Page 193 - Proceedings of the State Natural History Museum. Issue 37 (Lviv, 2021)

192 Zamoroka A. M.

Demonax transilis Bates, 1884, Plagionotus christophi (Kraatz, 1879). Grebennikov and al.
also indicated that Clytini is non-monophyletic [25]. Their study based on sequences of COI
from 11 species. In contrast, Nie et al. considered monophyly of Clytini based on complete
mitochondrial sequences of three species (Clytobius davidis (Fairmaire, 1878), Xylotrechus
magnicollis (Fairmaire, 1888), Chlorophorus diadema) [50]. However, deep branching of
Clytini clade on their tree indicates paraphyly. Phylogenetic analysis of Clytini with the most
comprehensive sample of taxa (27 species) is presented in Lee & Lee [39]. Their tree, based
on six genes (2 mitochondrial COI and 16S r RNA; 4 nuclear 18S rRNA, 28S rRNA, wingless
and CAD), clearly showed that Clytini is a paraphyletic group. It should be noted that the
group of genera Plagionotus – Chlorophorus – Demonax is well separated from another
group of genera Neoclytus – Perissus – Clytus – Xylotrechus on their tree. In general, Clytini
and Anaglyptini form a monophyletic clade that is well separated from other Cerambycinae
clades. However, Lee & Lee indicate that to resolve intergeneric relationships within Clytini
and Anaglyptini it needs to be tested with broader taxon sampling [39].
In the current study, I performed a five genes phylogenetic analysis of the most
comprehensive sample of taxa (79 species) of Clytini and Anaglyptini. My findings clearly
showed that Clytini is nonmonophyletic, but consists of two evolutionary lineages, which
could be recognized as separate tribes Clytini, trib. sensu nov. and Chlorophorini, trib. nov.
The monophyly of the large clade of Anaglyptini, Clytini and Chlorophorini, for which the
status of supertribe Chlorophoritae, supetrib. nov. is proposed, was confirmed. I proposed
new nomenclature acts including 1 new supertribe, 1 new tribe, 4 new subtribes, 3 new
genera, 4 new subgenera, 3 new statuses, 22 new combinations, 2 new synonyms. In addition,
I redescribed 1 tribe and 3 genera.

Materials and methods

I used publicly available DNA partial sequences of five genes (79 species of target group
and 4 species of outgroup) including three mitochondrial genes: 12S ribosomal RNA (12S
rRNA) and 16S ribosomal RNA (16S rRNA) and cytochrome c oxidase I (COI) and two
nuclear genes: 18S ribosomal RNA (18S rRNA) and 28S ribosomal RNA (28S rRNA)
generated from GenBank as a FASTA file. I also produced consolidated sequences for COI
and 28S rRNA from the sets of separate specimens of the same species. This allowed to avoid
the statistical noises caused by multiple point mutations which spread within the different
populations of the certain species. The genes were assembled in the matrix as follows: 12S
rRNA – 16S rRNA – COI – 18S rRNA – 28S rRNA with the total length 5.327 kilobase (kb).
While the species set with complete 12S rRNA + 16S rRNA + COI + 18S rRNA + 28S rRNA
sequences was limited, I filled the gaps of missing species with partial sequences of
mentioned genes, which overlap at least 50% of their length (fig. 1).
Multiple alignments were generated using the Muscle software in the environment of
SeaView 5 [24]. Alignments were provided with unlimited iterations and were edited
manually to correct regions containing missing data and to exclude unalignable positions.
Phylogenetic trees were constructed using maximum-likelihood (ML) and Bayesian
methods with PhyML [22]. Analyses were performed following a general time-reversible
(GTR) model of sequence evolution. We performed an approximate likelihood-ratio test
(aLRT) for branch support based on the Log Ratio between the likelihood value of the current
tree and that of the best alternative [2, 23]. The values of branch support were considered: 1-
0.90 – very strong, 0.70-0.89 – strong, 0.50-0.69 – moderate and less than 0.50 – weak
support. The optimal tree's structure was estimated using the best combination of nearest-
neighbour interchange (NNI) and Subtree Pruning Regrafting (SPR) algorithms. We also
used the neighbour-joining algorithm (BioNJ) optimizing trees topology for estimation of
branch distances [19].