Genetic and morphological evidence for a new species of the Maculatus Group of Anopheles subgenus Cellia (Diptera: Culicidae) in Java, Indonesia

Background Anopheles maculatus, a species of the Maculatus Group of subgenus Cellia (Diptera: Culicidae), is an important vector of human malarial protozoa in Java, Indonesia. However, the identity of this species in Indonesia has been questionable because published reports and records are based mainly on morphological identification, which is unreliable for distinguishing members of the Maculatus Group due to overlapping characters. Methods We performed morphological assessments, metaphase karyotype preparations, phylogenetic analyses of ITS2 and cox2 sequence data and cross-mating experiments to determine whether the Javanese form and An. maculatus (s.s.) from Thailand were conspecific. Results The adults of the Java strain are similar to those of An. maculatus (s.s.), but the larvae and pupae exhibit significant differences. The metaphase karyotype of Javanese specimens includes a long acrocentric X chromosome and a small telocentric Y chromosome, which are distinct from other members of the Maculatus Group. Cross-mating of the Java strain with An. maculatus (s.s.) revealed genetic incompatibility. Phylogenetic analysis of ITS2 and cox2 sequences revealed that the Java strain forms a single clade that is distinct from clades of other members of the group (Kimura 2-parameter, K2P, genetic distances 3.1–19.2% and 1.6–9.6%, respectively). Conclusions This study provides evidence that the Javanese form of An. maculatus is not conspecific with An. maculatus (s.s.) and constitutes a previously unrecognized species of the Maculatus Group.

as ponds, lakes, stream margins, stream pools, ground pools, sand pools, rock pools, marshes, river beds, rice fields, etc. The species have overlapping and disparate distributions and are variously involved in the transmission of human malarial parasites in tropical and subtropical areas of Asian countries. Among members of the group, the nominotypical species (An. maculatus) appears to be the most widely distributed species, ranging from Afghanistan, Pakistan and India eastward to the western Pacific islands, including Taiwan, the Indonesian archipelago and Timor Leste. It is absent from the Philippines where An. dispar and An. greeni are present [6,13,14].
Human malarial parasites have been detected in An. maculatus across its distribution, but its role in transmission to humans is well documented only in peninsular Malaysia [14,15]. Anopheles willmori is a primary vector of human malarial protozoa in the highlands of Nepal [16]. Anopheles pseudowillmori is a vector in southern Tibet [17], a secondary vector in western Thailand [18] and a suspected vector in Bhutan [19]. Anopheles sawadwongporni may be a secondary vector in Thailand and Vietnam [20][21][22]. Anopheles dispar and An. greeni are regarded as secondary vectors in the Philippines [13]. It is not known whether An. dravidicus, An. notanandai and An. rampae may play a role in malaria transmission.
Malaria is a significant public health problem throughout the Indonesian archipelago. Over 200,000 malaria cases were reported in Indonesia in 2016 [23]. Of about 80 Anopheles species recorded in Indonesia, 21 have been incriminated as primary and secondary vectors of malaria, but their roles in malaria transmission vary across the archipelago [24][25][26][27]. In Java, An. maculatus is recorded as one of the important vectors, but it is of little or no medical importance on other islands of Indonesia due to its more zoophilic behavior and normally low human-biting densities [25,28]. Other malaria vector species in Java include An. aconitus Dönitz, An. balabacensis Baisas, An. flavirostris (Ludlow), members of the An. sundaicus and An. subpictus complexes, and An. vagus Dönitz [25].
Published reports and records of An. maculatus in Java and other islands of Indonesia are based mainly on morphological identifications; hence, it is not possible to know with certainty if it is conspecific with continental populations of the species [9][10][11]. For many years, it was pointed out that the taxonomic status of An. maculatus in Indonesia is questionable and may represent an undescribed species in the Maculatus Group [6,25,29].
In this study, we performed morphological assessments, metaphase karyotype preparations, cross-mating experiments and phylogenetic analyses of DNA sequence data to provide, for the first time, clear evidence that An. maculatus in Java is a distinct species of the Maculatus Group.

Mosquitoes and morphological identification
Specimens of a stenogamous laboratory strain of An. maculatus from Java (Java strain) [29] maintained at the Vector and Reservoirs Research Institute in Salatiga, Central Java, were used to establish a colony in the Department of Parasitology, Hasanuddin University, Indonesia. This strain was used for examination of metaphase karyotypes and cross-mating with specimens from a laboratory colony of An. maculatus (s.s.) (Thailand strain), which was established with specimens collected in Mae Sarieng District in Mae Hong Son Province of northwestern Thailand and maintained at the Office of Disease Prevention and Control, Region 1, Chiang Mai two years prior to the present study. The specimens of the Thailand strain exhibit the characters that define all life stages of An. maculatus (s.s.) [1,6,11], and their ITS2 sequences fall in the clade with sequences previously generated for this species (see the "Results" below). Specimens of both strains were transferred to the insectary of the Department of Parasitology, Faculty of Medicine, Chiang Mai University, Thailand for rearing and study. The insectary was maintained at about 27 °C and 70% relative humidity with a photoperiod of 12:12 h L:D. Females of the Thailand and Java strains and feral females collected in a cowshed in Kulonprogo, Central Java, were used for DNA sequencing.
Morphological features of adult mosquitoes were examined under a stereomicroscope. Mature larvae were killed with hot water (60-65 °C) and preserved in 80% ethanol. Larval and pupal exuviae were collected within 24 h after molting or emergence and preserved in 80% ethanol. They were mounted on microscope slides with Hoyer's medium (Neo-shigaral, Shiga Konchu Fukyusha, Tokyo, Japan) or Euparal (Waldeck, Germany). Eggs and larval and pupal setae were examined under a bright-field compound microscope (Olympus CX31) using 10× and 40× objective lenses. Available morphological keys [1,6,30] were used to confirm identification of the species. Photographs were taken with a digital camera (Olympus E-330). The morphological terminology and abbreviations follow the Anatomical Glossary of the online Mosquito Taxonomic Inventory [31].

Metaphase karyotypes
Metaphase chromosomes were prepared from the brain ganglia of 10 fourth-instar larvae of the Java strain, using techniques described by Saeung et al. [32]. Identification of karyotypic forms followed the method of Baimai et al. [4].

Cross-mating experiments
Reciprocal crosses between the Java and Thailand strains were carried out to determine genetic compatibility. Pupae were sexed by observing the shape and size of the genital lobe ( [31], http://mosqu ito-taxon omic-inven tory. info/simpl etaxo nomy/term/6691) and kept separately until emergence of adults. Virgin blood-fed females were mated with males using the artificial mating technique [33]. Following mating, each female was isolated in an oviposition cup and provided with a cotton plug wetted with a 10% sucrose solution. Eggs were counted and allowed to hatch. Following oviposition, females were dissected to check for spermatozoa in their spermathecal capsules, and eggs from un-inseminated females were discarded. Newly hatched larvae from each egg batch were counted and placed in rearing trays. They were reared in dechlorinated water and fed with a finely ground fish food until pupation. Pupae were removed daily, sexed and placed separately in cups until emergence of adults. The emergent rates of F 1 hybrid adults were noted. The testes and ovaries of hybrids were dissected to check fertility. The fertility and viability of hybrids was determined by backcrosses with the Java strain.
Egg batches with no or little hatching were allowed to stand for another 3 days, and afterwards examined for the development of embryos. To check embryonation, both hatched and unhatched eggs were transferred onto a drop of water on a microscope slide, covered with a coverslip, gently pressed with the blunt end of a pen to break the egg chorion, and examined under a microscope for embryo formation.

DNA extraction, amplification and sequencing
Genomic DNA was extracted from whole mosquitoes or legs of individual adult mosquitoes using Pure Link ™ Genomic DNA Mini Kit (Invitrogen by Thermo Fisher Scientific, USA) according to manufacturer's instructions. The specimens, without legs, were retained for morphological examination. A product of approximately 450 bp of the ITS2 region of rDNA was amplified by PCR using the primers 5.8F (5′-TGT GAA CTG CAG GAC ACA TG-3′) and 28R (5′-ATG CTT AAA TTT AGG GGG TA-3′) [11]. Each PCR reaction was carried out in a 20 µl volume containing 1 µl of DNA, 0.8 µM of each primer, 1.2 mM MgCl 2 , 1.6 mM dNTP mix and 0.08 U of Platinum ® Taq DNA polymerase in 1× PCR buffer. Thermals cycling conditions for ITS2 included an initial denaturation at 95 °C for 2 min, 35 cycles at 95 °C for 1 min, 55 °C for 1 min, and 72 °C for 2 min and a final extension step at 72 °C for 5 min.
The mitochondrial cox2 gene was amplified using the primers SCTL2-J-3037 (5′-ATG GCA GAT TAG TGC AAT GA-3′) and TK-N-3785 (5′-GTT TAA GAG ACC AGT ACT TG-3′) [34]. PCR reactions were carried out in a 20 μl volume containing 0.4 U of Platinum ® Taq DNA polymerase, 1× of PCR buffer, 1.5-3.0 mM of MgCl 2 , 0.2 mM of each dNTP, 0.2 μM of each primer and 1 μl of extracted DNA. The amplification profile comprised initial denaturation at 95 °C for 5 min, 35 cycles at 95 °C for 30 s, 45 °C for 30 s and 72 °C for 30 s, and a final extension step at 72 °C for 5 min. The amplified products were electrophoresed in 2% agarose gels and stained with ethidium bromide. PCR products were purified using the illustra ™ ExoProStar ™ 1-Step (GE Healthcare Life Sciences, UK) and sequenced using a 23 ABI 3730XLs sequencer (Macrogen, South Korea).

Phylogenetic analysis
The ITS2 and cox2 sequences obtained during this study were compared with those of An. maculatus (s.s.) and other members in the Maculatus Group on GenBank using the Basic Local Alignment Search Tool (BLAST, http://blast .ncbi.nlm.nih.gov/Blast .cgi). Anopheles stephensi Liston was used as the outgroup taxon. Sequences of ITS2 were manually aligned using the CLUSTALW [35] and edited manually in MEGA version 7.0. Gap sites were excluded from the analysis. Construction of Neighbor-joining trees and the bootstrap test with 1000 replications were conducted with MEGA version 7.0 [36]. Genetic distances were estimated by using the Kimura two-parameter (K2P) method [37].

Morphology
Fifty-two females and 27 males of the Java strain, and six wild caught females from Java were examined. In general, females and males of the strain look similar to An. maculatus (s.s.), in particular abdominal terga II-IV are without scales; terga V and VI are without scales or have sparse pale falcate scales on the posterior margins; terga VII and VIII are largely or posteriorly covered with narrow pale spatulate scales; the posterolateral corners of terga VI-VIII, and rarely tergum V, bear patches of dark scales (Fig. 1a, b). Variation of wing scaling has been found in both the laboratory Java strain and feral females, in particular, vein R 2+3 has one dark spot or two dark spots on at least one wing (Fig. 2a-c; Table 1). The two spots occasionally join to form a single large spot, or are reduced to a few dark scales near the furcation. In both feral and laboratory females, the accessory sector pale spot is confined to vein R 1 and is never present on the subcosta and costa. Palpomere 3 is completely black-scaled.
Ten larvae of the Java strain were examined and compared with An. maculatus (s.s.) (Thailand strain). In general, the Javanese larvae exhibit characters of the Maculatus Group, but differ from An. maculatus (s.s.) as follows: seta 2-C has long, strong lateral barbs (aciculae) arising near the base (on basal 0.21, 0.11-0.31) (Fig. 3a), but that of An. maculatus (s.s.) has short, weak lateral barbs arising far from the base (on basal 0.39, 0.34-0.43) (Fig. 3b); the dorsomentum of the head has 5 teeth on either side of the median tooth (total 11 teeth) (Fig. 3c), whereas An. maculatus (s.s.) has 4 teeth on either side of the median tooth (total 9 teeth) (Fig. 3d); seta 1-P has 11-20 branches and the stem of seta 1-P is usually weaker than the stem of seta 2-P (Fig. 4a), but in An. maculatus (s.s.) seta 1-P has > 20 branches and the stem of seta is as strong as that the stem of seta 2-P (Fig. 4e); seta 3-P is separated from or proximal to the tubercle supporting seta 2-P, but that of An. maculatus (s.s.) is borne on a small tubercle joined to the tubercle supporting 2-P; the basal stem of seta 4-M is 3-5 times as long as its width (Fig. 4b), whereas it is > 5 times as long in An. maculatus (s.s.) (Fig. 4f ); most leaflets of abdominal palmate seta 1-II have distinctly serrated shoulders and a long filament (Fig. 4c), whereas those of An. maculatus (s.s.) rarely have distinctly serrated shoulders (Fig. 4g); leaflets of abdominal seta 1-III-VI have long slender, sharply pointed filaments that are one-third to one-half the length of the blade (Fig. 4d), whereas those of An. maculatus have short filaments that are one-fourth the length of the blade (Fig. 4h).
Ten pupal exuviae of the Java strain were examined. In general, they are similar to those of An. maculatus (s.s.) and no clear differences in setal branching were observed. Both strains have seta 9-II very short, blunt and transparent; seta 9-III is slightly longer than seta 9-II, blunt and darkly pigmented. However, a significant difference was observed in the development of abdominal seta 9-IV, which is short and blunt in An. maculatus (s.s.) (Fig. 5a) and horn-like, pointed or long and sharply pointed in the Javanese pupae (Fig. 5b, c). The ratio of lengths of seta 9-IV/9-V (Fig. 5b-e) of the

Metaphase karyotype
The metaphase karyotype was observed in the F 6 progeny of the Java strain compared with the Thailand strain.
The Java strain appears to comprise one genotypic form (Fig. 6a, b). The X chromosome is acrocentric, slightly longer that the long arm of autosome III and consists of a large block of centromeric heterochromatin and an euchromatic portion that is slightly shorter than the heterochromatic block. The heterochromatic Y chromosome is obviously small, telocentric and slightly shorter than the length of the heterochromatic block of the X chromosome. The autosomes possess a very limited amount of pericentric heterochromatin. By contrast, the X and Y chromosomes of the Thailand strain are large submetacentric (Fig. 6c, d), and similar to X 3 and Y 2 chromosomes of An. maculatus (form B) as described by Baimai et al. [4].

Cross-mating
Cross-mating between the Java and Thailand strains revealed genetic incompatibility. The resulting fertility and viability are summarized in Table 2. Reciprocal crosses in both directions yielded F 1 hybrids, but egg hatchability and adult emergence were low. The control cross (Java × Java) resulted in normal fertility and viability. All hybrid females examined appeared to have normal ovaries. Dissection revealed that all hybrid males were completely sterile, but the accessory glands looked normal (Fig. 7a). Therefore, these hybrid males were not used for backcrossing. Backcrosses between F 1 hybrid females and males of the Java strain resulted in low egg hatchability and adult emergence. All males that emerged were partially sterile, having abnormal spermatozoa with enlarged heads (Fig. 7b) and inactive spermatozoa, while the ovaries of females appeared to be normal.

DNA sequences and phylogenetic analysis
The ITS2 region was sequenced from five females of the Java strain, two without a distal dark spot on vein R 2+3 and three with such a spot. Three wild-caught females from Java were also sequenced, but unfortunately without noting the wing scales. The cox2 region was sequenced for the three feral females and four of the five females of the Java strain. The ITS2 region was sequenced from two females of the Thailand strain. The sequences are deposited in the DDBJ/EMBL/GenBank nucleotide sequence database under the accession numbers MK204640-MK204649 (ITS2) and MK236365-MK236371 (cox2).
The ITS2 sequences (334 bp) of the Java strain, with and without a distal dark spot on vein R 2+3 , and those of the wild-caught females were identical or very similar (K2P genetic distance 0-0.3%), whereas all of the cox2 sequences (620 bp) were identical. The ITS2 sequences of two specimens of the Thailand strain were also identical to one another. Phylogenetic analysis of the ITS2 and cox2 sequences (Figs. 8,9) revealed that specimens of the Java strain form a single clade that is distinct from An. maculatus (s.s.) (Thailand strain) and other species of the Maculatus Group (K2P, ITS2: 3.1-19.2%; cox2: 1.6-9.6%). Intraspecific variation of the ITS2 and cox2 sequences of An. maculatus (s.s.) from continental Asian countries, as indicated in Figs. 8 and 9, was very low (K2P, ITS2: 0-0.6%; cox2: 0-0.3%). Comparison of the ITS2 sequences reveals that the Java clade seems to be most closely related to An. dispar (K2P 3.1-3.4%), followed by An. maculatus (s.s.) (K2P 4.6-5.6%) and An. greeni (K2P 5.0-5.3%). However, comparison of the cox2 sequences indicates that the Java clade is most closely related to An.

Discussion
The results of the present study provide unambiguous evidence that An. maculatus in central Java is not conspecific with An. maculatus (s.s.) and is a distinct species of the Maculatus Group. Morphologically, the adult and larval stages have morphological characters that overlap with those of various species of this group. The adults have pale scaling and patches of dark scaling on the posterolateral corners of abdominal terga VI-VIII that is similar to the scaling of An. dispar, An. greeni and An. maculatus (s.s.) [6]. The wings of Javanese adults either have or lack a distal dark spot near the furcation of vein R 2+3 . Specimens without the spot are difficult to separate from An. dispar, An. greeni and An. maculatus (s.s.).
However, about half of specimens of An. dispar and An.
greeni have an accessory sector pale spot on the costa and subcosta, which is absent in Javanese specimens and other members of the Maculatus Group. The presence of a distal dark spot on vein R 2+3 is common in species of the Sawadwongporni Subgroup, i.e. An. notanandai, An. rampae and An. sawadwongporni, but those species differ in having abdominal terga II-VIII densely covered with pale spatulate scales and posterolateral corners of terga II and IV usually with a few black scales [1,38]. We consider that the two wing forms found in Javanese specimens are conspecific because both forms have similar or identical ITS2 and cox2 sequences. The absence of the distal dark spot on vein R 2+3 was found in about 15% of wild-caught adults of An. rampae [38] and noted in An. sawadwongporni [1]. Therefore, variation of the wing scaling in Javanese specimens is considered to be due to  intraspecific variation that can confound morphological identification. However, as only six wild-caught females were available for study, additional specimens are needed to assess the variable presence of the distal dark spot in Javanese populations. Larvae of the Java strain differ mainly from other members of the Maculatus Group in having longer and stronger lateral barbs (aciculae) near the base of seta 2-C (Fig. 3a), and in having the dorsomentum with 11 teeth (Fig. 3c). As stated by Christophers [39], seta 2-C of An. maculatus (s.s.) is very finely frayed on the distal two-thirds, and the dorsomentum has four teeth on either side of median tooth (total of 9 teeth) [6]. Compared with An. maculatus (s.s.) [6], the stem of seta 1-P of Javanese larvae is usually weaker with fewer branches, and seta 3-P is not borne on a tubercle joined to the tubercle supporting seta 2-P. Rattanarithikul et al. [30] used seta 4-M to characterize species of the Sawadwongporni Subgroup: the basal stem of this seta is short and no longer than four times its width, but it is five or more times longer in other members of the Maculatus Group. The basal stem of this seta in the Javanese larvae is usually four to five, occasionally six, times its width, which may confuse identification. However, in species of the Sawadwongporni Subgroup, as well as An. maculatus (s.s.) and An. dravidicus of the Maculatus Subgroup, the filaments of the leaflets of palmate seta 1-III-VI are distinctly short. In contrast, the filaments are long in the Javanese larvae, as well as larvae of An. dispar, An. greeni, An. pseudowillmori and An. willmori, which can confuse identification, particularly if Javanese specimens have seta 4-M with a slightly long basal stem. The leaflets of palmate seta 1-II have distinct shoulders in larvae of the Java strain, An. dispar, An. greeni and An. pseudowillmori, but the shoulders are not as distinctly developed in the other members of the Maculatus Group [6,30]. Therefore, identification of the Javanese form is possible using the combination of larval characters. Primary differences in the adults and larvae of the Java strain and the Sawadwongporni Subgroup are summarized in Table 3.
Pupae of Javanese specimens differ from pupae of An. maculatus (s.s.) in having seta 9-IV longer and more sharply pointed. This seta is shorter and blunt in pupae of An. maculatus (s.s.). The ratio of the lengths of seta 9-IV/9-V in Javanese pupae (0.38-0.68, mean 0.48) is greater than in pupae of An. maculatus (s.s.) (0.22-0.38, mean 0.28), which is nearly to the same as a previous report (0.18-0.38, mean 0.27) [6]. However, this ratio is close to that of An. dispar (0.30-0.71, mean 0.52) and overlaps with that of An. greeni (0.56-0.83, mean 0.72) [6]. Thus, variation in the length of seta 9-IV makes it Fig. 9 Phylogenetic tree derived from cox2 sequences of specimens from the laboratory colony of the Java strain, wild-caught females from Java and GenBank sequences of other members of the Maculatus Group, with An. stephensi as the outgroup difficult to separate the pupae of the Javanese strain and these two species.
Metaphase karyotypes of members of the Maculatus Group were described by Baimai et al. [4]. Similar to the Java strain, telocentric or acrocentric sex chromosomes are only found in An. dravidicus, An. pseudowillmori and An. sawadwongporni, whereas those of the other members of the group are large submetacentric chromosomes. The acrocentric X chromosome of the Java strain appears to be longer than the X chromosome of those three species. The small telocentric Y chromosome is shorter than the Y chromosome of An. dravidicus and An. sawadwongporni, but is similar to that of An. pseudowillmori. The two autosomes of the Java strain are similar to those of An. dispar, An. notanandai and An. maculatus (s.s.) in having a very limited amount of pericentric heterochromatin, but these species have large submetacentric sex chromosomes. Therefore, the metaphase karyotype appears to be diagnostic for the Java strain.
The results of the cross-mating experiments clearly indicate genetic incompatibility between An. maculatus from Java and An. maculatus (s.s.) from Thailand. Hybrid males are completely sterile, while hybrid females appear to be normal. This may indicate a moderately high divergence between the two taxa. Greater divergence was observed when hybrid females and males exhibited both sterility and unviability, such as the crosses between An. rampae and An. pseudowillmori/An. dravidicus [8].
Unfortunately, it was not possible, due to unavailability of colonies, to determine relative levels of divergence from other species of the Maculatus Group. However, the results of the phylogenetic analyses of ITS2 and cox2 sequences conducted during the current study suggest that the Java strain, based on lowest K2P genetic distances, is more closely related to the Philippine An. dispar and the continental An. maculatus (s.s.) Additionally, based on morphological observations, the Java strain more closely resembles An. dispar, An. greeni and An. maculatus (s.s.) than the other members in the Maculatus Group, suggesting that the Java strain, An. dispar and An. greeni have a closer affinity with the Maculatus Subgroup than with the Sawadwongporni Subgroup as reported previously [10,40].

Conclusions
In conclusion, based on the morphological, cytogenetic, cross-mating and molecular evidence gleaned from the current study, An. maculatus in Java is not conspecific with continental An. maculatus and is a distinct species of the Maculatus Group, which is far more diversified than previously thought. This finding raises a question about whether there may be other undescribed species of the group on other islands of the Indonesian Archipelago, in particular Borneo, Sumatra and Sulawesi. Employing a combination of approaches to include traditional morphology, genetic and molecular methods of investigation is essential for recognizing and distinguishing closely related and morphologically similar species.