**3. Results**

524 A Bird's-Eye View of Veterinary Medicine

found in cattle affected with classical BSE. Moreover, in vitro studies showed that the possibility to infect neurosphere cultures with scrapie prion is linked to an over-expression of PrPC (Giri et al., 2006). Although the determinism of prion disease is multifactorial (prion strain and prion protein sequence, see Doherr, 2003), it seems that a high expression of PrPC

In numerous organisms, post-transcriptional gene regulation involves small (around 18-25 nucleotides long) non-coding RNA molecules, the microRNAs (miRNAs). They recognise specific target sequences in the 3'UTR of some transcripts, mediating their silencing (Fabian et al., 2010) by inducing their degradation or by inhibiting their translation. According to the accuracy of the complementarity between both sequences, each miRNA can regulate up to hundreds of genes. Using microarrays, Saba et al. (2008) evidenced 15 miRNAs, potentially controlling the transcript amount of more than 100 genes. These micro-RNAs are de-

In their comparative analysis of *Prnp* organization in human, mouse and sheep, Lee et al. (1998) showed the high conservation of the 3'-UTR regions and suggested their role in mRNA stability. Moreover, they evidenced insertions of transposable elements in the sheep gene. In their more recent work, Premzl and Gamulin (2007) could easily align the part close to the poly-adenylation signal in a wide range of mammals. Taken together, these data suggest that alignments of mammalian 3'UTRs contain reliable information. The regulatory sequences borne by the 3'-UTR are involved in mRNA processing, transport, stability, and translation. As the 3'-UTRs harbor recognition sites for diverse RNA-binding proteins that regulate gene expression as well as active microRNA target sites, our strategy was to compare the 3'UTR of the *Prnp* gene, from the sequence terminus codon to the polyadenylation signal in different Mammal lineages. We expect to find in the most frequently families affected by TSE, i.e., Bovidae and Cervidae, oddities in their potential miRNA targets. Three types of results are predicted to be obtained. In the first type, based on the miRNA-related gene silencing, we predict that Bovidae and Cervidae families lack some targets widely present in other Mammals. In the second type, we predict that target sequences are present in DNA from most mammalian lineages, including the both critical families, but a loss of corresponding miRNA may occur. The last type, although more unlikely, would be to detect targets only harboured by Bovidae and Cervidae, which would lead to the activation of gene expression, as shown by Vasudevan et al. (2007) in a model of cellular stress. In brief, we search for potential targets that could affect repression mechanisms of *Prnp* gene expression, and thus enhance the sensitivity of Bovidae and

A Blast search of 3'UTR sequences in traces (Whole Genome Shotgun) of NCBI genome database was undertaken. We restricted the species sampling to 2 to 3 species for main division within Eutherians, and the outgroup was constituted by 2 Marsupials (*Macropus eugenii* and *Didelphis domestica*). The alignment of sequences was performed with Dialign software (Morgenstern, 2004), and then manually edited. The table 1 gathers the 27 analyzed sequences and their accession numbers. The repeat elements were identified with RepeatMasker site (http://www.repeatmasker.org/cgi-bin/WEBRepeatMasker). The target

gene is necessary (Krejciova et al., 2011).

regulated in the brain of scrapie affected mouse.

Cervidae to the disease.

**2. Material and methods** 

#### **3.1 Position of insertions and deletions**

The length of the 3'UTR region is highly variable as the murine sequence comprises about 1250 pb whereas the bovine one is more than 3500 pb long and possesses two potential polyadenylation signals that are separated by about 1300 bp. Such a result is also observed in sheep and mule deer. The difference is mainly explained by a series of transposable

Analysis of 3'UTR of Prnp Gene in Mammals:

Possible Role of Target Sequences of miRNA for TSE Sensitivity in Bovidae and Cervidae 527

Fig. 2. Position of insertions (loops above the thick blue line) and deletions (wide red rectangles below) in the 550 pb conserved part of Eutherians. The hypothetical ancestral sequence for Eutherians was taken as reference. Short vertical lines are drawn every 100 pb. The most interesting features are a 177 pb long deletion specific to Cervidae-Bovidae, and a 470 pb long insertion in African Elephant. No indel particular to human could be recorded.

The most conserved sites are between positions 70 and 115 downstream the stop codon. In the part corresponding to the deletion shared by Bovidae and Cervidae, we observed little variations in Eutherians. From the 550th position on, due to a lowest identity between

We focused on 3 regions: (i) between the stop codon and the last nucleotide before the large deletion affecting Bovidae-Cervidae (proximal region), (ii) inside this large deletion, and (iii)

The table 2 gathers the target sites present in more than one species, in the region proximal to codon stop. It should be noted that several potential targets are overlapping (fig. 4-5). In two cases, the distribution of target is identical in the studied mammals and they were considered as identical (Mir-569-155 on one hand and Mir-3918-509 on the other). In contrast, the presence of other overlapping targets varies among mammals (Mir-3102, 376, 3918/509, 519 and 3432). Several targets are shared by only Bovidae-Cervidae species (Mir-569/155, 376, 15b), or Bovidae-Cervidae and other mammals (Mir-3102 with *M. musculus* and *S*. *scrofa*, Mir-3918/509 with *H. sapiens*, *M. putorius*, and *O. princeps*, Mir-519 with *C* 

**3.2 Profile of site change rates along the 3'UTR** 

**3.3 Position of potentially target sites of miRNAs** 

species, the sequences cannot be significantly aligned (fig. 3).

in the transposable elements shared by Bovidae-Cervidae.

elements (SINE/RTE-BovB of about 380 pb length, SINE/BovA about 160 bp length and DNA/TcMar-Mariner with at least 1220 bp). All are inserted in the 3' end of the bovine sequence, and also shared by sheep, red deer, and mule deer (Fig. 1). A striking common organization is indicative of events that occurred in the ancestors of Bovidae and Cervidae. In spite of extensive searches, these elements were not retrieved in the other studied mammals. Upstream the transposable elements, we observe a good overall conservation of the sequences, allowing to clearly evidence insertion and deletion events in some species.

Fig. 1. Organization of the *Prnp* 3'UTR in several Mammals. The coding sequence of the gene is boxed in thin black. The inserted elements are boxed in thick blue. There are two poly-adenylation signals (Poly-A) in Bovidae-Cervidae, and only one in human and porcine. Short vertical lines are drawn every 1000 pb.

elements (SINE/RTE-BovB of about 380 pb length, SINE/BovA about 160 bp length and DNA/TcMar-Mariner with at least 1220 bp). All are inserted in the 3' end of the bovine sequence, and also shared by sheep, red deer, and mule deer (Fig. 1). A striking common organization is indicative of events that occurred in the ancestors of Bovidae and Cervidae. In spite of extensive searches, these elements were not retrieved in the other studied mammals. Upstream the transposable elements, we observe a good overall conservation of the sequences, allowing to clearly evidence insertion and deletion events in some species.

Fig. 1. Organization of the *Prnp* 3'UTR in several Mammals. The coding sequence of the gene is boxed in thin black. The inserted elements are boxed in thick blue. There are two poly-adenylation signals (Poly-A) in Bovidae-Cervidae, and only one in human and porcine.

Short vertical lines are drawn every 1000 pb.

Fig. 2. Position of insertions (loops above the thick blue line) and deletions (wide red rectangles below) in the 550 pb conserved part of Eutherians. The hypothetical ancestral sequence for Eutherians was taken as reference. Short vertical lines are drawn every 100 pb.

The most interesting features are a 177 pb long deletion specific to Cervidae-Bovidae, and a 470 pb long insertion in African Elephant. No indel particular to human could be recorded.

### **3.2 Profile of site change rates along the 3'UTR**

The most conserved sites are between positions 70 and 115 downstream the stop codon. In the part corresponding to the deletion shared by Bovidae and Cervidae, we observed little variations in Eutherians. From the 550th position on, due to a lowest identity between species, the sequences cannot be significantly aligned (fig. 3).

#### **3.3 Position of potentially target sites of miRNAs**

We focused on 3 regions: (i) between the stop codon and the last nucleotide before the large deletion affecting Bovidae-Cervidae (proximal region), (ii) inside this large deletion, and (iii) in the transposable elements shared by Bovidae-Cervidae.

The table 2 gathers the target sites present in more than one species, in the region proximal to codon stop. It should be noted that several potential targets are overlapping (fig. 4-5). In two cases, the distribution of target is identical in the studied mammals and they were considered as identical (Mir-569-155 on one hand and Mir-3918-509 on the other). In contrast, the presence of other overlapping targets varies among mammals (Mir-3102, 376, 3918/509, 519 and 3432). Several targets are shared by only Bovidae-Cervidae species (Mir-569/155, 376, 15b), or Bovidae-Cervidae and other mammals (Mir-3102 with *M. musculus* and *S*. *scrofa*, Mir-3918/509 with *H. sapiens*, *M. putorius*, and *O. princeps*, Mir-519 with *C* 

Analysis of 3'UTR of Prnp Gene in Mammals:

background: conservation in ≥ 50 % sequences.

**140** X X

*Cervus elaphus*

*Homo sapiens*

**376** X X X

**569/155** X X X X

**<sup>9</sup>**X X X X X X X

**15b** X X X

Table 2. Micro RNAs targets common to several species between the stop codon and the

Within the sequence corresponding to the large deletion region, several targets (Mir-4763, 1587, 744, 15b, and 369) are overlapping (fig. 6-7). In two couples of targets (Mir 569/155 and Mir-3918/509), the distributions in mammals are identical and each couple was considered as single. The remaining overlapping targets have their own distribution within mammals, and they were considered separately (table 3). It appears that the number of putative targets common to at least 2 species varies among the different mammals. There is none in *M. musculus*, but four in *E. caballus* and even six in *F. catus*. There is no target common to all orders of mammals, but two are shared by at least four groups: Mir-4763 is

*Canis familiaris*

**Mir name** 

**3918/50**

large deletion.

*Bos taurus* 

Possible Role of Target Sequences of miRNA for TSE Sensitivity in Bovidae and Cervidae 529

Fig. 5. Alignment of targets for Mirs in the proximal region of stop codon in mammals. Grey background: stop codon, green background: conservation in ≥ 75 % sequences; yellow

> *Mus musculus*

**3102** X X X X X X

**519** X X X X X X **3432** X X X

*Mustela putorius*

*Ochotona princeps*

*Odocoilus hemionus*

*Ovis aries* 

*Sus scrofa* 

*familiaris*, *M. putorius* or *S. scrofa*). Mir-3432 and Mir-140 have targets shared by *C. familiaris* and several species, including *H. sapiens*, *M. musculus* or *S. scrofa*. In summary, the widest distribution of a given target concerns Mir-3918/509, but not any target is common to all mammals excluding the Bovidae-Cervidae. If the region encompassing targets to Mir-3918/509-519 is taken as a whole, it should be noted that most mammals (except murine) share this region. Otherwise, no target was found to be shared by *F. catus* and non-Felidae mammals, or by *E. caballus* and non-Equidae mammals. In contrast, there are several targets shared by only Bovidae-Cervidae.

Fig. 3. Profile of site change rates within a 600 pb long segment downstream the stop codon. The numbers in abscissa correspond to the hypothetical ancestral sequence for Eutherians. The violet thick lines correspond to the position of putative targets for Mir-509 group and Mir-369 group. The green thick line corresponds to the position of the large deletion in Bovidae-Cervidae.

Fig. 4. Position of the putative targets for Mirs in the proximal region. The arrows indicate stop codon position. The grey boxes correspond to targets present in human.

*familiaris*, *M. putorius* or *S. scrofa*). Mir-3432 and Mir-140 have targets shared by *C. familiaris* and several species, including *H. sapiens*, *M. musculus* or *S. scrofa*. In summary, the widest distribution of a given target concerns Mir-3918/509, but not any target is common to all mammals excluding the Bovidae-Cervidae. If the region encompassing targets to Mir-3918/509-519 is taken as a whole, it should be noted that most mammals (except murine) share this region. Otherwise, no target was found to be shared by *F. catus* and non-Felidae mammals, or by *E. caballus* and non-Equidae mammals. In contrast, there are several targets

Fig. 3. Profile of site change rates within a 600 pb long segment downstream the stop codon. The numbers in abscissa correspond to the hypothetical ancestral sequence for Eutherians. The violet thick lines correspond to the position of putative targets for Mir-509 group and Mir-369 group. The green thick line corresponds to the position of the large deletion in

50 150 250 350 450 550 650

Fig. 4. Position of the putative targets for Mirs in the proximal region. The arrows indicate

stop codon position. The grey boxes correspond to targets present in human.

shared by only Bovidae-Cervidae.

Bovidae-Cervidae.

0 0.2 0.4 0.6 0.8 1 1.2

Fig. 5. Alignment of targets for Mirs in the proximal region of stop codon in mammals. Grey background: stop codon, green background: conservation in ≥ 75 % sequences; yellow background: conservation in ≥ 50 % sequences.


Table 2. Micro RNAs targets common to several species between the stop codon and the large deletion.

Within the sequence corresponding to the large deletion region, several targets (Mir-4763, 1587, 744, 15b, and 369) are overlapping (fig. 6-7). In two couples of targets (Mir 569/155 and Mir-3918/509), the distributions in mammals are identical and each couple was considered as single. The remaining overlapping targets have their own distribution within mammals, and they were considered separately (table 3). It appears that the number of putative targets common to at least 2 species varies among the different mammals. There is none in *M. musculus*, but four in *E. caballus* and even six in *F. catus*. There is no target common to all orders of mammals, but two are shared by at least four groups: Mir-4763 is

Analysis of 3'UTR of Prnp Gene in Mammals:

*Equus caballus*

**1587 X X** 

**3613 X X X** 

**4723 X X** 

*familiaris*

**Mir name** *Canis* 

**4. Discussion** 

Possible Role of Target Sequences of miRNA for TSE Sensitivity in Bovidae and Cervidae 531

**4763 X X X X** 

**15B/369 X X X X** 

**603 X X** 

**744 X X X** 

Table 3. Micro RNAs targets common to several species in the large deletion region

SINE/BovA

TcMar-Mariner

Table 4. Putative targets for Mirs, common to *B. taurus*, *O. aries* and *O. hemionus*.

The most interesting feature in the comparative analysis of 3'UTR region of mammalian *Prnp* genes is the occurrence of 2 events common to Bovidae and Cervidae, i.e., a 177 pblong deletion and 3 insertions of transposable elements. These insertions were previously described in sheep by Lee et al. (1998), but not in Cervidae. To simplify the discussion, the large deletion common to Bovidae-Cervidae will be abbreviated as LDCBC in the following. The mammalian distribution of insertions and LDCBC fit with the phylogeny within Artiodactyla, since Bovidae and Cervidae were shown to be sister families by Kuznetsova et al. (2005) from 12S and 16S rRNA analyses. In contrast, the more basally genera *Lama* (Camelidae), *Sus* (Suidae) and *Tursiops* (Delphinidae) within Cetartiodactyla do not possess the LDCBC and the 3 insertions. Unfortunately, none studied species harboured either the LDCBC or the 3 insertions. The sequencing of *Giraffa camelopardalis* could bring an

**Transposable elements miRNAs** 

SINE/RTE-BovB miR-4311

*Felix catus*  *Lama pacos* 

miR-137

miR-596

miR-2277-3p

miR-2284d miR-701

miR-3143

miR-2294

*Mustela putorius*

*Myodes glareolus* 

*Sus scrofa* 

*Homo sapiens* 

common to *E. caballus*, *F. catus*, *L. pacos* and *S. scrofa*, and Mir-15B/369 to *E. caballus*, *H. sapiens*, *F. catus*, and *S. scrofa*.

Fig. 6. Position of putative targets within the large deletion common to Bovidae-Cervidae. The grey box corresponds to the target present in human.

Fig. 7. Alignment of overlapping target sequences in the region of the large deletion common to Bovidae-Cervidae. Green background: conservation in ≥ 75 % sequences; yellow background: conservation in ≥ 50 % sequences.

Bovidae and Cervidae are the only species in which transposable elements are present in the transcript of *Prnp.* The poly-adenylation site common with other mammals is situated downstream from Mariner element. However, there is an extra potential poly-adenylation site, downstream to the SINE/BovA. Few targets for Mirs are present in these regions and common to *B. taurus*, *O. aries* and *O. hemionus* (Table 4).


Table 3. Micro RNAs targets common to several species in the large deletion region


Table 4. Putative targets for Mirs, common to *B. taurus*, *O. aries* and *O. hemionus*.
