Specimen size sorting, DNA extraction, PCR and
sequencing
We processed all samples using a semiquantitative metabarcoding protocol
as described in Lim et al. 2022. Large arthropod specimens contribute
considerably more biomass and hence DNA to a sample than small ones, so
to reduce biases by large arthropod specimens in the sample, we sorted
all specimens into four size categories aimed at approximately
distributing biomass: 0 – 2 mm, 2 – 4 mm, 4 – 7 mm, and
> 7 mm. One particularly abundant arthropod group in
Hawaiian wet forest is Collembola. We thus sorted Collembola out as a
fifth category. All size sorted samples were stored in 2 ml tubes. After
addition of a five mm stainless steel ball, lysis buffer and proteinase
K, specimens were finely ground on a Qiagen Tissuelyzer (Qiagen, Hilden,
Germany) and then DNA was extracted from each size sorted sample using
the Qiagen Puregen kit according to the manufacturer’s protocol.
A 365 bp fragment of the mitochondrial COI gene was then amplified
separately for each sample and size category using the primer pair
mlCOIintF/HCOdeg (Leray et al. 2013, Yu et al. 2012). PCRs were run
using the Qiagen Multiplex PCR Kit, according to the manufacturer’s
protocol, with 32 cycles and at an annealing temperature of 46 °C. PCR
success was checked using agarose gel electrophoresis. After the first
round of PCR, an indexing PCR was performed, with five cycles and an
annealing temperature of 55 °C, in which Illumina Truseq adapters and
indexes were incorporated on each fragment (Lange et al. 2014). The
resulting libraries were quantified using agarose gel electrophoresis
and pooled in approximately equimolar amounts. The final sample was
purified using 1 X AMpure Beads XP (Beckman Coulter, Brea, CA, USA) and
sequenced on an Illumina MiSeq using V3 chemistry and 600 cycles
(Illumina, San Diego, CA, USA). No-template control PCRs and negative
control extraction samples were also amplified and sequenced alongside
the arthropod samples.