Phylogenetic analyses with proximal promoter from skeletal and cardiac calsequestrins

Abstract

Calsequestrin (CASQ) is the main calcium binding protein of the sarcoplasmic reticulum. In mammalian muscles, it exists as a skeletal isoform (CASQ1) found in fast- and slow-twitch skeletal muscles, and a cardiac isoform (CASQ2) expressed in the heart and slow-twitch muscles. Evolutionary biology studies have a great impact to problems in medicine because evolutionary thinking do not displace other approaches to medical science, such as molecular medicine and cell and developmental biology, but that evolutionary insights can combine with and complement established approaches. Thus, the authors have followed evolutionary biology approach by studying Casq promoters evolution. Phylogenetic analyses were performed, using the minimum promoter of CASQ1 and CASQ2. Emerged topological discordances from the comparison between CASQ promoter based phylogenetic tree and amino acid based phylogenetic tree showed similar, but not identical evolutionary pathways.

Headnote

Abstract:

Calsequestrin (CASQ) is the main calcium binding protein of the sarcoplasmic reticulum. In mammalian muscles, it exists as a skeletal isoform (CASQ1) found in fast- and slow-twitch skeletal muscles and a cardiac isoform (CASQ2) expressed in the heart and slow-twitch muscles. Evolutionary biology studies have a great impact to problems in medicine because evolutionary thinking do not displace other approaches to medical science, such as molecular medicine and cell and developmental biology, but that evolutionary insights can combine with and complement established approaches. Thus, we followed evolutionary biology approach by studing Casq promoters evolution. Phylogenetic analyses were performed using the minimum promoter of CASQ1 and CASQ2. Emerged topological discordances from the comparison between CASQ promoter based phylogenetic tree and amino acid based phylogenetic tree showed similar, but not identical evolutionary pathways.

Keywords: Sarcoplasmic reticulum; Phylogeny; Molecular clock.

1. Introduction

Calsequestrin is the most abundant calcium (Ca2+) binding protein in the sarcoplasmic reticulum (SR) of cardiac and skeletal muscles with moderate affinity (Kd ~ 1 mM) and high capacity (60-80 mol 21 t 2+ -. . . Ca /mol calsequestrin) for Ca [1]. Localization in a sub-compartment of the SR (terminal cistemae, TC), is accomplished by anchoring to the membrane through interaction with trans-membrane proteins such as triadin (Trd) and junctin (Jet). CASQ polymerizes in response to rising Ca2+ concentrations in the lumen of SR, to form dimers and polymers. Post-traslational modifications such as phosphorylation and glycosylation are important for CASQ localization and function [1]. Beyond its i i * f* 2+ i # capability of Ca buffering, CASQ has additional i i 21 roles in muscle contraction such as regulator of Ca release through the channel Ryanodine Receptor (RyR) either directly or through interactions via triadin/junctin [2]. Finally CASQ is coordinately expressed with mayor SR and contractile protein during development and is important for the structural organization of the SR in adult [3]. Two CASQ isoforms have been described in mammals, each one encoded by a single gene. The "skeletal" isoform (encoded by Casql gene) is found in the fast-twitch skeletal muscle (like soleus), whereas the "cardiac" isoform (encoded by Casq2 gene) is expressed in the heart and in the slow-twitch skeletal muscle (like EDL). Cardiac muscle expresses exclusively the Casq2 gene. Both protein isoforms show considerable sequence and structure similarity. For example share a high nucleotide and amino acid homology, 84 and 80%, respectively in humans [4]. At this time there are few studies that point to differences in the physiological role of the different CASQ isoforms [5] thus their roles is sometimes considered equivalent.

CASQ1 interacts with many SR proteins and is proposed that these interactions and the molecular ratio between CASQ1 and Trd/Jct play a role in SR biogenesis. A quantitative link between CASQ1 and Trd is described in different mouse knock out (KO) models and also in physio-pathological conditions: CASQ 1-null mice [3] show a decrease in Trd content; pan Trd KO [6] show decrease in CASQ1 expression, but if the equilibrium between CASQ1 and Trd is due to specific gene regulation or is an effect of posttranslational mechanisms is not known. Another protein whose content is strictly coordinated with CASQ1 is SERCA1 [7]. Also opposite regulation of CASQ1 and CASQ2 is reported in different models and processes: is well known that CASQ1 and CASQ2 are co-expressed during early myotube formation in vitro, but as myogenesis progresses, CASQ2 is replaced by CASQl. In mouse and rat embryos, CASQ2 is detected in fetal heart and skeletal muscles, whereas CASQ1 transcripts are only found in fetal skeletal muscles [8]. Similarly, the expression of rabbit CASQ2 in the fast-twitch skeletal muscle declines progressively with development, whereas the expression of CASQ1 increases and totally replaces CASQ2 in adults. In contrast, the cardiac muscle expresses exclusively CASQ2 at all stages [9]. Recent work shows that KO of Sixl and Six4 homeoproteins alterates expression of CASQ 1/2 and SERCA 1 and leads to difficult in differentiation of fast skeletal muscles [10]. In models of chronic low frequence stimulation, CASQ1 decreases and CASQ2 increases in fast muscle [11]. Also denervation causes a considerable reduction in Ca - ATPase (SERCA) and calsequestrin in EDL, making it resemble the control of slow muscle like soleus. A recent work describes different effects on CASQ1 and SERCA 1 expression due to different trainings [12].

It appears that both CASQ1 and CASQ2 are upor down-regulated in many different models, pathologies and during development implying that the two proteins are not interchangeable or are submitted to quite different regulators, but the mechanisms that operate at genic and protein level are still obscure. Evolutionary biology studies have a great impact to problems in medicine because evolutionary thinking do not displace other approaches to medical science, such as molecular medicine and cell and developmental biology, but that evolutionary insights can combine with and complement established approaches to reduce suffering and save lives [13]. Thus, we followed evolutionary biology approach by studing Casq promoters evolution, in order to resolve such biomedical problems.

2. Material and Methods

2.1 Phylogenetic Analyses

CASQ amino acid and coding cDNA sequences were found in GenBank (www.ncbi.nlm.nih.gov/genbank/) and ENSEMBL (http://www.ensembl.org) databases, while all Casq promoter region sequences were retrieved from ElDorado annotation database of Genomatix Software Suite [14]. All respective sequences were aligned using T-Coffee multiple sequence alignment software package [15]. To find the best-fit model of molecular evolution we used ProtTest3 [16] on CASQ amino acid sequence alignment and jModelTest 0.1.1 [17] on Casq coding and promoter region sequence alignments. Phylogenetic analyses were conducted using the maximum likelihood (ML) and Bayesian inference (BI) methods. The ML (including bootstraps) of phylogeny was conducted on promoter region sequence alignment using PhyML [18]. Non parametric bootstrap analysis was performed using PhyML software (1000 replicates). BI of phylogeny was performed on all sequence alignments using BEAST vl.7.0 software package [19] with a tree prior of Yule spéciation process under a strict clock model. Four independent runs, each one with four simultaneous Markov Chain Monte Carlo (MCMC) chains, were performed for 10,000,000 generations sampled every 1000 generations. TreeAnnotator vl.7.0 software was used to identify the posterior probabilities of the nodes in the target tree (maximum clade credibility tree) generated by BEAST vl.7.0. after burning 500 trees. FigTree v 1.3.1 software was used to display the annotated phylogenetic trees. We used BEAST vl.7.0. also for clock rate estimation. Trace files generated by Bayesian MCMC runs in BEAST vl.7.0. were analyzed using Tracer 1.5.

2.2 Statistical Analyses

In three parameter data statistical comparison, one-way ANOVA was performed (Tukey, Bonferroni and Sceffè tests) using Origin 8.6. Student's t test was used for comparisons between two parameter data like substitution rates of mammals and all analyzed species promoter sequences. Statistical significance was set at P < 0.05.

3. Results and Discussion

3.1. Molecular Clock Tests

The molecular clock has become an indispensable tool within evolutionary biology, enabling independent timescales to be placed on evolutionary events. Despite these valuable contributions, date estimates derived from molecular data have not been without controversy. In particular, when molecular clocks have been employed to estimate the timing of recent events already tentatively dated on the basis of paleontological, archaeological or biogeographic sources, con icting dates are frequently obtained. In its most extreme form, the molecular clock hypothesis postulates that homologous stretches of DNA evolve at essentially the same rate along all evolutionary lineages for as long as they maintain their original function [19]. It was shown that the substitution rate of mitochondrial encoded proteins has increased in the order of fishes, amphibians, birds, and mammals and that the rate in mammals is at least six times, probably an order of magnitude, higher than that in fishes [19]. The higher evolutionary rate in birds and mammals than in amphibians and fishes was attributed to relaxation of selective constraints operating on proteins in warmblooded vertebrates and to high mutation rate of bird and mammalian mitochondrial DNAs. Since the assumption of rate constancy is violated even within Mammalians, a truly universal molecular clock that applies to all organisms cannot be assumed to exist. In phylogenetics, the unrooted model of phylogeny and the strict molecular clock model are two extremes of a continuum. Despite their dominance in phylogenetic inference, it is evident that both are biologically unrealistic and that the real evolutionary process lies between these two extremes. Local molecular clocks are another alternative to the global molecular clock. A local molecular clock permits different regions in the tree to have different rates, but within each region the rate must be the same. This new method conveniently allows a comparison of the strict molecular clock against a large array of alternative local molecular clock models [19]. A Likelihood Ratio Test (LRT) was performed in order to know which was the best-fit model to analyze CASQ protein sequence evolution. The LRT (LR=2[lnL(HA) - In L(HO)]) was conducted with n-2 degrees of freedom, where n is the number of considered taxa in the phylogeny. Likelihood scores were estimated using in BEAST vl.7.0 on JTT + G + F matrix which was determined previously by ProtTest3 software application [16] as the best model (-lnL= -5306.53) with a gamma shape value (four rate categories) of 1.057. We chose to use the constant rate birth-death process as it is probably the most popular homogeneous model. A birth-death process is a stochastic process which starts with an initial species. A species gives birth to a new species after exponential (rate X) waiting times and dies after an exponential (rate p) waiting time. A special case of the birth-death process is the Yule model where p=0. This birth-death model is implemented in BEAST vl.7.0. The estimated likelihood scores for both the null (H0: random local molecular clock) and the alternative hypotheses (HA: strict molecular clock) were -5381.14 and -5427.36, respectively. Following a chi-squared distribution with 18 degrees of freedom, the alternative hypotheses was accepted for P < 0.05; analyzed CASQ molecular evolution is based on the strict molecular clock model. It is similar to subunits 4 and 5 of the enzyme NADH-dehydrogenase and three tRNAs evolution, which appear as molecular clock examples in several phylogenetic software releases [20] and find good support for a clock among the anthropoids, but no support is found for random local clock. We used the strict molecular clock model in the phylogenetic tree constructions and substitution rate estimations.

3.2 Phylogenetic Tree Construction

All species promoter sequences present in ElDorado database were used in our bioinformatic analyses. ElDorado is the Genomatix genome annotation. It is based on the publicly available reference genome assemblies of 31 different organisms, from which only 13 organisms Casq promoter region sequences are well known (Table 1).

These promoter sequences were aligned using TCoffee in combined libraries of local and multiple alignments, which are known to induce high accuracy and performance in sequence alignments [15]. The residue consistency mean score of the all sequence alignment reported by T-Coffee aligner was very low (SCORE=29) demonstrating that Casq promoter sequences alignment is a low quality alignment. One possible reason for the low quality of the alignment would be the fact that the transcription in CASQ1 occurs in the opposite direction relative to CASQ2. For this reason Casq2 promoter were reversecomplemented in order to match the direction Casql transcription, and then were aligned together using T- coffee. Surprisely the score was identical to the previous alignment score. These results can be linked to the high number of observed nucleotide differences between the two classes of promoters in the comparison between them (Casql vs Casq2). However, each promoter class alignment (Casql and Casq2 promoter alignments) showed higher scores than previous alignments (Casql: SCORE=46 and Casq2: SCORE=53). jModelTest 0.1.1 software [17] determined the HKY+G model as being the best-fit model of promoter DNA sequence evolution with a gamma shape value (four rate categories) of 4.02, using two statistical criterion, Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) (-lnL=12098.567). Phylogenetic relationships of all these different organisms Casq promoter sequences were determined using the two most powerful statistical methods: ML and BI. The best phylogeny generated by the BI method is depicted in Figure 1A.

First we used a BEAUti vl.7.0 generated input file with no out-group sequences; then we used as outgroup all «ow-Mammalian sequences. The "outgroup" maximum clade credibility tree is the best of two trees generated by BEAST vl.7.0, because it is characterized of statistically significant lower values of likelihood (InL) (-12165.057±0.0996 for out-group tree vs -12163.746±0.1131 for non out-group tree; P < 0.05) and auto-correlation time (ACT) (4805.239 for out-group tree vs 5243.1218 for non out group tree) than the other tree. The BI phylogenetic reconstruction was highly representative of the phylogeny generated by ML. As shown in Figure 1A, the major branches were highly supported both in BI and ML trees, even BI phylogenetic tree was more resolved than ML one. Another difference between BI and ML, was the different position of rabbit (Oryctolagus cuniculus) Casq2 promoter sequence in the two trees; this branch was higher supported in BI phylogenetic tree (100% posterior probability) than in ML one (71% bootstrap value) (not shown).

Both Casql and Casq2 promoter sequences of mammals grouped into two separate and highly consistent clades (100% posterior probabilities in both cases), even rabbit Casql promoter sequence was an exception, because it's position was in the mammals "big" clade, but it was outside from each of two clades. In order to know if Casq promoter and CASQ proteins had followed the same pathway of evolution, a tree topology comparison between CASQ promoter and amino acid sequences [21] phylogenetic trees was conducted. Infante and colleagues [21] performed the phylogenetic relationships among the predicted sequence of Senegalese sole CASQ isoforms and the corresponding deduced amino acid sequences from other vertebrates using the BI method (Figure IB). In that phylogenetic tree, mammals and frog {Xenopus tropicalis) CASQ1 were grouped together into a clade, while mammals and chicken {Gallus gallus) CASQ2 were grouped into the other one. As it is shown in Figure 1A, in Casq promoter based phylogeny, the chicken Casq2 was positioned far outside the mammals Casq2 clade; only frog Casql position was identical to that observed in the CASQ amino acid based phylogeny [21]. Probably, this emerged topological discordance from the comparison between CASQ promoter based phylogenetic tree and amino acid based [21] phylogenetic tree show similar, but not identical evolutionary pathways and may be linked to possible different substitution rates present in the comparison between Casq promoter and protein sequences. As shown in Figure 1A, the Casq promoter sequences positions in each of the two mammals calsequestrin isoform clades were identical to those observed in the amino acid based phylogenetic tree (Figure IB). To confirm if the pathways of evolution were identical in mammals, we performed a comparison between estimated Casq promoter and CASQ protein, substitution rates.

3.3 Substitution Rate Comparison

As it is shown in Figure 1A, the Casq promoter positions in each of the two mammals CASQ isoform clade are identical to those observed in the amino acid based phylogenetic tree (Figure IB). Thus, first we estimated nucleotide substitution rates of mammals Casq promoter sequences using BEAST vl.7.0; then we estimated nucleotide and amino acid substitution rates of mammals Casq coding cDNA and protein sequences, respectively. Previously, the jModelTest 0.1.1 software determined HKY+G and TrN+G as being the best-fit models (four rate categories) of promoter and coding sequences evolutions (-InL = 8653.62 and 6713.66) with a gamma shape value of 2.20 and 0.38, respectively. ProtTest3 was used for determination of amino acid sequence evolution bestfit model. The best resulted to be the JTT + G + F model (-lnL= -2777.9 ) with a gamma shape value (four rate categories) of 0.46. By comparison of estimated substitution rate mean values (Figure 2A), statistical significant (P < 0.05, One-Way ANOVA tests) differences among mean rates of the analyzed categories were found.

In fact, as shown in Table 2, the promoter substitution rate mean value (1.169±0.008) was notably higher than the coding and protein sequence substitution rates (0.288±0.004 and 0.179±0.003).

We estimated also all ElDorado Casq promoter sequence substitution mean rate (Figure 2B). It resulted to be significantly higher than the mammals promoter mean rate (2.4521±0.012 to 1.169±0.008, P < 0.05 by unequal variances student's t test). However this difference was not so elevated like that between mammals and coding (or protein) substitution mean rates (Table 2). These results suggest different evolutionary pressure operating onto the CASQ gene (protein) and promoter sequences. Although substitution rate differences were present between promoter and amino acid sequences, mammals sequence phylogenetic relationships remained the same at the phylogenetic topology of promoter and amino acid sequence based trees (Figure 1A and B, respectively). This suggested that the majority of substitutions didn't change the structure/ function of the Casq promoters and Casq genes.

Though the Casq promoters mean rate was almost 2 times higher than coding region and protein mean rates, the Casq promoter structure /function remained unaltered like Casq gene. Thus, it could have been a consequence of the higher frequency of "synonymous" (no altered structure/function) substitutions in the Casq promoter than in Casq gene. This suggested that negative selection could have operated with higher degree onto the Casq promoter region than on Casq gene or CASQ protein, responsible for Casq promoter sequence conservation. Negative selection prevents deleterious mutation from reaching common frequencies and so should produce an excess of rare variation. It is acceptable that deleterious mutations are present also in non-coding regions [22]. In conclusion our analysis indicates that mammals Casq promoter (non-coding) region could be a perfect platform where the negative or purifying selection might operate to conserve structurally and/or functionally important nucleotide motifs.

4. References

References

1. Sanchez EJ, Lewis KM, Danna BR, Kang C: High-capacity Ca2+ binding of human skeletal calsequestrin. Journal of Biological Chemistry 2012, 287: 11592-11601.

2. Boncompagni S, Thomas M, Lopez JR, Allen PD, Yuan Q, Kranias EG: Triadin/junctin double null mouse reveals a differential role for triadin and junctin in anchoring CASQ to the jSR and regulating One 2012, 7: e39962. Ca2+ homeostasis. PLoS

3. Tomasi M, Canato M, Paolini C, Dainese M, Reggiani C, Volpe P: Calsequestrin (CASQ1) rescues function and structure of calcium release units in skeletal muscles of CASQ 1-null mice. American Journal of Physiology - Cell Physiology 2012,302: 575-586.

4. Yang A, Sonin D, Jones L, Barry WH, Liang BT: A beneficial role of cardiac P2X4 receptors in heart failure: rescue of the calsequestrin overexpression model of cardiomyopathy. American Journal of Physiology 2004, 287: 1096-1103.

5. Milstein ML, Houle TD, Cala SE: Calsequestrin isoforms localize to different ER subcompartments: evidence for polymer and heteropolymer-dependent localization. Experimental Cell Research 2009, 315: 523-534.

6. Chopra N, Yang T, Asghari P, Moore ED, Huke S, Akin B: Ablation of triadin causes loss of cardiac Ca2+ release units, impaired excitation-contraction coupling, and cardiac arrhythmias. Proceedings of the National Academy of Sciences 2009,106: 7636-7641.

7. Murphy RM, Larkins NT, Mollica JP, Beard NA, Lamb GD: Calsequestrin content and SERCA determine normal and maximal Ca2+ storage levels in sarcoplasmic reticulum of fast- and slow-twitch fibres of rat. Journal of Physiology 2009, 587: 443-460.

8. Park KW, Goo JH, Chung HS, Kim H, Kim DH, Park WJ: Cloning of the genes encoding mouse cardiac and skeletal calsequestrins: expression pattern during embryogenesis. Gene 1998, 217: 25-30.

9. Sacchetto R, Volpe P, Damiani E, Margreth A: Postnatal development of rabbit fast-twitch skeletal muscle: accumulation, isoform transition and fibre distribution of calsequestrin. Journal of Muscle Research and Cell Motility 1993,14: 646-653.

10. Richard AJF, Demignon J, Sakakibara I, Pujol J, Favier M, Strochlic L, Le Grand F: Genesis of muscle fiber-type diversity during mouse embryogenesis relies on Sixl and Six4 gene expression. Developmental Biology 359, 359: 303-320.

11. Donoghue P, Doran P, Dowling P, Ohlendieck K: Differential expression of the fast skeletal muscle proteome following chronic low-frequency stimulation. Biochimica et Biophysica Acta 2005,1752: 166-176.

12. Kinnunen S, Mänttäri S: Specific effects of endurance and sprint training on protein expression of calsequestrin and SERCA in mouse skeletal muscle. Journal of Muscle Research and Cell Motility 2012, 33: 123-130.

13. Steams SC: Evolutionary medicine: its scope, interest and potential. Proceedings of the Royal Society B - Biological Sciences 2012, 280: 1750.

14. Cartharius K, Frech K, Grote K, Klocke B, Haltmeier M, Klingenhoff A: Matlnspector and beyond: promoter analysis based on transcription factor binding sites. Bioinformatics 2005, 21: 2933-2942.

15. Notredame C, Higgins DG, Heringa J: T-Coffee: A novel method for fast and accurate multiple sequence alignment. Journal of Molecular Biology 2000, 302:205-217.

16. Darriba D, Taboada GL, Doallo R, Posada D: ProtTest 3: fast selection of best-fit models of protein evolution. Bioinformatics 2011, 27: 1164-1165.

17. Posada D: jModelTest: Phylogenetic Model Averaging. Molecular Biology and Evolution 2008, 25: 1253-1256.

18. Guindon S, Gascuel O: A simple, fast, and accurate algorithm to estimate large phytogenies by maximum likelihood. Systematic Biology 2003, 53: 696-704.

19. Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 2007, 7: 214.

20. Yang Z: PAML 4: phylogenetic analysis by maximum likelihood. Molecular Biology and Evolution 2007, 24: 1586-1591.

21. Infante C, Ponce M, Manchado M: Duplication of calsequestrin genes in teleosts: molecular characterization in the Senegalese sole (Solea senegalensis). Comparative Biochemistry and Physiology Part B - Biochemistry and Molecular Biology 2011,158:304-314.

22. Wakeley J: Substitution rate variation0 among sites in hypervariable region 1 of human mitochondrial DNA. Journal of Molecular Evolution 1993, 37: 613-623.