Dunaliella was originally described by Teodoresco 1905. Since then, taxonomic studies among Dunaliella have identified several new species. However, even today differentiation among halophilic and carotenogenic Dunaliella species in both green and red stages is difficult and time consuming. In addition, some strains and species in culture collections are misnamed and have given rise to unnecessary strains and species names 1. This confusion of strains and species names makes comparison of results by different authors difficult 2. D. salina and D. salina/bardawil are the only reported β-carotene hyperproducer species of the genus (at least 10%), that grows in salt saturated lagoons 3. However, a controversy still exists about identification of D. salina/bardawil as different species of D. salina 2. Molecular identification provides a useful tool to distinguish between inter and intra-specific morphologically similar species 45 and mixed populations 67. Intra-species identification of community members without cultivation, avoid some selective biases associated with pure culture methods 8. Specie-specific oligonucleotides could be useful to identify either species from culture collections or from natural environments 910. Slight phylogenetic and taxonomic differences in Dunaliella species can conceal profound differences in their potential for production of metabolites such as carotenoids 4. To study these organisms one must discriminate them from their presumed non carotenogenic congeners and from among species with similar morphological features 11.

D. salina strains have been found around the world (Australia, Chile, India, Israel, México, etc), however, D. salina/bardawil classified by Dr. Ben-Amotz has not been reported from other places than Bardawil lagoon. In this work, using intron-sizing method, we demonstrated that each hyperproducer species has an exclusive 18S rDNA fingerprint profile. D. salina/bardawil species with the same 18S rDNA fingerprint and same amount of β-carotene was found in salt saturated lagoons in Baja California Mexico (Fig. 1). Non β-carotene hyperproducer species of Dunaliella were easily discriminated with our methodology, even in the green stage, avoiding long cleaning, purification and growing process. The intron-sizing method provides a novel and very powerful DNA-fingerprinting technique to accomplish a specific, rapid and sensitive identification of carotenogenic Dunaliella species. In this work we demonstrated that auxiliary methods, such as the one proposed, could be useful obtaining well characterized and certified international culture collection.

Microscopic examination of Dunaliella salina 19/18 sample showed two different species of Dunaliella; one red and one green. Molecular fingerprinting determination using MA1–MA2 18S rDNA conserved primers from 19/18 DNA sample, gave two PCR products: one band of ~2100 bp that belongs to β-carotene hyperproducer species of D. salina, and a second product of ~1700 bp from Dunaliella species that never turns red (Fig. 2). These results agree with microscopic determination where two species of Dunaliella were observed (data not shown). Furthermore, using 19/18 DNA sample and DSs-MA2 specific primers we obtained a PCR product of ~700 bp from red species and no band from green species (Fig. 2). With these results we established the fingerprinting profile of 19/18 β-carotene hyperproducer strain of D. salina as (MA1–MA2 = ~2100 bp/DSs-MA2 = ~700 bp). Interestingly, the same fingerprinting was shown by D. salina M84320 isolated from Chile and reported by Wilcox and coworkers 12. Furthermore, D. salina BCO2 strain isolated in Mexico, also presented the same 18S rDNA fingerprint 513 as did D. salina strain found in India by Raja and coworkers 14. A common characteristic from these D. salina strains in addition to their fingerprinting profile was their β-carotene hyperproduction capacity.

Taking into account that D. salina strains isolated from different continents presented the same fingerprinting profile and β-carotene levels, we can be assured that the intron-sizing method developed by our group is trustworthy, rapid and sensitive, for specifically identifying these hyperproducer strains of D. salina. We can conclude that D. salina strains with the following fingerprinting profile of MA1–MA2 = ~2100 bp and DSs-MA2 = ~700 bp, belong to the same species, are distributed worldwide and their β-carotene hyperproduction capacity is well conserved. Thus, in order to facilitate identification, we suggest a sub-classification calling these strains as "D. salina var Teod".

On the other hand, both species classified as D. salina/bardawil purchased from UTEX (LB2538) ten years ago and Dunaliella strain isolated from Baja Mexico (this work), turned red and reached 10% dry weight of β-carotene. In addition, both strains amplified a PCR product of ~2500 bp using MA1–MA2 primers (Fig. 3). Specific 18S rDNA fingerprints showed a PCR product of ~1000 bp using DBs-MA2 primers in both strains and no bands were amplified with DSs-MA2 (Fig. 3). With these results we established the fingerprinting profile of D. salina/bardawil as MA1–MA2 = ~2500 bp/DBs-MA2 = ~1000 bp. D. salina/bardawil species has not been reported from other places than Bardawil lagoon. In this sense, a controversy still exists about the authenticity of D. salina/bardawil as new species 2. In addition, microscopic differentiation between D. salina and D. salina/bardawil species is difficult and time consuming. However, molecularly, D. salina/bardawil has an exclusive fingerprinting profile of MA1–MA2 = ~2500 bp/DBs-MA2 = ~1000 bp, different from the one presented by "D. salina var Teod" (MA1–MA2 = ~2100/DSs-MA2 = ~700). Certainly, more work must be done by taxonomic experts to definitively classify D. salina/bardawil as "D. bardawil" or as "D. salina var Bardal" or with another name. However, our methodology works well to differentiate between the two most important carotenogenic species of Dunaliella and to make a rapid and precise identification of them. Most importantly, this methodology is helpful in the beginning of the isolation process even in the green stage. This information is important both for commercial and scientific applications.

Recently, the National Center for Biotechnology Information (NCBI) has initiated a D. salina genome sequencing project (ID: 18607). In this sense, it is important to know specific identity of target species, particularly if several groups from several countries are involved in this ambitious and time consuming task. Confidence knowing specific identity of target species can avoid wasting time and money.

Additionally, demands for natural products are growing in importance and β-carotene is not an exception 115. Rapid and specific identification and certification of commercial strains or species can improve production yields and help in the competitiveness by avoiding contamination with undesired species. Cross-contamination even with other hyperproducer strains still can affect the productivity due different environmental and nutritional conditions required for each strain. As an example, in May of 2008 we purchased LB2538 strain from UTEX again, but surprisingly, a fingerprinting profile of "D. salina var Teod" was obtained (MA1–MA2 = ~2100, DSs-MA2 = ~700) instead of the fingerprinting profile of D. salina/bardawil (MA1–MA2 = ~2500, DBs-MA2 = ~1000) which we were expecting (data not shown).

Moreover, contamination with non hyperproducer strains is more dangerous because these strains grow faster than hyperproducers, due to the fact that there are no metabolic efforts required for β-carotene production. This was the case for 19/18 samples where hyperproducer and non hyperproducer species of Dunaliella were identified by our fingerprinting methodology (Fig. 2).

D. salina 19/30 supposedly was the strain isolated by Dr. Ben-Amotz (D. salina/bardawil), however PCR amplification with MA1–MA2 presented neither the ~2500 bp product associated with the original D. salina/bardawil species 11, nor the ~2100 bp amplification associated with D. salina var Teod species. Nevertheless, the 19/30 strain amplified a PCR product of ~1700 bp using MA1–MA2 conserved primers (Fig. 3) and the species itself never turns red, indicating β-carotene is not hyperproduced by this strain. In addition, Gene Bank had reported a 19/30 strain with ~1700 bp 18S rDNA gene size (EF473749.1) and another 19/30 strain was reported with ~2100 bp 18S rDNA gene size (DQ447648.1). The first report comes from Italy and the second comes from France. This is another good example of cross-contamination that researchers and even certified cultured collection banks can undergo, due to almost indistinguishable taxonomic characteristics between Dunaliella species.

Auxiliary methods like the one we are proposing could be useful for characterizing and certifying international culture collections as well. Taxonomic characterization is important to identify and differentiate species; however, differentiation between subspecies, ecotypes or similar species with taxonomy tools is a difficult and time consuming task and sometimes impossible. For this reason, the 18S rDNA intron-sizing method proposed in this work provides a novel and powerful DNA-fingerprinting methodology to accomplish a specific, rapid and sensitive identification of carotenogenic Dunaliella species.

On the other hand, D. salina LB1644 strain obtained from UTEX did turn to red; however it produced only 1% dry weight of β-carotene. LB1644 strain amplified a ~1700 bp PCR product with MA1–MA2 conserved primers as reported by our group 5 and recently by gene bank (DQ009765.1), but did not amplify any band with DSs-MA2 or DBs-MA2. All results; 18S rDNA gene size, fingerprinting profile and pigment production capacity, corroborated that LB1644 strain it is not "D. salina var Teod" n or "D. salina/bardawil". In this case, more taxonomic and molecular research must be done to identify this Dunaliella specie. Absence of introns within 18S rDNA gene is an inconvenience for Dunaliella species fingerprinting determination and identification. Fortunately, all hyperproducer species known carry I or II introns within the 18S rDNA gene. Gómez and González 4, have been working with ITS sequences and some interesting results have been presented that could help in primer design. In addition, it seems to be the rule for hyperproducer species of Dunaliella to have introns within 18S rDNA, thus "D. salina var Teod" has I, D. salina/bardawil has II and D. parva has II 11.

However, we did find a non carotenogenic strain in hypersaline environments that has two introns within the 18S rDNA (data not shown). According to literature this strain could be D. viridis. Gene Bank sequence analysis showed a D. viridis strain (DQ009776) with an 18S rDNA gene size of ~2500 bp approximately, which corresponds to the size of the strain that we found. Species-specific primer design must be done to molecularly characterize this strain and, wheatear it is possible accomplish a fingerprint profile of it.

Using our intron-sizing method we demonstrated that "D. salina var Teod" and D. salina/bardawil are two different species of Dunaliella that could easily be differentiated by size and sequence of the 18S rDNA gene (Fig. 5). These two species of Dunaliella produce nearly same amounts and profile of carotenoids, but they are difficult to distinguish phenotypically. More work in taxonomic identification must be done to classify D. salina/bardawil as D. bardawil as "D. salina var Bardal" or with another name.

Table 1, shows three groups of Dunaliella strains. The first group belongs to Dunaliella with two introns and approximately 2500 bp 18S rDNA gene size. Four strains had species-specific sequence differentiation between them and as a consequence a species-specific fingerprinting profile for each one (data shown only for D. salina/bardawil). Table 1 also shows two groups of Dunaliella strains with approximately 2100 bp 18S rDNA gene size. According to our results strains with that size contain only one intron within the 18S rDNA gene, and are classified as D. salina strains. Group I D. salina strains (M84320 group) were called "D. salina var Teod", because all strains share 100% sequence conservation with DSs specific oligonucleotide, growth on hypersaline environments and hyperproduce β-carotene. In addition, these strains had been found around the world and all share the same fingerprinting profile (see results and discussion section).

A second subgroup of D. salina strains called group II (DQ324002 group), was identified from Gene Bank sequence analysis (data not shown). More work must be done to known their β-carotene production capacity. Additionally, even when strains of group II had the same 18S rDNA gene size, as group I did, they only maintain 66% sequence conservation with "D. salina var Teod" specific oligonucleotide (DSs).

On the other hand, D. salina/bardawil obtained ten years ago from UTEX it is not the same strain found presently in the culture collection bank. However, a D. salina/bardawil strain with the original fingerprinting profile was found in San Quintín, B.C. Mexico ten years ago and also during this work.

It is important to point out that although culture collection banks may have some cross-contamination, as do collections of some researchers around the world (results and discussion), it is not our intention to point out cross contamination or incorrect affiliation of Dunaliella species. Nevertheless, we want to provide an auxiliary method for rapid, precise and specific identification of Dunaliella species.

Since our first report in 2002 on differentiation of carotenogenic Dunaliella species by the introns sizing method 11, we have refined, corroborated and deeply advanced the molecular differentiation of Dunaliella species by the fingerprinting profile method. Actually, it is the only method that has been tested on both collection bank strains as well as field samples, that allows proper identification of the physiologically plastic strains of Dunaliella at the level of species.

The authors declare that they have no competing interests.

JOS and JPM conceived the study, performed the analysis and wrote the manuscript. LOS sampled and obtained isolates. RCF performed some experimental work.