Molecular Identification of Diaspididae and Elucidation of Non-Native Species Using the Genes 28s and 16s

Armored scale insects pose a serious threat to habitat conservation across the globe because they include some of the most potent invasive species in the world. They are such a serious concern because their basic morphology, small size, and polyphagous feeding habits often allow them to exist undetected by growers and quarantine experts. In order to provide a potential solution to the problem, we have attempted to elucidate the effectiveness of molecular identification techniques using ribosomal 28s and endosymbiotic 16s rRNA. Sequence data was obtained from many field-collected insects to test the feasibility of identification techniques. A protocol for quick species determination based on sequence data is provided.


Introduction
The introduction of invasive scale insects has caused serious economic and ecological disasters documented. Some of the largest problems have included the introduction of Icerya purchasi to the USA resulting in the near collapse of the citrus industry and the invasion of cassava mealybug Phenacoccus manihoti to West Africa threatening a food staple of 200 million people [1]. It is estimated that there are 1019 scale insect species in the United States of which approximately 255 are invasive [2]. Of the invasive species, 75% are considered pests compared to only 7% of indigenous scales that are considered pests [2]. The group of scale insects that potentially present the largest threat are the armored scale insects, Diaspididae. According to current estimates, there are approximately 132 invasive armored scale insects in the United States alone [2].
Diaspididae are ecological and agricultural pests that cause serious damage to native plants by excessively feeding on phloem and quickly reproducing to large population sizes [3]. They are called armored scales because of the waxy secretion the female extrudes via her pygidium, which creates a protective covering [4]. Diaspidids display extreme sexual dimorphism, however, males are difficult to collect because of their brief time as adults and the lack of males in parthenogenetic species [3,5,6]. Due to the lack of males, all identification is based on female morphology. Female armored scale insects are entirely sessile as adults and will spend their lives underneath their scale covering feeding on the host plant [7]. They lack most distinguishable morphological characteristics such as a distinctive head, thorax, and abdomen, and their eyes and antennae are primitive [8]. The lack of distinctive morphology causes extremely problematic identification of invasive or pest species for non-expert scale insect taxonomists [4].
Currently, Diaspididae comprises nearly 2500 described species [9], some of which are extremely destructive [10]. Many scales are invasive species, and their small size, problematic detection and difficult identification poses a serious challenge to customs and quarantine workers as well as conservation biologists worldwide [1,2,11]. Polyphagy further exacerbates the problem of becoming serious invasive species as some armored scales can feed on up to 100 host families [12]. Evidence of numerous morphologically cryptic species among diaspidids increases the issues surrounding identification, and phylogenetic data has shown inconsistencies in previous methods of armored scale classification [13][14][15][16]. This combined with the fact that armored scales are potent agricultural pests has made the need for effective scale insect identification dire.
One critical part of the solution to growing problems with invasive and pest species among armored scale insects will require accurate species level identification. Current measures to identify scales based on morphology are becoming ineffective due to the lack of taxonomic experience among new scientists and the continued retirement of scale insect taxonomy experts [1]. Despite this, DNA sequencing and phylogenetic analysis has opened up a new mode of molecular identification for armored scale insects [13,15]. This study is designed to determined the efficacy of molecular identification of armored scale insects using 28s rRNA gene and bacterial 16s rRNA from the armored scale insect primary endosymbiont Candidatus Uzinura diaspidicola [17]. Provided is a protocol to allow molecular identification of unknown species.

Experimental Section
Samples of armored scales were obtained in the summer of 2010 and 2011 from the southeastern United States including Florida, Georgia, Alabama, Mississippi, Louisiana, Missouri, Oklahoma and Texas. Genomic DNA was extracted using a modified DNeasy Kit (Qiagen, Valencia, CA, USA) protocol in which a single specimen is punctured in a location that is not critical for microscopic identification to release DNA and preserve the exoskeleton for slide mounting [13,14]. Mounted scale insects were examined to confirm molecular identification and are currently stored in the Penn State Behrend Entomology Collection.
PCRs were performed on a TC-300G thermocycler (Techne, Minneapolis, MN, USA) using the typical three-step PCR protocol with a 52 C annealing temperature for 16S rRNA bacterial loci and 48 C for 28S rRNA insect loci. 16S amplification was done using the following two primers, the universal forward primer 27F-AGAGTTTGATCMTGGCTCAG and a reverse primer designed from Candidatus Uzinura diaspidicola a1271DIASP-CATTGTAGCACGTGTGTAGCCCAAG [17]. 28S amplification was completed using universal insect 28S rRNA primers, 28S3a-AGTACGTG AAACCGTTCAGG and 28Sb-TCGGAAGGAACCAGCTAC. For both loci, twenty-five microliter reactions contained, 12 µL Hot Start Taq Mastermix (Amresco), 1 µL each primer, 1-3 µL genomic DNA and 10-12 µL ultrapure water. All PCRs were run with negative controls and results were visualized via agarose gel electrophoresis.
Cleaned PCR products were sent to the High Throughput Genomics Center (Seattle, WA, USA) where they underwent Sanger sequencing. Sequences were then trimmed and annotated using the software package Geneious (Biomatters Ltd., Aukland, New Zealand). The original 28s and 16s sequences were combined with forty and thirty-six sequences with E-values less than 10 −20 taken from Genbank, respectively [18]. Original data has been uploaded to Genbank and accession numbers are available in Table 1. Genbank sequence accession numbers are provided in Table 2. Phylogenetic trees using Bayesian analysis were created where each tree was run for 10 million generations with a subsampling frequency of 100,000 using Geneious software and the MrBayes plugin [19].

Results and Discussion
Results of phylogenetic tests are displayed in Figure 1a,b. Inconsistencies in species identification between the two genes were seen in samples AS037 and AS087 though the AS087 16s posterior probability is insignificant and generic ID is consistent in both cases. Species invasiveness is listed with identification in Table 1 according to Miller et al., 2005 [10]. Seven different invasive species were detected in our samples of the eleven significant species detected. Samples AS013, AS015, AS017, AS023, AS026, AS044, AS053, AS056, AS057, and AS058 were identified as Pseudalacaspis cockerelli. Species AS030 and AS045 were identified as Pinaspis piperis. Samples AS035, AS038, AS039, AS040, AS041, AS073, AS079, AS089, and AS090 were identified as Melanaspis obscura. Specimen AS042 was identified though a 28s sequence as Unaspis euonymi. AS050 was identified as Parlatoria sp. via 28s phylogeny. AS074 was identified as Chionaspis pinifoliae. AS077 was identified as Diaspidiotus osborni. AS087 was identified via 16s phylogeny as Diaspidiotus osborni with a posterior probability of 52 while the 28s phylogeny identified it as Dynaspidiotus californicus with a posterior probability of 96. AS084 was identified as Hemiberlesia rapax. AS016, AS043, and AS027 had insignificant probability values to be identified as falling within a diaspidid phylogeny. Morphological analysis determined that these samples are likely in the family Halimococcidae. AS054 yielded an unknown 28s sequence. AS085 yielded an unknown 16s sequence. These samples will be further analyzed as to determine the exact species. ---

Figure 1. (a) Phylogeny of 28s utilizing Bayesian analysis. AS---numbers indicate specimen sample numbers to correlate with collections at the Behrend Entomology
Collection. Taxa provided with species names are from insects previously identified and sequenced and Genbank accession numbers are found in Table 2. An asterisk (*) indicates invasive status. Branch label indicates posterior probabilities. (b) Phylogeny of 16s utilizing Bayesian analysis. AS---numbers indicate specimen sample numbers to correlate with collections at the Behrend Entomology Collection. Taxa provided with species names are from insects previously identified and sequenced and Genbank accession numbers are found in Table 2. An asterisk * indicates invasive status and a cross † indicates unknown invasive status. Branch label indicates posterior probabilities.
(a) Because cryptic diversity among armored scales still poses a challenge to growers, undertrained quarantine professionals and conservation biologists across the globe, alternative identification methods will improve diagnosis and control of such organisms. With the advent of cheaper sequencing and molecular methods, the technique we have presented here offers a potential method for accurate armored scale insect identification for individuals lacking a background in armored scale insect taxonomy. We offer the workflow below, as a guide for molecular identification of armored scale insects: (1) Obtain DNA sequence of 28S and 16S from either in a lab utilizing the methods outlined above or by sending it to a contract lab.
(2) When accurate DNA sequences for one or both of the loci outlined above have been obtained, sequences should be imported into an already prepared, unaligned data file (FASTA file) in the form of a text document. The file is available for download through corresponding author's faculty webpage and the unknown diaspidid sequences in question can be added after the end of the last sequence listed in the document. The sequence should be entered in this format: <28Suknown1, i.e., left arrow, followed by the sequence name without spaces followed by a single return and the DNA sequence in all lowercase letters.
Data files for both 16S and 28S genes can be copied from the following site, [20], under the publications tab, directly below citation for this publication.
(3) The entire data file should then be pasted in the data window on the online sequence alignment server, MAFFT [21]. In the segments below the window, the following settings should be selected. (a) UPPERCASE/lowercase-Amino Acid….. (b) Direction of nucleotide sequences-Adjust direction according to the first sequence (accurate for most cases) (c) Output order-aligned. An email address should be provided to MAAFT as well.
(4) Species identification can now by estimated three ways. (a) The output in MAFFT has two files to view. One is the aligned data set, which will have the most similar sequences next to each other. Viewing and comparing original data to provided data in the aligned data set can elucidate similarities between sequences. A high enough similarity infers a potential species match. (b) MAFFT also allows for viewing a rough phylogenetic tree. This tree depicts the sequences most closely related to the original data, and therefore provides an estimate of species identity. (c) The aligned data set can be exported from MAFFT and subjected to a rigorous phylogenetic analysis by uploading the data set into an online phylogenetics utility site such as CypresPortal and obtain a tree with higher specificity.
Once the identity of an armored scale is known, biologists can take appropriate measures to control it. Often such measures require species identification because the treatments are species specific, such as biological control efforts utilizing parasitoid wasps or pheromone traps [22,23]. A workable, easy protocol to identify common but taxonomically problematic scale insects will prove to be an important tool for many biologists lacking taxonomic expertise.
Beyond control of invasive or pest species, if those carrying out this protocol will share their data by uploading it to online databases such as Genbank [19], their collaborative efforts will help to elucidate the cryptic diversity of armored scale insects. As molecular systematists continue to work in this area, it is becoming increasing apparent that many of the cosmopolitan, invasive species of scale insects are actually many cryptic species making up a species complex [1,13,14]. Such complexes are of evolutionary significance and the data can be very useful across disciplines [24]. In this regard, scientists investigating evolutionary founded questions as well as those in applied fields such as agriculture can share data and collaborate in ways to move both fields in positive directions. of CO1 for armored scales is significantly more difficult and less reliable, thus in order to provide a workflow for non-molecular biologists to ID armored scale insects, we have chosen to pursue 28s and 16s rRNA. Identification of armored scale insects via molecular biology technology looks hopeful, however, there are obstacles to overcome. Diversity is prevalent within most major genera of scales. With more sequences, identification of cryptic species within these genera should be elucidated. With the advent of new sequencing technologies and the continued expansion of online bioinformatics tools and databases, we hope to see an increase in contribution to scale insect DNA sequencing and identification that can lead to better diagnostics and control of invasive armored scale insects.