Molecular Ecology Resources (2014) 14, 1271–1280 doi: 10.1111/1755-0998.

12275

‘Direct PCR’ optimization yields a rapid, cost-effective,

nondestructive and efficient method for obtaining DNA
barcodes without DNA extraction
WING HING WONG,* YWEE CHIEH TAY,* JAYANTHI PUNIAMOORTHY,* MICHAEL BALKE,†‡
PETER S. CRANSTON§ and R U D O L F M E I E R * ¶
*Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, Singapore 117543, Singapore,
†GeoBioCenter, Ludwig-Maximilians-University, Richard-Wagner-Str. 10, 80333 Munich, Germany, ‡Zoological State Collection,
Münchhausenstr 21, D-81247 Munich, Germany, §Evolution, Ecology and Genetics, Research School of Biology, The Australian
National University, Canberra, ACT 0200, Australia, ¶University Scholars Programme, National University of Singapore, 14
Science Drive 4, Singapore 117543, Singapore

Abstract
Macroinvertebrates that are collected in large numbers pose major problems in basic and applied biodiversity
research: identification to species via morphology is often difficult, slow and/or expensive. DNA barcodes are an
attractive alternative or complementary source of information. Unfortunately, obtaining DNA barcodes from speci-
mens requires many steps and thus time and money. Here, we promote a short cut to DNA barcoding, that is, a non-
destructive PCR method that skips DNA extraction (‘direct PCR’) and that can be used for a broad range of
invertebrate taxa. We demonstrate how direct PCR can be optimized for the larvae and adults of nonbiting midges
(Diptera: Chironomidae), a typical invertebrate group that is abundant, contains important bioindicator species, but
is difficult to identify based on morphological features. After optimization, direct PCR yields high PCR success rates
(>90%), preserves delicate morphological features (e.g. details of genitalia, and larval head capsules) while allowing
for the recovery of genomic DNA. We also document that direct PCR can be successfully optimized for a wide range
of other invertebrate taxa that need routine barcoding (flies: Culicidae, Drosophilidae, Dolichopodidae, Sepsidae;
sea stars: Oreasteridae). Key for obtaining high PCR success rates is optimizing (i) tissue quantity, (ii) body part, (iii)
primer pair and (iv) type of Taq polymerase. Unfortunately, not all invertebrates appear suitable because direct PCR
has low success rates for other taxa that were tested (e.g. Coleoptera: Dytiscidae, Copepoda, Hymenoptera: Formici-
dae and Odonata). It appears that the technique is less successful for heavily sclerotized insects and/or those with
many exocrine glands.
Keywords: Chironomidae, DNA barcodes, DNA extraction, macroinvertebrates
Received 7 February 2014; revision received 5 May 2014; accepted 5 May 2014

‘Any study in community ecology requires, ideally, difficult, time-consuming and/or error-prone (Meier
that the individuals in a sample be properly et al. 2008; Valentini et al. 2009; Schlick-Steiner et al.
counted and identified to species‘ Gotelli 2004 2010). This is often the case for invertebrate taxa that are
small and very abundant. Such hyperabundant yet hard-
to-identify taxa are the main obstacle to species-based
Introduction bioassessment, biodiversity monitoring and community
ecology. For these taxa, DNA barcoding can provide an
DNA barcoding based on COI (50 cytochrome c oxidase alternative or complementary method to morphological
subunit 1) has been promoted as a method for species identification (Pfenninger et al. 2007). However, this
identification for many taxa (Hebert et al. 2003; Vences requires that the barcodes accurately reflect species
et al. 2005; Ward 2009; Dinc
a et al. 2011; Zou et al. 2011; boundaries for at least most species in a sample and that
Keskin & Atar 2013). Arguably, DNA barcodes are the sequences can be obtained rapidly and at low cost.
particularly useful when morphological identification is We here address the latter two issues by promoting the
Correspondence: Rudolf Meier, Fax: +65-67792486, E-mail: use of ‘direct PCR’, a method that makes it cheaper and
[email protected] faster to obtain DNA barcodes.

© 2014 John Wiley & Sons Ltd

1272 W . H . W O N G E T A L .

In direct PCR, the time and cost for obtaining DNA (Kerans & Karr 1994; Cranston 2000; Rabeni & Wang
barcodes are reduced by placing tissue directly into a 2001; Nijboer et al. 2005; Roque et al. 2010) although
PCR master mix without DNA extraction prior to ampli- recent studies have shown that inclusion improves accu-
fication of the target gene. This aids in reducing the time racy and signal-to-noise ratio (Ferrington 2008; Raunio
taken (4–16 h depending on extraction protocol) and cost et al. 2011; Brodin et al. 2013; Milosevic et al. 2013). In
by eliminating the usage of kits and decreasing man- addition, the high abundance of chironomids in urban
power needs. Direct PCR procedures were proposed as aquatic systems can degrade the recreational value of
early as 1993 (Panaccio et al. 1993), but success rates were waterfront areas. Thus, their mass occurrence can have
either unreported or low. For example, Panaccio et al. serious economic implications and correct species-level
(1993) and Grevelding et al. (1996) did not mention identification for their control is crucial (Cranston et al.
amplification success rates nor did they discuss whether 2013).
the amplified specimens were suitable for morphological Given the problems with morphology in this case,
study and vouchering. Rochlin et al. (2007) evaluated the DNA barcodes are an obvious alternative or complemen-
method by placing one Culex mosquito leg into the well tary tool as long as one can demonstrate that at least
containing PCR reagents and distilled water but reported most midge species have discrete barcodes. Currently,
low success rate (<50%) that were not likely to encourage the chironomid literature indicates that this is the case.
the use of this method. Yet, our experience with the Most barcoded species are separated by genetic distances
method indicates that it can yield consistently high suc- of >4% at the DNA barcode region while intraspecific
cess rates (>80%) for a wide range of taxa including flies variation rarely exceed 2% (Ekrem et al. 2007, 2010;
(Culicidae, Drosophilidae, Dolichopodidae, Sepsidae) Sinclair & Gresens 2008; Cranston et al. 2013). Beyond
and sea stars (Oreasteridae). Key to success is the optimi- their use in routine identification, DNA barcodes have
zation of a few parameters. We thus consider direct PCR other important uses in Chironomidae. First, species are
a convenient method for taxonomists and molecular traditionally described based on adults, rendering larvae
ecologists who routinely have to obtain DNA barcodes unidentifiable until they are associated with adults via
for taxa, which are used widely for bioassessment or rearing or DNA barcoding. Making such associations is
studies of genetic variability (e.g. Chironomidae, Culici- important because larvae are the more important stage
dae, Drosophilidae). for freshwater assessment (Ferrington 2008; Marziali
et al. 2009; Luoto 2011; Nandi et al. 2012 Milosevic et al.
2013) and nowadays DNA barcodes are the choice
Direct PCR of Chironomidae
method for this purpose (Carew et al. 2005; Ekrem et al.
We here use nonbiting midges (Diptera: Chironomidae) 2007). Second, DNA barcodes have been important for
to describe how to optimize direct PCR. These midges the discovery of putatively cryptic species (Ekrem et al.
are extremely abundant in many freshwater samples and 2007; Sinclair & Gresens 2008).
routinely used as key bioindicator taxa for water quality Here, we describe how an optimized direct PCR pro-
assessment (Ferrington 2008; Luoto 2011; Milosevic et al. cedure without prior DNA extraction can yield success
2013) as different species have distinct ecological prefer- rates of 80–100% for a range of taxa. Furthermore, we
ences (Punti et al. 2009; Krosch & Cranston 2012; Molozzi document that direct PCR preserves the cuticle such that
et al. 2012). However, accurate identification of chirono- diagnostic morphological features can be studied and
mid larvae is difficult because the morphological features allows for the recovery of genomic DNA for the amplifi-
are often minuscule and easily obscured (Epler 2001; cation of additional mitochondrial and nuclear genes.
Sharley et al. 2004; Hajibabaei et al. 2011; Kim et al. 2012). We also discuss other invertebrate taxa for which we
These factors can lead to high error rates when the larvae were able to optimize direct PCR successfully.
are identified by parataxonomists (Cranston & Hillman
1992; Krell 2004), thus impacting subsequent water qual-
Materials and methods
ity analyses. In addition to identification problems, the
cost of morphological identification is high because
Chironomidae
microscopic slides have to be prepared (Cranston 1994;
Epler 2001; Carew et al. 2007; Cranston et al. 2013), which Larvae and adults of various sizes and species were
requires approximately 15–20 min per specimen. Yet, it collected from three freshwater habitats (Bedok
is not unusual to find thousands of chironomid larvae in Reservior—1°200 32″N 103°550 30″E, Upper Seletar Reserv-
one environmental sample. It is therefore not surprising ior—1°240 04″N 103°480 14″E, Pandan Reservior—1°180 50″
that the high cost of morphological identification has led N 103°440 30″E) in Singapore via sediment grab sampling
to a debate whether chironomid larvae should be (larvae) or light trapping (adults) and preserved in
included in bioassessment of aquatic environments 70–100% ethanol or isopropanol (Cranston et al. 2013).

© 2014 John Wiley & Sons Ltd

 D I R E C T P C R W I T H O U T D N A E X T R A C T I O N 1273

Larvae and adults were presorted into three size classes mens (see below). In addition to tissue amount and
(Fig. 1). For adults, we consider ‘small’ midges to range source, the type of polymerase influenced PCR suc-
from (anterior to posterior) 1.5–2.4 mm, ‘medium’ from cess rates. Four different polymerases were tested
2.5 to 3.4 mm and ‘large’ >3.5 mm. The corresponding (TaKaRa ExTaq, Qiagen HotStar Taq, KAPA Taq and
size classes for larvae are small (3.0–4.4 mm), medium a homemade Taq). The total reaction volume was
(4.5–6.4 mm) and large (6.5–8.0). 20 lL, with 2 lL of Taq 109 buffer, 1.5 lL of 2 mM
dNTPs, 1 lL of 10 lM primers and varying volume of
Taq- and DNase-free sterile water depending on the
Direct PCR optimization
Taq brand (0.15 lL TaKaRa ExTaq; 0.5 lL Qiagen
We initially used whole specimens for direct PCR. HotStar Taq; 0.5 lL KAPA Taq; 1 lL homemade Taq).
The success rates with primer pair LCO 1490/HCO The highest success rates were achieved using TaKa-
2198 (Folmer et al. 1994) were as low as reported in Ra ExTaq while KAPA Taq and the homemade Taq
the literature (<50%; Rochlin et al. 2007; small adults, gave successful amplifications, but the products were
n = 8, 87.5% successful; medium adults, n = 4, 0%; often ‘smeary’ under UV light. Qiagen HotStar Taq
large adults, n = 4, 0%; small larvae, n = 8, 50%; large did not yield successful amplifications. Given that the
larvae, n = 8, 37.5%). Exceptions were the smallest removal of legs requires time, we tested also if whole
adults (1.5–2.4 mm) that yielded high success rates. specimens for midges and larvae in the mid- and
We thus adjusted the tissue amount for medium- and large-sized classes could be used if a larger reaction
large-sized adults (2.5–3.4 mm and >3.5 mm, respec- volume (40 lL) was used, but success rates were not
tively), and all three size classes of larvae (3.0– as consistent and predictable as with 20 lL and the
4.4 mm, 4.5–6.4 mm and 6.5–8.0 mm, respectively). tissues described above. Lastly, the addition of 1 lL
High success rates were obtained when using (1 mg/mL) of bovine serum albumin (BSA) did not
(i) three legs from medium-sized adult and two legs influence success rates.
from large-sized adults (n = 4: 100% successful),
(ii) the dissected anterior body segments from larvae
Optimized protocol for direct PCR and sequencing
of all three size classes (small larvae, n = 5: 80% suc-
cessful; medium larvae, n = 10: 90% successful; large After optimizing direct PCR for chironomids, we tested
larvae, n = 5: 60% successful). We also experimented the procedure with a new and larger batch of specimens
with another tissue source (dissected abdomen of (Table 2: 30–60). Whole specimens (small adults), three
mid- and large-sized adults; n = 4: 0% successful). legs (medium adult), two legs (large adult), anterior half
After these trials with small numbers of specimens, of larval body (small larvae), or head capsule and 2–3
we applied the methods to larger numbers of speci- anterior ‘segments’ (total length = ~1 mm) for the

(A) (B) (C) Fig. 1 Sizes of specimens and the corre-

sponding parts (indicated by arrows)
used for direct PCR. Larvae are shown in
the top row: (A) small-sized larvae, (B)
medium-sized larvae and (C) large-sized
larvae. Adults are shown in the bottom
row: (D) small-sized adults, (E) medium-
sized adults and (F) large-sized adults.

(D) (E) (F)

© 2014 John Wiley & Sons Ltd

1274 W . H . W O N G E T A L .

medium- and large-sized larvae were used. Relevant agent. For medium- and large-sized adult, this test was
body parts intended for direct PCR were removed, dried not needed because the diagnostically important body
briefly, then placed into the PCR wells and dried further parts were not used during direct PCR and even the legs
for 1 min using a Speed-Vac. PCR reaction mixtures were could be retrieved after direct PCR.
prepared (20 lL volume: 2 lL of 109 ExTaq Buffer,
1.5 lL of 2 mM dNTP mixture, 0.15 lL of TaKaRa ExTaq
Recovery of genomic DNA
polymerase, 1 lL of 10 lM primers for the forward and
reverse direction, DNase-free sterile water) and subse- We tested two methods for recovering genomic DNA
quently pipetted into the wells containing the dried spec- from the small adults that were exposed to direct PCR.
imens. The DNA barcoding region consisting of a 658-bp First, the specimens from the PCR wells were taken out
fragment of COI was amplified using the general inverte- and subjected to DNA extraction using the DNeasy
brate primers LCO 1490—50 GGT CAA CAA ATC ATA Blood and Tissue Kit (Qiagen) according to the manufac-
AAG ATA TTG G 30 and HCO 2198 50 TAA ACT TCA turer’s protocol. For the second method, we performed
GGG TGA CCA AAA AAT CA 30 (Folmer et al. 1994). genome amplification for 14 COI PCR products obtained
Cycling conditions were: denaturation at 94 °C for 30 s, with direct PCR earlier. The DNA was successfully
annealing at 51 °C for 30 s and extension at 72 °C for amplified using the Repli-G Mini Kit (Qiagen) according
1 min. The cycle was repeated for 41 times. After PCR, to manufacturer’s protocol but with halved reaction vol-
the body parts inside the PCR wells were removed, pre- umes (from 50 to 25 lL). Subsequently, PCRs were per-
served as vouchers in 70% or 100% ethanol and stored in formed for three genes using the genomic DNA obtained
20 °C. The amplification results were checked with by both methods (COI: nonbarcode segment; 18s: nuclear
electrophoresis on a 1% agarose gel stained with GelRed rDNA; AATS: segment 1, nuclear protein-encoding).
(Biotium Inc.). Successfully amplified products were These genes were used previously in phylogenetic stud-
purified with SureCleanTM (Bioline) according to the ies of Chironomidae (Cranston et al. 2010, 2012; Dahle
manufacturer’s instructions. Purified products of small- 2012; Krosch & Cranston 2012). Primer sequences and
and large-sized adults and larvae were sequenced in PCR cycling conditions are provided in Table 1.
both directions with BigDye (PerkinElmer) terminator
reactions and analysed on the ABI Avant 3130xl Genetic
Direct PCR of other taxa
Analyzer. Medium-sized adults and larvae were
sequenced in single direction (forward) after we found We subsequently used the same optimization procedures
that single direction sequencing can yield >500 bases described here in an attempt to optimize direct PCR for
which is sufficient for routine species delimitation. other taxa (Copepoda; Coleoptera: Dytiscidae; Hyme-
The chromatograms were assembled and edited using noptera: Formicidae and Odonata; Diptera: Culicidae,
Sequencher ver. 4.6. Alignment of sequences was per- Dolichopolidae, Drosophilidae and Sepsidae; Valvatida:
formed in MAFFT version 7 with default options. The Oreasteridae).
alignments were translatable to amino acid sequences Specimens of Drosophilidae and Sepsidae were
free of stop codons and the sequences checked via obtained from reared cultures at Temasek Life Sciences
MEGABLAST in GenBank and compared to sequences Laboratory and Evolutionary Biology Laboratory
in our own local database for chironomids of Pandan, (National University of Singapore), respectively. Culici-
Bedok, and Upper Seletar Reservoirs in Singapore dae specimens were either obtained from Malaise traps
(Cranston et al. 2013). The sequences were assigned to or from cultures at Duke-NUS Graduate Medical School
species clusters as described in Meier et al. (2006) using insectary, while all specimens of Dolichopolidae and
uncorrected pairwise distances (4%) as recommended in Formicidae were collected from Malaise traps. The
Srivathsan & Meier (2012). reared culture specimens were preserved in 100% etha-
nol, while the Malaise trap specimens were preserved in
70% ethanol. The Oreasteridae were sampled at Chek
Retrieval of body parts for morphological study
Jawa, Pulau Sekudu, Cyrene Reef, Pulau Semakau and
Adult chironomids are identified based on male genita- Beting Bronok, while the Odonata samples came from
lia, structures of the thorax and wing patterns, while lar- Golf Link Marsh, Kent Ridge Park, Bishan Park, Nee
vae are distinguished based largely on characters of the Soon Swamp Forest and Pulau Ubin and the Copepoda
head capsule (Cranston 1994; Epler 2001; Cranston et al. samples from Pandan and Upper Seletar Reservoirs. The
2013). It was thus important to ascertain that specimens specimens were preserved in 100% ethanol. All locations
that had undergone direct PCR could still be used as mentioned are in Singapore. The Dytiscidae specimens
vouchers. Slide mounts of specimens that had undergone were collected in Brandenburg, Germany, and preserved
direct PCR were prepared using Hoyer’s mounting in 96% ethanol.

© 2014 John Wiley & Sons Ltd

 D I R E C T P C R W I T H O U T D N A E X T R A C T I O N 1275

Table 1 Primers and cycling conditions used for PCRs of genome-amplified PCR products

Locus/Primer Primer sequence Reference Cycling condition

COI Denaturation: 94 °C, 30 s

s2183 50 CAA CAT TTA TTT TGA TTT TTT GG 30 Simon et al. (1994) Annealing: 51–54 °C, 30 s
Extension: 72 °C, 1 min 30 s
a3014 50 TCC AAT GCA CTA ATC TGC CAT ATT A 30 Simon et al. (1994) Final extension: 72 °C, 3 min
Number of cycles: 41

18S Denaturation: 94 °C, 1 min

18S_ai 50 CCT GAG AAA CGG CTA CCA CAT C 30 Whiting et al. (1997) Annealing: 55–58 °C, 1 min
Extension: 72 °C, 1 min 30 s
18S_bi 50 GAG TCT CGT TCG TTA TCG GA 30 Whiting et al. (1997) Final extension: 72 °C, 3 min
Number of cycles: 41

AATS Denaturation: 94 °C, 30 s

A1-92F 50 TAY CAY CAY CAN TTY TTY GAR ATG 30 Regier (2008) Annealing: 51–54 °C, 30 s
Extension: 72 °C, 1 min
A1-244R 50 ATN CCR CAR TCN ATR TGY T 30 Feng-Yi Su et al. (2008) Final extension: 72 °C, 3 min
Number of cycles: 41

Skuse, 1889; Polypedilum leei Freeman, 1961; Polypedi-

Results
lum (Pentapedilum) nodosum Johannsen, 1932; Polypedi-
lum griseoguttatum Kieffer, 1921; Polypedilum sp. 1 and
Amplification success rate
Chironomus circumdatus Kieffer, 1916.
DNA barcodes were sequenced for three different size For the nonchironomid taxa tested, the initial amplifi-
classes of chironomid adults and larvae. We tested cation success rates were low, but rose to >80% for most
≥30 specimens for each size class and stage. The taxa after optimizing tissue quantity, tissue type, primer
direct PCR success rates were high, ranging from pair and Taq polymerase (Table 3). Once optimized for a
90% to 100% for all size classes and stages (Table 2; particular taxon, the same recipe worked consistently
see Fig. 2A). DNA sequence quality was comparable and could be used for subsequent samples. However,
to what would have been obtained using traditional certain taxa appeared unsuitable for direct PCR. Regard-
techniques involving DNA extraction (Fig. 2B). less of body parts used (head, legs and thorax), Formici-
Sequence clustering at 4% using SpeciesIdentifier ver. dae yielded low success rates of <20%, possibly due to
1.7.9 (Meier et al. 2006) revealed that the adult DNA glands that may produce PCR inhibitors (Billen 2009;
barcodes belonged to eight molecular operational tax- Schrader et al. 2012). Optimization was also unsuccessful
onomic units (MOTUs), while larval barcodes for the heavily sclerotized Dytiscidae and Copepoda. Of
belonged to 10 MOTUs. According to our morpholog- the taxa with less sclerotized legs/bodies, only Odonata
ical and molecular databases for Singapore’s reser- failed to yield high success rates. The taxa that failed to
voirs, the MOTUs belong to the following species/ amplify via direct PCR could be amplified with extracted
genera: Tanytarsus oscillans Johannsen, 1932; Tanytarsus DNA when using normal PCR recipes with the same pri-
ovatus Johannsen, 1932; Cladotanytarsus sp.1, Cladotany- mer pairs thus suggesting that direct PCR was the source
tarsus sp. 2, Paratanytarsus sp.1, Polypedilum nubifer of the problem.

Table 2 Chironomidae: number of direct PCR reactions and associated success rates. Refer to Dyrad doi in data accessibility for full
specimen information

Morphospecies Sample size Successful attempts Success rate for first amplification attempt (%) GenBank Accession no.

Small adult 62 59 95.2 KJ530765–KJ530823

Medium adult 33 31 93.9 KJ530824–KJ530854
Large adult 31 28 90.3 KJ530855–KJ530882
Small larvae 30 30 100 KJ530883–KJ530912
Medium larvae 30 29 96.7 KJ530913–KJ530941
Large larvae 30 28 93.3 KJ530942–KJ530969

© 2014 John Wiley & Sons Ltd

1276 W . H . W O N G E T A L .

Fig. 2 (A) Agarose gel electrophoresis of

(A)
direct PCR product from small-sized
larvae 1 (L1–L8) using LCO-1490 and
HCO-2198. WL, DNA weight ladder; N,
negative control, (B) DNA sequence with
BigDye nucleotide peaks visualized in
DNA Sequencher (position 78 to position
224).

(B)

with extracted DNA: COI: 64.3% (9/14), AATS1: 0%

Recovery of morphological features and genomic DNA
(0/14), 18s rDNA: 21.4% (3/14).
in Chironomidae
Direct PCR does not impair the morphologically impor-
Discussion
tant diagnostic characters. Male genitalia characters such
as anal tergite bands, medium volsella, superior volsella, Direct PCR has high success rates for a wide range
and gonostylus shapes, among others were preserved of taxa once a few key parameters are optimized. The
during direct PCR. Similarly, the characters of larvae most critical factor is the amount of tissue because
such as the mentum, mandible, antenna etc were intact too little or too much template released during the
after the procedure (Fig. 3). initial heating step of PCR often leads to PCR failure,
It is shown that genomic DNA can be recovered presumably due to a suboptimal reagent ratio (Kra-
after direct PCR via one of the two methods tested. mer & Coen 2001). Fortunately, nowadays PCR is
Of the two genomic DNA recovery methods successful across a fairly wide range of reagent ratios
described above, genome amplification of the direct so that, for example, for chironomids we only have
PCR products proved more effective because the suc- to differentiate between three different size classes in
cess rates for PCR of three nonbarcoding markers order to obtain high success rates. Note that high
were significantly higher: COI: 100% (14/14), AATS1: success rates for direct PCR are clade-independent
100% (14/14), 18s rDNA: 78.6% (11/14) than those because the tested specimens belong to five genera

Table 3 List of taxa and success rates for direct PCR

Number of
Life Body Type of successful Success
Taxa stages parts used* Primer Pairs (COI) Sources polymerase trials rate (%)

Culicidae Adult 2–3 legs LEP F1/LEP R1 Hebert et al. (2004) TaKaRa 27/30 90
ExTaq
Drosophilidae Adult 2–3 legs/ LCO-1490/HCO-2198 Folmer et al. (1994)/ TaKaRa 25/30 83.3
whole body or LEP F1/LEP R1 Hebert et al. (2004) ExTaq
Dolichopodidae Adult 1–3 legs LCO-1490/HCO-2198 Folmer et al. (1994) homemade 17/20 85
Taq
Sepsidae Adult 1–3 legs mtD4/mtD9 Simon et al. (1994) TaKaRa ExTaq 77/95 81.1
Oreasteridae Adult 1 mm3 tube foot tRNAasn42F/ Crandall et al. (2008) homemade Taq 78/83 93.9
ValvaCOI-770R

*Depends on the size of specimens, c.f. Chironomidae’s size.

© 2014 John Wiley & Sons Ltd

 D I R E C T P C R W I T H O U T D N A E X T R A C T I O N 1277

(A) Fig. 3 Morphology of adults and larvae

(B)
after direct PCR (A) Hypopygium mor-
phology of Tanytarsus sp. ap, anal point;
iv, inferior volsella; sv, superior volsella;
gs, gonostylus. (B) Larval head capsule.
mt, mentum; vmp, ventromental plate;
mdb, mandible.

(Tanytarsus, Cladotanytarsus, Paratanytarsus, Polypedilum, tions, however, and we routinely use direct PCR
Chironomus). Adult chironomids like most other ar- successfully for Malaise trap material preserved initially
thropods are particularly convenient tissue sources for in 70% ethanol (Dolichopodidae).
direct PCR because (i) legs are easily removed thus
providing a convenient way for scaling tissue quan-
Direct PCR: why now?
tity, (ii) they rarely contain large glands that may
harbour PCR inhibitors, and (iii) they are bilaterally Given that direct PCR procedures have been known for
symmetrical so that diagnostic features can be more than 20 years, one has to wonder why they have
retained on the specimen by removing only one leg not been widely adopted. We believe that there are three
on either side of the body. Should initial direct PCR reasons. First, early publications on direct PCR reported
fail, additional pairs of leg are available for a second low or no success rates which are unlikely to encourage
trial. Note that the legs can be retrieved after the pro- the adoption of the method. We find that such discourag-
cedure without significant damages to the structures. ingly low rates are normal during initial rounds of direct
A second factor that influences direct PCR success PCR optimization. We believe that some laboratories
rates is the choice of Taq polymerase. Over the years, we have tried direct PCR, but abandoned it following low,
have observed that some taxa require expensive and initial success rates. These laboratories never went to the
high-fidelity Taq polymerases (e.g. Chironomidae) while optimization stage.
high success rates can be attained for other taxa even Second, scientists may feel uneasy about direct PCR
with low-cost and homemade enzymes (e.g. Sepsidae). procedures that use whole specimens because the initial
Given that simultaneous optimization of tissue quantity barcode may reveal that a specimen belongs to a particu-
and enzyme is time-consuming, we recommend first larly interesting species. Sequencing of additional genes
optimizing the former with a high-fidelity Taq polymer- is then desirable, but it remained untested whether geno-
ase. Once PCR success rates are high, a cheaper and/or mic DNA can be recovered from a specimen that had
more versatile Taq polymerase can be tested. A third fac- already been used in direct PCR. We here show that this
tor that influences PCR success rates is the source of tis- is possible as long as a genome amplification of the origi-
sues. For insects, legs generally work well while nal amplification product is performed. Using the ampli-
abdominal tissues should be avoided. Low success rates fied genome as template, we were able to sequence
for the latter suggest that they may contain PCR inhibi- additional mitochondrial and nuclear markers. Initially,
tors that could come from the digestive tract (Juen & we worried that such procedure would overamplify the
Traugott 2006). If abdominal tissues cannot be avoided COI sequence because it is already enriched in the direct
(e.g. chironomid larvae), success rates increase when the PCR product, but the primers used by the Repli-G Mini
anterior-most abdominal segments are used. In addition, Kit appear to be sufficiently diverse that genomic DNA
we have recently increased success rates further using is amplified effectively. Note that the recovery of tem-
more specific primer pairs (LEP F1/LEP R1 instead of plate DNA via genome amplification is needed only for
LCO 1490/HCO 2198 for insect taxa). Lastly, not surpris- small specimens where the whole body is used as tem-
ingly, storage conditions of specimens influence success plate in direct PCR. For most specimens, direct PCR uti-
rates. Fresh specimens that are used immediately after lizes only legs so that most of the specimen remains in
preservation and/or are stored at low temperatures pristine condition and can be used for DNA extraction.
amplify easily, whereas older material with inferior pres- Lastly, based on previous studies on direct PCR, it
ervation and storage conditions can fail. There are excep- was unclear whether the morphology of specimens

© 2014 John Wiley & Sons Ltd

1278 W . H . W O N G E T A L .

undergoing direct PCR remained intact for vouchering. the next best choice would be DNA barcodes. Again, for
In this study, we found that even sensitive and small these laboratories, a cheaper and faster technique for
body parts are well preserved. Re-identification and generating barcodes will help with implementing itera-
description of specimens that have undergone direct tive procedures for validating parataxonomists’ morpho-
PCR can be performed. species boundaries. Moreover, the DNA barcodes
generated can serve as libraries that are subsequently
used for NGS-based bioassessment.
Direct PCR: still relevant?
Molecular ecology and bioassessment are quickly
Conclusions
adopting next-generation-sequencing (NGS) (Hajibabaei
et al. 2011; Carew et al. 2013) and one may wonder Direct PCR is an old technique neglected by most molec-
whether there is still a need for a technique such as ular ecologists. However, our study demonstrates that
direct PCR. However, the analysis of NGS data gener- for many taxa, high success rates can be achieved with-
ally involves the identification of species (e.g. food out substantial damage to specimens once PCR condi-
species) via barcode-like markers (Porazinska et al. tions have been optimized. This reduces overall
2009; Zhou et al. 2013). Hence, DNA barcode databases sequencing cost by saving time via skipping the DNA
are still very much needed. Unfortunately, these data- extraction step. Based on several years of experience, we
bases remain extremely species-poor for many groups only occasionally encounter taxa for which direct PCR
(Harris 2003; Ekrem et al. 2007; Begerow et al. 2010; optimization fails. Note that further reductions in cost
Kwong et al. 2012; Kvist 2013). Improving species cov- can be achieved through one-directional sequencing,
erage will require a large amount of species-specific which is sufficient for routine identifications.
Sanger sequencing in order to obtain new DNA bar-
codes for additional species. Given that many very
Acknowledgements
common species have already been barcoded, most
new barcodes will have to be generated for rare spe- We would like to express our gratitude to Rayson Lim, Audrey
cies and these are likely to be produced by laborato- Heyzer (Tropical Marine Science Institute) and Jeremy Woodford
(Environmental Health Institute) for providing the chironomid
ries that specialize in particular taxa and that thus
specimens. We also wish to thank Dr. Ian Mendenhall and the
have an incentive for optimizing direct PCR for these
insectary at Duke-NUS Graduate Medical School for providing
taxa. Indeed, much of the species coverage for COI the material for Culicidae. The images were prepared with the
sequences in GenBank already comes from projects help of Ang Yuchen (Evolutionary Biology Laboratory, National
that are unlikely to be associated directly with DNA University of Singapore). We also thank Diego Pitta de Araujo
barcoding studies (Kwong et al. 2012). and Amrita Srivathsan (Evolutionary Biology Laboratory,
Direct PCR will be particularly attractive for two National University of Singapore) for their input in the prepara-
tion of manuscript. Funding was provided by a grant from the
types of laboratories. The first type is invertebrate sys-
Public Utility Board (Singapore), WBS: R-154-000-526-490.
tematics laboratories. Such laboratories routinely acquire
large numbers of specimens in target groups. Many such
laboratories have already adopted sample processing
References
procedures that consist of rough sorting based on mor-
phology with subsequent testing of the morphological Begerow D, Nilsson H, Unterseher M, Maier W (2010) Current state and
units via COI sequencing for exemplars from the puta- perspectives of fungal DNA barcoding and rapid identification proce-
dures. Applied Microbiology and Biotechnology, 87, 99–108.
tive morpho-species (Tan et al. 2010). Species identifica- Billen J (2009) Occurrence and structural organization of the exocrine
tion is ultimately achieved through the evaluation of glands in the legs of ants. Arthropod Structure & Development, 38, 2–15.
multiple data sources (Tan et al. 2010; Yeates et al. 2011). Brodin Y, Ejdung G, Strandberg J, Lyrholm T (2013) Improving environ-
mental and biodiversity monitoring in the Baltic Sea using DNA bar-
This procedure requires that many specimens are bar-
coding of Chironomidae (Diptera). Molecular Ecology Resources, 13,
coded and any shortcut to the process (e.g. direct PCR) is 996–1004.
very welcome. The second type of laboratories that Carew ME, Pettigrove V, Hoffmann AA (2005) The utility of DNA mark-
should be interested in direct PCR is those that employ ers in classical taxonomy: using cytochrome oxidase 1 markers to dif-
ferentiate Australian Cladopelma (Diptera: Chironomidae) midges.
parataxonomists in bioassessment. These laboratories are
Annals of the Entomological Society of America, 98, 587–594.
increasingly aware that the accuracy of the sorting by Carew ME, Pettigrove V, Cox RL, Hoffmann AA (2007) DNA identifica-
paratoxonomist can be improved through a feedback tion of urban Tanytarsini chironomids (Diptera: Chironomidae). Jour-
loop that provides information on the validity of the nal of the North America Benthological Society, 26, 587–600.
Carew ME, Pettigrove V, Metzeling L, Hoffmann AA (2013) Environ-
morphospecies that were delimited by parataxonomists
mental monitoring using next generation sequencing: rapid identifi-
(Stribling et al. 2008). Ideally, the verification source is cation of macroinvertebrate bioindicator species. Frontiers in Zoology,
taxonomic experts, but they are often not available and 10, 45.

© 2014 John Wiley & Sons Ltd

 D I R E C T P C R W I T H O U T D N A E X T R A C T I O N 1279

Crandall ED, Jones ME, Munoz MM, Akinronbi B, Erdmann MV, Barber butterfly Astraptes fulgerator. Proceedings of the National Academy of
PH (2008) Comparative phylogeography of two seastars and their ecto- Sciences of the United States of America, 101, 14812–14817.
symbionts within the Coral Triangle. Molecular Ecology, 17, 5276–5290. Johannsen OA (1932) Chironominae of the Malayan Subregion of the
Cranston PS (1994) Systematics. In: The Chironomidae: Biology and Ecology Dutch East Indies. Archiv fiir Hydrobiologie, Supplement, 11, 503–552.
of Non-biting Midges (eds Armitage PD, Cranston PS, Pinder LCV), Juen A, Traugott M (2006) Amplification facilitators and multiplex
pp. 31–61. Chapman and Hall, London, Glasgow, New York, Tokyo, PCR: tools to overcome PCR-inhibition in DNA-gut-content analy-
Melbourne, Madras. sis of soil living invertebrates. Soil Biology and Biochemistry, 38,
Cranston PS (2000) Monsoonal tropical Tanytarsus van der Wulp (Dip- 1872–1879.
tera: Chironomidae) reviewed: new species, life histories and signifi- Kerans B, Karr JR (1994) A benthic index of biotic integrity (B-IBI) for riv-
cance as aquatic environmental indicators. Australian Journal of ers of the Tennessee Valley. Ecological Application, 4, 768–785.
Entomology, 39, 138–159. Keskin E, Atar HH (2013) DNA barcoding commercially important fish
Cranston PS, Hillman P (1992) Rapid assessment of biodiversity using species of Turkey. Molecular Ecology Resources, 13, 788–797.
‘biological diversity technicians’. Australian Biologist, 5, 144–155. Kieffer JJ (1916) Tendipedides (Chironomides) de Formose conserves au
Cranston PS, Hardy NB, Morse GE, Pusledink L, McCluen SR (2010) Museum National Hangrois de Budapest et determines par J. J. KIEF-
When molecules and morphology concur: the ‘Gondwana’ midges FER. Annales historic-naturales Musei nationalis Hungarici, 14, 81–121.
(Diptera: Chironomidae). Systematic Entomology, 35, 636–648. Kieffer JJ (1921) Nouveaux representants du genre Polypedilum (Chirono-
Cranston PS, Hardy NB, Morse GE (2012) A dated molecular phylogeny mides). Annales de la Societe Scientifiques Bruxelles, 41, 97–101.
for the Chironomidae (Diptera). Systematic Entomology, 37, 172–188. Kim S, Song K, Ree H, Kim W (2012) A DNA barcode library for Korean
Cranston PS, Ang YC, Heyzer A et al. (2013) The nuisance midges (Dip- Chironomidae (Insecta: Diptera) and indexes for defining barcode gap.
tera: Chironomidae) of Singapore’s Pandan and Bedok reservoirs. The Molecules and Cells, 33, 9–17.
Raffles Bulletin of Zoology, 61, 779–793. Kramer MF & Coen DM (2001) Enzymatic amplification of DNA by PCR:
Dahle S (2012) Phylogenetic analysis of the genus Paratanytarsus (Diptera: standard procedures and optimization. In: Current Protocols in Molecu-
Chironomidae). Master’s Thesis, Norwegian University of Science and lar Biology, Unit 15.1, October 2001 (ed. Harkins B) 15.1.1–15.1.14. John
Technology. Available from ntnu.diva-portal.org/smash/get/diva Wiley & Sons Inc., Hoboken, New Jersey, 2003.
2:608830/FULLTEXT01.pdf. Krell FT (2004) Parataxonomy vs taxonomy in biodiversity studies – pit-
Dinc
a V, Zakharov EV, Hebert PDN, Vila R (2011) Complete DNA bar- falls and applicability of ‘morphospecies’ sorting. Biodiversity and Con-
code reference library for a country’s butterfly fauna reveals high per- servation, 13, 795–812.
formance for temperate Europe. Proceedings. Biological Sciences/The Krosch MN, Cranston PS (2012) Non-destructive DNA extraction from
Royal Society, 278, 347–355. Chironomidae, including of fragile pupal exuviae, extends analyzable
Ekrem T, Willassen E, Stur E (2007) A comprehensive DNA sequence collections and enhances vouchering. Chironomus Newsletter on Chiro-
library is essential for identification with DNA barcodes. Molecular nomidae Research, 25, 22–27.
Phylogenetics and Evolution, 43, 530–542. Kvist S (2013) Barcoding in the dark?: a critical view of the sufficiency of
Ekrem T, Stur E, Hebert PDN (2010) Females do count: documenting Chi- zoological DNA barcoding databases and a plea for broader integra-
ronomidae (Diptera) species diversity using DNA barcoding. Organ- tion of taxonomic knowledge. Molecular Phylogenetics and Evolution, 69,
isms Diversity and Evolution, 10, 397–408. 39–45.
Epler JH (2001) Identification manual for the larval Chironomidae (Dip- Kwong S, Srivathsan A, Meier R (2012) An update on DNA barcoding:
tera) of North and South Carolina. North Carolina Department of low species coverage and numerous unidentified sequences. Cladistics,
Environmental and Natural Resources; Division of Water Quality, Spe- 28, 639–644.
cial publication SJ2001-SP13. Luoto TP (2011) The relationship between water quality and chironomid
Feng-Yi Su K, Narayanan Kutty S, Meier R (2008) Morphology versus distribution in Finland – a new assemblage-based tool for assessments
molecules: the phylogenetic relationships Sepsidae (Diptera: Cyclorrh- of long term nutrient dynamics. Ecological Indicators, 11, 255–262.
apha) based on morphology and DNA sequence data from ten genes. Marziali L, Armanini DG, Cazzola M et al. (2009) Responses of Chirono-
Cladistics, 24, 902–916. mid larvae (Insecta, Diptera) to ecological quality in Mediterranean
Ferrington LC (2008) Global diversity of non-biting midges (Chironomi- river mesohabitats (South Italy). River Research and Applications, 26,
dae: Insecta-Diptera) in freshwater. Hydrobiologia, 595, 447–455. 1036–1051.
Folmer O, Black M, Hoeh W, Lutz R, Wrijenhoek R (1994) DNA primers Meier R, Kwong S, Vaidya G, Ng PKL (2006) DNA barcoding and taxon-
for amplification of mitochondrial cytochrome c oxidase subunit 1 omy in Diptera: a tale of high intraspecific variability and low identifi-
from diverse metazoan invertebrates. Molecular Marine Biology and Bio- cation success. Systematic Biology, 55, 715–728.
technology, 3, 294–299. Meier R, Zhang GY, Ali F (2008) The use of mean instead of smallest
Freeman P (1961) The Chironomidae (Diptera) of Australia. British Musuem interspecific distances exaggerates the size of “Barcoding Gap” and
(Natural History), London, 731 pp. leads to misidentification. Systematic Biology, 57, 809–813.
Gotelli NJ (2004) A taxonomic wish-list for community ecology. Philo- Milosevic D, Simic V, Stojkovic M et al. (2013) Spatio-temporal pattern of
sophical Transactions of the Royal Society of London. Series B, Biological Sci- the Chironomidae community: toward the use of non-biting midges in
ences, 359, 585–597. bioassessment programs. Aquatic Ecology, 47, 37–55.
Grevelding CG, Kampk€ otter A, Hollmann M, Sch€afer U, Kunz W (1996) Molozzi J, Feio MJ, Salas F, Marques JC, Callisto M (2012) Development
Direct PCR on fruitflies and blood flukes without prior DNA extrac- and test of a statistical model for the ecological assessment of tropical
tion. Nucleic Acids Research, 24, 4100–4101. reservoirs based on benthic macroinvertebrates. Ecological Indicators,
Hajibabaei M, Shokralla S, Zhou X, Singer GAC, Baird DJ (2011) 23, 155–156.
Environmental barcoding: a next-generation sequencing approach Nandi S, Aditya G, Saha GK (2012) Nutrient conditions and chironomid
for biomonitoring applications using river benthos. PLoS One, 6, productivity in Kolkata, India: assessment for biomonitoring and eco-
e17497. logical management. Journal of Limnology, 71, 320–329.
Harris DJ (2003) Can you bank on GenBank? Trends in Ecology and Evolu- Nijboer RC, Verdonschot PFM, Van Der Werf DC (2005) The use of indi-
tion, 18, 317–319. cator taxa as representatives of communities in bioassessment. Fresh-
Hebert PDN, Cywinska A, Ball SL, deWaard JR (2003) Biological identifi- water Biology, 50, 1427–1440.
cations through DNA barcodes. Proceedings of the Royal Society – Bio- Panaccio M, Georgesz M, Hollywell C, Lew A (1993) Direct PCR from
logy, 270, 313–321. solid tissue without DNA extraction. Nucleic Acids Research, 21, 4656.
Hebert PDN, Stoeckle MY, Zemlak TS, Francis CM (2004) Ten species in Pfenninger M, Nowak C, Kley C, Steinke D, Streit B (2007) Utility of
one: DNA barcoding reveals cryptic species in the neotropical skipper DNA taxonomy and barcoding for the inference of larval community

© 2014 John Wiley & Sons Ltd

1280 W . H . W O N G E T A L .

structure in morphologically cryptic Chironomus (Diptera) species. Tan DSH, Ang Y, Lim GS, Ismail MR, Meier R (2010) From ‘cryptic
Molecular Ecology, 16, 1957–1968. species’ to integrative taxonomy: an iterative process involving
Porazinska DL, Giblin-Davis RM, Faller L et al. (2009) Evaluating high- DNA sequences, morphology, and behaviour leads to the
throughput sequencing as a method for metagenomic analysis of nem- resurrection of Sepsis pyrrhosoma (Sepsidae: Diptera). Zoologica
atode diversity. Molecular Ecology Resources, 9, 1439–1450. Scripta, 39, 51–61.
Punti T, Rieradevall M, Prat N (2009) Environmental factors, spatial vari- Valentini A, Pompanon F, Taberlet P (2009) DNA barcoding for ecolo-
ation, and specific requirements of Chironomidae in Mediterranean gists. Trends in Ecology and Evolution, 24, 110–117.
reference streams. Journal of the North America Benthological Society, 28, Vences M, Thomas M, Bonett RM, Vieites DR (2005) Deciphering
247–265. amphibian diversity through DNA barcoding: changes and challenges.
Rabeni CF, Wang N (2001) Bioassessment of streams using macroinverte- Philosophical Transactions of the Royal Society of London. Series B, Biologi-
brates: are the Chironomidae necessary? Environmental Monitoring and cal Sciences, 360, 1859–1868.
Assessment, 71, 177–185. Ward RD (2009) DNA barcode divergence among species and genera of
Raunio J, Heino J, Paasivirta L (2011) Non-biting midges in biodiversity birds and fishes. Molecular Ecology Resources, 9, 1077–1085.
conservation and environmental assessment: findings from boreal Whiting MF, Carpenter JC, Wheeler QD, Wheeler WC (1997) The Strep-
freshwater ecosystems. Ecological Indicators, 11, 1057–1064. siptera problem: phylogeny of the holometabolous insect orders
Regier JC (2008) Protocols, concepts, and reagents for preparing DNA inferred from 18S and 28S ribosomal DNA sequences and morphol-
sequencing templates. Version 12/4/08. Available from www.umbi. ogy. Systematic Biology, 46, 1–68.
umd.edu/users/jclab/PCR_primers.pdf. Yeates DK, Seago A, Nelson L, Cameron SL, Joseph L, Trueman JWH
Rochlin I, Santoriello MP, Mayer RT, Campbell SR (2007) Improved (2011) Integrative taxonomy, or iterative taxonomy? Systematic Ento-
high-throughput method for molecular identification of Culex mos- mology, 36, 209–217.
quitoes. Journal of the American Mosquito Control Association, 23, Zhou X, Li Y, Liu S et al. (2013) Ultra-deep sequencing enables high-fidel-
488–491. ity recovery of biodiversity for bulk arthropod samples without PCR
Roque FO, Siqueira T, Bini LM et al. (2010) Untangling associations amplification. GigaScience, 2, 4.
between chironomid taxa in Neotropical streams using local and land- Zou S, Li Q, Kong L, Yu H, Zheng X (2011) Comparing the usefulness of
scape filters. Freshwater Biology, 55, 847–865. distance, monophyly and character-based DNA barcoding methods in
Schlick-Steiner BC, Steiner FM, Seifert B, Stauffer C, Christian E, Crozier species identification: a case study of Neogastropoda. PLoS One, 6,
RH (2010) Integrative taxonomy: a multisource approach to exploring e26619.
biodiversity. Annual Review of Entomology, 55, 421–428.
Schrader C, Schielke A, Ellerbroek L, Johne R (2012) PCR inhibitors –
occurrence, properties and removal. Journal of Applied Microbiology,
W.H.W. and R.M. conceived the study and designed the
113, 1014–1026.
Sharley DJ, Pettigrew V, Parsons YV (2004) Molecular identification of experimental setups with input from M.B. and P.S.C.
Chironomus spp. (Diptera) for biomonitoring of aquatic ecosystems. Y.C.T. and J.P. performed direct PCR optimization on
Australian Journal of Entomology, 43, 359–365. Oreasteridae and Dolichopodidae respectively, while
Simon C, Frati F, Beckenbach A, Crespi B, Liu H, Flook P (1994)
optimization of other nonchironomid taxa was per-
Evolution weighing, and phylogenetic utility of mitochondrial gene
sequences and a compilation of conserved polymerase chain formed by W.H.W. Analysis of data and writing of man-
reaction primers. Annals of the Entomological Society of America, 87, uscript were done by W.H.W., with input from R.M.,
651–701. M.B. and P.S.C.
Sinclair CS, Gresens SE (2008) Discrimination of Cricotopus species (Dip-
tera: Chironomidae) by DNA barcoding. Bulletin of Entomological
Research, 98, 555–563.
Skuse FAA (1889) Diptera of Australia. Part VI. The Chironomidae. Pro-
ceedings of the Linnean Society of New South Wales, 2, 215–311. Data Accessibility
Srivathsan A, Meier R (2012) On the inappropriate use of Kimura-2-
parameter (K2P) divergences in the DNA barcoding literature. Cladis- DNA sequences are uploaded to NCBI and have the fol-
tics, 28, 190–194. lowing GenBank accession nos KJ530765–KJ530969.
Stribling JB, Pavlik KL, Holdsworth SM, Leppo EW (2008) Data quality, Sequence alignment and an Excel file containing individ-
performance, and uncertainty in taxonomic identification for biological
ual specimen information were deposited in the Dryad
assessments. Journal of the North American Benthological Society, 27,
906–919. database (doi: 10.5061/dryad.4q4kf).

© 2014 John Wiley & Sons Ltd