Introduction

Allantoicase (EC 3.5.3.4) belongs to the purine degradation pathway and catalyzes the hydrolysis of allantoic acid to ureidoglycollate and urea. Uric acid is the first common intermediate of adenine, guanine, and inosine nucleotide catabolism, while its further degradation varies from species to species (Keilin 1959). Uric acid is degraded to allantoin by uricase (urate oxidase; EC 1.7.3.3), the first step in the uricolytic pathway. This does not occur in humans, birds, and reptiles, where uricase is absent (Friedman et al. 1985), but it happens in most mammals that excrete allantoin in the urine as the end product of purine metabolism. Purine catabolism proceeds further in lower vertebrates (i.e., amphibians and fishes) since allantoin is degraded by allantoinase (EC 3.5.2.5) to allantoic acid, which in turn is hydrolyzed to urea and ureidoglycollate by allantoicase; ureidoglycollate is degraded by ureidoglycollate lyase (EC 4.3.2.3) to urea and glyoxylate. Urea is then degraded to CO2 and NH3 by the urease (EC 3.5.1.5) present in intestinal flora of vertebrates. Since the purine degradation is much less complete in higher animals, it appears that certain enzymes have been lost during evolution (Hayashi et al. 2000).

Allantoicase most likely appeared very early during evolution (Keilin 1959). In bacteria and fungi, the characterization of the enzymatic activity was complemented by the molecular cloning of the allantoicase gene (Lee and Moss 1990; Lee et al. 1990; Stover et al. 2000; Finan et al. 2001). In contrast, in the vegetal kingdom, allantoic acid degradation is still debated, but it is probably due to the activity of allantoate amidohydrolase (EC 3.5.3.9) (Munoz et al. 2001), although the presence of allantoicase was also proposed (Shelp and Ireland 1985). The allantoicase activity is widespread in the animal kingdom (Hayashi et al. 2000), but the protein sequences were unknown until recently, when the allantoicase cDNA was cloned in the anuran Xenopus laevis (Vigetti et al. 2000a). The sequence similarity with microrganism allantoicases has shown that, in animals, the allantoicase activity is due to an enzyme of the allantoicase family.

Allantoicase activity appears to have been lost during vertebrate evolution (Keilin 1959), as it is present in most fish and amphibians but not in amniotes (Urich 1994). The absence of allantoicase activity in mammals, birds, and reptiles, however, does not imply the absence of the allantoicase gene or of its transcripts. Recently, in fact, mRNAs coding for allantoicase have been found in mouse (Vigetti et al. 2001) and in human (Vigetti et al. 2000b, 2002a). While the human transcripts seemed to derive from an improper start site and resulted incomplete or were mutated at the initial ATG codon, the mouse transcript appeared to be capable of coding for a 46-kDa polypeptide sharing 54% identity with X. laevis allantoicase. To determine whether mouse allantoicase cDNA encodes a functional protein, we produced the recombinant enzyme in E. coli and showed that it was competent in catalysis; however, the affinity for allantoate of this recombinant enzyme is 50-fold lower than that of the X. laevis allantoicase (Vigetti et al. 2002b). In mammals, however, the absence of uricolytic pathway enzymes is well documented and generally accepted (Fujiwara and Noguchi 1995). In humans, this lack is exploited in the diagnosis of several diseases such as gastrointestinal motility disorders (Ruemmele et al. 1997) and Helicobacter pylori infections (Nakamura 2001) by the so-called urea breath test.

The total absence of allantoicase activity in mammals together with the presence of an apparently well-conserved and functioning gene raises interesting questions on the meaning of allantoicase gene modifications during vertebrate evolution.

Recently, new statistical methods have been developed to infer adaptive evolution at the molecular level (for a review see Yang and Bielawsky 2000). Although varying in the different assumptions about the substitution rates used in the model, all the methods available rely on the nonsynonymous/synonymous ratio to detect neutrality of the observed nucleotide substitutions. The ratio ω = d N/d S, thus, represents an indicator of whether a protein coding DNA sequence has undergone selective pressure or the observed polymorphism is due to random fixation of neutral mutations. An ω value >1 is taken as evidence of positive selection, an ω value <1 is taken as evidence of purifying selection, and an ω value = 1 is taken as evidence of neutral selection (Yang and Nielsen 2002).

In this paper, we estimate ω from comparison of sequences of the allantoicase gene from 10 species, spanning the phylogenetic tree from Ciona intestinalis to Homo sapiens, to address the issue of the apparent contradiction between the absence of enzymatic activity and the presence of potentially active genes in higher vertebrates.

Materials and Methods

Animals

Xenopus tropicalis, Xenopus mulleri, Bufo marinus, and Litoria infrafrenata were purchased from a local dealer (Rettili, Varese, Italy). Animals were quickly anesthetized by hypothermia and killed by head concussion. Amphibian liver was quickly extracted and conserved at −80°C until use. For Rattus norvegicus allantoicase sequence, testes were used instead of liver since allantoicase transcript was expressed mainly in this tissue (Vigetti et al. 2001).

RT-PCR

Total RNA was extracted from the frozen tissues using TRIzol Reagent (Invitrogen) following the manufacturer instructions. Two micrograms of total RNA was retrotranscribed using 200 U of MMLV reverse transcriptase (Life Technologies) and dT(16) primer at a concentration of 500 µg/ml. The reaction was carried out in 50 mM Tris–HCl, pH 8.3, 75 mM KC1, 3 mM MgCl2, 10 mM DTT, and 500 µM dNTP mix at 42°C for 50 min. PCR amplification of the cDNA molecule was performed using the first-strand cDNA synthesis mix, 25 pmol of degenerated primers (5′-GGI TGG ATG GIT GGG AGA C-3′) and (5′-CAG ICT TCT TGC IGT TTC CC-3′), 200 µM dNTPs, and 1 U of Advantage 2 DNA polymerase mix (Clontech) in its own buffer. Reaction mixtures were subjected to 8 touchdown cycles with annealing temperature from 58 to 54°C and subsequently 25 cycles using the following parameters: denaturation at 94°C for 30 s, annealing at 54°C for 30 s, an elongation at 72°C for 40 s. Ciona intestinalis allantoicase sequence derived from PCR screening (using the same conditions described above) of a C. intestinalis tadpole cDNA library kindly gifted by Dr. Margherita Branno of the Stazione Zoologica, Naples, Italy. To minimize errors, a high fidelity DNA polymerase enzyme was used in all PCRs and both strands of the obtained cDNA fragments were sequenced.

Cloning and Sequencing

PCR products were resolved by agarose gel electrophoresis and visualized by ethidium bromide staining. The amplified bands were extracted from the gel using the MinEluted Gel Extraction Kit (Qiagen) and cloned in pGEM-T Easy vector (Promega) following the commercial kit instructions and the standard laboratory procedures (Sambrook et al. 1989). DNA sequencing was carried out by BMR (Padova, Italy). The obtained sequences have been deposited in the GenBank database under accession numbers AF480058 (Xenopus tropicalis) AF480059 (Bufo marino), AF480060 (Ciona intestinalis), AF480061 (Litoria infrafrenata), AF480062 (Rattus norvegicus), and AF480063 (Xenopus mulleri). Other sequences used in this work were AF153230 (Xenopus laevis), AF278712 (Mus musculus), and AF395820 (Homo sapiens) that were previously characterized in our lab.

Japanese pufferfish (Fugu rubripes) allantoicase sequences were obtained performing BLAST searches against the Fugu v3.0 (genome; sequence scaffold 880) and Fugu v3.0 (transcript; sequence FRUT00000146535) databases (Aparicio et al. 2002; http://genome.jgi-psf.org/ ).

Sequence Analysis

The DNA sequences have been aligned by ClustalX and clipped to the same length of 432 nucleotides, corresponding to the 144 amino acids of the N-term, to reduce the fraction of noninformative sites due to alignment ambiguities in the following region and to restrict the analyses to the first two functional conserved domains of the protein. The phylogenetic reconstruction has been carried out by PHYLIP; sequences were bootstrapped 1000 times, distance matrices were built by DNADIST, and clustering was done by the neighbor-joining procedure of NEIGHBOR. To estimate the nonsynonymous (d N) and synonymous (d S) substitution rates for all the branches of the tree, the program Codeml of the PAML package (Yang and Nielsen 2000) was used. Pairwise ω (d N/d S) values between all the species studied were estimated by the program YN00 in the same software package. According to the analysis performed by the authors, this maximum-likelihood method will be less biased when the sequences studied may have transition/transversion rate bias or base/codon frequency bias, especially when long sequences (>100 bp) are concerned. Another advantage of this method is to allow for multiple comparisons between sequences taking into account their phylogenetic relationship as determined in other ways or already known.

Results

Sequence Alignment

To obtain a set of comparable allantoicase sequences from different animal species, we designed a pair of degenerate primers containing inosine for RT-PCR reactions. To design the primers, we have taken into consideration the sequences of the three known vertebrate allantoicases (i.e., those of Homo sapiens, Mus musculus, and Xenopus laevis; Fig. 1) and we have chosen two very well-conserved regions approximately 585 nucleotides apart. RT-PCR experiments performed on total liver RNA from Xenopus tropicalis, Xenopus mulleri, Bufo marino, and Litoria infrafrenata, and from total testis RNA of Rattus norvegicus produced cDNA fragments of the expected size (results not shown). The corresponding portion of Ciona intestinalis allantoicase sequence was obtained by screening a cDNA library using the same degenerate primers.

Figure 1
figure 1

Schematic representation of the allantoicase protein structure. The conserved amino acid residues are boxed and the corresponding cDNA sequences of the three vertebrate allantoicases are aligned below. Arrows indicate the positions of the degenerated primers used in this work. Asterisks indicate identical nucleotides and the numbers beside the sequences indicate their position from the ATG starting codon. Mm, Mus musculus; Hs, Homo sapiens; Xl, Xenopus laevis.

Full size image

To obtain a comparable allantoicase fragment from fish, we tried the same RT-PCR approach on liver sea bass Dicentrarchus labrax, trout Salmo trutta, catfish Ictalurus melas, goldfish Carassius auratus, and rudd Scardinius erythrophthalmus with no results. Recently, however, since the complete Fugu rubripes genome was published (Aparicio et al. 2002), an F. rubripes genomic sequence with a high similarity to X. laevis allantoicase cDNA became available. Performing a BLAST search against the F. rubripes transcript database, we have identified the sequence FRUT00000146535, which showed 47.5% identity with that of X. laevis and, hence, is likely to represent the pufferfish allantoicase mRNA. Therefore, we have included in our dataset the portion of this F. rubripes allantoicase sequence corresponding to the other fragments isolated by the degenerate RT-PCR approach (more precisely, from nucleotide positions 44 to 625 of the sequence FRUT00000146535).

We have designed other degenerate primers (GAG TTT GGI AAA TGG ITG G and GCI GTT TCC CAI CCI TCT), based upon the F. rubripes allantoicase sequence, to obtain allantoicase sequences from fish using the same RT-PCR strategy, but with no success.

BLAST searches and ClustalW alignments performed on the obtained sequences showed a high similarity to other known allantoicases present in public databases (data not shown). The obtained cDNA sequences have been aligned with the corresponding portion of previously characterized allantoicase cDNAs (Vigetti et al. 2000a, 2001, 2002a). We did not detect any spurious termination signal, thus lending robustness to the DNA sequencing procedure. We virtually translated all cDNAs for the coding region corresponding to the N-term lobe of the protein and we aligned the putative amino acid sequences again (Fig. 2). The most evident feature of the alignment is the presence, in the murine putative protein, of a double glutamic residue (positions 53–54 of the alignment) that is absent, either partially (H. sapiens and R. norvegicus) or totally, in the other proteins. Apart from that, the amino acid sequences show a remarkable homology, especially in the regions corresponding to the motifs DGWETRR and PDGG.

Figure 2
figure 2

Alignment of the in silico translations of the allantoicase cDNAs isolated for this work by RT-PCR (Bm, Bufo marino; Xm, Xenopus mulleri; Xt, Xenopus tropicalis; Ci, Ciona intestinalis; Li, Litoria infrafrenata; Rn, Rattus norvegicus) or previously characterized (Xl, Xenopus laevis; Mm, Mus musculus; Hs, Homo sapiens; Fr, Fugu rupribes) using ClustalX program. Dashes indicate gaps introduced for optimal alignment; an asterisk indicates identical or conserved residues in all sequences; a colon conserved substitutions; and a period semi-conserved substitutions.

Full size image

Phylogenetic Analysis

For the phylogenetic analysis, we used only the approx. 432-bp-long nucleotide sequences (the length depending on the species considered) coding for the N-terminal lobe of the allantoicase protein (corresponding to the first 144 amino acids shown in Fig. 2). We started by carrying out an analysis by PHYLIP (Felsenstein 1985) by means of DNADIST (Kimura two-parameter) and NEIGHBOR (UPGMA) on the original dataset bootstrapped 1000 times. The phylogenetic reconstruction yielded the tree shown in Fig. 3. The clustering of Xenopus spp. (i.e., X. tropicalis, X. mulleri, and X. laevis) apart from the other amphibian and tunicate species (i.e., B. marino, L. infrafrenata, and C. intestinalis), from F. rubripes, and from the mammalian cluster (i.e., H. sapiens, R. norvegicus, and M. musculus) is supported by bootstrap values of 100%.

Figure 3
figure 3

UPGMA tree of the species used in this study, based upon the sequence comparison of 432 nucleotides, corresponding to the 144 amino acids of the N-terminal of the allantoicase gene. The original dataset was bootstrapped 1000 times and each node is supported by a bootstrap value of 100% (not shown). Along the branches the d N/d S ratios are reported as obtained using the Codeml program (n.d., not determined). Four clusters are defined (A, B, C, and D) and the numbers in boldface at the nodes indicate putative ancestor sequences as discussed in the text.

Full size image

The relative number of nonsynonymous differences (d N) compared with the number of synonymous differences (d S) was determined for each pairwise species comparison. The d N/d S ratio, called ω, is often taken as a measure of the degree of selection that a coding sequence is experimenting (Messier and Stewart 1997). The values of d N were then plotted against the d S values (Fig. 4). In the plot, all the pairwise comparisons display an ω < 1, that is, considered evidence of a certain amount of purifying selection acting on the region. Moreover, if we define as A (mammals), B (F. rubripes), C (Xenopus spp.), and D (B. marino, L. infrafrenata, and C. intestinalis) the four major clusters of the phylogenetic tree (Fig. 3) and take into account the possible comparison among these clusters, a pattern appears (Fig. 4). The pairwise comparisons between B and all other organisms (A vs B, C, and D in Fig. 4) display the same d N values; the same also applies to the A vs C and A vs D (A vs C and D in Fig. 4) and C vs D (C vs D in Fig. 4) comparisons. It seems therefore that what makes the difference between groups of related species is d S and d N, thus lending support to the hypothesis of selection still acting on the gene.

Figure 4
figure 4

Plot of the synonymous (d S) and nonsynonymous (d N) substitution rates in the region of the allantoicase sequences in the species used in this study and calculated by the maximum-likelihood method of Yang and Nielsen (2000). Filled diamonds represent pairwise comparisons between species of cluster B and species of clusters A, C, and D; filled squares represent pairwise comparisons between species of cluster A and species of clusters C and D; and filled triangles represent pairwise comparisons between species of cluster C and species of cluster D. The diagonal line indicates d S = d N. The nonsynonymous rates are always lower than the synonymous rates for all the pairwise comparisons.

Full size image

We have also estimated, by the maximum-likelihood method, the ω = d N/d S performing a simultaneous comparison of the sequences by the free-ratio model (one ratio for each branch of the tree) taking into account their phylogenetic relationship as revealed by the former analysis, ω values are reported along the branches of the tree in Fig. 3 and range from 0.0000 (C. intestinalis vs others) to 0.611 (ancestral node 14 to ancestral node 15), with the exception of the 4.626 value found for the branch leading to B. marino.

The alignment of the amino acid sequences, obtained in silico from both real DNA sequences and the DNA sequences reconstructed by maximum-likelihood corresponding to ancestral nodes of the phylogenetic tree is shown in Fig. 5. The reconstructed ancestral sequence corresponding to node 14 was taken as the sequence of reference to better highlight differences among the clusters. Conservation of the amino acid residues is evident for the N-terminal region (DGWETRR) and for the region spanning amino acids 115 and 130 and comprising the PDGGI motif (Vigetti et al. 2001). The intermediate regions display a higher variability.

Figure 5
figure 5

Alignment of the putative allantoicase sequences of the four clusters identified. The cDNA fragments analyzed were virtually translated and aligned using GeneDoc program (Nicholas and Nicholas 1997). “Node” sequences represent the ancestral sequences, as reconstructed by maximum likelihood, constituting the nodes in Fig. 3.

Full size image

Several differences are evident when comparing the four groups revealed by phylogenetic analysis. First, the lengths are different, the C cluster sequences and the F. rubripes one are 142 aa long, H. sapiens and R. norvegicus are 143 aa long, M. musculus is 144 aa long, and the sequences of the D cluster are 141 aa long. Second, many similarities are present between the F. rubripes and the mammalian sequences.

The reconstruction by maximum likelihood of the putative progenitor sequences to those analyzed also allows us to pinpoint other mutations of interest. In fact, most important in our view are those differences between the “functional” and the “nonfunctional” sequences. In particular, note the substitutions occurring at amino acid positions 23 (proline to glutamine), 34 (arginine to serine), 50 (cysteine to asparagine), 52 (lysine to serine), 92 (lysine to proline), 113 (lysine to glutamine or phenilalanine), 133 (phenilalanine to leucine), and 138 (isoleucine to threonine) in the comparison between clusters C and D on one side and clusters A and B on the other. Putative phosphorylation sites, as indicated by serine, threonine, and tyrosine, appear in mammals at positions 34 and 138, but one is lost at position 33. Finally, differences peculiar to H. sapiens are present at position 19, where glutamic acid becomes arginine, at position 100, where tyrosine becomes asparagine, and at position 102, where glutamic acid is substituted by alanine.

Discussion

During vertebrate evolution, the purine catabolism pathway was differently truncated, with the possible advantage of preserving water. Consequently, the activity of the enzymes in this pathway as well as its end products varies from species to species (Keilin 1959).

As far as concerns allantoicase, one of the enzymes in this pathway, the only vertebrates that maintain its activity are some fish and all of the amphibians (Urich 1994). And it is from the amphibian Xenopus laevis that the first vertebrate allantoicase cDNA was cloned (Vigetti et al. 2000a). Surprisingly, however, the allantoicase transcript is detectable also in several mammalian tissues (i.e., mouse testis [Vigetti et al. 2001], rat testis [this work], bovine testis [GenBank accession No. BF599485], and human kidney and testis [Vigetti et al. 2000b]), although these tissues do not seem to possess any allantoicase activity. The presence of deleterious mutations in mRNAs is one of the possible mechanisms of gene silencing as in the case of primate uricase, another enzyme in this catabolic pathway (Wu et al. 1989; Yeldandi et al. 1990a, b; Wu et al. 1992; Oda et al. 2002). However, this does not seem to be the case for mouse allantoicase transcript since it can produce an enzyme competent in catalysis, although with a reduced affinity (Vigetti et al. 2002b). The biological significance of such mRNAs, apparently coding for unnecessary functions, is still unknown.

Are we witnessing the formation of a relict gene or the change of a function? In this paper, we have made an attempt to understand the reasons for the presence of the allantoicase gene in higher vertebrates and we have resorted to bioinformatic techniques to infer the degree of selection that could act on the coding sequences of this gene.

To perform the analyses, we have considered a comparable length from 10 allantoicase sequences, all from Chordata. Those from X. laevis, mouse, and human were available from previous work (Vigetti et al. 2000a, b, 2001). We have obtained those of Xenopus tropicalis, Xenopus mulleri, Bufo marino, Litoria infrafrenata, and Rattus norvegicus were obtained by RT-PCR, and that from Ciona intestinalis by a cDNA library screening. Instead, we have inferred the fish sequence from Fugu rubripes databases (Aparicio et al. 2002). In fact, we have always failed to amplify fish allantoicase sequences by means of RT-PCR. This can be ascribed to two factors: first, although amino acid sequences are fairly similar between F. rubripes and the other species analyzed, this does not hold true at the nucleotide level; second, the fish allantoicase gene could be regulated at the transcriptional level and be expressed only under particular conditions, such as environmental constraints, as demonstrated for other urea producing enzymes (Randall et al. 1989). Although this transcript is only hypothetical, we think it is a good prediction. In fact, alignments revealed that the F. rubripes allantoicase gene maintains the same general exon/intron organization as the corresponding mammals gene (i.e., 11 exons and 10 introns), but for an additional small exon localized at the 5′ end of the gene (data not shown). As previously reported for a number of other genes, we have noted a dramatical decrease in the intron lengths resulting in a overall gene size of about 3.5 kb, vs the 20 kb of the previously characterized mammal allantoicase gene (Vigetti et al. 2002a).

The number and the “quality” of the species sampled to build the phylogenetic tree can introduce bias into the reconstruction itself, either because recent speciation events are not represented or because of sympatric events leading to the creation of rare species (Nee et al. 1994; Berlocher 1998). However, in our case the general correspondence of the phylogenetic tree built on a region of a single gene (Fig. 3) with the accepted phylogeny indicates that the genes sampled are indeed orthologous, therefore suited for subsequent analyses on substitution rates. Noteworthily, as far as concerns allantoicase sequence, F. rubripes proved to be closer to the mammals than to lower vertebrates. However, the allantoicase activity distribution in the compositum group of fish is still undetermined (Urich 1994).

The genomic era has spawned a lot of analytical methods based on sequence analysis (alignment, comparison, phylogenetic reconstruction), some of which have originated the field of functional bioinformatics. In particular, several functional bioinformatic techniques have been developed to understand the dynamics of evolution at the molecular level. The estimation of synonymous and nonsynonymous substitution rates is particularly important in this aspect (Kimura 1983; Ohta 1995). We have found that for all branches of the tree the values of ω = d N/d S are always less than 1 (but for B. marino), meaning that the regions analyzed are under some kind of functional constraint. A very high ω value (4.626) was estimated for the branch leading to B. marino. Usually, this is taken as evidence for positive selection, but possible stochastic effects, due to the relatively short length of the sequences analyzed, could provide an explanation for this ω value (Holbrook et al. 2000). A large number of the cases of ω > 1 detected so far regard genes involved in a response to rapidly evolving conditions. In a large-scale search for genes on which positive selection operates, Endo et al. (1996) found that the majority of such genes were surface antigens. One of the most studied genes exhibiting ω > 1 is the abalone lysin, probably answering to rapidly evolving receptors carried on the egg surface (Vacquier et al. 1997; Yang and Bielawsky 2000). However, this does not seem to be this case for B. marino allantoicase and we do not think that positive selection is acting on this gene.

A closer view of the plot in Fig. 4 reveals a certain degree of clustering among amphibians and mammals. Other factors discriminate amphibians from mammals, such as the overall variability of the amino acid sequences, higher in mammals, and the ω ratio, again, higher in mammals. All these features concur to indicate that the functional constraint must be lower in mammals.

In human, alternative starts of transcription (Vigetti et al. 2002a), mutations at the starting codon, and peculiar amino acid substitution seem to tell a different story about the allantoicase gene. For example, we have noted that many residues which are target of phosphorylation activities have been mutated in the evolution from amphibians to mammals, thus suggesting that regulation of the activity of the protein could have changed. In fact, several examples indicate that a few amino acid substitutions can completely change the substrate specificity for a given enzyme (You and Arnold 1996; Knecht et al. 2002).

Another enzyme in the uricolytic pathway, ureidoglycollate lyase, catalyzing the degradation of ureidoglycolate to glyoxylate and urea, originally thought not to be present in mammals, was found in rat liver (Fujiwara and Noguchi 1995). Also in this case, the enzymatic properties were different, displaying a lower affinity for the substrate than the enzyme found in fish, and a putative new activity was proposed by the authors. It does not seem unreasonable, therefore, that a similar process is taking place in mammals for the allantoicase. We cannot rule out the possibility that the gene is maintained in mammals by a background selection acting on the chromosomal regions where allantoicase maps (i.e., on human chromosome 2p25 and on mouse chromosome 12 at 4.4 cR from the molecular marker D12Mit107 [Vigetti et al. 2002a]). There are at least two genes in these syntenic regions (i.e., myelin transcription factor 1-like and thyroid peroxidase) that are both closely linked to allantoicase and perform an important function. We can safely assume that both genes undergo a certain degree of selection. This means that the low nonsynonymous rates observed in our work could be explained by genetic hitchhiking associated with a background selection (Johnson 1999; Kim and Stephan 2000). Moreover, the presence of a transcripted gene could be justified by the recent observation (Oliver et al. 2002; Spellman and Rubin 2002) that groups of adjacent and coregulated genes that are not otherwise functionally related can be identified by expression profiling in Drosophila. This could suggest the involvement of chromatin domains in the control of genes within a genomic neighborhood.

Conclusions

In this work, we have shown that a certain degree of purifying selection is acting on the allantoicase sequences of “anamnia” (i.e., C. intestinalis, F. rubripes, and amphibians), where the enzyme is known to have a catabolic role, as well as on amniotes (i.e., reptiles, birds, and mammals), where the protein activity is known to be lost. Hence, we discuss the possibility that this apparently redundant gene could be the source for a new and still unknown function.