Review TRENDS in Biotechnology Vol.23 No.7 July 2005

Food and forensic molecular identification: update and challenges

Fabrice Teletchea, Celia Maudet and Catherine Ha¨ nni

Centre de Ge´ne´ tique Mole´ culaire et Cellulaire, CNRS UMR5534, UCB-Lyon I, 16, Rue Raphael Dubois, 69622 Villeurbanne Cedex, France

The need for accurate and reliable methods for animal development of these methods should protect both species identification has steadily increased during past consumers and producers from frauds, and protect animal decades, particularly with the recent food scares and the species from over-exploitation or illegal trafficking. Mol- overall crisis of biodiversity primarily resulting from the ecular authentication or molecular traceability, which is huge ongoing illegal traffic of endangered species. A basedonthePCRamplificationofDNA,hasbeen relatively new biotechnological field, known as species developed in the past ten years and offers promising molecular identification, based on the amplification and solutions for these issues. Furthermore, this field will analysis of DNA, offers promising solutions. Indeed, probably experience tremendous development because: despite the fact that retrieval and analysis of DNA in (i) Most DNA methods developed have proved useful on processed products is a real challenge, numerous almost all organic substrates and will certainly technically consistent methods are now available and become the new legal standards for identification. allow the detection of animal species in almost any (ii) More regulations have introduced safety standards organic substrate. However, this field is currently facing into the chain of food production (e.g. European a turning point and should rely more on knowledge Regulations such as the 2000/104/EC, which estab- primarily from three fundamental fields – paleogenetics, lishes that fish products can enter the commercial molecular evolution and systematics. circuit only if the commercial name, method of production and capture area are clearly labelled or the 2002/1774/EC, which bans the intra-species recycling of animal by-products). Introduction (iii) During the past decade, molecular identification tests Recent food scares (e.g. BSE, avian flu, foot-and-mouth have only been developed for a few species but it is disease, etc.), malpractices of some food producers, likely that this number will steadily increase, religious reasons, food allergies and GMOS have tremen- particularly among fish (Box 1). dously reinforced public awareness regarding the compo- Therefore, it is now crucial to reassess the different sition of food products. However, because labels do not molecular methods available for animal species identifi- provide sufficient guarantee about the true contents of a cation, particularly in light of three fundamentals fields: product, it is necessary to identify and/or authenticate the paleogenetics, molecular evolution and systematics. We components of processed food, thus protecting both are convinced that these three fields could improve the consumers and producers from illegal substitutions [1]. potential of analysing DNA in degraded substrates, help In addition, trade of endangered species has contrib- to choose the most appropriate molecular markers and uted to severe depletion of biodiversity. Approximately highlight some common problems encountered in sys- 10–20% of all vertebrates and plant species are at risk of tematics that could result in erroneous identification. extinction over the next few decades (IUCN; http://www. Here we concentrate mainly on species identification and redlist.org and CITES http://www.cites.org). Wildlife and not in the recognition of distinct populations of the same their products represent the third greatest illegal traffic species because the latter question requires different after drugs and arms [2] and one of the most serious concepts at some point. threats to the survival of animal populations is poaching. Each year, millions of endangered animals are illegally killed or captured for private zoo collections, hunt DNA from food and forensic samples trophies, ornamental objects (e.g. elephant ivory [3]), Fresh food products or forensic samples without proces- human consumption (e.g. sea turtles and their eggs [4])or sing are suitable for many types of molecular or protein traditional medicine (e.g. tiger [5,6], rhinoceros [7]). analyses (traditional biochemical approaches based on Food authentication and protection of biodiversity both proteins used either electrophoretic, chromatographic or require reliable and accurate methods for determining, immunological techniques; reviewed in [8]). Unfortu- without ambiguity, the animal species in a wide array nately, because most foodstuffs and forensic samples are of degraded and processed substrates (Table 1). The processed, DNA is usually altered. However, several

Corresponding author: Ha¨nni, C. ([email protected]). research fields had already worked with such DNA: Available online 31 May 2005 ancient DNA studies or paleogenetics (studying DNA www.sciencedirect.com 0167-7799/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibtech.2005.05.006 360 Review TRENDS in Biotechnology Vol.23 No.7 July 2005

Table 1. Examples of forensic or food substrates for molecular analysis Samples analyzed Species targeted Extraction methods Fragment targeted Fragment size (bp) Refs Food products Canned products Tuna Chloroform, methanol, Cyt b 171 [58] water Canned products Sardine Chelexa, phenol, chloroform, Cyt b 152 [59] isoamyl alcohol Meat-and-bone meal in Beef, sheep, pig, chicken Guanidium thiocyanate tRNALys, ATPase 145–313 [60] compound feeds (GUSCN) Blood or meat meal, pet Ruminant, avian, fish and Dneasy tissue kitb 12s rRNA, tRNAval, 104–290 [61] food, baby food pig 16S rDNA ‘Mortara’ salami Goose Genomic prep kitc Cyt b 350 [62] Goat cheese Beef Dneasy tissue kitb D-loop 413 [63] Foie gras Goose and mule duck Phenol, chloroform, isoamyl 5S rDNA 250–1000 [56] alcohol Caviar Sturgeon species Not indicated Cyt b, 16S, 12S Not indicated [55] Forensic products Dried, salted and unfrozen Whale Not indicated D-loop 155–378 [64] strips of meat Formalin fixed paraffin Human Chelexa, QIAmpb Microsatelittes, 128–250 [65] embded tissues Amelogenin Skin, tanned hide, scales Chinese alligator Phenol/chloroform Cyt b 180 [66] Pills and plasters made with Tiger Chelex1, phenol, chloroform Cyt b 165 [6] tiger’s bone Elephant tusk (Ivory) African elephant QIAamp kitb 12S, cyt b, micro- 70–251 [3] satellites Faeces Tiger Guanidium thiocyanate Cyt b 510 [67] (GUSCN) Muscle, blood, eggs, skin Sea turtles Phenol-chloroform; Dneasyb Cyt b 875–876 [4] Soup, dried fin, cartilage Shark Phenol-chloroform; Dneasyb Cyt b 155–188 [68] pills aBioRad (http://www.bio-rad.com). bApplied Biosystems (http://www.appliedbiosystems.com). cAmersham Biosciences (http://www1.amershambiosciences.com). from fossil bones or ancient organic remains), human genetic code and the presence of many non-coding regions, forensics (studying DNA from hairs, saliva, blood etc. from DNA provides much more information than proteins crime scenes), non-invasive ecological studies (studying do. However, in processed products, DNA is altered and DNA from animal faeces or hairs found in the field) and displays several particular features that must be taken more recently, food authentication. Taken together, these into account. fields have demonstrated that, (i) despite being altered, DNA is more resistant and thermostable than proteins Substrates, DNA quality and contaminations are and it is still possible to PCR amplify small DNA Short DNA fragments. First, during production processes, fragments (with sufficient information to allow identifi- food products might be subject to thermal treatments cation) and (ii) DNA could potentially be retrieved from (cooking, pasteurization, etc.), high pressure, pH modifi- any substrate because it is present in almost all cells of an cations, irradiation, drying and so on. For example, many organism. In addition, molecular evolution and phyloge- food products are heated up to 1008C for 10–60 min and netics have shown that, because of the degeneracy of the are exposed to a pH!4. Consequently, molecular

Box 1. Review of species identified

To evaluate the species for which at least one method is currently O1200 species are fished (http://www.fishbase.org/search.cfm)and available, we attempted to collate comprehensively, but not exhaus- w220 are farmed (both numbers include shellfish) [72]. Nevertheless, tively, all studies published in the past decade. It seemed that almost as with the two previous groups, we observed a strong bias towards all of these studies (w100 reviewed) had focused on a few species several species, such as tuna (20%), salmon (16%) and sturgeon (6%). belonging to three main groups of vertebrates: mammals (36%), By contrast, the most valuable fish have hardly been studied, such as actinopterygian fishes (34%) and birds (20%). sardines [73,59] and cod [74], which are the first (22 472 563 metric † Within mammals, more than one-third (41%) of the studies only tons in 2002) and second (8 392 479 metric tons in 2002) most dealt with one or all of the same four livestock species (cattle, pig, important groups ‘harvested’ worldwide (http://www.fao.org), sheep and goat). The remaining mammal species studied are usually respectively. endangered ones, such as seven species of animals for bushmeat [69], † Finally, we found several studies (10%) that focused on other rhinoceroses [7] or tigers [5]. groups, such as clam species [75] or sea turtles [4]. † Within birds, two-thirds (60%) only dealt with chicken and/or Consequently, despite the relatively high number of studies turkey, few dealt with other birds such as goose and duck [70] or published, the number of species for which a method of detection ostrich [71]. is currently available is still small, probably less than 100 animal † Within fishes, the number of species studied is much higher than species overall. Thus it seems likely that in the future the number in the two previous groups. This was expected, because the number of of species studied will increase significantly, particularly within exploited species is far higher than either in mammals or birds. Indeed the fish. www.sciencedirect.com Review TRENDS in Biotechnology Vol.23 No.7 July 2005 361

laboratory [physically isolated from those where unal- Box 2. Example of simultaneous detection of mixed species tered DNA (from fresh samples) and amplified DNA are in foodstuff handled] and appropriate negative controls (extraction DNA was first extracted from a pet food can labelled ‘meat flavour’ and PCR blanks) should be processed on every run. (Figure I, [76]) and was amplified using trans-vertebrates (enabling PCR inhibitors. Many food or forensic constituent the amplification of all vertebrate species) or ‘universal primers’ designed by Kocher et al. [26]. The amplified DNA (a 380 bp portion products could be co-purified with the target DNA and of the Cytochrome b gene) was then cloned using a commercial are known to be inhibitors of DNA amplification. These cloning kit (TOPO TA cloning, Invitrogen; http://www.invitrogen. include organic and phenolic compounds, polysaccharides com/) to separate each molecule of DNA, and 30 clones were (or products of the Maillard reaction), glycogen, fats, milk sequenced using a direct sequencing method. Sequences obtained proteins, collagen, iron, cobalt or fulvic acids (reviewed in were compared to those in the sequence bank GenBank (http://www. ncbi.nlm.nih.gov/Genbank/index.html). A mix of five species [13] and [14]). Other more widespread inhibitors include sequences was observed: seven clones of beef (Bos taurus), four bacterial cells, non-target DNA and exogenous contami- clones of pig (Sus scrofa), five clones of duck (Cairina moschata), nants [13]. Therefore, the absence of amplification in food seven clones of chicken (Gallus gallus), and seven clones of sea trout or forensic samples (i.e. a negative readout) could be (Salmo trutta). Today, this approach is the only one (with DNA chips) explained by the inhibition of the PCR amplification that enables broad identification of animal species used in a food product, even if, to our knowledge, it has rarely been used. rather than by an insufficient amount or the absence of the target DNA. Thus, to detect a species in a food product, DNA isolation methods should allow removal of inhibitors (e.g. [15]). Despite the numerous protocols already described (examples are given in Table 1) to date, no general extraction method has proved useful with all the Canned pet food different matrices encountered [16]. Therefore, each new substrate requires the development of new extraction and DNA isolation amplification methods or the adaptation of existing ones. Nevertheless, several ancient DNA studies have shown PCR amplification using trans-vertebrates primers that the use of the PTB (N-phenecylthiazolium bromide) allows cleaveage of cross-links between proteins and DNA Cloning and sequencing of 30 clones (Maillard reaction) and thus enhances the subsequent PCR reaction by permitting the movement of the Taq polymerase on the DNA molecule [17,18]. Moreover, it is also possible to neutralize some of these inhibitors by the addition of bovine serum albumin (BSA) or spermidine to the PCR reagents [19,20].

TRENDS in Biotechnology Chemical modifications and PCR artefacts Figure I. These are probably the least known features of DNA in food and forensic applications but paleogenetic studies showed that environment (acidity, UV light, moisture, identification methods from such highly degraded sub- etc.) could induce chemical modifications of the DNA strates should be based on the analysis of very short DNA molecule. In this case, modified DNA molecules might fragments, preferably between 100–200 base pairs display breaks or artifactual mutations (reviewed in [21]). (e.g. researchers were not able to amplify fragments Consequently, absence of DNA degradation from a food longer that 200 bp from canned tunas [9] or in processed product, particularly for extensively processed ones, animal meals [10,11]). should be checked before using methods based on few Low amounts of DNA. DNA is not only degraded but is variable sites. These DNA modifications can indeed also present in small quantities that considerably produce misidentification of species because the DNA reduce the number of DNA fragments with suitable sequence obtained is slightly different from the reference size for molecular analysis. Thus, paleogenetics and one. Although these chemical degradations have never non-invasive molecular studies have shown that been studied in food products, Ram et al. [22] found increasing the number of PCR cycles, up to 45 or 55 inexplicable variations in DNA extracted from canned cycles, is often required to get enough amplified DNA tuna; this variability could be because of chemical for subsequent analysis. However, DNA from other modifications of the DNA molecule studied. components, fraudulently or accidentally included, and Paleogenetic and non-invasive studies have also shown from minority constituents, could be present in very that PCR amplification starting from tiny amounts of small quantities. Species detection methods should altered DNA could induce artifactual results (e.g. allelic therefore be very specific to provide reliable results. drop-out in microsatellite analysis [23] or chimeric Contaminations. Because PCR is such a sensitive molecules produced by jumping PCR [24]). method, sometimes a single exogenous DNA molecule could be preferentially amplified instead of the degraded Mix of individuals and species one (reviewed in [12]). Thus, food and forensic samples DNA extracted from a food product is a mix of DNA of should be manipulated (before PCR) in a dedicated many origins: bacterial, vegetable, animal and fungi. www.sciencedirect.com 362 Review TRENDS in Biotechnology Vol.23 No.7 July 2005

Furthermore, several individuals of the same species More precisely, a suitable DNA marker for identifi- (sometimes from different lineages) could be used in a food cation at the species level should be sufficiently variable product. However, unfortunately almost none of the between species (particularly between the closest ones) available methods allow the simultaneous identification and display either low or no intra-specific variations of a wide array of species (or individuals) mixed in food across the geographic distribution area. In addition, this products. This simultaneous detection of several species is marker should be widely studied for a large number of certainly one of the greatest challenges in the field, but species to enable comparison of the nucleotide sequence still remains unresolved. Two methods have the potential from an unknown sample with reference sequences in a to offer convincing and reliable solutions – cloning and database. The gene encoding cytochrome b satisfied most sequencing (Box 2) or DNA chips. DNA chips (also known of these criteria and is by far the most studied gene for as DNA macroarray or DNA microarray) allow the phylogeny. It is certainly used in more than half of the examination of complex mixtures of PCR products and, phylogenetic studies published in the past decade (more potentially, the identification of hundreds or thousands of than 50 000 sequences available in GenBank in 2004). species simultaneously. Such methods have already been This gene also displayed both (i) conserved regions used in numerous fields such as in ecology (e.g. for the allowing the determination of primers such as ‘universal’ simultaneous detection of five marine fish pathogens [25]) primers published by Kocher et al. [26], which could be but not extensively for species identification in food and used to amplify a wide range of vertebrate species (see forensic samples (F. Teletchea et al., unpublished data). Box 2), and (ii) regions with a high level of variability, Despite all of these technical constraints and difficul- which allows the evolutionary studies of even closely ties, it has been shown that DNA could be retrieved and related species. Nevertheless, in the cases of identification analysed with sufficient size and information to allow the of breeds, geographic origins, or individual assignments, correct identification of species from a wide array of markers should be different and those showing an substrates by using different extraction protocols (Table 1, important intra-specific variability will be very useful Box 3). (e.g. D-loop [27]; microsatellites [28,29]; or coding region [30]). Thus, in some cases, a strong haplotypic structure DNA markers for species identification within a species can allow allocation of an individual to a DNA retrieved from food and forensic products displays particular geographic population. The large number of several specific features that complicate its detection. reconstructed phylogeography using mtDNA genes clearly These characteristics underline why, for species identifi- illustrated that infraspecific information can be used to cation in these kinds of products, the mitochondrial DNA improve identification and potentially to identify geo- genome (mtDNA) is preferentially targeted compared to graphic origins of new invasive species. nuclear DNA. Indeed, because there are several copies of Thus, as expected, among approximately one hundred mtDNA inside a cell (w1000!the copies of nuclear DNA) molecular identification studies (collated from the litera- it is more likely to amplify a fragment within this genome ture comprehensively but not exhaustively in the past rather than within the nuclear genome. Besides, this decade) almost half of them targeted the cytochrome b small circular genome (w16 kb in most vertebrate species) (44%), then the 12S (11%), 16S (8%) and D-loop (8%). displays maternal inheritance in most animal species, is Lastly, about fifteen other markers have occasionally been haploid, and does not undergo recombination – character- used, mostly located within mtDNA [e.g. cytochrome c istics that make its study easier and more straightfor- oxidase II (COX2)]. Besides, despite the technical problem ward. Lastly, mtDNA generally evolves much faster than of analyzing DNA in these highly degraded samples (as nuclear DNA and thus enables even closely related species mentioned above), few studies targeted nuclear markers to be differentiated and identified. such as microsatelittes (Table 1).

Box 3. Examples of food or forensic molecular identification

Three molecular identification methods today represent O90% of all most direct means for obtaining information from PCR products, studies published: PCR-RFLP, Species-specific PCR and PCR-FINS. (i.e. by sequencing). Compared to the two previous methods, this PCR-RFLP and species-specific PCR successes could be explained by method presents two main advantages: (i) it is not too sensitive to the short development time necessary to design molecular markers, intraspecific variations; and (ii) by using cloning (Box 2) it is possible to low cost and straightforward results. detect several species in the same product. Je´roˆme et al. [59] used PCR-RFLP (restriction fragment length polymorphism). For example, direct sequencing of a 103-bp diagnostic sequence to differentiate 14 Hold et al. [77] differentiated 36 fish species by digesting a 464 bp sardine species used in food preparation. This method was success- cytochrome b fragment with six different restriction enzymes. They fully applied to 45 out of 47 commercial canned sardine and sardine- showed that this method was still useful with processed products (1008C type products tested. during 15 min) and also when several species were mixed. New methods such as real time PCR (see [78] for a detailed Species-specific PCR. Wand and Fang [5] analyzed tiger DNA in presentation) and DNA chips have not been developed significantly meat, faeces and dried skin and therefore showed that this protected for species identification to date. This low interest is particularly animal was still illegally traded. Moreover, using a multiplex of because of longer developing times and costs compared with the species-specific primers Dalmasso et al. [61] proposed to identify the previous methods. However, in our laboratory, we have developed a presence of the most used vertebrate groups in feedstuff products DNA chip that allows the simultaneous detection of numerous (ruminant, poultry, fish and pork). vertebrate species in processed food and forensic samples PCR-FINS (forensically informative nucleotide sequencing) is the (F. Teletchea et al., unpublished data). www.sciencedirect.com Review TRENDS in Biotechnology Vol.23 No.7 July 2005 363

Caution with mtDNA: the Numts gene (447 bp) differed by 3% between the English and USA The vast majority of identification methods developed in breeds of turkey (Meleagris gallopavo). the past decade were based on mitochondrial sequences. Two valid species could be genetically closely related. However, we would like to emphasize that mitochondrial While studying the complete cytochrome b gene varia- sequences could be obtained from a nuclear copy of the bility within the bear family, Talbot and Shields [38] found mtDNA rather than from endogenous mitochondrial DNA. that brown bears (Ursus arctos) inhabiting the islands of Indeed, translocated pieces of mitochondrial DNA (known southeastern Alaska seemed to be more closely related to as Numts) are sometimes integrated into the nuclear polar bears (Ursus maritimus) than to any other brown genome. Thus, during analysis, these Numts might be bear populations. Similar conclusions were obtained amplified inadvertently by the PCR in addition to or even between the Greenland cod Gadus ogac and the Pacific instead of the authentic target mtDNA and therefore led cod Gadus macrocephalus, in which sequences of partial to erroneous results [31]. For example, while studying the cytochrome b gene (401 bp) and COX1 (495 bp) were snapping shrimp genus Alpheus, Williams and Knowlton identical [39]. The authors concluded that these two [32] found multiple copies of the COX1 gene in at least ten species are in fact the same and should be synonymized. species. They observed that within a single animal, Hybridization between closely related species (intro- differences between the real mtDNA and pseudogene gression) can occur if these species share overlapping sequences ranged from 0.2–18.8%. Thalmann et al. [31] habitats or sometimes through human intervention found similar results within gorillas. Bensassson et al. [33] (e.g. during captive breeding). Within the group of the provided a detailed census of Numts in eukaryotic groups Bovini (comprising cattle-like species) several hybrids (they found Numts in 82 different species) and described have been described (e.g. [40–42]). Similarly, several fish methods for detecting them (e.g. checking for unique species show interspecific mitochondrial introgression changes and odd substitution patterns). However, current (charr [43,44]; sturgeon [45]). These introgressions could practices do not preclude inadvertent analysis of Numts sometimes disturb molecular identification of closely andmuchcautionshouldbeexercisedwhenusing related species [46]. mitochondrial sequences for identification purposes – For all these reasons, more caution should be taken in explicit measures need to be taken to authenticate this area. First with the use of scientific names, indeed mtDNA sequences [31]. most authors were quite elusive about the species they studied, for example neither the name of the author of authority who identified the species (e.g. Gadus morhua Species concepts Linnaeus, 1758) nor geographical distribution and bio- The literature about species concepts might be more logical information were indicated; and sometimes only extensive than that about any other subject in evolution- common names were given (e.g. [47,48]). Second, research- ary biology and therefore it is obviously out of the scope of ers should be reminded that scientific names could refer to the present review to discuss the validity of all these different contents according to the progress of science [49] concepts (see [34]). Rather, we illustrate the three main and that discrepancies could exits between taxonomical problems usually encountered in systematics that could conclusions obtained from morphological and molecular result in erroneous molecular identification methods if characters even in well known groups, as shown here. current practices do not take them into account. Therefore identification methods cannot only rely on one or several sequences per species as was the case in the Complexes of cryptic species, closely related species and methods reviewed here. All the problems highlighted here introgression are particularly critical within fish, where there are The same scientific names could refer to highly divergent numerous examples of such unclear species boundaries molecular groups. While studying the mitochondrial and species complexes [50]. In addition, because this sequence variation (927 pb fragment of the ATPase and group is likely to become the most studied one in the next COX3 genes) within and between tuna species (Thunnus decades, it is likely that these kinds of situations will also spp.), Takeyama et al., [35] found two divergent groups become more and more common in the molecular identi- (diverging by w4,5%) within the northern bluefin tuna, fication field (for a detailed review of species already Thunnus thynnus, (i.e. T. t. thunnus, living in the Atlantic analyzed see Box 1). ocean and T. t. orientalis, in the Pacific ocean). Thus, they showed that a re-evaluation of previous restriction Recommendations methods was required to avoid inconsistent profiles and Whatever the species, a thorough analysis should be made erroneous tuna identification. This result has considerable each time a new group is to be studied (ideally with consequences as Thunnus thynnus is one of the most taxonomists of each group) and several samples from the highly prized fish worldwide and Atlantic populations are full distribution range should be taken into consideration on the red list. Similarly, Ludt et al. [36] found two distinct to validate the method [51,52]. Voucher specimens should groups of red deer diverging by 5–6% (complete cyto- ideally be preserved, and submitters of sequences should chrome b gene) and concluded that the Cervus elaphus conform to the ICZN (International Commission on species is clearly subdivided giving two valid species Zoological Nomenclature; http://www.iczn.org/) rules of Cervus canadensis (occurring in Asia and North America) nomenclature [51]. and Cervus elaphus (inhabiting Europe). Hird et al. [37] Finally, because sequences from international banks also found that the sequences of the partial cytochrome b are usually taken as reference, we would like to emphasize www.sciencedirect.com 364 Review TRENDS in Biotechnology Vol.23 No.7 July 2005

generally in implementing the recent call for a ‘DNA Box 4. What is DNA taxonomy, and why is it so taxonomy’ (Box 4). controversial?

What is DNA barcoding? Conclusion Herbert et al. [79] recently proposed a DNA-based barcoding system for all animal species, similar in practice to a supermarket barcode. Despite these technical and conceptual challenges, mol- According to the proponents of this system, the basic procedures of ecular species identification in food products and forensic DNA barcoding would be straightforward [80]. A tissue sample is samples is likely to increase exponentially. Indeed, taken from a collected individual and DNA is extracted from this. This numerous reliable methods now exist for identifying DNA serves as the reference sample from which one or several gene species in almost all kinds of substrates (Table 1 and regions are, as a first approximation, an identification tag for the species from which the respective individual was derived. This Box 3) and some approaches have already produced sequence is then made available via appropriate references, against significant and interesting results, for example in gourmet which sequences from sampled individuals can be compared (http:// food (such as species identification in caviar [55] or in foie www.barcodinglife.com/). Put simply, given the long history of use gras [56]) and on forensic samples made from endangered of molecular markers for molecular identification of species Moritz species (e.g. tiger [5] or rhinoceros [7]). Molecular and Cicero [46] considered that there is nothing fundamentally new in this DNA barcoding concept, except increased scale and proposed identification has already proven useful in court [4,57] standardization. and some methods are currently used in industry [1,76]. Nevertheless, to become more accurate and reliable, this Why is it so controversial? biotechnological field will have to take into account the However, the main problem with the DNA barcoding initiative is that technical and conceptual knowledge of its nearest funda- the proponents of it propose not only to use DNA for identifying species (the ‘DNA barcode’, which seems to be generally accepted) mental fields. but also for defining species, and therefore they want to give DNA a central and mandatory role in taxonomy (the so-called ‘DNA Acknowledgements taxonomy’). Indeed they think that this is the only way to overcome We thank our team for constructive discussions, Brian B. Rudkin and the current impediment of taxonomy and particularly they consider Vincent Laudet for critical reading of the article, and CNRS UCB-Lyon1, that (i) it is impossible to describe biological diversity with traditional Re´gion Rhoˆne-Alpes and Ministe`re de l’Education Nationale, de approaches [81], (ii) the current system depends heavily on l’Enseignement Supe´rieur et de la Recherche for financial support. specialists, whose knowledge is frequently lost when they retire [80], and (iii) species identification based on morphological char- acters has several significant limitations [79]. References Such remarks have already received outright condemnation, 1 Pascal, G. and Mahe, S. (2001) Identity, traceability, acceptability and chiefly from taxonomists, on both the theoretical (e.g. [49,82,83]) substantial equivalence of food. Cell Mol. Biol. 47, 1329–1342 and practical level (e.g. [46,84]). Among the numerous criticisms, the 2 Manel, S. et al. (2002) Detecting wildlife poaching: identifiying the main one is that DNA sequences, as morphological characters, are origin of individuals with Bayesian assignment test and multilocus only data and could not serve to define a species on their own. In genotypes. Conserv. Biol. 16, 650–659 fact, what defines a species is an intractable debate that cannot be 3 Comstock, K. et al. (2003) Amplifying nuclear and mitochondrial DNA resolved satisfactorily using part of a single gene. The circumscrip- from African elephant ivory: a tool for monitoring the ivory trade. tion of a species will therefore always remain an opinion based on all Conserv. Biol. 17, 1840–1843 data available [49,84]. Finally, molecular identification, using 4 Moore, M. et al. (2003) Use of restriction fragment length polymorph- mtDNA, can be misled by Numts, introgression (as we have isms to identify sea turtle eggs and cooked meats to species. Conserv. illustrated here) but also by incomplete lineage sorting, retention Genet. 4, 95–103 of polymorphism or absence of polymorphism (see [85] for a more 5 Wan, Q.H. and Fang, S.G. (2003) Application of species-specific complete review). polymerase chain reaction in the forensic identification of tiger For all these reasons, most taxonomists consider that relegating species. Forensic Sci. Int. 131, 75–78 taxonomy, rich in theory and knowledge, to a high-tech service 6 Wetton, J. et al. (2004) An extremely sensitive species-specific ARMs industry would decidedly be a setback for science. They argued that, PCR test for the presence of tiger bone DNA. Forensic Sci. Int. 140, instead of discarding more than 250 years of knowledge, new 139–145 generations of taxonomists must be trained in initiatives such as the 7 Hseih, H.M. et al. (2003) Species identification of rhinoceros horns Partnerships for Enhancing Expertise in Taxonomy, PEET program using the Cytochrome b gene. Forensic Sci. Int. 136, 1–11 ([86]; http://web.nhm.ku.edu/peet/). 8 Civera, T. (2003) Species identification and safety of fish products. Vet. Res. Commun. 27 (Suppl. 1), 481–489 9 Quinteiro, J. et al. (1998) Use of mtDNA direct polymerase chain reaction (PCR) sequencing and PCR-restriction fragment length that they could contain false sequences. For example, polymorphism methodologies in species identification of canned Forster [53] found that half of all published studies of tuna. J. Agric. Food Chem. 46, 1662–1669 human mtDNA sequences contain mistakes, not to 10 Frezza, D. et al. (2003) A competitive polymerase chain reaction-based approach for the identification and semiquantification of mitochon- mention Numts (as described above). Therefore, Harris drial DNA in differently heat-treated bovine meat and bone meal. [54] proposed several solutions to check the quality of the J. Food Prot. 66, 103–109 published sequences and the ‘simplest’ is to re-sequence 11 Prado, M. et al. (2004) Application of a polymerase chain (PCR) them. Another solution would be to use as many reference method as a complementary tool to microscopic analysis for the sequences as possible resulting from different studies detection of bones and other animal tissues in home-made animal meals. J. Sci. Food Agric. 84, 505–512 (when available) and never base a method on only one 12 Cooper, A. and Wayne, R. (1998) New uses for old DNA. Curr. Opin. reference sequence. Consequently, even though there is a Biotechnol. 9, 49–53 great need for molecular identification in numerous fields, 13 Wilson, I.G. (1997) Inhibition and facilitation of nucleic acid protection of the biodiversity, food traceability and also amplification. Appl. Environ. Microbiol. 63, 3741–3751 14 Scholz, M. et al. (1998) A polymerase chain reaction inhibitor of ecology [46], much more caution should be taken in ancient hard and soft tissue DNA extract is determined as human developing molecular identification methods and more collagen type I. Anal. Biochem. 259, 283–286 www.sciencedirect.com Review TRENDS in Biotechnology Vol.23 No.7 July 2005 365

15 Ha¨nni, C. et al. (1995) Isopropanol precipitation removes PCR 43 Redenbach, Z. and Taylor, E.B. (2003) Evidence for bimodal hybrid inhibitors from ancient bone extract. Nucleic Acids Res. 23, 881–882 zones between two species of charr (Pisces: Salvelinus) in north- 16 Woolfe, M. and Primrose, S. (2004) Food forensics: using DNA western North America. J. Evol. Biol. 16, 1135–1148 technology to combat misdescription and fraud. Trends Biotechnol. 44 Doiron, S. et al. (2002) A comparative mitogenomic analysis of the 22, 222–226 potential adaptive value of Arctic charr mtDNA introgression in brook 17 Hofreiter, M. et al. (2001) Ancient DNA. Nat. Rev. Genet. 2, 353–359 charr populations (Salvelinus fontinalis Mitchill). Mol. Biol. Evol. 19, 18 Poinar, H.N. et al. (1998) Molecular coproscopy: dung and diet of the 1902–1909 extinct ground sloth Nothrotheriops shastensis. Science 281, 402–406 45 Ludwig, A. et al. (2003) Nonconcordant evolutionary history of 19 Pa¨a¨bo, S. et al. (1988) Mitochondrial DNA sequences from a 7000-year maternal and paternal lineages in Adriatic sturgeon. Mol. Ecol. 12, old brain. Nucleic Acids Res. 16, 9775–9787 3253–3264 20 Orlando, L. et al. (2002) Ancient DNA and the population genetics of 46 Moritz, C. and Cicero, C. (2004) DNA barcoding: promise and pitfalls. cave bears (Ursus spelaeus) through space and time. Mol. Biol. Evol. PLoS Biol. 2, 1529–1531 19, 1920–1933 47 Sun, Y.L. and Lin, C.S. (2003) Establishment and application of a 21 Poinar, H.N. (2002) The genetic secrets some fossils hold. Acc. Chem. fluorescent polymerase chain-reaction fragment length polymorphism Res. 35, 676–684 (PCR-RFLP) method for identifying porcine, caprine, and bovine 22 Ram, J.L. et al. (1996) Authentication of canned tuna and bonito by meats. J. Agric. Food Chem. 51, 1771–1776 sequence and restriction site analysis of polymerase chain reaction 48 Saez, R. et al. (2004) PCR-based fingerprinting techniques for rapid products of mitochondrial DNA. J. Agric. Food Chem. 44, 2460–2467 detection of animal species in meat products. Meat Sci. 66, 659–665 23 Goossens, B. et al. (1998) Plucked hair samples as a source of DNA: 49 Seberg, O. et al. (2003) Shortcuts in systematics? A commentary on reliability of dinucleotide microsatellite genotyping. Mol. Ecol. 7, DNA-based taxonomy. Trends Ecol. Evol. 18, 63–65 1237–1241 50 Sweijd, N.A. et al. (2000) Molecular genetics and the management and 24 Pa¨a¨bo, S. et al. (1990) DNA damage promotes jumping between conservation of marine organisms. Hydrobiologia 420, 153–164 templates during enzymatic amplification. J. Biol. Chem. 265, 51 Ruedas, L.A. et al. (2000) The importance of being earnest: what, if 4718–4721 anything, constitutes a “specimen examined? Mol. Phylogenet. Evol. 25 Gonzalez, S.F. et al. (2004) Simultaneous detection of marine fish 17, 129–132 pathogens by using multiplex PCR and a DNA microarray. J. Clin. 52 Comesana, A.S. et al. (2003) Molecular identification of five commer- Microbiol. 42, 1414–1419 cial flatfish species by PCR-RFLP analysis of a 12 rRNA gene 26 Kocher, T.D. et al. (1989) Dynamics of mitochondrial DNA evolution in fragment. J. Sci. Food Agric. 83, 752–759 animals: amplification and sequencing with conserved primers. Proc. 53 Forster, P. (2003) To err is human. Ann. Hum. Genet. 67, 2–4 Natl. Acad. Sci. U. S. A. 86, 6196–6200 54 Harris, D.J. (2003) Can you bank on GenBank? Trends Ecol. Evol. 18, 27 Gongora, J. et al. (2004) Phylogenetic relationships of Australian and 317–319 New Zealand feral pigs assessed by mitochondrial control region 55 DeSalle, R. and Birstein, V. (1996) PCR identification of black caviar. sequence and nuclear GPIP genotype. Mol. Phylogenet. Evol. 33, Nature 381, 197–198 339–348 56 Rodriguez, M.A. et al. (2003) Identification of goose, mule duck, 28 Nielsen, E.E. et al. (2001) Fisheries. Population of origin of Atlantic chicken, turkey, and swine in foie gras by species-specific polymerase cod. Nature 413, 272 chain reaction. J. Agric. Food Chem. 51, 1524–1529 29 Cunningham, E.P. and Meghen, C.M. (2001) Biological identification 57 Whitler, R.E. et al. (2004) Forensic DNA analysis of pacific salmonid systems: genetic markers. Rev. Sci. Tech. 20, 491–499 samples for species and stock identification. Environmental biology of 30 Maudet, C. and Taberlet, P. (2002) Holstein’s milk detection in cheeses fishes 69, 275–285 inferred from melanocortin receptor 1 (MC1R) gene polymorphism. 58 Terol, J. et al. (2002) Statistical validation of the identification of tuna J. Dairy Sci. 85, 707–715 species: bootstrap analysis of mitochondrial DNA sequences. J. Agric. 31 Thalmann, O. et al. (2004) Unreliable mtDNA data due to nuclear Food Chem. 50, 963–969 insertions: a cautionary tale from analysis of humans and other great 59 Jerome, M. et al. (2003) Direct sequencing method for species apes. Mol. Ecol. 13, 321–335 identification of canned sardine and sardine-type products. J. Agric. 32 Williams, S.T. and Knowlton, N. (2001) Mitochondrial pseudogenes are pervasive and often insidious in the snapping shrimp genus Food Chem. 51, 7326–7332 Alpheus. Mol. Biol. Evol. 18, 1484–1493 60 Krcmar, P. and Rencova, E. (2003) Identification of species-specific 33 Bensasson, D. et al. (2001) Mitochondrial pseudogenes: evolution’s DNA in feedstuffs. J. Agric. Food Chem. 51, 7655–7658 misplaced witnesses. Trends Ecol. Evol. 16, 314–321 61 Dalmasso, A. et al. (2004) A multiplex PCR assay for the identification 34 Sites, J.W. and Marshall, J.C. (2003) Delimiting species: a renaissance of animal species in feedstuffs. Mol. Cell. Probes 18, 81–87 issue in systematic biology. Trends Ecol. Evol. 18, 462–470 62 Colombo, C. et al. (2002) Identification of the goose (Anser anser)in 35 Takeyama, H. et al. (2001) Mitochondrial DNA sequence variation italian “Mortara” salami by DNA sequencing and a polymerase chain within and between tuna Thunnus species and its application to reaction with an original primer pair. Meat Sci. 61, 291–294 species identification. J. Fish Biol. 58, 1646–1657 63 Maudet, C. and Taberlet, P. (2001) Detection of cows ‘milk in goats’ 36 Ludt, C.J. et al. (2004) Mitochondrial DNA phylogeography of red deer cheeses inferred from mitochondrial DNA polymorphism. J. Dairy (Cervus elaphus). Mol. Phylogenet. Evol. 31, 1064–1083 Res. 68, 229–235 37 Hird, H. et al. (2003) Rapid detection of chicken and turkey in heated 64 Baker, C.S. et al. (2003) www.DNA-surveillance: applied molecular meat products using the polymerase chain reaction followed by taxonomy for species conservation and discovery. Trends Ecol. Evol. amplicon visualisatoin vista green. Meat Sci. 65, 1117–1123 18, 271–272 38 Talbot, S.L. and Shields, G.F. (1996) Phylogeography of brown bears 65 Legrand, B. et al. (2002) DNA genotyping of unbuffered formalin fixed (Ursus arctos) of Alaska and paraphyly within the Ursidae. Mol. paraffin embedded tissues. Forensic Sci. Int. 125, 205–211 Phylogenet. Evol. 5, 477–494 66 Yan, P. et al. (2004) Identification of chinese alligators (Alligator 39 Carr, S.M. et al. (1999) Molecular systematics of gadid fishes: sinensis) meat by diagnostic PCR of the mitochondrial Cytochrome b implications for the biogeographic origins of pacific species. Can. gene. Biological conservation J. Zool. 77, 19–26 67 Verma, S.K. et al. (2003) Was elusive carnivore a panther? DNA typing 40 Nijman, I.J. et al. (2003) Hybridization of banteng (Bos javanicus)and of faeces reveals the mystery. Forensic Sci. Int. 137, 16–20 zebu (Bos indicus) revealed by mitochondrial DNA, satellite DNA, 68 Hoelzel, A.R. (2001) Shark fishing in fin soup. Cons Genet 2, 69–72 AFLP and microsatellites. Heredity 90, 10–16 69 Kelly, J. et al. (2003) Forensic identification of non-human mammals 41 Kikkawa, Y. et al. (2003) Phylogenies using mtDNA and SRY provide involved in the bushmeat trade using restriction digests of PCR evidence for male-mediated introgression in Asian domestic cattle. amplified mitochondrial DNA. Forensic Sci. Int. 136, 378 Anim. Genet. 34, 96–101 70 Calvo, J.H. et al. (2001) Random amplified polymorphic DNA 42 Verkaar, E.L. et al. (2003) Paternally inherited markers in bovine fingerprints for identification of species in poultry pate. Poult. Sci. hybrid populations. Heredity 91, 565–569 80, 522–524 www.sciencedirect.com 366 Review TRENDS in Biotechnology Vol.23 No.7 July 2005

71 Abdulmawjood, A. and Bu¨lte, M. (2002) Identification of ostrich meat 78 Lockley, A.K. and Bardsley, R.G. (2000) DNA-based methods for food by restriction fragment length polymorphism (RFLP) analysis of authentification. Trends Food Sci. Technol. 11, 67–77 Cytochrome b gene. J. Food Sci. 67, 1688–1691 79 Hebert, P.D. et al. (2003) Biological identifications through DNA 72 Naylor, R.L. et al. (2000) Effect of aquaculture on world fish supplies. barcodes. Proc. R. Soc. Lond. B. Biol. Sci. 270, 313–321 Nature 405, 1017–1024 80 Tautz, D. et al. (2003) A plea for DNA taxonomy. Trends Ecol. Evol. 18, 73 Je´roˆme, M. et al. (2003) Molecular phylogeny and species identifi- 70–74 cation. J. Agric. Food Chem. 51, 43–50 81 Blaxter, M. (2003) Molecular systematics: Counting angels with DNA. 74 Comi, G. et al. (2004) Molecular methods for the differentiation of Nature 421, 122–124 species used in production of cod-fish can detect commercial frauds. 82 Lipscomb, D. et al. (2003) The intellectual content of taxonomy: a Food Contr. 16, 37–42 comment on DNA taxonomy. Trends Ecol. Evol. 18, 65–66 75 Fernandez, A. et al. (2002) Genetic differentiation between the 83 Mallet, J. and Willmott, K. (2003) Taxonomy: renaissance or tower of clam species Rudipates decussatus (grooved carpet shell) and babel? Trends Ecol. Evol. 18, 57–59 Venerupis pullustra (pulletcarpetshell).J. Sci. Food Agric. 82, 84 Will, K.W. and Rubinoff, D. (2004) Myth of the molecule: DNA 881–885 barcodes for species cannot replace morphology for identification and 76 Donne-Gousse´,C.et al. (2002) Method of determining the existence of classification. Cladistics 20, 47–55 animals or vegetals mixtures in organic substances. Patent 85 Ballard, J.W.O. and Whitlock, M.C. (2004) The incomplete natural FR2826022 history of mitochondria. Mol. Ecol. 13, 729–744 77 Hold, G.L. et al. (2001) Development of a DNA-based method aimed at 86 Rodman, J.E. and Cody, J.H. (2003) The taxonomic Impediment identifying the fish species present in food products. J. Agric. Food overcome: NSF’s parterships for enhancing expertise in taxonomy Chem. 49, 1175–1179 (PEET) as a model. Syst. Biol. 52, 428–435

Elsevier celebrates two anniversaries with gift to university libraries in the developing world In 1580, the Elzevir family began their printing and bookselling business in the Netherlands, publishing works by scholars such as John Locke, Galileo Galilei and Hugo Grotius. On 4 March 1880, Jacobus George Robbers founded the modern Elsevier company intending, just like the original Elzevir family, to reproduce fine editions of literary classics for the edification of others who shared his passion, other ’Elzevirians’. Robbers co-opted the Elzevir family’s old printer’s mark, visually stamping the new Elsevier products with a classic old symbol of the symbiotic relationship between publisher and scholar. Elsevier has since become a leader in the dissemination of scientific, technical and medical (STM) information, building a reputation for excellence in publishing, new product innovation and commitment to its STM communities. In celebration of the House of Elzevir’s 425th anniversary and the 125th anniversary of the modern Elsevier company, Elsevier will donate books to 10 university libraries in the developing world. Entitled ‘A Book in Your Name’, each of the 6 700 Elsevier employees worldwide has been invited to select one of the chosen libraries to receive a book donated by Elsevier. The core gift collection contains the company’s most important and widely used STM publications including Gray’s Anatomy, Dorland’s Illustrated Medical Dictionary, Essential Medical Physiology, Cecil Essentials of Medicine, Mosby’s Medical, Nursing and Allied Health Dictionary, The Vaccine Book, Fundamentals of Neuroscience, and Myles Textbook for Midwives. The 10 beneficiary libraries are located in Africa, South America and Asia. They include the Library of the Sciences of the University of Sierra Leone; the library of the Muhimbili University College of Health Sciences of the University of Dar es Salaam, Tanzania; the library of the College of Medicine of the University of Malawi; and the libraries of the University of Zambia, Universite du Mali, Universidade Eduardo Mondlane, Mozambique; Makerere University, Uganda; Universidad San Francisco de Quito, Ecuador; Universidad Francisco Marroquin, Guatemala; and the National Centre for Scientific and Technological Information (NACESTI), Vietnam. Through ‘A Book in Your Name’, the 10 libraries will receive approximately 700 books at a retail value of approximately 1 million US dollars. For more information, visit www.elsevier.com www.sciencedirect.com