSystems 125 (2014) 22–31 Seligmann, H., 2010b. Mitochondrial tRNAs as light strand replication origins: similarity between anticodon loops and the loop of the light strand replication origin predicts initiation of DNA replication. Biosystems 99, 85–93. Seligmann, H., 2011a. Two genetic codes, one genome: Frameshifted primate mitochondrial genes code for additional proteins in presence of antisense antitermination tRNAs. Biosystems 105, 271–285. Seligmann, H., 2011. Mutation patterns due to converging mitochondrial replication and transcription increase lifespan, and cause growth rate-longevity tradeoffs. In: DNA Replication-Current Advances, Seligmann H. (ed.), Chapter 6, INTECH, http://www.intechopen.com/books/dna-replication-current-advances/. Seligmann, H., 2012a. Overlapping genes in the 30 -to-50 -direction in mitochondrial genes and 30 -to-50 polymerization of non-complementary RNA by an ‘invertase’. J. Theor. Biol. 315, 38–52. Seligmann, H., 2012b. An overlapping genetic code for frameshifted overlapping genes in Drosophila mitochondria: Antisense antitermination tRNAs UAR insert serine. J. Theor. Biol. 298, 51–76. Seligmann, H., 2012c. Overlapping genetic codes for overlapping frameshifted genes in Testudines: and Lepidochelys olivacea as a special case. Comput. Biol. Chem. 41, 18–34. Seligmann, H., 2012d. Coding constraints modulate chemically spontaneous mutational replication gradients in mitochondrial genomes. Curr. Genomics 13, 37–54. Seligmann, H., 2012e. Putative mitochondrial polypeptides coded by expanded quadruplet codons, decoded by antisense tRNAs with unusual anticodons. Biosystems 110, 84–106. Seligmann, H., 2013a. Polymerization of non-complementary RNA: systematic symmetric nucleotide exchanges mainly involving uracil produce mitochondrial transcripts coding for cryptic overlapping genes. Biosystems 111, 156–174. Seligmann, H., 2013b. Systematic asymmetric nucleotide exchanges produce human mitochondrial RNAs cryptically encoding for overlapping protein coding genes. J. Theor. Biol. 324, 1–20.

31

Seligmann, H., 2013c. Triplex DNA:RNA, 30 -to-50 inverted RNA and protein coding in mitochondrial genomes. J. Comp. Biol. 20, 660–671. Seligmann, H., 2013. Putative protein-encoding genes within mitochondrial rDNA and the D-Loop region. Ribosomes: molecular structure, role in biological functions and implications for genetic diseases, chapter 4, Lin Z and Liu W (eds), Novapublishers. Seligmann, H., 2013e. Pocketknife tRNA hypothesis: anticodons in mammal mitochondrial tRNA side-arm loops translate proteins? BioSystems 113, 165–176. Seligmann, H., 2014a. Putative anticodons in mitochondrial tRNA sidearm loops: pocketknife tRNAs. J. Theor. Biol. 340, 155–163. Seligmann, H., 2014b. Species radiation by DNA replication that systematically exchanges nucleotides? J. Theor. Biol. 363, 216–222. Seligmann, H., Krishnan, N.M., Rao, B.J., 2006a. Possible multiple origins of replication in primate mitochondria: alternative role of tRNA sequences. J. Theor. Biol. 241, 321–332. Seligmann, H., Krishnan, N.M., Rao, B.J., 2006b. Mitochondrial tRNA sequences as unusual replication origins: pathogenic implications for Homo sapiens. J. Theor. Biol 243, 375–385. Seligmann, H., Labra, A., 2013. Tetracoding increases with body temperature in Lepidosauria. Biosystems 114, 155–163. Yang, W., Wu, J., Bi, A., Ou-yang, Y., Shen, H., Chirn, G., Zhou, J., Weiss, E., Holman, E.P., 2013. Possible formation of mitochondrial-RNA containing chimeric or trimeric RNA implies a post-transcriptional and post-splicing mechanism for RNA fusion. PloS one 8, e77016. Zhang, Z., Schwartz, S., Wagner, L., Miller, W., 2000. A greedy algorithm for aligning DNA sequences. J. Comput. Biol. 7, 203–214. Zhou, Z.-H., Akguen, E., Jasin, M., 2001. Repeat expansion by homologous recombination in the mouse germ line at palindromic sequences. Proc. Natl. Acad. Sci. U. S. A. 98, 8326–8333.