This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Gubernatrix cristata, the Yellow Cardinal, is a rare bird species from the Pampas grassland. It has a restricted geographical distribution in southern South America and is unique to this Biome (Uruguay, Argentina, and Southern Brazil). Fourteen field trips were made from April 2006 to March 2009 in an attempt to find individuals and collect samples for genetic analysis. A total of 72 yellow cardinal samples was taken in this study, 59 were from contemporary specimens and 13 samples were from museum specimens. We accessed the species genetic diversity through ten polymorphic nuclear microsatellite loci of Gubernatrix cristata and a ND2 mtDNA fragment. We found only three haplotypes with nucleotide and haplotype diversity equals to Ï€ = 0.00277 and Hd = 0.6219. The Fst (0,00340) and Nm (73,18) shown a weak structuring with little genetic differentiation. Fu's Fs statistic was significantly different from zero (2,248) and Tajima's D was positive (2,17506). Both tests showed likely evidence indicating a decrease in population size and/or balancing selection. Analyses using the "no admixture model" and a larger burn-in (500000) yielded no clustering of individuals. Due to this result no individual assignment to any geographical area was possible. The present study assists in understanding the conservation needs for Yellow Cardinal by providing information on the genetic diversity and population structuring along its distribution range. Although we found no structuring within our study area, further study is needed, examining the genetic diversity and population structuring throughout the species' range, adding more samples from wild animals from La Pampa population and Corrientes.
Keywords: Endangered species, Genetic diversity, Gubernatrix cristata, Mithochondrial DNA, Microsatellites, ND2 gene, Pampas grassland.
Gubernatrix cristata, the Yellow Cardinal, is a rare bird species from the Pampas grassland. It has a restricted geographical distribution in southern South America and is unique to this Biome (Uruguay, Argentina, and Southern Brazil) (Azpiroz, 2003). This species has undergone a dramatic population decline across its range and its conservation status is defined as endangered (BirdLife International, 2000; Fontana et al. 2003). In Argentina, it is now rare except very locally (Fraga, 1997). In Uruguay, it was historically known from 13 departments, but recently from only Paysandú, Río Negro, Florida and Rocha (BirdLife International, 2000; Azpiroz, 2003). In Southern Brazil (Rio Grande do Sul State) it was already considered rare in the 1970s and 1980s (Belton, 1985). Since then it suffered a dramatically decline and now is believed it is confined to restricted spots in the SE hills, near Uruguayan border and at W tip of Rio Grande do Sul (Fontana et al. 2003). Once the lack of yellow cardinal sightings in the last few years and its bad situation in the adjacents areas in Uruguay, the SE hills population seems to be isolated or even extinct. (Martins-Ferreira, 2007). The remaining populations are small and isolated (BirdLife International, 2000). The only breeding wild population known in Brazil of G. cristata seems to be limited to Espinilho State Park (Parque Estadual do Espinilho) and surroundings (Damiani et al., 2009). The destruction and fragmentation of its habitat have been enumerated as possible additional threats for the species, as well as the extreme pressure of capture to supply the traffic of wild animals (Chebez 1994; BirdLife International, 2000). The main reason behind the traffick is that bird is a colorful and singing passerine. The male has a bright yellow chest and back with black crest, eye line and throat. It has the full eyebrow and malar stripe yellow while the female are white with a gray breast (Belton, 1994; Sick, 1997). During the breeding time the male defends its territory aggressively atacking other males. That behavior turns to be a way how poachers capture them using a captive male as decoy (Fontana et al. 2003). Hudson (1920) already atested that its qualities as beautiful bird with loud and musical voice made it a favourite cage bird. The growing record of hybrids among yellow cardinal and common diuca finch (Diuca diuca) in Argentina seems to be a response to the shortage of male individuals in the population of the species (Bertonatti & Guerra, 1997, 2001). With the null hypothesis that Yellow Cardinal has not a population structure throughout its distribution range, we decide to investigate that. Our aim was to determine the population status of Gubernatrix cristata across its distribution range, taking in consideration the population structure, genetic diversity and gene flow. Only with such information in hand we would be able to propose a specific conservation strategy to the species.
Fourteen field trips were made from April 2006 to March 2009 in an attempt to find individuals and collect samples for genetic analysis. In Rio Grande do Sul (the southernmost Brazilian state), just four Yellow Cardinals were found in a single locality (Espinilho State Park). In Uruguay, four separate field trips to different localities found only six individuals in Minas, Lavalleja Department, five in one farm and one in another land. All of them were ring banded with rings provided by CEMAVE/ICMBio (National Research Center for the Conservation of Wild Birds). In Argentina, a visit was made to the Museo de Ciencias Naturales Bernardino Rivadavia from which tissue samples of eight skins were collected, one from each province where the species has ocurred, and one from Uruguay. Five more samples came from museum specimens: one from Museu de Zoologia da Universidade de São Paulo (MZUSP, São Paulo, Brazil), two from Museu de Ciências Naturais da Fundação Zoobotânica do Rio Grande do Sul (MCN/FZBRS, Porto Alegre, Brazil) and two from Royal Ontario Museum (ROM, Toronto, Canada) (Fig. 1). The rest of the samples came from specimens seized by the Environmental Police and are of unknown origin. A total of 97 yellow cardinal samples was taken in this study (Appendix A.1). Some of which having been maintained (but not bred) in captivity by zoos, conservation institutions, or private individuals. Of these, 25 presented a poor DNA quality that wasn't possible to proceed with any analysis. From 72 samples that worked properly 59 were from contemporary specimens. The other 13 samples were from museum specimens (years 1905-1975) selected from across the range, but with a special emphasis on the population in Argentina since many samples with unknown geographical data were supposed to come. Blood samples were stored in FTA Cards (Whatman) and feathers were placed in 95% ethanol solution. Museum tissues samples had their DNA extracted using the DNeasy Tissue Kit (QIAGEN) with the appropriate protocol according to the manufacturer for each type of tissue (Mundy, N. I. et al., 1997). Contamination with modern DNA or PCR products was monitored by including two extraction blanks in every extraction round and prevented by performing all museum samples extractions in a dedicated 'clean' laboratory, kept free of good quality DNA and PCR products.
Figure 1. Map showing the points of collected samples from wild individuals or from museum skins across G. cristata distribution range. Polygon used with permission from Birdlife.
A microsatellite-enriched genomic DNA library of Gubernatrix cristata was constructed and ten polymorphic nuclear microsatellite loci (with 2-bp repeats) were selected on the basis of polymorphism levels and size of amplification products. All samples were genotyped using a polymerase chain reaction targeting different loci (Table 1). Primers and PCR conditions are given in Molecular Ecology Resources Primer Development Consortium (2010). PCR products were purified, cycle sequenced on both strands, and run on an ABI 3730 sequencer.
We extracted DNA from feather samples using the Alkali protocol for DNA extraction and attempted to sequence several mitochondrial DNA (mtDNA) loci using degenerate avian primers from Sorenson (2003) to determine if sufficient genetic variability was present for analysis. The control region, ATPase8, and ND2 were chosen, as these regions have proven useful in population genetic studies of birds (Burg and Croxall, 2001; Sorenson, 2003; Oyler-McCance et al., 2005). We successfully amplified portions of the ND2, but were unable to amplify control region and ATPase8. Given the lower quality of DNA extracts from many samples we decided to amplify the gene in two pieces and run four sequencing reactions (both strands of each PCR product). The primers used were ND2H 5' CCT TGA AGC ACT TCT GGG AAT CAG A 3' (Tavares et al., 2006) - ND2F_Passerine 5' CCA YCC ACG AGC YAT TGA AGC 3' (new primer designed), and MetLTF 5' AAG CTA TCG GGC CCA TAC CCG 3' (Tavares et al., 2006) - ND2R_Passerine 5' GCC ATG CRT TGG TYA TGC TNG AG 3' (new primer designed). Those primers were used to amplify a 349 base-pair region that contained the variable sites in ND2. We amplified DNA via the polymerase chain reaction (PCR) in 12,5 Î¼l reaction volumes containing 2.0 Î¼l DNA, 7.92 Î¼l water, 0.5 Î¼l (10 pmol) of each primer, and 1.58 Î¼l PCR master mix, consisting of a 1.25 Î¼l buffer (10X:100 mM Tris-HCl, pH 8.3, 500 mM KCl, 25 mM MgCl2, 0.1% gelatin), 0.28 Î¼l dNTPs (10 mM), and 0.05 U Taq polymerase (Amersham Pharmacia Biotech). Reactions were performed under the following conditions: initial denaturation at 94Â°C for three minutes, followed by 36 cycles of 94Â°C for 45 sec, 50Â°C for 45 sec, and 72Â°C for 1 minute and 30 sec, and a final extension of 72Â°C for 7 minutes. The PCR products were then electrophoresed on a 1% agarose gel containing ethidium bromide and visualized under UV light. Amplified segments were purified by excising bands from agarose gels and centrifuging each through a filter tip. The sequencing reactions were then electrophoresed on an ABI Prism 3100 automated sequencer (Applied Biosystems). We sequenced in both directions and sequences were aligned and assembled using ChromasPro V. 1.49 (Technelysium Pty Ltd) and aligned using the CLUSTAL W algorithm with default options, implemented in MEGA 4.0.2 (Tamura et al., 2007). Alignments were checked and edited by hand when necessary. The haplotype sequences were deposited in GenBank under accession numbers HQ15712 to HQ15714.
Table 1. Gubernatrix cristata microsatellite loci used in this study.
Significant departures from HWE: * P < 0.001
We assessed the bottleneck history of these populations using Bottleneck v.1.2.02 (Cornuet & Luikart, 1996). This program was used to detect a heterozygote excess for individual populations, considering the two-phased model (TPM) of microsatellite mutation, a 70% stepwise-mutation model (SMM) and 30% infinite alleles model (IAM), and 1000 replications. Several other combinations of the SMM:IAM ratio were tested to establish the sensitivity of these data to the mutational mechanism. The Wilcoxon signed-rank test was used to determine if the allele frequency distribution for a population exhibited significant heterozygote excess relative to model expectations. Bayesian clustering with software Structure v.2.3.2 (Pritchard et al., 2000) was used to assign individuals to populations (K) based on posterior probabilities where K is unknown. The number of groups was set to 1-7 with 3 runs per K. Posterior probabilities were calculated for all K hypothetical populations. All analyses were based on 1,000,000 Markov Chain Monte Carlo iterations following a burn in of 50,000 iterations.
To measure mtDNA diversity, both haplotype diversity, Hd, and nucleotide diversity, Ï€, and their standard deviations were estimated using DnaSP v5.10.00 (Librado & Rozas, 2009). To test for evidence of recent population expansion we calculated Fu's Fs (Fu, 1997) and Fu and Li's (1993) D* and F* statistics to compare with Fu's Fs. Thus, if Fs is significant and F* and D* are not, it is an indication of population expansion, while the opposite indicates selection (Fu, 1997). We also calculate Tajima's D (Tajima, 1989). We used ARLEQUIN 3.11 (Schneider et al., 2000) and DnaSP v5.10.00 (Librado & Rozas, 2009) to perform these calculations. We created a mismatch distribution of pairwise differences using DnaSP v5.10.00 (Librado & Rozas, 2009) to compare the expected distribution for a population. The distribution tends to be multimodal when populations are at equilibrium and unimodal in cases of recent demographic expansion or reduction (Rogers and Harpending, 1992). To graphically display the observed mismatch distribution compared to the expected distributions for populations in equilibrium and expansion, we used Roger's method of moments (Rogers, 1995) as calculated in DNASP. The topological relationship between the haplotypes was estimated using the program Network 188.8.131.52 (http://www.fluxux-engineering.com) with median joining approach, for the three data sets (Bandelt et al., 1999).
Allelic variation for the ten nuclear microsatellite loci ranged between 4 and 14 alleles with an average of 7.5 alleles per locus. The observed heterozygosity (HO) varied from 0.126 to 0.893 and the expected (HE) heterozygosity from 0.130 to 0.930, with means and standard errors of 0.6824 + 0.1144 and 0.7132 + 0.0272 respectively (Table 1). Four loci exhibited a departure from Hardy-Weinberg expectation that may be the result of a sampling bias. Since the species is rare and most of the material used was from seizures made by the Environmental Police (Table 2), we cannot determine if all of the samples represent the same population.
Table 2. Geographical origin and source of the samples processed in our study.
Results of bottlenecks detection using Wilcoxon signed-rank and sign tests under Infinite Allele Model (IAM), Two-Phase mutation (TPM) and Step-wise mutation (SMM) are presented in Table 3. In bottlenecked populations, the observed gene diversity exceeds the expected equilibrium gene diversity under the assumption of mutation-drift equilibrium. The null hypothesis tested for heterozygosity excess using Wilcoxon sign-rank test provided (P < 0. 05273), (P > 0.09668) and (P < 0.42285) probabilities under the IAM, TPM, and SMM respectively. The null hypothesis is accepted under the SMM model only, implying that the Gubernatrix cristata has not experienced any recent genetic bottlenecks. The estimated values of the heterozygosity excess and their probabilities in sign test were 5.89 (P < 0.15033) for the IAM, 5.90 (P < 0.15148) for the TPM and 5.91 (P < 0.38936) for SMM. The null hypothesis of mutation-drift equilibrium was not also accepted based on the sign test under mutation models.
Not all of the 97 sampled individuals were successfully sequenced for the target ATPase 8 or Control Region fragment. Only 67 were succesfully sequenced for the target ND2 fragment (haplotype sequences were deposited at GenBank under accession HQ15712 to HQ15714). A total of 349 bp were aligned in G. cristata samples contained two variable and parsimony informative sites and a total of three haplotypes (H1 - H3). Among H2 and H1 haplotypes 75 mutational steps were found whilst 34 among H2 and H3 (Fig. 2). Estimates of nucleotide and haplotype diversity in the sample were relatively high, with Ï€ = 0.00277 and Hd = 0.6219. The genealogical relationships between haplotypes were estimated in the median-joining networks illustrated in Figure 2. The network revealed no structure pattern at all. Haplotype 2 had the highest probability of represent ancient forms because of its central position. The Fst (0,00340) and Nm (73,18) shown a weak structuring with little genetic differentiation. The Fst analysis was done following Nei's (1978) classification where a value is considered low when Fst < 0,05; medium when 0,05 < Fst < 0,15, and high when Fst > 0,15. Lack of support to a demographic expansion hypothesis was either obtained by different tests made to test the neutral theory of molecular evolution: Fu and Li's D* test statistic (0,71916) showed no statistical significance (P > 0.10) as well not significant was Fu and Li's F* test statistic (1,34814) (P > 0.10). Fu's Fs statistic was 2,248 which was significantly different from zero and is a likely evidence indicating a decrease in population size and/or balancing selection. Tajima's D was positive (2,17506) what signifies low levels of both low and high frequency polymorphisms, indicating a decrease in population size and/or balancing selection.
Table 3. Genetic bottleneck detection using Wilcoxon signed-rank sign tests under Infinite Allele (IAM) , Two-Phase (TPM), and Step-wise (SMM) mutations models of microsatellite evolution
Wilcoxon signed-rank test:
Probability of Heterozygosity excess
P < 0.05273
P > 0.09668
P < 0.42285
Number of loci with heterozygosity excess
P < 0.15033
P < 0.15148
P < 0.38936
Figure 2. Genealogical relationships between haplotypes estimated in the median-joining networks. H_1, H_2 and H_3 - Haplotyes; 75 - mutational steps between H2 and H1; 34 - mutational steps between H2 and H3. Yellow - unknown geographical origin; Red - individuals from Argentina; Green - individuals from Brazil; Blue - individuals from Uruguay
As the exact geographical origin of many (45%) of the sampled Yellow cardinal individuals is unknown or uncertain (many of the individuals sampled in zoos or at privates, or got from police seizures were not associated to reliable information on their exact location of capture), the structure analysis would be particularly relevant because it would allows clustering of the individuals without the need for a priori geographical information. Three independent runs for one to seven populations using an admixture model and correlated allele frequencies among populations (Falush et al., 2003), were run. Figure 3 shows the results for k equals to 3 populations as bar plot and triangle plot. Analyses using the "no admixture model" and a larger burn-in (500000) yielded very similar results (data not shown). The results for all runs were practically the same suggesting no clustering of individuals (Fig. 3). Due to this result no individual assignment to any geographical area was possible.
The mismatch distribution of observed haplotype variation was unimodal revealing a departure from expected variation under a Constant population size (Fig. 4) (k = 0,965, variance of k = 0,7205, Raggedness statistic r = 0,1319).
Bar plot showing no structure for k = 3 populations.
Triangle plot showing no structure for k = 3 populations.
Figure 3. Results of Structure run. Here is depicted when k = 3 populations. a) bar plots and b) triangle plot.
Figure 4. Mismatch distribution of the observed haplotype variation from 67 Yellow Cardinal mtDNA samples compared to the expected under a population expansion.
Endangered birds usually have small and fragmented populations. A commonly concern about population studies is a proper sample size to rightly infer genetic diversity measures. Despite the relatively small sample size of this study it has been recommended that at least 20 to 30 individuals be sampled in microsatellite studies (Pruett and Winker, 2008). Our 72 sample size is more than enough to run studies about genetic diversity and population structure estimations. The findings about no structuring appear to be related to other issues.
The lack of population structure showed by microsatellites with low Fst and large Nm apparently tell us the history about only one huge population across Yellow cardinal range with an intense gene flow. The same is truth for ND2 mitochondrial DNA gene.
The fact that both Fu's Fs and Tajima's D indicates a decrease in population size and/or balancing selection shows us a possible explanation of what is happening.
The yellow cardinal has suffered an intense hunting pressure by almost one hundred years now (Hudson, 1920). This time period is not sufficient to be detected at the molecular level, however is enough to drastically reduce the species population size or even driven to extinct some subpopulations (Martins-Ferreira, 2007).
The species had a wide distribution area in the past that has been reduced over time, both due to changes in habitat, as the hunting pressure. Today it seems to be limited to a narrow continuous irregular arc-shaped band in Argentina, from San Luís, passing through La Pampa, Córdoba, Santa Fé, Entre Ríos and Corrientes (Marchiori, 2004; Martins-Ferreira, 2007). Following the center south of Uruguay to the west end of Rio Grande do Sul, Brazil. That reduction in its distribution range to only the optimal environment for the species, also ended up reducing their genetic variability and, possibly, the haplotype diversity. The Whooping Crane (Grus americana) experienced a severe decline in their population size in early 1900. Reducing its population to 14 individuals in 1938, resulted in the loss of two thirds of its haplotype diversity. Only three haplotypes were identified in the current population (Glenn et al., 1999).
As shown in Whooping Cranes, a severe reduction in population size can cause a genetic bottleneck, resulting in loss of genetic diversity (Nei et al., 1975). However, the impact of a bottleneck on population genetic diversity depends on how rapidly the population declines, the size of the bottleneck population and the duration of the bottleneck. When a population suffers a genetic bottleneck, rare alleles initially are lost, but the population can maintain diversity if the population recovery is rapid (Allendorf , 1986). If population abundance remains low, diversity can continue to be lost due to genetic drift (Allendorf ,1986). There are several possible explanations for the relatively high genetic diversity observed in Yellow Cardinals along Southern South America. First, the Yellow Cardinal decline may have been less severe than that of the Whooping Crane. The population status in the early 1900s is unknown, but never drop low before than today. Secondly, population recovery may have been sufficiently rapid so as not to have lost diversity due to genetic drift. Although Yellow Cardinal in Uruguay and Brazil has recently experienced several episodes of population decline, each time population recovery appears to have been rapid. Finally, recolonization by individuals that emigrated from areas where genetic diversity remained high, such as Argentina, could have contributed to the high genetic diversity. The observed genetic similarity within the region is consistent with an open population, as gene flow is expected to homogenize populations (Charlesworth, 2003). We detected no pattern in the distribution of haplotypes to suggest that genetic structuring exists in Yellow Cardinals along its range.
Thus, our results support managing Yellow Cardinals along its distribution range as a single breeding population. It is possible that the Yellow Cardinal may form one large panmictic population. The existence of one large population could explain the high level of genetic diversity found in the population by allowing the effective population size to remain high during periods of population fluctuation and serve as a source of immigrants. Although the Yellow Cardinal's demographic history does not show drastic reductions in genetic diversity, the population size fluctuations appear to have had no detectable effects on the differentiation and frequency of haplotypes as evidenced by the mismatch distribution and Fu's Fs. The minimum spanning network (Fig. 2) shows that the 3 haplotypes are not closely related (within 75 and 34 mutational steps). A triangle-shaped phylogeny, as shown in Fig. 2, tell us only that a population has experienced a severe decline on its genetic diversity. The Yellow Cardinal population still appears smaller than historical accounts suggest. Historical accounts described Yellow Cardinal as common in Argentina in the 1900s (Hudson, 1920; Wetmore, 1926). Thus, it seems that the population may have being suffering a decline in its numbers from the last century until now.
The present study assists in understanding the conservation needs for Yellow Cardinal by providing information on the genetic diversity and population structuring along its distribution range.
Although we found no structuring within our study area, further study is needed, examining the genetic diversity and population structuring throughout the species' range, adding more samples from wild animals from La Pampa population and Corrientes.
The main implication for Yellow Cardinal conservation with these results is that a captive breeding program could be established on an international level, without much concern about preserving specific haplotypes, since all populations share the same ones.
We thank to Oliver Haddrath for the tremendous help in conducting the lab work at ROM. We thank to Carla Lopes, Paula Rorato, and Nelson Fagundes for helping us with the software packages and valuable discussions. We thank to Darlise Lopes for the last-minute help in the lab. We thank CNPq for providing the doctorate scholarship to CMF, to CAPES for the Sandwich Scholarship, and to FBPCN for financing the project. Laboratory work was funded by NSERC operating grant to AJB.