2
Introduction
This draft report describes protocols and equipment that will enable single laboratories to achieve
production rates of 50K sequences per year. However, most suggestions are also relevant to labs
with lower production goals as they seek simply to minimize the costs of analysis and speed its
execution.
Most steps of the analysis (specimen to PCR product) can be carried out in facilities with modest
infrastructure ($20K). However, the establishment of a sequencing facility capable of analyzing
50-100K samples per year is much more expensive (circa $1M). As a result, it may often be
appropriate to funnel PCR products from satellite laboratories to a central sequencing facility.
Although commercial kits are available for varied stages of the analytical chain (e.g. DNA
extraction, PCR amplification, product detection), home-made reagents can lower costs
dramatically. We describe both approaches in many cases.

1. Specimen Collection/Preservation
Whenever possible, specimens should be killed in a DNA-friendly fashion (freezing, cyanide,
immersion in ethanol), avoiding even brief exposure to killing/preservation agents such as ethyl
acetate or formalin that damage DNA. DNA in dried specimens ordinarily remains in good
condition for at least a year, but degradation becomes increasingly problematic as time passes.
DNA in frozen specimens (especially those held in cryogenic conditions) remains stable
indefinitely, but DNA in ethanol-preserved material often degrades due to acidification. As a
result, barcode analysis should follow collection as quickly as possible.

2. Tissue Sampling/Handling
All specimen samples should be handled on a clean working surface and all instruments should
be acid or flame sterilized between each sample. A Bunsen burner flame is convenient for
sterilization; small propane tanks are ideal for settings where gas is not on-line.
In any laboratory that seeks high production rates, it is critical to carry out all stages of barcode
analysis in 96-well plates. Care must be taken when loading these plates with samples to avoid
cross-contamination between wells. This risk can be reduced by covering each plate with strip
caps and opening just one row at a time. Soaking dry insect legs in ethanol for a few minutes
before extraction is also helpful as it prevents specimen flying due to static electricity.

3. Genomic DNA Isolation/Purification

3.1.

F
AST
DNA

E
XTRACTION
-

L
OW
C
OST

C
HELEX
100

(D
RY
R
ELEASE
)
Fresh or frozen specimens are ordinarily an easy target for barcode analysis, allowing the use of
rapid Chelex resin protocols (Walsh et al. 1991) for DNA isolation. Chelex extraction (Jaulnac et
al. 1998) can be combined with proteinase K treatment to create a simple, cheap and efficient 96-
well protocol for DNA extraction. This protocol has been used successfully with arthropods, fish,
birds (including feathers) and mammals (including skin and hairs). It requires only a small amount
of tissue (1-3 mm
3
), making even a single insect leg sufficient for several DNA extractions.
3
Some Chelex protocols involve grinding tissue in liquid nitrogen (Gregory & Rinderer 2004), an
approach that is not readily compatible with 96-well format extractions. By combining Chelex
extraction with proteinase K treatment, the need for tissue disruption is eliminated. Tissue
samples usually dissolve completely after overnight incubation, while chitinous parts remain
intact, but DNA is released. Therefore, the Chelex/Proteinase K combination can be used for
non-destructive DNA extraction from small invertebrates (e.g. collembolans, rotifers). In this
case, the entire specimen is placed in the solution and removed at the end of the procedure.
Chelex-based extraction is not suitable for samples with high levels of PCR inhibitors (e.g.
haemoglobin) or for samples where DNA is degraded. A second disadvantage of this method lies
in the fact that the extracted DNA is relatively impure and is, hence, often unstable for more than
a few weeks.
Two commercial kits provide very fast options for DNA extraction (Sigma-Aldrich Extract-N-Amp*
PCR* Kit, Genereleaser), but are more costly and are not currently available in a 96-well format.

3.2

S
ILICA OR
S
ILICA
M
EMBRANE
DNA

E
XTRACTION

Various silica and silica-membrane based protocols produce relatively pure DNA. These
approaches are also more effective in extracting DNA, a factor that makes them particularly
useful for studies on specimens with degraded DNA. These approaches rely on DNA binding to
silica in the presence of a high concentration of chaotropic salt (Boom et al. 1990; Hoss and
Paabo 1993). This class of methods provides an alternate non-destructive approach for extracting
DNA that involves soaking samples in guanidinum-thiocyanate (GuSCN) with subsequent
sorption of DNA to silica (Rohland et al. 2004).
We have tested four commercially available systems that employ silica-binding.
a) Sigma-Aldrich GenElute
TM
Mammalian Genomic DNA Miniprep Kit
b) QIAGEN DNeasy tissue kit (DNeasy 96 tissue kit)
c) Promega Wizard® SV 96 Genomic DNA Purification System
d) Macherey-Nagel Nucleospin® 96 tissue
The GenElute kit is sensitive, but it is not available in a 96-well format and is relatively slow to
use. The other three kits are available in 96-well formats. A multi-channel pipettor is required to
effectively perform 96-well DNA extractions with any of these kits. For those that are considering
very high volume production, most of these kits can be automated on robotic liquid handling
stations.

3.3

M
AGNETIC
B
EAD
-B
ASED
DNA

E
XTRACTION

Dynal Biotech's Dynabeads® DNA DIRECT
TM
and Dynal MPC® -auto 96 Magnet Station
This system might be effectively applied in cases where a robotic liquid handling device is
available.

4. Genomic DNA Quantitation
It is not usually necessary to quantify genomic DNA extracts because even a few copies of the
target gene are sufficient for PCR amplification. However, the quantity of extracted DNA can be
determined with a plate reader.

4
5. PCR Amplification of Barcode Region

5.1

P
RIMERS

Primer design is critical and minor adjustments can have large impacts on barcode recovery. The
first phase of any study on a new group should involve a serious effort to identify optimal primers.
Whenever we have done this, we have gained very high success in barcode recovery. For
example, the 2 primer sets that we routinely employ for lepidopterans recover the barcode region
from more than 99% of species and our 2 primer sets for fishes have about 97% success.
D
EGENERATE
P
RIMERS
,

M
ODIFIED
B
ASES
(
E
.
G
.

I
NOSINE
)
Single bp mismatches at the 3-end of a primer usually prevent PCR amplification (Simsek &
Adnan 2000). This problem can often be solved by the use of degenerate or inosine containing
primers (Batzer et al. 1991; Shultz & Regier 2000; Candrian at al. 1991; Christopherson et al.
1997). Sorenson et al. (1999) suggest that primers with appropriate degenerate sites are also
less likely to preferentially amplify nuclear pseudogenes because they accommodate usual
differences between nuclear and mtDNA sequences (e.g. 3rd positions changes in the mtDNA
copy).
Primers with 2-4 degenerate positions will often rescue barcodes from recalcitrant specimens.
Early results with primers containing inosine show that they are also effective in amplifying
difficult samples.

5.2

R
EACTION
M
INIMIZATION
,

R
EACTION
M
IXES AND
PCR

E
NHANCERS

Although Chelex-based DNA extracts sometimes resist amplification because of the presence of
inhibitors, this can usually be overcome by incorporating amplification facilitators such as bovine
serum albumin (BSA), betaine or DMSO (Al-Soud & Rådström 2000) in the PCR mix. Betaine
exerts its effect by stabilizing AT base pairs while destabilizing GC base pairings, resulting in a
net specific destabilization of GC-rich regions (Rees et al. 1993, Henke et al. 1997). Most
commercial PCR-enhancing buffers contain betaine (Frackman et al. 1998). The addition of
amplification enhancers also improves the specificity of PCR and allows the amplification of GC-
rich templates.
To reduce costs we have lowered reaction volumes; we regularly employ 10 µl (versus standard
25-50 µl). In order to accurately dispense such small volumes, it is useful to make up a larger
volume master mix. This can be dispensed into 96-well plates and stored frozen till use, but
freezing is not possible without a cryoprotectant. Such pre-mixing of PCR reagents speeds an
otherwise time-consuming step and aids quality assurance.
Trehalose is widely used as a cryoprotectant (Franks 1990; Spiess et al. 2004). It also acts as a
potent PCR enhancer by both lowering the DNA melting temperature and stabilizing Taq
polymerase (Spiess et al. 2004). Because of these properties, trehalose is ideal to stabilize frozen
PCR mixes and to overcome the effect of inhibitors that may be present in Chelex extracts,
re