General introduction
Eukaryotic genomes contain a large amount of repetitive sequences, the vast majority derives from transposable elements. Mobile elements were first discovered by Barbara McClintock in the maize genome (McClintock 1950; McClintock 1953; McClintock 1956; McClintock 1984). In the beginning they were called “junk DNA”(Ohno 1972; Pagel and Johnstone 1992), because they did not seemed to have a useful function for the host. Later, they were called “selfish genes” (Dawkins 1976; Doolittle and Sapienza 1980; Orgel and Crick 1980) and “genomic parasites” (Yoder, Walsh, and Bestor 1997), suggesting that their only purpose is to multiply themselves in the host genome. However, more recent studies have shown that they can cause diseases (Kazazian 1998), but they can also have a beneficial effect by increasing the genomic diversity (J. D. Boeke and Pickeral 1999; Nekrutenko and Li 2001). Today it is clear that mobile elements are more than “junk DNA” and they influence the host genomes and phenotypes (Hancks, Kazazian, and Jr. 2010; Beck et al. 2011; Hancks and Kazazian 2012; Richardson, et al. 2015).
DNA transposons and RNA transposons are the major types of mobile elements. The present thesis focuses on RNA transposons, also called retrotransposons. This type or mobile elements replicate through a RNA intermediate and generate a new copy in the genome (Moran et al. 1996; Feng et al. 1996). Due to their copy-and-paste behavior, retrotransposons accumulate in the genome and have a large impact on genomic composition.
SVA and LAVA, together with PVA and FVA, are part of the VNTR-composite retrotransposons. They are specific to hominoid primates. SVA is the only VNTR-composed retrotransposon that had expanded in hominids (Fam. Hominidae), and the LAVA, PVA and FVA elements are found only in the gibbon genome. The mobilization of these elements involves proteins encoded by L1 elements (Eric M Ostertag et al. 2003; Carbone et al. 2014; Ianc and Ochis et al. 2014). However, the mechanism of its mobilization is not elucidated to date.
During my PhD, I studied the structural requirements in mobilization of VNTR-composed elements. I worked on human SVA and gibbon LAVA elements, with specific interest in structural/sequence features which allows them to be efficiently mobilized.
1.2. Classification of mobile elements
The completion of the first human genome sequencing revealed that 45%, of its sequence is represented by transposable elements (Lander et al. 2001; Venter et al. 2001).
Figure 1.1 illustrates the different types of transposons with suggestive exemples from mammalian and human genome.
Based on their mobilization process, mobile elements can be classified on: RNA transposons (class I) and DNA transposons (class II) (Pace and Feschotte 2007). Elements from class I are mobilized via a RNA intermediate using a reverse transcriptase and the integration of class II elements are DNA mediated and they use a transposase instead.
Figure1.1. General classification of interspersed elements with suggestive examples of mobile elements, adapted from (Ayarpadikannan and Kim 2014).
1.2.1 DNA transposons
These mobile elements, first identified by Barbara McClintock in maize genome (McClintock 1950; McClintock 1953; McClintock 1956; McClintock 1984), are mobilized via a DNA intermediate by “cut-and-paste” mechanism, also called transposition.
The general structure of a DNA transposon is illustrated in the Figure 1.2. They encode a transposase which help the element in self-excision (Smit and Riggs 1996) and are flanked by inverted repeats. The transposase seems to not be specific to oane element, it can mobilize in trans any sequence which is flanked by a transposase recognition signal.
The enzyme recognize the flanking inverted repeats and then catalyze the excision and, after a target site cleavage, the insertion of the element from one genomic place to another.
Figure 1.2. Schematic representation of a human DNA transposon.They encode a transposase enzyme (pink rectangle) that is flanked by terminal inverted repeat sequences (pink arrows).
Class II of transposon, occuping ~3% of the human genome (Lander et al. 2001), is considered to be fossil in the human genome (Pace and Feschotte 2007). But it is active in others genomes, like plants or bats (Lander et al. 2001; Mitra et al. 2013).
1.2.2. RNA transposons
These mobile elements, also called retrotransposons, are the largest class of mobile elements in the human genome.
They are mobilized by a “copy-and-paste” mechanism, via an RNA intermediate (Cordaux and Batzer 2009). Their mobilization mechanism involves transcription of the element, reverse transcription of the RNA into cDNA and integration of the cDNA copy into a new genomic location after target site cleavage. Thus, an active RNA transposon is multiplied into the genome. As a mark of their mobilization, retrotransposons are flanked by TSDs (target-site-duplications), terminate in a polyA tail and can be truncated or inverted and can transduce genomic sequences during retrotransposition process (Eric M Ostertag et al. 2003).
From a structural point of view, retrotransposons are divided into two large groups based on the presence or absence of long terminal repeats (LTRs), namely LTR retrotranspososn and non-LTR retrotransposons (Ayarpadikannan and Kim 2014). Although they have a different structure, these groups share some characteristics, like a functional 5’ promoter, encode for a reverse transcriptase (Hata and Sakaki 1997)
Based on mobilization mechanism, there are two types of retrotransposons: autonomous, which encode factors necessary for their own mobilization, and non-autonomous, which use the mobilization machinery of the autonomous retrotransposons (Eric M. Ostertag and Kazazian Jr 2001).
1.2.2.1.Autonomous retrotransposons
The autonomous retrotransposons can be or not flanked by long terminal repeats (LTR).
Figure 1.3. Schematic representation of human autonomous retrotransposons, adapted from Beck et al. 2011. A) Representation of LTR-retrotransposon (HERV). It is flanked by long terminal repeats (LTRs) and contains genes similar to retroviral gag and pol and env (not functional). B) Representation of non-LTR retrotransposon (LINE 1). The 5' untranslated region (5' UTR), containing polymerase II promoter activity (marked with arrow), is followed by two ORFs, separated by a connector spacer, a 3' UTR that contains a polyA signal (An). ORF1 encodes a nucleic acid binding protein and contains: CC – coiled-coil domain, RRM – RNA recognition motif and CTD – carboxyl-terminal domain. ORF2 encodes an endonuclease (EN), reverse transcriptase (RT), and also contains a cysteine-rich domains (C).
1.2.2.1.1. Autonomous LTR-retrotransposons
They are also called retrovirus-like elements or endogenous retroviruses. They have a retroviral-like structure and mechanisms, being flanked by Long Terminal Repeats sequence, of 300-1000bp long (J. Boeke and Stoye 1997), but they don’t have a functional envelope gene (eng gene). In this condition their life cycle is only intracellular and cannot infect other cells (Esnault et al. 2008).
The human endogenous retroviruses (HERVs), elements of about 9.5kb (Shin et al. 2013) (represented in Figure 3A), are part of this category, representing ~8% of the human genome (Lander et al. 2001). They have retroviral origin, as their name suggest. The original retroviruses which infected a host, became prisoner in the host germline, beeing unable to reinfect and remain captured within host genome (John L Goodier and Kazazian 2008). They typically encode a reverse transcriptase and have analogous pol and gag genes to retroviruses, but they lack a functional envelope gene (Ono, Kawakami, and Takezawa 1987), mainly due to mutations which inactivated their open reading frames and produced a non-functional protein (H.-S. Kim 2012).
HERV elements are grouped in 3 classes based on the similarities to the retroviral genera. Although there are some polymophic loci in the human genome, no de novo HERV insertion was discovered to date, it is considered an inactive element (Shin et al. 2013).
1.2.2.1.2. The autonomous non-LTR retrotransposons
These elements are not flanked by long terminal repeats (LTRs).
Based on their endonuclease domain, the autonomous non-LTR retrotransposons are divided into two groups: RE-type (with restriction enzyme-like EN domain) and APE-type (with apurinic-apyrimidinic endonuclease domain). They encode one (RE-type) or two open reading frames (APE-type) and end with a polyA sequence. The RE-type elements are considered to be oldest non-LTR retrotransposons (Malik, Burke, and Eickbush 1999).
The main autonomous non-LTR retrotransposon in human genome is the Long INterspersed Element (LINEs), which is an APE-type non-LTR retrotransoson. It comprise ~21 % of the genome (Beck et al. 2011). Its structure is represented in Figure 1.3B.
1.2.2.1.2.1. L1 element
LINE1 (L1) is the only active autonomous transposable element in the human genome. It is very active in germline cells (Branciforte and Martin 1994), but it was demonstrated that L1 retrotransposition occurs also in somatic cells, for example brain cells (Muotri et al. 2005; Coufal et al. 2009; Upton et al. 2015), generating also an intra-individual genetic diversity.
L1 has more than 500 000 copies in the human genome but less than 100 copies were predicted to be active (Brouha et al. 2003). The major part of L1 elements contain mutations or are 5’ or 3’ truncated and therefore inactive.
A full-length element (having about 6 kb in length) contains a 5' UTR that contains an endogenous RNA polymerase II promoter (Swergold 1990), two open reading frames separates by a spacer and a 3’ UTR which contains a polyadenylation signal (Dombroski et al. 1991) (Figure 1.3B). The existence of a 155 bp endogenous L1 promoter (Swergold 1990) was confirmed later by Minakami et al. 1992.
The ORF1 encodes a nucleic acid binding protein (Hohjoh and Singer 1996, 1997), which also have a chaperone activity (Martin and Bushman 2001). It has three domains: a coiled-coil domain, which binds to L1 RNA to form the ribonucleoprotein particles (RNPs) (Hohjoh and Singer 1996, 1997), a central RNA recognition motif (RRM), which mediate the nucleic acid binding (Januszyk et al. 2007; Khazina and Weichenrieder 2009) to the carboxyl terminal domain (CTD).
The ORF2 encodes ORF2p with endonuclease and reverse transcriptase activity (Feng et al. 1996; Mathias et al. 1991). The ORF2p endonuclease cleaves single stranded DNA at 5’ TT/AAAA 3’ consensus sequence (with bar is indicated the cleavage site) (Feng et al. 1996; Jerzy Jurka 1997), this nick expose a 3’ hydroxyl group that is used by ORF2p to prime the reverse transcription of L1 RNA (or other cellular RNAs). The ORF2p RT use L1 RNA and some others RNAs as template in target-primed reverse transcription (TPRT) (Kulpa and Moran 2006), the mechanism through which new L1 copies are integrated in a new genomic location. The carboxyl terminal domain of ORF2p contains a cysteine-rich region with unknown function. It was shown that mutations occurring in this region have a negative effect on L1 retrotransposition (Moran et al. 1996). Also, in a recent study was showed that this cysteine-rich domain is involved in RNA binding (Piskareva et al. 2013).
The L1 insertions have a great influence on the host genome by creating genomic diversity. L1 elements contain multiple splicing sites on both strands (Belancio, Hedges, and Deininger 2006), and also polyadenylation signals (Perepelitsa-Belancio and Deininger 2003), which can cause splicing and premature polyadenylation of genes that contains L1 insertions. In addition, because the 3’UTR polyadenylation signal of L1 element is a week one, it can often be bypassed, leading to transcription of the flanking sequence (Moran, DeBerardinis, and Kazazian 1999), as promoters, regulatory sequences or exons; consequently those would be moved into a new genomic location and would influence the genic expression in that insertion area.
L1 mobilization machinery have a cis-preference, meaning that they prefer to mobilize their RNA to the detriment of other cellular RNAs or those of non-autonomous retrotransposons (SINEs, SVA, etc.) (Wei et al. 2001).
1.2.2.2. The non-autonomous retrotransposons
The non-autonomous retrotransposons do not encode any proteins and their mobilization is possible only with LINE proteins. In this category take part: SINEs elements, VNTR-composed elements and processed pseudogenes.
1.2.2.2.1. SINEs – Alu element
SINE elements, Short INterspersed Elements, are the most abundant class of retrotransposons in the human genome, comprising 13% of the genome (Lander et al. 2001). Between them, Alu element represents 11 % (Lander et al. 2001), being found mostly in introns, intergenic regions and untranstated regions of genes (Batzer and Deininger 2002).
Figure 1.4. Representation of Alu element (adapted from Cordaux and Batzer 2009). It is formed by two 7SL-derived monomers (shown in dark green) connected by an A-rich region. The left-monomer contains A and B boxes of PolIII-promoter (shown in pink). Alu ends in a polyA tail.
As non-autonomous retroelements, they also use L1 machinery for their reverse transcription and integration (Dewannieux, Esnault, and Heidmann 2003).
The Alu elements are of 300 bp length, and contains an internal PolII-promoter (Singer 1982; Schmid 1996) (Figure 1.4). Their name come from the AluI restriction enzyme site which was identified in these elements (Houck, Rinehart, and Schmid 1979). Alu has two monomers originating from 7SL RNA (Ullu and Tschudi 1984), component of the signal recognition particle (SRP), which target nascent secretory proteins to endoplasmic reticulum (Walter, Gilmore, and Blobel 1984). It is considered that Alu was generated from 7SL RNA by several steps: first the deletion of 7SL S-domain, followed by the acquisition of the polyA tail and next the dimerization (Dewannieux, Esnault, and Heidmann 2003).
The Alu elements were classified in three subclasses based on differences relative to the 7SL sequence: subfamily J, subfamily S and subfamily Y (J Jurka and Smith 1988). The oldest subfamily, AluJ, is considered fossil because it is inactive in the human genome. Instead, the AluY subfamily is the youngest one and it is very active in the human genome, being responsible for lots of diseases (reviewed in Kim et al. 2016). It was estimated that a new Alu insertion appears at ~200 births (Batzer and Deininger 2002).
1.2.2.2.2. VNTR-composed elements
VNTR-composite elements are non-autonomous, non-LTR retrotransposons, which are composed, as the name suggests, by different domains around of the central VNTR (variable-number-of-tandem-repeats) region. In this category take part: SVA (Shen et al. 1994; Wang et al. 2005), PVA (Hara et al. 2012), FVA (Ianc and Ochis et al. 2014) and LAVA (Carbone et al. 2012) elements, which are specific to the hominoid primates. They share a similar 5’ and central region and differs by their 3’ part (Figure 1.5).
The 3’ end of the VNTR-composed elements was acquired by splicing to the VNTR region, the case of LAVA and FVA, and to the SVA2, the case of SVA and PVA (Ianc and Ochis et al. 2014).
Figure 1.5. Schematic representation of VNTR-composite retrotransposons, adapted from Ianc and Ochis et al. 2014. The elements share a central VNTR domain and the 5’ end: CT hexameric region and Alu-like domain. They differs at the 3’end: SINE-R- sequence derived from HERV-K, PTGR2- exon 4 and 5′-part of intron 4 of the gene encoding Prostaglandin Reductase 2, FR- partial FRAM (Free Right Alu Monomer) sequence, Alu L1- partial AluSz and L1ME5 sequences. At the 3’ end all elements have a polyadenylation signal (not represented). An-polyA tail.
1.2.2.2.2.1. SVA element
The SVA element was first identified by Shen et al. and was named after its structural components 3’ SINE-R – VNTR – Alu 5’ (Shen et al. 1994). The SVA was shown to be mobilized by L1 machinery in trans (Dustin C. Hancks et al. 2011; Raiz et al. 2012).
The element has ~2 kb and is composed at its 5’ end by a hexameric-repeat region (CCCTCT) and an Alu-like region which contains two antisense Alu fragments and a sequence with unknown origin. Follows a variable number of 30-50bp tandem repeats (VNTR) region (Wang et al. 2005) and a SINE-R region (490bp) at the 3’ end, which originate from the 3’end of the env gene and the 3’ part of LTR of the HERV-K10 (Ono, Kawakami, and Takezawa 1987). The element ends by a polyadenylation signal (AATAAA) (Wang et al. 2005) and a polyA tail (Figure 1.5).
Even though at their 5’ end contains Alu-like sequence, SVA elements have not an RNA polymerase III internal promoter, ant it seems to be transcribed by RNA polymerase II like L1 elements (Wang et al. 2005), but until today the endogenous polI promoter have not been identified.
Figure 1.6. Evolution of SVA family in primates, adapted from Wang et al. 2005. The estimated copy number is shown below primate name.
In the rhesus macaque genome it was identified a precursor of the VNTR region present in SVA elements. This precursor was also identified in the human genome (K. Han et al. 2007) and was called SVA2 (it is formed by a VNTR region, followed by unique 3’ region and a polyA tail). About 40 SVA2 elements were found in rhesus genome (K. Han et al. 2007). So, from evolutionary point of view, it can be said that SVA originate before the divergence of hominid primates, but after the divergence of hominid and Old World primates (Wang et al. 2005) (Figure 1.6). The same study shows that SVAs had two expansion periods. First, before the divergence of Great Apes (orangutan, gorilla, chimpanzee, human) SVA-A and SVA-B were expanded, and the second expansion was after the divergence of orangutan when SVA-D and SVA-C amplified (Wang et al. 2005) (Figure 1.6). These two major expansions were followed by lineage specific expansions in chimp and human genome. In this way appeared SVA-E and SVA-F human specific subfamily and PtA chimp specific subfamily (Wang et al. 2005) (Figure 1.6). No SVA elements were identified in Old World monkeys (green monkey and Rhesus macaque) but gibbons genome reveled 29 copies of SVA element (Ianc and Ochis et al. 2014).
Figure 1.7. SVA copy number in primates, taken from (Wang et al. 2005), estimated by qPCR based on copy number in the human genome (2762). In the gibbon genome 29 SVA copies were found (not represented in graphic).
SVA elements are very abundant in human, chimp and gorilla compared to orangutan which has ~900 copies (Wang et al. 2005) and gibbon 29 copies of SVA (Ianc and Ochis et al. 2014) (see copy number represented in Figure 1.7). In the human genome were identified 2762 SVA copies and ~63% of them are full-length elements (Wang et al. 2005).
Figure 1.8. SVA subfamilies in the human genome (taken from Wang et al. 2005). The percentage of each subfamily is given relative to the all SVA elements in the human genome (left). Median-joining network of the SVA subfamilies, constructed using S part of the consensus of each subfamily (right). The number of substitution on each line is indicated on the top of the lines. The circles size represent the relative size of the subfamily
SVA elements are divided into six subfamilies based on diagnostic mutations compared to HERV-K10 sequence: named with A to F, from older to younger ones. The SVA-E and SVA-F are human specific (Wang et al. 2005) (Figure 1.8).
SVA is still active in the human genome and can create diseases (Kobayashi et al. 1998; Wilund et al. 2002; Eric M Ostertag et al. 2003; Callinan and Batzer 2006) together with L1 and Alu elements (Eric M. Ostertag and Kazazian Jr 2001; Batzer and Deininger 2002). They also cause inter-individual variation by the presence/absence polymorphism in humans. Wang et al. 2005 had estimated that 37.5% of SVA-E and 27.6% of SVA-F elements are polymorphic, and that the human genome has 56 SVA insertion polymorphism (E. A. Bennett et al. 2004).
1.2.2.2.2.2. LAVA element
LAVA (3’ L1/Alu/VNTR/Alu-like 5’) element was recently discovered in gibbon genome (Carbone et al. 2012). It is a gibbon specific VNTR-composed element.
As all elements from this category, LAVAs contain a CT hexameric region, followed by a sequence with homology to Alu elements, a VNTR central domain and ends by a region which originate from intron 2 of hydroxysteroid (17-beta) dehydrogenase 3 gene (HSD17B3 gene) from Macaca mulatta (in rhesus macaque genome on chromosome 15) – this portion of intron 2 contains sequence of the the AluSz and L1ME5 elements. The AluSz fragment is flanked by two unique sequences (named U1 and U2) (represented in Figure 1.5 and 1.9).
Figure 1.9. Representation of LAVA elements. A full-length element contains 4 regions: at 5’ end a CT-hexameric repeat region, followed by a Alu-like region, a VNTR region and at 3’ end contains a sequence from intron 2 of HSD17B3 gene of resus macaque. In detail, the 3’ part have 2 unique sequences (marked with U1 and U2), which flank a sequence with similarity to AluSz element, followed by a sequence with similarity to L1ME5 in antisense orientation. The element ends through a polyA tail. LAVA’s are grouped into 6 families, named from LAVA_A to LAVA_F. A specific deletions in the Alu-like region of LAVA_F is represented.
A total of 1797 LAVA copies were identified in gibbon genome, based on the the 3’ end of the elements, and were classified in 22 subfamilies which can be grouped into six larger families, named from LAVA_A to LAVA_F (Figure 1.9) (Carbone et al. 2014).
The older family is the LAVA_A, considered a fosil element, which have the longest Alu-like region and it presents high sequence similarity with Alu-like region of the old SVA_A family (Carbone et al. 2014; Damert A, unpub data). The other LAVA families are more sequence related to each other and contain specific Alu-like deletions (for example LAVA_F represented in Figure 1.9) or a insertion (in the case of LAVA_D) (Carbone et al. 2014; Damert A, unpub data). The younger family is considered to be LAVA_F which have a large deletion in the Alu-like region (Figure 1.9) and it is currently an active element in the gibbon genome (Carbone et al. 2014).
The LAVA element is still active in the gibbon genome. Carbone et al. in 2014 showed that 4 LAVA loci in the Nomascus leucogenis genome were polymorphic and other three loci are specific to only one individual out of 7 of N. leucogenis. This suggest a relative recent retrotransposition event (Carbone et al. 2014).
The largest family is LAVA_B, which comprise about 50% of the LAVA elements from the gibbon genome, and the smallest one is LAVA_D (about 5% of total LAVAs), and the rest of families are represented in aproximatively equal proportions (Figure 1.10) (Carbone et al. 2014).
Figure 1.10. Relationships among LAVA subfamilies showed by a median-joining network, taken from Carbone et al. 2014. The circles represent LAVA subfamilies and their size is proportional to the number of elements in the subfamily. Black dots represent hypothetical sequences connecting adjacent subfamilies. Branch lengths are not drawn to scale.
If analyzing the evolutionary history of LAVAs in primates, the LAVA prototype should derive from intron 2 of the HSD17B3 gene of M. mulatta, and must have been assembled before the split of gibbons and great apes. This assumption is based on the fact that 5’ truncated LAVA elements – containing only the VNTR and LA part – were identified in the gorilla, chimpanzee and human genome. No such intermediate was found the the orangutan genome and nor in the old world monkeys, and the expansion of LAVA subfamilies should be after the split of gibbons and great apes (Damert A, unpub data). Only the younger family, LAVA_F, seems to be amplified after the divergence of the four gibbon genera, and is Nomascus leucogenis specific (Carbone et al. 2014).
1.2.2.2.2.3. FVA and PVA elements
Those elements have the same structure as SVA and LAVAs, being VNTR-composed retroelements, and having a different 3’ region (Ianc and Ochis et al. 2014). Both elements have hallmarks of mobilization by L1 machinery, like TSDs. As all VNTR-composed elements they acquired the 3’region through a splicing mechanisms (Ianc and Ochis et al. 2014).
Figure 1.11. Detailed representation of PVA and FVA retrotransposons, adapted from Ianc and Ochis et al. 2014. The elements share a central VNTR domain and the 5’ end: CT hexameric region and Alu-like domain. The 3’end of PVAs (PTGR2-marked in bright pink) contains the exon 4 and 5′-part of intron 4 of the gene encoding Prostaglandin Reductase 2. The FVA 3’ end contains a sequence of Free Right Alu Monomer (FRAM –marked in dark orange)which is flanked by 2 non-repretitive sequences (U1- U2, marked with white boxes).
The PVA (PTGR2/VNTR/Alu-like) contains at the 3’ end: the full exon 4 and 5′part of intron 4 of a gene homologue to the human Prostaglandin Reductase 2 encoding gene, preceding a sequence of 31 bp of the 5′part of the SVA2 unique 3′sequence (Hara et al. 2012; Ianc and Ochis et al. 2014). The elements ends with a polyA sequence which contains the AATAAA polyadenilation signal (Hara et al. 2012) (see Figure 1.11).
Because this element contain a sequence of the SVA2 (also present in older SVA_A family and in SVA_NLE – SVA from Nomascus leucogenys), made us speculated that PVA elements derive from SVAs through a splicing mechanism (Ianc and Ochis et al. 2014). A total of 143 copies of PVA were identified in the gibbon genome (Ianc and Ochis et al. 2014).
We had discovered in the gibbon genome the FVA retroelements (FRAM/VNTR/Alu-like) (Ianc and Ochis et al. 2014). These elements are characterized by a 3′ end which contains fragments from FRAM (Free-Right-Alu-Monomer) element flanked by non-repetitive sequence (see Figure 1.11). Only 11 copies of this element ware identified (Ianc and Ochis et al. 2014).
No family structure were identified for these elements (Ianc and Ochis et al. 2014). Their Alu-like region have a high similarity to the Alu-like region of the human SVA_A, SVA_NLE and to LAVA_A elements (Ianc and Ochis et al. 2014).
1.2.2.2.3. Processed pseudogenes
Processed pseudogenes or retropseudogenes are cellular RNA which was reverse transcribes by RT ORF2p of L1 element and inserted to a new genomic location (Esnault, Maestre, and Heidmann 2000).
Figure 1.12. Representation of a processed pseudogene, adapted from Richardson et al. 2015: spliced and reverse transcribed cellular mRNA , followed by a polyA tail and flanked by TSDs
There are more than 8 000 of processed pseudogenes in the human genome (Zhang et al. 2003), which lacks promotors and introns, but contains a polyadenylation tail (Ding et al. 2006), and are flanked by TSDs (target-site-duplications) as mark of retrotransposition (Figure 1.12). They do not encode functional proteins, because they had acumulated mutations over time, but there exist some exceptions. Vinckenbosch and collegues (Vinckenbosch, Dupanloup, and Kaessmann 2006) had estimated that ~160 of processed pseudogenes can be translated into functional proteins, called retrogenes (Ding et al. 2006)