Background

Insect-borne diseases are major global health problems, causing hundreds of thousands of deaths per year [1], and various mosquito species are among the most important vectors of disease, being responsible for the highest transmission of human and animal parasites and pathogens. Globally widespread, mosquito-transmitted diseases include malaria, lymphatic filariasis, dengue, chikungunya, Rift Valley fever, Japanese encephalitis, West Nile virus and Usutu virus [2]. Hematophagy is present in most females of all mosquito species: they require a blood meal to produce mature eggs [3]. The extent of vector contact with an infected host or reservoir is an important factor in the transmission of mosquito-borne diseases.

Different mosquito species have wide and variable host preferences, the most important of which include humans, cattle and other mammals, but also include birds and reptiles, amphibians, worms, leeches and fish [3]. The host preference of each mosquito species shows an innate and specific pattern and may also be influenced by factors such as environmental conditions and host characteristics [2]. By detecting the source of blood in blood-feeding mosquitoes, important information can be obtained, such as the degree of contact between vector and host populations and the insect’s host preferences under natural conditions. The anthropophagic index (percentage feeding on humans) is used as component of measuring vectorial capacity for human diseases, while knowledge of other hosts can provide a measure of the relative importance of animal reservoirs of vector-borne zoonotic or enzootic infections [4]. Thus, by knowing the feeding patterns of mosquitoes on a particular host, it is possible to understand their life history, as well as the effect of host selection on their survival and reproduction and the ecology of diseases transmitted by mosquitoes. These data are also vital for operations associated with entomological surveillance and vector-monitoring, especially in the field of diseases whose etiology has an environmental component (environmental disorders) [3].

Various methods are available for determining the source of a mosquito blood meal. Previously, arthropod blood-feeding was detected by serological assays, such as the enzyme-linked immunosorbent (ELISA) method and the precipitin test [5,6,7]. While these assays are still used, they are insufficiently specific when attempting to identify the sources of a mosquito blood meal and often limited to determination of the order and family of the mosquitoes' host under study. New applications of PCR-based methods and the increasing amount of vertebrate DNA sequence data available in the public domain have allowed an increase in the specificity of blood-meal recognition down to the species or individual host level.

The efficacy and the success of PCR amplification are affected by several factors, including time since blood meal, PCR amplicon size, locus sequence/gene copy number, DNA extraction procedure, preservation method, storage time and blood-meal size [8,9,10,11,12,13]. Among these factors, longer time, increased amplicon size and low gene/locus copy number have inverse effects on the success of the PCR, as has been demonstrated extensively in studies designed to identify blood meals in hematophagous arthropods [14, 15].

Different DNA-based molecular markers have been used to identify ingested blood in arthropods. Mitochondrial genes (mtDNA) have been used extensively as reliable common markers to identify the source of the blood meal ingested by arthropods [16,17,18,19,20,21]. Consequently, mtDNA has been used widely in entomological, forensic and genetic studies because of their valuable features, such as lack of recombination, presence of very high copy numbers in each cell, maternal inheritance, absence of introns, existence of single-copy orthologous genes and high mutation rate [22,23,24,25]. Among the mtDNAs, cytochrome b (cytB) and 16S ribosomal RNA (rRNA) are the most applicable and are therefore appropriate for use in arthropod blood-meal identification in terms of determining the host species in phylogenetic and biodiversity studies and in identifying human individuals in forensic investigations [26,27,28,29,30,31,32,33,34].

Single-copy nuclear genes can also be used to identify the source of an arthropod blood meal [35]. However, when working with such a small amount of starting material, the amplification of target DNA can be more challenging with single-copy genes, thus a preferred approach is the use of Alu elements, which are transposable elements (TEs) that exist only in primates [36]. Alu repeats consist of short interspersed nuclear elements (SINEs) that replicate through LINE (long interspersed nuclear element)-mediated reverse transcription of an RNA polymerase III transcript [37]. They are the most abundant individual feature in the human genome, forming 10% of the human genome mass, with over one million copies per genome; each Alu element is approximately 280 bp long and always comes after a poly(A) tail of varying length. Although Alu repeats are not known to have a specific biological function, they have been extensively studied due to their many branches and copies in the human genome [37, 38]. There have been many reports on the use of Alu-based PCR amplification as a very sensitive and powerful tool for human genomic DNA identification and quantitation in forensic laboratories [39].

The objectives of this study were to compare the effects of elapsed time, copy number and amplicon size on the effective use of three different loci (mitochondrial cytB and 16S rRNA genes, and nuclear Alu-repeat elements) in PCR assays for tracking human DNA in blood meals of the Anopheles stephensi mosquito. Anopheles stephensi was selected for study is one of the most important malaria vectors in Asian countries and, currently, in the Horn of Africa.

Methods

Mosquito rearing

Anopheles stephensi mosquitoes were maintained at 28 °C ± 2 °C, 60 ± 10% humidity, under a 12:12-h light:dark photoperiod, and were fed only on a 10% sucrose solution before the experiments. The colony of An. stephensi used in this study was maintained in the insectarium at the School of Public Health, Tehran University of Medical Sciences. Eggs were hatched in about 1 l of tap water that was continuously supplemented with a few flakes of fish food until the larvae pupated. Pupae were collected and separated according to age. Adult mosquitoes aged 5–7 days were blood-fed, either artificially on membrane blood feeders or directly on male or female human volunteer hands and forearms in a cage under laboratory conditions [40].

Sample collection

The exclusively human blood-fed mosquitoes were transferred to a cage supplied with 10% sucrose solution to reduce mortality. Individual mosquitoes were randomly chosen, anesthetized and killed by freezing at 0, 6, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120 and 132 h post-feeding. Knockdown and killing times of < 10 s was achieved for each specimen. Starved male mosquitoes and ddH2O were used as negative controls. Specimens were stored dry in micro-tubes at − 20 °C. Prior to the experiment on PCR success rate, the head and legs were removed from each dead mosquito to reduce potential PCR inhibitors [41, 42] and to avoid high concentrations of nontarget DNA, respectively.

DNA extraction and quantification

Genomic DNA was extracted using the protocol of the Vivantis GF-1 nucleic acid extraction kit (Vivantis Technologies, Singapore). The final elution was carried out twice in the same 20 μl of elution buffer to maximize the amount of recovered DNA. Positive controls were 0.1–3 μl human blood samples, roughly equal to the minimum and maximum amount, respectively, of a blood meal in a mosquito′s midgut [43]. These samples were taken with a lancet from a female and male donor separately and poured directly into a 1.5-ml micro-tube. Using the Vivantis GF-1 extraction kit, DNA was extracted from the male mosquito (negative controls), the human blood (positive controls) and 10 individual female mosquitoes from each post-feeding time point. DNA was quantified on a Nano-Drop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

Primers and PCR reactions

Three human-specific primer pairs (Table 1) for the cytB, 16S rRNA and Alu-repeat loci were used to confirm the absence of human DNA from ddH2O and male mosquitoes (the negative controls) and the presence of human DNA in the fed female mosquitoes and in the blood stains (positive controls). Various volumes (0.1, 1.0 or 2.0 μl) of DNA from human blood stains, as positive controls, were used to test the sensitivity and reproducibility of the three loci amplification. To test possible inhibitory substances (exocuticle and thorax) that might affect the PCR amplification, a serial dilution of human blood was performed in volumes of 0.1, 1.0, and 2.0 μl, plus a male mosquito in each sample.

Table 1 List of human-specific primers and thermal cycles used in the study
Full size table

PCR amplifications for all genes or loci were performed according to the protocol for Taq DNA Polymerase Master Mix (2×) (Ampliqon A/S, Odense M, Denmark). The PCR reaction was carried out in a 25-μl reaction volume consisting of 12.5 µl of Master Mix, 1 µl of each primer and 10 μl of template or ddH2O. The Master Mix is a premixed solution containing HotStarTaq DNA polymerase, PCR buffer and deoxynucleotide triphosphates (dNTPs), with a final concentration of 2 mM MgCl2 and 200 mM of each dNTP. Details of the thermal cycling of each protocol are listed in Table 1. All amplicons were visualized under ultraviolet light using a UV-transilluminator after electrophoresis at 85 V in a 3% agarose gel and staining with Safe stain. Negative and positive controls were included in each PCR.

The success rates of the PCR assay for each locus were subjected to regression analysis, and the median detection period (T50) [44] was designed for each amplicon using the regression equation y = mx + c, where x = time and m and c are constants. T50 is defined as the point at which 50% of amplification is expected to be successful, as predicted by the regression line. In this experiment we calculated regression lines only for critical time points, which included the last two time points with 100% success amplification, followed by the time points with < 100% success up to the lack of amplicons (0%). Regression lines were subsequently subjected to comparison by analysis of covariance (ANCOVA; general linear model [GLM]) using SPSS version 26 software package (SPSS IBM Corp., Armonk, NY, USA) to compare the rate of waning of the slopes of the three amplicons. To control the experiment, a subset of PCR products from each locus was randomly sequenced by the same forward and reverse primers used for the PCR and their identity was then determined by BLASTn searches against the GenBank nucleic acid sequence database (http://www.ncbi.nlm.nih.gov/BLAST/).

Results

The sensitivity and reproducibility tests of the PCR for the three loci showed that all of the DNA samples (0.1, 1.0 and 2.0 µl) were positive for the expected bands of 226, 228 and 157 bp for the Alu-repeat, cytB and 16S rRNA, respectively, validating the DNA extraction process and showing the sensitivity and reproducibility of the amplification of the three loci using the minimum and maximum blood volumes (0.1 and 2.0 µl, respectively; Additional file 1: Fig. S1). No PCR product was obtained with the negative controls. A test was performed for the presence of potential PCR inhibitors in the mosquito DNA and for the frequency of false-negative PCR reactions in DNA from blood-fed mosquitoes. All the samples (a mix of DNA from human blood and a decapitated and amputated male mosquito) produced bands of the expected sizes corresponding to the three loci, confirming that any possible PCR inhibitors had been removed (Additional file 2: Table S1). In addition, BLASTn searches against the GenBank nucleic acid sequence database showed that the blood meals identified from mosquitoes were identical (> 99%) to human sequences.

The results of this experiment showed that the success rate of the amplification of human DNA decreased with increasing elapsed time post-feeding. Specifically, human DNA could be traced in the blood meals of An. stephensi using the cytB gene for up to 84 h post-feeding; the 16S rRNA locus could be used up to 120 h post-feeding; and the Alu-repeats in human DNA could be detected up to 120 h post-feeding (Figs. 1, 2, 3). The maximum period of detection was similar (120 h) for the shortest amplicon (16S rRNA, 157 bp) and the amplicon with the highest copy number (Alu-repeat, 226 bp).

Fig. 1
figure 1

Effect of the elapsed time post-feeding on the success of PCR amplification of cytB (228 bp) from human blood meal in Anopheles stephensi. The numbers above the well are the elapsed time intervals (h) post-feeding. Amplification was successful up to 84 h post-feeding. CytB, Cytochrome b; M, 100-bp marker; N, Negative control; P, positive control

Full size image
Fig. 2
figure 2

Effect of the elapsed time post-feeding on the success of PCR amplification of the Alu-repeat (226 bp) from human blood meal in Anopheles stephensi. The numbers above the well are the elapsed time intervals (h) post-feeding. Amplification was successful up to 120 h post-feeding. M, 100-bp marker; N, negative control; P, positive control

Full size image
Fig. 3
figure 3

Effect of the elapsed time post-feeding on the success of PCR amplification of 16S rRNA (157 bp) from human blood meal in Anopheles stephensi. The numbers above the well are the elapsed time intervals (h) post-feeding. Amplification was successful up to 120 h post-feeding.M, 100-bp marker; N, negative control; P, positive control; rRNA, ribosomal RNA

Full size image

The success of PCR amplification was 100% for up to 84-, 84- and 72-h post-feeding for 16S rRNA, the Alu-repeat and the cytB gene, respectively. The time points at which 50% of the PCR amplification is expected to be successful (T50), as calculated by the regression lines for the critical time points, were 117 h for the Alu-repeat (226 bp), 113 h for the 16S rRNA (157 bp) and 86.4 h for the cytB (228 bp) amplicons (Fig. 4). A comparison between the regressions for the two mitochondrial amplicons (cytB and 16S rRNA), which are assumed to have the same copy number in the cells, showed the PCR success rate of the smaller amplicon (16S rRNA) to be significantly higher (GLM, P = 0.008, F = 11.497, R2 = 82.5%). A comparison between the regressions for the cytB and Alu-repeat amplicons showed that the PCR success rate of the higher copy-number amplicon (Alu-repeat) was also significantly higher than that for the cytB amplicon (GLM, P = 0.008, F = 11.687, R2 = 79.5%). However, the comparison between the regressions for the 16S rRNA and Alu-repeat amplicons showed that the PCR success rate of the higher copy-number amplicon (Alu-repeat) was only slightly higher than that of the 16S rRNA amplicon, with the difference lacking statistical significance (GLM, P = 0.638, F = 0.236, R2 = 79.7%). The logistical regression found that the probability of obtaining successful amplification for cytB, 16S rRNA and Alu-repeat loci was, respectively, 30%, 18.3% and 17.4% less for every 12-h increase in post-feeding interval at their critical time points.

Fig. 4
figure 4

Regression lines for time course and successful amplification of human DNA within the midgut of Anopheles stephensi for three human-specific molecular markers (mtDNA cytB and 16S rRNA, and nuclear Alu-repeat) with 228-, 157- and 226-bp amplicons, respectively. CytB, Cytochrome b; rRNA, ribosomal RNA

Full size image

Discussion

Here we propose a method for identifying human DNA in the blood meal of An. stephensi mosquitoes that involves amplification of the human-specific regions of cytB, 16S rRNA and Alu repeats, without the need to sequence the PCR products. We show that the success of the amplification over time is mainly dependent on amplicon size and locus copy number, with smaller amplicons (16S rRNA) and higher copy number loci (Alu-repeats) being detectable for longer periods of time after blood-feeding.

The PCR amplification of either mitochondrial 16S rRNA and or nuclear Alu-repeats in this study proved to be very efficient for identifying the origin of mosquito blood meals. We were able to trace human DNA in An. stephensi mosquito blood meals up to 120 h post-feeding. The use of this mitochondrial marker thus has the potential to be extremely useful for use in species determination and in individual human identification, for epidemiological entomology studies and also for forensic purposes [25, 46, 47]. In addition, the Alu repeats are a naturally amplified source of human genetic information and robust tools for sensitive human DNA detection and quantitation, as has previously been demonstrated in a forensic study [39].

In the present study, we demonstrated that human DNA can potentially be amplified by PCR, even when it has been degraded by digestion in the mosquito gut, because human DNA was detectable in An. stephensi for up to 84–120 h post-feeding, depending upon the locus used. The stage of blood digestion progressed from the fully fed period (fresh) to the semi-gravid period and then to fully gravid period within the first 2–3 days (48–72 h) post-feeding under the laboratory conditions (28 °C ± 2 °C), and visual validation of blood in the mosquito abdomen was also possible during the fresh and semi-gravid periods. However, our results show that the source of the blood meal could be identified from DNA, even when the blood-meal residue in the abdomen was no longer distinguishable by eye. This result is consistent with the findings reported by Replogle et al. [48], who showed that it was possible to successfully genotype individual human blood donors using mtDNA obtained from excrement of the pubic louse (Pthirus pubis L.) (Phthiraptera, Pthiridae), even when no DNA from the blood meals remained in the gut, thus demonstrating the potential of the method for the amplification of DNA degraded by digestion.

We showed here that short amplicons with high copy number can be recovered from a mosquito blood meal up to 120 h (5 days) post-feeding. This is consistent with a short amplicon (103–118 bp) assay for the identification of sand fly blood meal, which could detect various small amounts of the host DNA up to 120 h after blood-feeding [49]. However, the use of longer amplicons generally permits detection of host DNA for 1–4 days post-feeding [8, 35, 50,51,52,53,54]. In forensic studies, short amplicons of short tandem repeat (STR) alleles (normally < 200 bp) are used to identify an individual from blood or tissue isolated from insects [37, 38]. However, the use of short STR amplicons has only been shown to allow DNA recovery for 15–88 h post-blood-meal [50, 55]. Our data suggests that exploring the use of shorter STR alleles with higher copy numbers could increase the recovery time of human DNA from mosquitoes for criminal investigation.

The results of this study demonstrate that longer amplicons are more subject to degradation than short amplicons (Fig. 4). In the An. stephensi blood meal, the DNA degradation rate of the short 16S rRNA amplicon is comparable with that of the longer, high copy number Alu-repeat, suggesting that the high copy number of the Alu-repeat compensates for its longer amplicon size.

In addition to the time post-blood meal, the locus copy number and the amplicon size, other factors may affect the success of blood-meal source identification using PCR. These include the PCR-inhibitory substances present in insect tissue, especially those in the cuticle, head and thorax [56, 57], as well as those in blood, such as heme [5]. In this study we used a combination of male mosquitoes and human blood to test the inhibitory effects of these materials on our PCR reactions; it is evident from the absence of false negatives that the PCR reactions were not hindered by those materials present in the insects or the blood. It appears that across the time course of the experiment, the combined activities of digestive enzymes and nucleases play the main roles in degrading the ingested human blood, so that human DNA could not be obtained after 5 days post-blood meal.

Conclusion

We describe here an assay with greatly enhanced sensitivity and specificity for identifying the source of the blood meals of hematophagous mosquitoes, which allows detection of human DNA in a single round of PCR. We demonstrate for the first time that the use of 16S rRNA and/or Alu-repeat elements permits the identification of human blood sources from An. stephensi mosquitoes up to 5 days post-blood meal. It is evident that selecting short amplicons and high copy number loci of the human host will increase the longer-term success rate of PCR. However, since DNA degradation over time is a concern, we recommend exploring combinations of high copy number loci and shorter amplicons to maximize the recovery of human DNA from hematophagous insects for longer times post-blood meal.