Construction of cDNA libraries and EST analysis

Total RNA was independently isolated from tissues (head, abdomen, thoracopods and photophores) dissected from specimens of E. superba collected at five different time points during 24 hours. The analysis of total RNA samples, performed by capillary electrophoresis, showed absence of genomic DNA contamination and a peculiar electropherogram as shown in Fig. 1. In particular, total RNAs show low molecular weight from 200 bp to 1 kb, perhaps as result of a partial RNA degradation.

Five independent tissue-specific cDNA libraries, named K01 and K05 (head), K06 (abdomen), K07 (photophores) and K09 (thoracopods), were produced from total RNA pools. For head we have sequenced only K05 cDNA library because it presented more recombinant clones compared to K01. Recombinant bacterial clones from each library were randomly picked and the EST were sequenced from the 5'-end. The average insert size for all libraries was estimated to be 412 bp.

EST assembly and construction of an Antarctic krill transcript catalogue

The number of ESTs in each cluster varies from 2 (93 clusters) to 88 (1 cluster). The average length of a cluster is 436 bp with the longest assembled sequence being 1,417 bp and the shortest 153 bp. The 91% of cluster consensus contains the 3'-end region of mRNAs as demonstrated by the presence of a poly-adenylation signal.

Each non redundant sequence was searched in the nucleotides database and UniProtKB database using Blast-N and Blast-X with an e-value cut off of < e^-40 and < e^-10, respectively. These values were empirically chosen considering the low amount of sequences data available for Euphausiacea and similar shrimp species and the need of stringency in providing a reliable catalogue of Antarctic krill genes. All annotations were further manually examined, in order to assign the best describing text to the correspondent cluster.

Overall, 70% of non-redundant sequences (708 out of 1,017), identified by about 50% of total produced ESTs, showed no or poor similarity matches, and they probably represent completely unknown Antarctic krill transcripts that could be characterized in future studies. Additional file 1 lists the 309 (30%) non-redundant sequences identifying known Euphausia genes or sequences showing significant similarity to genes from arthropods (63.4%) and other species (36.6%), such as Homarus americanus (3.2%), Aedes aegypti (12.0%), Drosophila melanogaster (3.9%), Bombyx mori (3.6%), Brachydanio rerio (4.9%), Mus musculus (3.9%), Rattus norvegicus (3.9%), Homo sapiens (4.5%). Antarctic krill sequences generally show a greater similarity to genes of insects (about 36%) than to genes of Malacostraca (about 23%) and only 7% were similar to the known sequences of Euphausiacea. This result could be due to the limited number of Malacostraca gene and protein sequences available in the public databases (gene sequences: 116,666 and protein sequences: 11,941 at November 2007) respect to insects (gene sequences: 1,869,511 and protein sequences: 336,246 at November 2007). The total collection of 1,770 E. superba 3'-EST has been deposited in the EBI-GenBank-DBJ database (GenBank accession numbers from ES542703 to ES544472).

One of the most noticeable features of our EST catalogue is that mitochondrial transcripts are quite abundant (about 10% of the total ESTs): 84 ESTs (about 4.7%) matched with nine different E. superba mitochondrial transcripts (ATPase6, COI, COII, COIII, cyt b, ND1, ND2, ND3, ND6) and about 5% ESTs (88 out of 1,770) identified the large mitochondrial 16S rRNA gene. In future experiments, to avoid the repetitive sequencing of this abundant mRNA, we plan to introduce, during cDNA insert amplification, interference primers specifically designed for the 16S rRNA [17]. Moreover, we found a small percentage of ESTs (1.6%), showing similarity with 18S and 28S rRNA. About 2% of ribosomal RNAs contamination is common to other systematic sequencing projects.

The E. superba mitochondrial gene sequences are very similar to those commonly found in the mitochondrial genomes of other arthropods, including 13 protein-coding genes, 19 tRNA and 2 rRNA genes [10]. Machida et al. [10] have demonstrated that mitochondrial protein-coding genes are transcribed from the same DNA strand, left to right, except for ND1, ND4L, ND4 and ND5 genes, while two ribosomal RNA genes are encoded by L strands.

We did not find highly represented tissue-specific mRNAs, while some ribosomal proteins like L22 (31 ESTs), S25 (31), P1 (16), L37A (15), L24 (13), P2 (13), S14 (9), S14 (10), L34 (9) were expressed at the same level in all analyzed krill tissues. The only genuine tissue-specific mRNA appears to be myosin light chain (8 ESTs, cluster KRC00032) that is highly represented in a strictly committed tissue such as skeletal muscle localized in abdomen (library K06). Moreover, we found 3 unknown transcripts among the 20 most expressed genes: KRC00118 (10 ESTs), KRC00407 (9 ESTs) and KRC00101 (7 ESTs) that could be interesting for future functional studies.

Since very few abundant transcripts were found that could hamper the identification of rare transcripts, it seems plausible that random sequencing of our Antarctic krill libraries would continue to represent an effective strategy for identifying novel E. superba mRNAs.

Functional categorization of E. superba ESTs

Other interesting krill transcripts that we were able to annotate are those involved in stress responses, proteolysis and immunoresponse (3.24%) like Hsp90, chaperones, cathepsine L-like cysteine protease, a lysosomal cysteine proteinase [22], peptidyl-prolyl cis-trans isomerase A1 (cyclophilin A1) and peptidylprolyl isomerase B (cyclophilin B). Cyclophilines are members of the immunophilin protein family, which play a role in immunoregulation and basic cellular processes involving protein folding and trafficking. ESTs with good similarity to hemocyanin are present in our collection: this protein has been recently reported to have antifungal and antiviral activities [23, 24].

Some krill ESTs identify histone 2A (KRC00431) and histone 3.3A (KRC00024) indicating the presence of unexpected polyadenylated histone transcripts displaying the polyadenylation signal and tail. In vertebrates, these evolutionary conserved housekeeping mRNAs are not polyadenylated, and this has been related to the high turnover of these transcripts in the dividing cells. Interestingly, polyadenylated H2A and H3 histone sequences were detected also in the systematic sequencing of 3'-end cDNA libraries obtained from brain and kidney of channel catfish Ictarulus punctatus [25, 26] and from various tissues (haemolymph, gills, digestive glands, mantles and adductor muscles) of the mussel Mytilus galloprovincialis [27]. The presence of polyadenylation signals in E. superba histone transcripts deserves a more detailed analysis. In fact, Eirin-Lopez et al. [28] have recently shown that all histone genes in the repetitive unit are characterized by two different mRNA termination signals in their 3' UTR: the typical stem-loop or hairpin-loop signal followed by a purine-rich element and a polyadenylation signal AATAAA located downstream to this last element. The presence of a double mRNA termination signal is unique to histone genes and common for other invertebrates such as Chironomus thummi [29], D. melanogaster [30], Chaetopterus variopedatus [31], M. galloprovincialis [32] and Crustacea [33]. Although in some invertebrates core histone transcripts (H2A, H2B, H3 and H4) include polyA tails, these sequences are among the most evolutionary conserved eukaryotic proteins [34].

Transcriptional signature of E. superba tissues

Gene expression profiling depends on the functional specificity of cells composing different tissues. So, the systematic sequencing of EST from unbiased cDNA libraries is a suitable approach for analyzing the gene expression profile of a given tissue [35]. In fact, the frequency of a given EST in the cDNA library can be related to the relative abundance of the corresponding mRNA in the source tissue.

To define tissue transcriptional signatures of E. superba, annotated ESTs obtained from the four tissue-specific cDNA libraries (head, abdomen, photophores, thoracopods) were separately grouped in 13 functional categories (Table 5) a further abundant category was created for those ESTs to which no function may be yet associated. Fig. 2 shows four different diagrams standing for ESTs distribution among functional categories in each cDNA library. The presence of highly represented functional categories is peculiar of strictly committed tissues such as abdomen and thoracopods in which transcripts involved in striated muscle contraction are very abundant (about 26% in abdomen and 7% in thoracopods). In the abdomen library, we were able to recognize the principal structural components of the sarcomeric contractile machinery (myosin heavy chain, myosin light chain 1, myosin-2, actin, alpha-tubulin, tropomyosin) and two subunits of the troponin complex (troponin T, troponin I), a key regulator of muscle contraction. About 10% of sequences produced from head and thoracopods libraries fall in functional categories related to metabolic processes (amino acid, fatty acid and carbohydrate metabolism). Interestingly, about 6% and 4% of ESTs respectively sequenced from head and thoracopods libraries identified structural constituents of cuticle (arthrodial cuticle protein AMP16.3, arthrodial cuticle protein AMP1A, calcification-associated peptide-1 precursor). This reflects the presence of cuticle traces in the head and thoracopods samples. In photophores and thoracopods transcripts displaying putative identity with ribosomal sequences are more abundant compared to other tissues (55% and 46%, respectively), indicating a relevant activity of the translation machinery.

#	Process	Head		Abdomen		Photophores		Thoracopods
		# ESTs	%	# ESTs	%	# ESTs	%	# ESTs	%
1	DNA replication	3	1.43	2	0.90	2	0.88	1	0.45
2	Transcription	3	1.43	6	2.71	5	2.19	2	0.90
3	Translation	69	32.86	65	29.41	126	55.26	103	46.40
4	Transport	5	2.38	12	5.43	7	3.07	8	3.60
5	Metabolic process	22	10.48	16	7.24	14	6.14	21	9.46
6	Proteolysis, protein folding and modification	12	5.71	6	2.71	14	6.14	16	7.21
7	Striated muscle contraction	11	5.24	57	25.79	12	5.26	15	6.76
8	Signal transduction	4	1.90	1	0.45	3	1.32	1	0.45
9	Structural constituent of cuticle	13	6.19	1	0.45	7	3.07	9	4.05
10	Cell growth, proliferation and adhesion	2	0.95	3	1.36	2	0.88	5	2.25
11	Ion Binding	4	1.90	0	0.00	1	0.44	0	0.00
12	Hypotetical protein	4	1.90	6	2.71	6	2.63	0	0.00
13	Mitochondrial genes	58	27.62	46	20.81	29	12.72	41	18.47
	Total	210	100.00	221	100.00	228	100.00	222	100.0

Table showing the number and the percentage of ESTs in each functional category for each tissue specific library

We have also identified from the head cDNA library a novel opsin sequence (ID ESTs: KRC00735, KRC00802), a light-sensitive membrane-bound G protein-coupled receptors mediating the conversion of a photon of light into an electrochemical signal in the visual transduction cascade. In insects there are at least four main spectral classes: long-wavelenght-sensitive (LWS), middle-wavelenght-sensitive (MWS) and two short-wavelenght-sensitive (SWS) groups. The opsin sequences available for E. superba (GenBank accession no. DQ852576–DQ852580) show a spectral sensitivity with short wavelength (496–501 nm, λ_max = 487) and cannot be aligned with our consensus [36].

Quantitative RT-PCR analysis was performed to quantify and validate the expression level of some genes presenting different EST countings in krill tissues. We selected ten genes (compound eye opsin BCRH1, myosin light chain, myosin heavy chain, arthrodial cuticle protein AMP16.3, tail muscle elongation factor 1 gamma, cellular retinoic acid/retinol binding protein, eukaryotic initiation factor 4A, transport protein SEC61 subunit gamma, chromodomain helicase DNA binding protein and voltage-dependent calcium channel) representative of different levels of transcript abundance.

The housekeeping gene 18S rRNA was used as endogenous control. As reported in the Additional file 2, the expression values obtained with the quantitative RT-PCR for the tested transcripts were in agreement with the EST counting in the four libraries. In particular, we have demonstrated that the compound eye opsin is strongly expressed in head compared to other tissues and myosin light chain and myosin heavy chain are highly expressed in abdomen and thoracopods confirming their key role in the contractile machinery. Instead, the eukaryotic initiation factor 4A is expressed at about the same level in all tested tissues.

Identification of microsatellite-containing ESTs

Among the 1,017 non-redundant sequences examined in this study, 41 (4%) consensus sequences containing ESTs were identified by using MISA software. Twelve of these consensus present 2 distinct simple sequence repeats interrupted by more than 100 bp for a total of 69 identified microsatellites (SSR). The majority of these sequences (72%) fall into the 3 bp repeat type class with a preponderance of GAA and GAT. After a manual inspection of redundancy, raw sequence, data quality and the presence of sufficient flanking sequences we designed 9 pairs of specific PCR primers. We obtained successful amplifications for 6 of these 9 pairs of primers. Assessment of polymorphism information content (PIC), observed and expected heterozygosity and other population genetics analysis will be performed in the near future. These markers will increase the currently available Euphausiacea SSR markers. In fact, only five microsatellite loci isolated from the northern krill Meganyctiphanes norvegica have been reported so far [37]. Since our novel microsatellite markers were developed on the basis of expressed sequences and they are presumably conserved across other Euphausiacea species, they could also be useful for comparative mapping and for a molecular approach to Antarctic krill ecology.

Tissues samples, RNA extraction and quality control

Antarctic krill (Euphausia superba) were fished from the Ross Sea (longitude: 167°28'81" E – 179°54'68" W, latitude 68°40'54" S – 77°01"81" S) in the January 2004 during the XIX Italian Antarctic Expedition. Specimens were collected at different time of the day (01:00, 06:00, 10:00, 15:00, 18:00), over a complete 24-hour cycle. Samples were frozen at -40°C in RNA stabilization solution (RNA-later, Ambion). For each fishing, selected tissues (head including compound eyes and brain, abdomen, thoracopods and photophores) from five animals were dissected individually in RNA later ice solution (Ambion). After dissection, tissues were rapidly rinsed in sterile water, weighed, frozen in Trizol reagent (Invitrogen) and stored at -80°C. A large excess of Trizol (15 ml for 0.5–1.5 g. of sample) was used in order to prevent RNA degradation by endogenous RNAse. Frozen tissues were minced and homogenized for 3–5 min using an ultra-turrax-T8.01 blender (IKA-Werke). Total RNA was isolated using the Trizol reagent (Invitrogen) following the manufacturer's instruction and further purified with LiCl in order to remove glucidic contaminants. All RNA samples were checked for quality by capillary electrophoresis (RNA 6000 Nano LabChip, Agilent Bioanalyzer 2100, Agilent Technologies). For each tissue (head, abdomen, thoracopods and photophores), equal amounts of total RNA (2 μg) extracted from every collection were pooled.

Construction of cDNA libraries

Five independent cDNA libraries, named K01 and K05 (head), K06 (abdomen), K07 (photophores) and K09 (thoracopods), were constructed.

We have developed a new method using a combination of SMART protocol (Clontech), ensuring almost full-length cDNA, and Gateway technology (Invitrogen), allowing unidirectional cloning without enzymatic digestion. In this protocol, only fully-transcribed first strand cDNA (ss cDNA) is tagged with a short sequence complementary to a modified SMART oligo (Fig. 3). The SMART oligo sequence (SMART-16attB1-T3: 5'-TACAAAAAAGCAGGCTAATTAACCCTCACTAAAGGG-3') and the overhang of the oligo(dT) primer (5'-GGGGACCACTTTGTACAAGAAAGCTGGGCGGCCGC [dT]₂₀VN-3') used for first strand synthesis include an attB1 and attB2 recombination site respectively.

First strand cDNA synthesis was performed from 1.5 μg of total RNA in a 15 μl reaction. Then, the reaction was then diluted 1:5 ratio and incubated at 72°C for 2 min. Second strand reaction mix was added to 1 μl of diluted first strand cDNA to give a final concentration of 1× BD Advantage 2 PCR reaction buffer (Clontech), 0.2 mM dNTPs, 120 nM primers (attB1-8T3: 5'-GGGGACAAGTTTGTACAAAAAAGCAGGCTAATTAACC-3' and attB2: 5'-GGGGACCACTTTGTACAAGAAAGCTGGG-3') and 1× of Advantage2 DNA polymerase mix (Clontech) in a volume of 50 μl. This second strand reaction mixture was incubated for 21 cycles of 15 sec 95°C, 30 sec 66°C and 3 min 68°C. Only those ss cDNAs having a SMART anchor sequence at the 5' end were used as template and exponentially amplified. The second strand reaction was glass fibre column purified and cDNA was size selected by Sepharose CL-4B SPUN COLUMN (GE Healthcare). The cDNA was inserted in the cloning vector by a recombination reaction performed at 25°C for 18 h with about 35 ng of attB-cDNA, 150 ng of pDONR221 (Invitrogen) and 2 μl of BP Clonase II (Invitrogen) in 10 μl final volume. 1/5 of the purified reaction was used to transform electrocompetent DH10B E. coli cells. Recombinant colonies were selected on agar SOB medium plus kanamicin. Individual library colonies were arrayed by manual picking on 96 well plates in liquid selective SOB medium plus 7.5% of glycerol for independent growth [40].

DNA sequencing

After lysis of the bacterial colonies, cDNA inserts were directly amplified with universal primers M13 forward (5'-TGTAAAACGACGGCCAGTCTTA-3') and M13 reverse (5'-CAGGAAACAGCTATGACCATGT-3'). The polymerase chain reaction (PCR) profile consisted of: (1) initial denaturation for 5 min at 95°C, (2) 35 cycles of 40-s denaturation at 95°C, 40-s annealing at 60°C and 1-min elongation at 72°C and (3) final extension for 5 min at 72°C; samples with a size over 0.5 Kb were selected for sequencing. Single pass DNA sequencing from plasmids was performed by using the vector specific primer attB1_seq (5'-CTTTGTACAAAAAAGCAGGCT-3') and a modified Sanger dideoxy terminator cycle sequencing chemistry, the ABI BigDye kit version 3.1, on a ABI 3730 48-capillary sequencer and 36 cm capillaries (Sequencing Service of Max-Plank Institute for Molecular Genetics, Berlin, Germany).

Computer Management of Data

Trace2dbest and Partigene [41] were used to process chromatograms, clusterize sequences, and build an annotation database. Trace2dbest extracts sequences and quality information from traces (Phred algorithm), removes vector contamination and poly(A), and performs the trimming of low quality sequences. Sequences shorter than 100 bp were discarded. Partigene reads all sequence files and performs an assembling process in two step: 1) CLOBB software [42] clusterizes sequences on the basis of BLAST similarity; 2) Phrap [43] makes a consensus from each cluster.

Each consensus, converted in FASTA format, was searched locally in nucleotides database, downloaded from NCBI [44] and UniProtKB database [45], using Blast-N and Blast-X, respectively. First 10 HSPs (High Scoring Pair) from each blast result were collected and stored in a local PostgreSQL table, as a collection of automatic annotations.

Each cluster annotation in our database was further manually examined to assign the best describing text to the correspondent cluster: matches with expectations values greater than e^-10 for protein (Blast-X) and e-⁴⁰ for nucleotide (Blast-N) were considered as poorly informative. Moreover, for each UniProt ID, taken from Blast-X description field, we associated specific Gene Ontology annotation, that integrates information about process, function, and component. Clusters, consensus and related similarity and gene ontology searches were electronically organized and stored in a dedicated PostgreSQL database.

Identification of microsatellite containing ESTs

The unique sequences were screened for microsatellites by using the MISA software [46]. Only di-, tri-, tetra-, penta- and esanucleotide repeats were targeted, since mononucleotide repeats are not useful for mapping or population genetics due to difficulties in their genotyping. Strings of oligonucleotide sequences were used to search for microsatellites: 6 repeats for dinucleotide; 5 repeats for trinucleotide; 5 repeats for tetranucleotide, pentanucleotide and esanucleotide. Primers were designed for the flanking regions of the SSR using a web-based software, "Primer3" [47], and based on the criteria of 50% GC content, a minimum melting temperature of 55°C, and absence of secondary structure. Primers ranged from 18–27 nucleotides in length and amplified products of 150–390 bp. The primers were synthesized with a 5'-KS-tail (KS sequence: 5'-cgaggtcgacggtatcg-3') allowing to amplify microsatellite alleles in combination with a 5'-fluorescent-labeled KS primer [48].

Quantitative RT-PCR

Quantitative RT-PCR was conducted for some genes using the same tissues tested (head, abdomen, thoracopods and photophores) to confirm the integrity and robustness of EST sequencing.

Three μg of total RNA from each tissue was used to perform three independent cDNA syntheses in a final volume of 10 μl, using random decamers and SuperScript II reverse transcriptase (Invitrogen). 1 μl aliquot of diluted first-strand cDNA was PCR amplified in 10 μl volume using SYBR Green chemistry, according to the manufacturer's recommendations (Finnenzymes). Gene-specific primers were designed using Primer Express^® Software (Applera) to amplify fragments of 120–180 bp in length, close to the 3'-end of the transcript. To avoid the amplification of contaminant genomic DNA, we treated total RNA samples with DNase I (Qiagen). The dissociation curve was used to confirm the specificity of the amplicon. PCR reactions were performed in a 7500 Real-Time PCR System (Applied Biosystems). Thermal cycling conditions were as follows: 15 min denaturation at 95°C; followed by 40 cycles of 30 sec denaturation step at 95°C, annealing and elongation steps for 1 min each at 60°C and a final 3 min elongation at 72°C. To evaluate differences in gene expression a relative quantification method was chosen where the expression of the target gene is standardized by a non-regulated reference gene; consequently, three replicates of each sample and endogenous control were amplified. 18S rRNA was used as an endogenous control because the level of rRNA remains essentially constant from sample to sample (QuantumRNA™ 18S Internal Standards, Ambion). To calculate the relative expression ratio, the 2^-ΔΔCt (RQ, relative quantification) method implemented in the 7500 Real Time PCR System software [49] was used. This method determines the change in expression of a nucleic acid sequence (target) in a test sample relative to the same sequence in a calibrator sample. In our experiments, the expression of ten targets was tested (compound eye opsin BCRH1, myosin light chain, myosin heavy chain, arthrodial cuticle protein AMP16.3, tail muscle elongation factor 1 gamma, cellular retinoic acid/retinol binding protein, eukaryotic initiation factor 4A, transport protein SEC61 subunit gamma, chromodomain helicase DNA binding protein and voltage-dependent calcium channel) which displayed differential expression in the head, thoracopods and photophores, compared with the abdomen.