Wings’ Descriptive and Geometric Morphometry Analyses Support Distinct Phlebotomus sergenti Parrot Populations in Central Morocco

Phlebotomus (Paraphlebotomus) sergenti Parrot, 1917, the proven vector of Leishmania tropica Wright, 1903, the causative agent of anthroponotic cutaneous leishmaniasis, is widely distributed in Morocco. Previous works using molecular markers (ITS2 rDNA and Cyt b mtDNA) have hy-pothesized the existence of multiple closely related populations of sandfly species (cryptic species) that would exhibit distinct vectorial capacities. In this work, descriptive characteristics of wings (size and shape of the right and left wings) were measured in samples collected from fourteen sta-tions in central Morocco. These analyses support the existence of distinct P. sergenti populations, enlightening significant phenotypic variations of P. sergenti’s wings, regarding their size and shape, depending on geographic origin. In addition, geomorphometric analyses of wing’s length, centroid size, alpha, and beta distances allowed clear discrimination of P. sergenti sub-populations. These data pinpoint the adaptative ability of P. sergenti to local environmental conditions. Additional studies are now required to further shed light on the genetic structure of P. sergenti populations in Morocco.


Study area
The studied area locates in central Morocco ( Figure. 1), where cutaneous leishmaniases (CL) are widespread [19]; the first CL case due to L. tropica was reported in 1986 in Azilal [20]. The environment is diverse, with altitudes varying from 491 m to 1650 m elevations. It encompasses the Tadla plain, the phosphate plateau, and the High and Middle Atlas, with altitudes ranging from 491 m to 1650 m elevation ( Figure 1, Table 1).

Sandflies sampling and identification
Sampling is carried out using sticky traps placed in domestic, peridomestic, and wild biotopes in fourteen stations between May and October 2017. Traps were left in the field a single night to avoid deteriorations of wings. In the laboratory, each specimen is dissected under a binocular. First, the wings are separated, then the head and genitalia of males or the spermathecae of females are dissected and mounted in Canada balsam. Males are identified by examining the morphology of the external genitalia. The examination concerned the spermathecae and the cibarial and pharyngeal teeth for females. Identification is performed using published keys and descriptions [22,8].

Wings preparation
Well-preserved wings of specimens identified as P. sergenti were prepared following the described protocol [23], with modifications for better visualization of veins and ribs. After treatment and coloring, wings are mounted in Canada balsam on marked slides and photographed with a digital camera (Leica microsystem CH) connected to a microscope (Leica) in the Museum of Natural History of Marrakech Cadi Ayyad University, Morocco.

Morphometry and Geomorphometry
The measurements were carried out using the Gryphax 12.0 software (Jenoptik). Pictures were processed with tps-UTIL 32 version 1.74 and tps-Dig 2 version 2.30 software [24]. In addition, the 17 landmarks already used in studies concerning other sandfly species were considered [17,18]. These landmarks are located at the intersections of the wing veins with their margin and the crossroads of the transverse veins with the prominent veins, as depicted in Figure 2. The size of P. sergenti wings is determined via the measure of the length (5-17), the width (1-9), and the centroid size (CS) (Figure 2). The latter is defined by the square root of the sum of the squared distances of a set of reference points to their centroid. The analysis of the wing shape is based on seven distances α, β, δ, ɛ, ɵ, π and γ and on the coordinates of the 17 landmarks [5,25]. First, we performed a descriptive analysis of the length, width, CS, and each of the seven distances. Then, the mean, standard deviation, median, mode, and variation ratio were calculated. Next, the normality and variance homogeneity assumptions were analyzed for each variable, using normalized plots. Finally, the correlation ratio was calculated to examine the link between a qualitative variable (wing side, station) and a quantitative one (size, shape). For each character, the Student t-test was used to verify the significance of the differences with a significance level of 0.05. All statistical analyses were performed with Microsoft Excel 10. Then, the variations in the wing's shape and size according to the geographical origin are characterized using canonical analysis, carried out with the STATISTICA software [26]. Finally, stations with low specimen numbers were grouped and considered an out-group.
Unique link hierarchical classification dendrograms were performed for wing size and shape, according to the method of Ward's, using the Past 3.23 software [27,28]. Visualization and comparison of the geometric configurations of P. sergenti wings are performed by the Procrustes superposition method or a Generalized Procrustes Analysis (GPA) [29], using the MorphoJ 1.07a software version 1.8.0_251 [30].

Results
This is the first study using descriptive and geomorphometric analyses of P. sergenti wings that investigates the size and shape of a set of stations from central Morocco differing in their ecological characteristics. Ninety-six wings of P. sergenti were collected in 14 stations of central Morocco (Table 1), 92 male wings (77 right and 15 left), and 4 right wings in females.

Wing size and shape measures
The wing's size measured as described in the Material and methods section is presented in Supplementary data ` Table S1. Overall, in males, the length of the right wings is 2137.01 ± 124.05 µm, the width is 738.17 ± 69.78 µm, and the CS is 3236.79 ± 160.66 µm. In the left wings, the length, width, and CS are 2293.27 ± 65.79 µm, 790.00 ± 40.17 µm, and 3484.06 ± 95.03 µm, respectively. The length of the right wings of females is 1897.50 ± 457.04 µm, the width is 687.56 ± 142.96 µm, and the CS is 3033.50 ± 342.53 µm. In males, the left-wing are bigger in size; significantly longer (student t-test = -11.36, p = 0.000), wider (t = -4.79, p = 0.000) and larger CS (t = -12.41, p = 0.000) than the right one ( Figure 3). The correlation ratios between wing's side and length, width, and CS are 0.73 and 0.46, and 0.84, respectively. Female wings are, in general, smaller than males. The normality test and the homogeneity distribution for the three characters' right and left male wings showed polymodal distributions, particularly of the length and CS responsible for subpopulations structuring. The shape of P. sergenti wings is characterized by seven distances; α, β, δ, ɛ, ɵ, π, and γ (Supplementary data Table S2). The left wings have longer distances α, β, δ, ɵ, ε and π than the right ones (Fig. 4). The differences are statistically significant for the distances α and β (t = -2.65, p = 0.019; t = -2.63, p = 0.020, respectively). The distance γ is the only one that appears to be smallest in left wings (t = -0.342, p = 0.013) (Figure 4). The correlation ratios between wing's side and distances α, β, δ, ɛ, ɵ, π and γ are 0.59, 0.55, 0.27, 0.26, 0.16, 0.14 and 0.15, respectively. The normality test and the homogeneity distribution show variations for each parameter. However, the wing shape varies particularly in the α and β distances, with correlation ratios exceeding 50%.

Spacial variations analyses
The size of the right-wing shows significant differences linked to their geographic origin ( Figure 5), as demonstrated by the correlation ratios computed between the origin and the length, width, or CS (0.68 and 0.56 and 0.74, respectively). Similarly, the seven distances characterizing the wing shape ( Figure 6) showed differences according to the stations (correlation ratio and Student's t-test), with β, ɵ, and γ as the most variable distances. Whether in size or shape, the wings of P. sergenti of Afourer were statistically significantly different from all other stations (Figures 5 and 6).
As for results from the descriptive analysis, canonical analyses of the size and shape of P. sergenti wings, illustrated in figure 5, point to substantial variation according to the geographical origin of the wings. Factorial analysis of the size ( Figure 7A) discloses that the factorial plane (1-2) explains 92.57% of the total inertia (canonical variate 1: 76.43% and canonical variate 2: 16.14%). These results indicate that the wing size differs from station to station according to its geographical origin. First, specimens from Afourer were located on the right of the canonical analysis, followed by those from Tiski, then from Bzou, Foum Jamaa, and the other stations (stations 2, 3, 4, 5, 7, 10, 12, 13, and 14). For the wing shape, the projection in a factorial plane (1-2) explains 63.08% of the total inertia (canonical variate 1: 43.23% and canonical variate 2: 19.85%) ( Figure 7B). As for the parameter of wing size, the canonical analysis reveals gradual variations according to the stations. The hierarchical single-link classification trees for the measured parameters of size ( Figure 8A) and wing shape ( Figure 8B) demonstrate a clustering according to the geographic origin, which again reinforces the results of the canonical analysis. But, again, two groups stand out; the populations of Afourer and Tiski with smaller wings according to measured parameters vary from those of Foum Jamaa, Bzou, or other stations. Likewise, the mean configurations of the right-wing shapes vary according to the geographical origin, with marked differences in the populations of Afourer and Tiski (Figure 9).

Discussion
Phlebotomus sergenti is a proven vector of L. tropica [1] in Saudi Arabia [6], Israel [31], and Morocco [7]. Previous studies have led to consider P. sergenti as a species complex, regarding the tremendous genetic diversity recorded over its distribution [9][10][11]32]. Nevertheless, no morphotype indicator is currently available to probe P. sergenti population diversity. Therefore, the wing's geo-morphometric analyses would help shed light on the population's diversity of P. sergenti. This has been used successfully on several sandfly species, like P. tobbi, P. papatasi, or P. ariasi [17,18,23,33,], and has provided evidence on the separation of P. papatasi populations between the southern and northern slopes of the High Atlas Mountains in Morocco [17]. This was further delineated via a genetic analysis of these populations [34]. Therefore, we applied GM analyses to probe P. sergenti population diversity in central Morocco for the first time.
Significant differences in P. sergenti wing's size and shape were recorded in populations we sampled in central Morocco. The size, length, width, and CS were significantly different depending on geographic origin, and the left wings were relatively bigger than the right ones. Such differences were absent between the right and left wings of P. papatasi originating from the northern or the southern slope of the Atlas Mountains, Morocco [17]. Despite the small number of P. sergenti females investigated in this study, we recorded that females presented smaller wings than males. Besides size variations, the shape of P. sergenti wings also showed significant variations, particularly at the two distances; α and β.
Furthermore, the right-wing shape was different from the left one. The size and shape of the P. sergenti wings showed spatial variations depending on geographic origin. Length, width, CS, and the seven distances characterizing the wing's shape differ significantly between stations, with correlation ratios higher than 50%. Population from Afourer locality being notably different.
Descriptive analyses point to a link between the size and shape and the geographical origin of the sample, which is further supported by multivariate analyses. Canonical studies reveal a gradual variation in size (92.57% of the total inertia) and shape (63.08% of the total inertia). The two populations of Afourer and Tiski are distinct with relatively smaller wings, following hierarchical classifications analyses. Superimpositions of the average configurations of the male right wings indicate significant variations in landmarks 2, 9, 10, and 14 positions. According to an altitudinal gradient, previous works have demonstrated a link between wing size and shape in P. tobbi and P. papatasi [17,23,35]. Even via an indirect effect, altitude can not be at the origin of the recorded gradual variation in P. sergenti. The individualized populations of Afourer and Tiski locate at the same altitudinal area as Bzou and Foum Jamaa, only about 60 km apart. At the same time, the slope effect described in P. ariasi female populations [33] cannot be involved in the variation we detected in P. sergenti. The stations compared are located on the same High Atlas Mountains slope. Therefore, it seems that the variations observed in P. sergenti in central Morocco should be correlated with the adaptation to a local environment. Moreover, ecologically, P. sergenti in Morocco is a constant species with an abundance and density directly related to the micro-habitat it occupies [13]. This adaptation phenomenon to a local environment is also reported in P. ariasi in the Oiselette massif in the south of France [33].
In conclusion, our results underline the diversity of P. sergenti in Morocco. It also points to P. sergenti to adapt to local environmental conditions. The correlation between the phenotypic differentiation and the genetic structure of P. sergenti in Morocco must be further investigated.

Supplementary
Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1: Morphometry of P. sergenti's wings collected in the study region in central Morocco, Table S2: Numerical characteristics of the distances α, β, δ, ε, ɵ, π and γ of the male (right and left) and right female wings of P. sergenti collected in central Morocco.