Genetic variation and cryptic lineage among the sergestid shrimp Acetes americanus (Decapoda)

Sampling of the biological material

The specimens used in this study were obtained through samplings and loans. All individuals of A. americanus from Penha/Santa Catarina, Ubatuba, Cananéia, São Vicente/São Paulo, Macaé/Rio de Janeiro, Baía Formosa/Rio Grande do Norte, and A. petrunkevitchi from Ubatuba/São Paulo were sampled using a fishing boat equipped with an otter trawl net with 2 m opening and 3 m in length. The mesh diameter (interknot distance) was five mm in the first half of the net and two mm in the final part. The obtained specimens were collected under field permit approval by Instituto Chico Mendes de Biodiversidade/ICMBio, number 23008-1 and Permanent License to FLM 11777-2) and deposited in the Crustacean Collection of the Department of Biology of the Faculty of Philosophy, Sciences, and Letters of Ribeirão Preto of the University of São Paulo, Brazil (CCDB/FFCLRP/USP) and in the Crustacean Collection of the Laboratory of Biology of Marine and Freshwater Shrimp, UNESP, Bauru, Brazil (CCLC/FC/UNESP) (Table 1 and Table S1).

All other specimens used in this study were loaned from scientific collections: Zoology Museum of the University of São Paulo, São Paulo, Brazil (MZUSP); National Museum of Natural History, Smithsonian Institution, United States (USNM); National Museum of the Federal University of Rio de Janeiro (MNRJ); Oceanographic Museum of the University of Pernambuco, Brazil (MOUFPE); Federal University of Rio Grande (FURG); Crustacean Collection from the Department of Biology of the Faculty of Philosophy, Sciences and Letters of Ribeirão Preto, University of São Paulo (CCDB); Collection of Crustaceans of the Federal University of Sergipe (UFS); Collection of Crustaceans from the Federal University of Espírito Santo (UFES); Collection of Crustaceans from the PUCRS Museum of Science and Technology (MCP); Carcinological Collection of the Institute of Scientific and Technological Research of the State of Amapá, Macapá, AP (IEPA); National Crustacean Collection, UNAM, Mexico (CNCR); Peabody Museum of Natural History Yale University, United States (YPM); University of Louisiana at Lafayette, USA (ULLZ); and Zoological Museum Kiel, Germany (ZMK) (Table 1 and Table S1).

Specimens were morphologically identified following the identification key proposed by Omori (1975), D’Incao & Martins (2000), Vereshchaka, Lunina & Olesen (2016a), and Vereshchaka, Olesen & Lunina (2016b), using the shape of the third thoracic sternite (thelycum) in females and the petasma shape in males.

The names currently used for extant taxa follow the most recent literature (Vereshchaka, Lunina & Olesen, 2016a; Vereshchaka, Olesen & Lunina, 2016b) and are currently adopted by WoRMS (2022). Genus Peisos is dealt with differently. A review of the shrimp genera Acetes, Peisos, and Sicyonella (Vereshchaka, Lunina & Olesen, 2016a) proposed to include Peisos within the genus Acetes using morphological and phylogenetic information. Later, in a global phylogeny, Vereshchaka (2017) proposed new families, including Acetidae, to accommodate all species of Acetes. WoRMS (2022) does not adopt the proposed changes, that is, Peisos is an accepted genus, and Acetidae does not exist. We can conjecture a lack of taxonomic reassignment with a proper nomenclatural act in the previous paper. Considering that we were not able to find a clear justification and to avoid taxonomic instability, we maintained the proposed taxonomic status of Peisos as part of Acetes following Vereshchaka, Lunina & Olesen (2016a).

DNA extraction and amplification

DNA extraction from the muscular abdominal tissue was performed following Mantelatto, Robles & Felder (2007); Mantelatto et al. (2009a); Mantelatto et al. (2009b) and Pileggi & Mantelatto (2010), with specific modifications (Carvalho-Batista et al., 2014). The regions of interest were amplified by the polymerase chain reactions (PCR) using the primers described in Table S2. Primers specific for this region were designed due to the difficulty in the amplification of COI (see Mantelatto et al., 2016 for details) (Table S2).

PCR reactions were performed with a total volume of 25 µL, containing 5 µL of betaine (5M) (Acros Organics), 4 µL of dNTPs (5 mM), 3 µl of MgCl₂ (25 mM), 3 µl of 10X Taq buffer with KCl (Thermo Scientific), 2 µl of 1% bovine albumin (Sigma), 1 µl of each primer (10 µM), 1 µl of resuspended DNA (50 ng/ ml), and 0.5 µl of recombinant Taq DNA Polymerase (Thermo Scientific). The missing volume (25 µl) was filled with ultrapure water. PCR was performed using an Applied Biosystems^© Veriti 96 well thermal cycler. PCR steps comprised an initial denaturation period of 3 min at 95 °C, followed by 40 thermal cycles [30 s of denaturation at 95 °C, 45 s for annealing at variable temperature (42–44 °C for 16S; 44–50 °C for COI), 1 min for the extension at 72 °C, and final extension for 10 min at 72 °C. The obtained results were observed by 1.5% agarose gel electrophoresis and photographed with Olympus^© C-7070 digital camera in a UV M20 UV transilluminator.

The PCR products were purified using the Sureclean purification kit following the manufacturer’s protocol. The PCR-purified products were sequenced bidirectionally in automated sequencers (ABI 3100 Genetic Analyzer) at the Department of Technology of the Faculty of Sciences Agricultural and Veterinary Sciences of Jaboticabal, São Paulo State University.

The generated sequences were confirmed and edited (sequence consensus obtained from sense and antisense) in BioEdit 7.0.7.1 software (Hall, 1999) and aligned using CLUSTAL W (Thompson, Higging & Gibson, 1994).

Each genetic sample obtained for the analyses was deposited in the scientific collection of origin (Table S1).

Molecular analyses

A total of 118 sequences used in this study were generated for this project. For the 16S rRNA region, 52 sequences of 518 bp were obtained, of which 29 were A. a. Brazil 1, nine were A. a. Brazil 2, five were A. a. carolinae, five were A. paraguayensis, and four were A. petrunkevitchi. For the COI region, 66 sequences of 589 bp were generated of which 36 were A. a. Brazil 1, 15 were A. a. Brazil 2, six were A. a. carolinae, six were A. paraguayensis, and three were A. petrunkevitchi.

However, an additional 20 sequences (COI) of A. sibogae, A. japonicus, A. serrulatus, and A. indicus, two sequences (16S and COI) of A. petrunkevitchi, and two sequences (16S and COI) of B. faxoni (outgroup) were retrieved from the GenBank database and used for phylogenetic and genetic distance analyses to complement this study (Table S1). As there are COI sequences for more Acetes species than 16S sequences available on GenBank, phylogenetic analysis with the COI gene generated a larger number of clades.

The mean nucleotide composition and genetic distances were estimated using MEGA 5.0 software (Tamura et al., 2011) and the Neighbor-Joining dendrogram based on the Kimura 2-parameter substitution model (Kimura, 1980).

The appropriate models of nucleotide evolution HKY + G for 16S and TPM1uf + G for COI were selected by Bayesian information criterion (BIC) in jModeltest 2.1.4 (Darriba et al., 2012). Selected models and estimated parameters (Table S3) were implemented in the Bayesian inferences and considered for the choice of the closest models in the maximum likelihood analyses. Belzebub faxoni (Borradaile, 1915) (Superfamily Sergestoidea, Family Luciferidae) was included as an outgroup following the most recent global phylogeny (Vereshchaka, Lunina & Olesen, 2016a; Vereshchaka, 2017). Bayesian inference was used to reconstruct the phylogenetic relationships of the species analyzed, with the two genes as distinct partitions in MrBayes v. 3.2.2 (Ronquist et al., 2012). Bayesian inference was performed with 30 million generations in two independent analyses, with five parallel chains each, one “cold” and four “hot”. The parameters were saved every 1,000 simulations. The analysis was completed on attaining stationarity (mean standard deviation <0.01) after the stipulated number of generations. The first quarter of the parameters and trees were discarded as burn-in (Ronquist, Van Der Mark & Huelsenbeck, 2009). The support values of the branches were obtained using the a posteriori probability method.

This analysis was performed for the 16S and COI genes separately and a concatenated matrix of both the trees was generated and edited in the program Figtree v.1.3.1 (Rambaut, 2007).

Population analyses

Only the COI gene was used for population analyses. The use of this gene has been shown to be efficient in decapod population studies (Gusmão, Lazoski & Solé-Cava, 2005; Laurenzano, Mantelatto & Schubart, 2013; Rossi & Mantelatto, 2013; Carvalho-Batista et al., 2014; Teodoro et al., 2015) because it is considered to be more variable (Schubart & Huber, 2006).

The number of haplotypes was calculated in the DnaSP 4.10.9 software (Rozas & Rozas, 1999). Haplotype networks were constructed using the median-joining method, using the network program (Bandelt, Forster & Röhl, 1999), based on the data prepared in the DnaSP. Haplotype and nucleotide diversities, analysis of molecular variance (AMOVA), and pairwise fixation indices (Fst) (Excoffier, Smouse & Quattro, 1992) were calculated using Arlequin 3.11 (Excoffier, Laval & Schneider, 2005).

Morphometric analyses

We used only females (281 individuals) of A. americanus for morphometric analyses. Previous observations focused on the difference in size and shape of the female thelycum between the subspecies, motivating us to obtain potential information to complement our study.

Individuals were sexed based on the presence of petasma (first pleopod) in males and the thelycum (third thoracic sternite) in females (Xiao & Greenwood, 1993).

Morphometric measurements were obtained using a stereo microscope Zeiss^© Stemi 2000C connected to an imaging system Zeiss^© AxioVision, with an error of up to 0.01 mm. We measured the height (HT) and length (LT) of the thelycum and carapace length (CL).

Following the methodology proposed by Marramà & Kriwet (2017), data were standardized by calculating the ratio between each measure and carapace length (CL), removing the effect of size. The log transformation of data (Log (X+1)) was used to overcome the issue of the non-normal distribution of data by unstretching large-scale values. Moreover, log transformation is useful to considerably reduce variation due to ontogeny (allometric effect) since we assumed that specimens of different developmental stages were subjected to our analyses. A Euclidean distance matrix was constructed using the log-transformed data. Multivariate analysis of variance (PERMANOVA) was applied to test similarities within subspecies and localities (P < 0.005) (Anderson, 2001) using PRIMER software (version 6; Clarke & Gorley, 2006). A Permanova pairwise post-hoc test was performed to further investigate the differences between the subspecies (P < 0.005). Principal component analysis (PCA) was performed to characterize the differences between groups using the measured parameters (HT/CL and LT/CL ratios). Similarity percentage tests (SIMPER) were used to evaluate which measure contributed more to the differentiation between subspecies.

Morphological identification of taxonomic entities

All specimens loaned from the United States of America presented characteristics of the subspecies A. a. carolinae (hereafter named A. a. USA) (Fig. 2C). However, Brazilian samples fell into two categories: (i) samples exhibiting characteristics of the subspecies A. a. americanus senso stricto (hereafter named “A. a. Brazil 1”; Fig. 2A) and (ii) samples that exhibited morphological divergence from the nominal subspecies (hereafter named “A. a. Brazil 2” (Fig. 2B)). The sampling localities of A. a. Brazil 1 and A. a. Brazil 2 are shown in Fig. 1.

Genetic distance

16S rRNA gene: The intraspecific distances varied from 0% (A. a. USA, A. a. Brazil 2, and A. petrunkevitchi) to 0.20% (A. paraguayensis) (Table 2). The interspecific distances between congeneric species varied from 0.99 to 11.8%. The distance to the outgroup was 25.9 to 28.5% (Table 3). Regarding the A. americanus subspecies, the lowest interspecific distance was between A. a. Brazil 2 and A. a. USA (0.99%). The highest was A. a. Brazil 1 and A. a. USA (2.26%) (Table 3).

COI gene: The intraspecific distances varied from 0.02% (A. a. Brazil 1) to 0.97% (A. paraguayensis) (Table 2). The interspecific distances between the congeneric species varied from 4.86 to 22.3%. The distance to the outgroup was 21.8 to 25.5% (Table 3). Regarding the A. americanus subspecies, the lowest interspecific distance was between A. a. Brazil 1 and A. a. USA (4.86%). The highest was A. a. Brazil 2 and A. a. USA (8.08%) (Table 3).

Phylogenetic analyses

The 16S rRNA phylogenetic analyses showed the following clades (Fig. 3): Clade 1, formed by A. paraguayensis sampled in north Brazil (Santarém and Porto de Moz/Pará); Clade 2, formed by A. petrunkevitchi sampled in southeast Brazil (São Vicente and Ubatuba/São Paulo); Clade 3, formed by A. a. USA sampled in the United States (Lumcon/LA and Horn Island/MS); Clade 4, formed by A. a. Brazil 2 sampled in southeast Brazil (Cananéia/São Paulo and Macaé/Rio de Janeiro); Clade 5, formed by A. a. Brazil 1 sampled in the south (Penha/Santa Catarina), southeast (Ubatuba, São Vicente and Cananéia/São Paulo, Macaé/ Rio de Janeiro, Anchieta/Espirito Santo) and northeast Brazil (Baía Formosa/Rio Grande do Norte).

Based on the COI sequences, the phylogenetic analysis revealed the following clades (Fig. 4): Clade 1, formed by A. a. Brazil 1 sampled in the south (Penha/Santa Catarina), southeast (Ubatuba and Cananéia/São Paulo, Macaé/Rio de Janeiro, Anchieta/Espirito Santo) and northeast Brazil (Baía Formosa/Rio Grande do Norte, Maceió/Alagoas); Clade 2, formed by A. a. USA sampled in the United States (Lumcon/LA and Horn Island/MS); Clade 3, formed by A. a. Brazil 2 sampled in southeast Brazil (Cananéia/SP and Macaé/RJ); Clade 4, formed by A. sibogae sampled in Peninsula Malaysia; Clade 5, formed by A. japonicus sampled in Peninsula Malaysia; Clade 6, formed by A. serrulatus sampled in Peninsula Malaysia; Clade 7, formed by A. paraguayensis sampled in north Brazil (Santarém and Porto de Moz/Pará); Clade 8, formed by A. petrunkevitchi sampled in southeast Brazil (Ubatuba/SP); Clade 9, formed by A. indicus sampled in Peninsula Malaysia.

The phylogenetic tree constructed based on the concatenated data (16S rRNA and COI) generated the same clades observed in the phylogenetic tree constructed with 16S rRNA and COI separately. In addition, A. a. USA and A. a. Brazil 2 were sister clades (Fig. 5), whereas COI, A. a. USA was a sister clade of A. a. Brazil 1 (Fig. 4).

The phylogenetic trees constructed (16S rRNA, COI, and concatenated data) resulted in the formation of two distinct clades of A. americanus sampled in Brazil with a high support value, in contrast to the clade formed by A. a. carolinae sampled in the United States. Furthermore, the A. americanus group appears to be a sister taxon to all other Acetes (Fig. 4).

Population analyses

The haplotype network exhibited a genetic structure in the three groups, corresponding to those observed in the phylogenetic trees (Fig. 6). Available COI sequences of A. americanus (N = 57) resulted in 12 haplotypes of which six were unique, that is, represented by a single individual. “Brazil 1″presented the lowest haplotype (h = 0.11) and nucleotide (π = 0.00022) diversities. This group included 34 individuals sharing one haplotype and two individuals with unique haplotypes. “USA” and “Brazil 2” presented similar high haplotype (h = 0.73 and h = 0.76, respectively) and nucleotide diversities (π = 0.00181 and π = 0.00277, respectively).

Analysis of molecular variance (AMOVA) did no detect genetic structuring of any subspecies among the location studied (p > 0.05) (Table 4).

Morphometric analysis

The ratio between the height and length of the thelycum was determined for several western Atlantic locations (Table 1). The ratio values were higher in A. a. USA and A. a. Brazil 2 than in A. a. Brazil 1 (Table 1). The studied A. americanus groups (BR1, BR2, and the USA) were morphologically different (P = 0.0002) (Table 5). Permanova Pairwise tests indicated that at least one of the taxonomic entities is different from the others (P = 0.0001) (Table 6). However, when comparing A. americanus groups with each other, Pairwise tests indicated that there was a statistically significant difference only between Brazil 1 and Brazil 2 (t = 3.1563; p = 0.0018).

The PCA visualization plot shows that A. a. Brazil 1 is more similar to A. a. Brazil 2 than to A. a. USA (Fig. 7). A simplified Simper test showed that the LT/CL RATIO was responsible for the differences between subspecies, contributing more than 73% for Brazil 1, more than 53% for Brazil 2 and more than 61% for the USA (Table 7; Fig. 7).

General phylogeny

A previous phylogenetic study of Acetes proposed that the Acetes clade without A. petrunkevitchi (former Peisos petrunkevitchi) never gained robust support, thus considering Peisos as a junior synonym of Acetes (Vereshchaka, Lunina & Olesen, 2016a). Although we carried out an analysis with a robust but limited number of samples, our results confirmed the phylogenetic positioning recovered by both 16S rRNA and COI genes and indicated that A. petrunkevitchi is part of the Acetes group, but in all analyses forming a single clade of “Acetes paraguayensis + A. petrunkevitchi”, with a high support value for 16S rRNA and concatenated data. As only mitochondrial genes were used, further studies adding nuclear markers should be carried out to test the topology recovered herein.

However, recent studies about the morphology of the male reproductive system and spermatophore of A. petrunkevitchi differed from those of A. americanus, A. marinus and A. paraguayensis, which remains open the discussion about the inclusion of Peisos in the Acetes group (Salti, 2020).

The global morphological phylogeny proposed for the superfamily Sergestoidea showed significant changes in taxonomy, with the description of three new families, particularly because the family Sergestidae was not considered monophyletic (Vereshchaka, 2017). As a result, A. americanus was classified in the new family Acetidae, as proposed by Vereshchaka (2017). The author proposed that a single clade of A. marinus and A. paraguayensis within Acetidae received high bootstrap support. Our concatenated molecular phylogeny indicated a close relationship between A. paraguayensis and A. petrunkevitchi in a separate clade, as mentioned above. Additional samples of these species, including A. marinus, confirm this hypothesis. This molecular reconstruction sheds light on the unsolved evolutionary relations between the species of the genus Acetes, which should be investigated using more comprehensive integrated studies and the addition of nuclear markers.