The fecal microbiomes analysis of Marabou storks (Leptoptilos crumenifer) reveals their acclimatization to the feeding environment in the Kampala urban areas, Uganda

Abstract

The Marabou stork (Leptoptilos crumenifer) is a typical scavenging bird and adapted to the Savannah environment, where they show a carnivorous feeding style. However, Marabou stork recently penetrated into the city areas and acclimatized to the urban environment, where they modified their feeding habits to an omnivorous type toward more carbohydrate. To reveal their adaptation to the variable feeding customs, this study compared the gut microbiomes and chemical compositions of feces of Marabou storks inhabiting two different locations in peri urban Kampala: one is a slaughter house floc that predicted their original carnivorous feeding, and the other is a landfill floc that adapted more to the omnivorous feeding. 16S rRNA gene sequencing analysis revealed more diverse gut microbiome, more enriched Lactobacilli, and less abundant Peptostreptococci in the landfill flock comparing to the slaughter house flock. Isolation work and predicted metagenome analysis confirmed more diverse Lactobacilli and more enriched functions for carbohydrate metabolism in the landfill flock. In addition, chemical composition of feces revealed higher ammonia in the former, which is consisting with higher Peptostreptococci and their practice of carnivorous feeding. These results highlighted their adaptation to the variable feeding environment, which presumably protects their health and ensure survival of species.

The Marabou stork (Leptoptilos crumenifer) is a typical scavenging bird in the Savannah environment, which contributes towards nutrient and energy recycling in the ecosystem [16]. Although the Marabou stork is originally a scavenging bird and adapted to the carnivorous feeding systems, many of them recently acclimatized to the urban environment by relying on wastes generated due to human activities [4], where they modified their feeding habits toward an omnivorous type. As shown in Fig. 1, this bird nests on the treetop in the city and feeds on city garbage. Changes in feeding habits may result into a fundamental modification of the gut microbiome as suggested from several microbiome studies comparing wild birds with their captive counterparts [11, 25], but the fecal microbiome of Marabou stork has never been investigated. The gut microbiome is now recognized as a part of the host [5] and the notion of “Super-organism” has been well accepted [24]. Accordingly, we hypothesized that the modification of gut microbiome is a key component of adaptation to this new feeding system in Marabou storks. To reveal the influence of different feeding habits on the gut microbiome, this study explored the fecal microbiomes of two groups of Marabou storks involved in diverse feeding lifestyles; one was feeding on the city garbage at a landfill managed by Kampala capital city authorities and the other mainly fed on wastes generated by a pig slaughter house found in peri urban Kampala. The feeding behavior of the slaughter house-stationed storks is relatively similar to that of the storks that feed on the carcasses of animals in the savanna. The unknown difference in the gut microbiome and fecal metabolite characteristics could explain their adaptation to the different feeding habits. Additionally, we hypothesized that there is a possibility of differences shown in the abundance of lactic acid bacterial population by 16S rRNA gene sequencing data, namely lactobacilli in the fecal microbiome of the two groups in order for them to adapt to the novel feeding environment. Thus, we tried to compare the distribution of Lactic acid bacteria, such as Lactobacillus spp. and Bifidobacterium spp., in the two flocks’ samples. Unfortunately, none of Bifidobacterium was isolated, therefore we compared the distribution of Lactobacillus spp. between two flocks.

Fig. 1.

Marabou storks (Leptoptilos crumenifer) in the urban environment. (A) At the city center (Makerere University), (B) at a pig slaughter house, (C) landfill area and (D) the resting place of the flocks of landfill area. All the pictures were taken by the authors; (A), (C) K Ushida; (B), (D) S Tsuchida.

MATERIALS AND METHODS

Feces collection

Fecal collection was made in a non-invasive manner. Fresh feces (27 samples in total) were collected from wild adult Marabou storks at two different locations; 17 samples were collected at a pig slaughter house located at Nakassozi, near Kampala (0°17’13.49” N, 32°29’18.21”E) and 10 samples were collected at an open space of a cement factory near Kampala (0°24’14.26”N, 32°35’01.48”E) in September 2019. The former place, at the outside of southwestern limit of Kampala, is the feeding place of a group of Marabou storks that lodge in a forest (about 0.6 km²) behind the slaughter house. This group of Marabou storks feed every day on the waste of slaughtered pigs such as pieces bones, the gut and the pig skin. The latter place, at the outside of northeastern limit of Kampala, is the resting place for Marabou storks that feed the garbage at the Kampala city landfill (0°24’36.70”N, 32°34’33.63”E). Their action ranges were informed by the workers in both sites. Since Marabout storks do not show clear sexual dimorphism and age difference except for young chicks, it was hard to identify the individual information about feces. Figure 1 shows the Marabou storks at city center and both sampling sites. The number of birds in the two flocks were about 20 in the pig slaughter house and about 80 at the landfill respectively. We carefully observed the birds within a distance (around 15 m) which did not modify their behavior. Upon visual recognition of defecation, we slowly approached the birds and carefully collected feces using sterile tweezers avoiding contamination of soil or other environmental materials. In principle, three portions of feces were introduced into a transport medium [17], a DNA conservation buffer [35] and 12% (w/v) perchloric acid [33]. All the samples were transported to the Laboratory at Central Diagnostic Laboratory of Makerere University promptly under ambient temperature.

16S rRNA gene sequencing

Total bacterial DNA was extracted by a QuickGene-Mini80 (Kurabo Industries, Tokyo, Japan) from the feces as indicated elsewhere [34]. A DNA quality check was done using a microvolume spectrometer (Denovix DS-11, Wilmington, DE, USA) and the samples were sent under frozen state with dry ice to Novogene Co. (Hong Kong, China) for bacterial 16S amplicon sequencing. Briefly, barcoded amplicon libraries targeting an ~300 bp fragment of the V3-V4 region of the bacterial 16S rRNA gene were prepared using the universal primer pair Pro341F and Pro805R [27] for the simultaneous detection of bacteria and Archea, and NEBNext® UltraTM DNA Library Prep Kit for Illumina Hiseq 2500 (Illumina, San Diego, CA, USA). The PCR products were mixed at equal density ratios and then purified using Qiagen Gel Extraction Kit (Qiagen, Hilden, Germany). The libraries were generated with NEBNext^® UltraTM DNA Library Prep Kit for Illumina Hiseq 2500. According to the quality checks, 10 samples from a landfill group and 7 samples from a slaughter house group were subjected to illumina Hiseq sequencing.

Gut microbial composition analysis

The sequenced reads were processed using QIIME 2^TM pipeline as indicated elsewhere [30]. In brief, taxonomic classification was performed using the GTDB (Genome Taxonomy Database, date: July 2019, https://gtdb.ecogenomic.org/#) and the Naïve-Bayes classifier. Genera, which were detected in a sole sample or only in two samples, and genera with a mean abundance below 0.01% were filtered [31]. Principal coordinates analysis (PCoA) with Weighted Unifrac distance and the calculation of Shannon-Wiener alpha-diversity index were performed by R (v3.6.3) [21] to examine the difference of overall microbiome characteristics between groups. The abundance of bacterial genera and Shannon-Wiener alpha-diversity index were determined to be either significantly increased or decreased using a two-sided Wilcoxon rank sum test of R (v3.6.3). In addition, a Benjamini-Hochberg false-discovery rate-corrected P value (q value) was estimated.

Bacterial Culturing, DNA extraction and 16S rRNA gene sequencing

Fecal specimens in transport medium were further diluted using a 10-fold dilution series to obtain a 1.0-MacFarland standard using the same transport medium. A portion (20 µL) was then applied on the plates made with BD Difco™ Lactobacilli MRS agar medium or BD BBL™ LBS agar medium (BD Difco BBL, Franklin Lakes, NJ, USA) in addition to non-selective media, BL medium (Nissui Pharmaceutical, Tokyo, Japan) and EG medium [18]. Anaerobic culture methods for the isolation and identification of lactic acid bacteria were performed as indicated elsewhere [29]. In brief, the plates were placed in anaerobic jars with an AnaeroPack™ (Mitsubishi Gas Chemical, Tokyo, Japan) to obtain anaerobiosis and incubated at 37°C for 24 hr. Developed colonies were picked and transferred to the same fresh media and incubated in the same manner for 24 hr. Purity of the isolates was checked microscopically before subjection to DNA extraction as indicated in a previous report with the combination of beads beating in Tris-HCl buffer. Bacterial 16S rRNA genes were amplified by PCR and sequenced as indicated elsewhere [30].

PICRUSt2 prediction of metagenome functions

Functional profiles of the fecal microbiomes of Marabou storks were predicted by PICRUSt2 [6]. MetaCyc pathway abundances were calculated in PICRUSt2 through structured mappings of EC (Enzyme Commission numbers) gene families to pathways. The abundance of pathway was determined to be either significantly increased or decreased using a two-sided Wilcoxon rank sum test of R (v3.6.3). In addition, a Benjamini-Hochberg false-discovery rate-corrected P value (q value) was estimated by R (v3.6.3).

Fecal organic acid and ammonia analyses

A known weight of feces was mixed with 12% perchloric acid. The mixtures were centrifuged at 10,000 rpm for 15 min at 4°C. The supernatants were recovered and quickly analyzed by an ion-exclusion HPLC organic analysis [33]. Concentrations of succinic acid, DL-lactic acid, formic acid, acetic acid, propionic acid, iso-butyric acid, butyric acid, iso-valeric acid and valeric acid were determined. Concentration of ammonia was determined by phenol-hypochlorite assay [36].

Ethics

This research project was approved by Uganda National Council for Science and Technology (A522, on 22th August 2016) and all the sampling and access to the birds in this study was permitted by Uganda Wildlife Authority (UWA/COD/96/05 on 12th February 2018).

RESULTS

Comparisons of gut microbiome between the two Marabou stork flocks

The composition of the gut microbiome at species, genus and family level was shown in Supplementary Tables 1, 2 and 3. The identified bacterial families were in total 90 in which 82 were detected from the landfill flock and 41 were from the slaughter house flock. In the case of the genus level identification, a total of 210 genera were identified of which 197 were detected from the landfill flocks and 81 were from the slaughter house flocks.

The fecal microbiome structure was different between two locations as shown in PCoA (Fig. 2A). Shannon-Wiener alpha-diversity index was higher in the landfill flock compared to the slaughter house flock (Fig. 2B, P=0.0001). The abundance of each genus was compared between two locations as shown in Figs. 3 and 4. Abundance of Lactic acid bacteria (LAB), such as Lactobacillus, Bifidobacterium, Streptococcus, Weissella, and Lactococcus, was significantly higher in the landfill flock than the slaughter house flock. In addition to LAB, abundance of Erysipelotrichales such as Erysipelothrix, Turicibacter and Erysipelactoclostridium was also significantly higher in the landfill flock than the slaughter house flock. Both groups of bacteria were larger in the landfill flock than the slaughter house flock. In reverse, abundance of Peptostreptococcus was significantly higher in the slaughter house flock than the landfill flock.

Fig. 2.

(A) Scatter Plot showing the result of principal coordinate analysis (PCoA) with Weighted Unifrac distance of Marabou storks at the landfill and the pig slaughter house, (B) box plot showing the IQR of the Shannon-Wiener alpha-diversity index of Marabout storks at the Landfill and the slaughter house. Intergroup differences were analyzed by the Wilcoxon rank sum test.

Fig. 3.

Comparison of the fecal microbiome of Marabou storks between flocks at the landfill area and those at a slaughter house. The X axis shows the log₂ of the median abundance in a slaughter house flock over the median abundance in a landfill flock. The Y axis shows the –log₁₀ of the q value analyzed by the Wilcoxon rank sum test with a Benjamini-Hochberg false-discovery rate-correction. The red horizontal line indicates q=0.05. The size of the circle shows the median abundance of each genus.

Fig. 4.

Comparison of the fecal microbiomes of Marabou stork flocks at the landfill area and at a slaughter house. The figure shows the interquartile range (IQR) of the relative abundance of the genera between a slaughter house flock and a landfill flock. Intergroup differences were analyzed by the Wilcoxon rank sum test with a Benjamini-Hochberg false-discovery rate-correction. Values in the figure show q-values. This box plot figure shows 14 genera, of which abundance was all significantly different between two flock

Comparisons of gut predicted metagenome between the two Marabou stork flocks

The abundance of pathways was compared between two locations as shown in Figs. 5 and 6. The volcano plot (Fig. 5) indicates that most of the pathways were present in both flocks. Abundance of 7 pathways were significantly higher in the landfill flock than the slaughter house flock as shown in Fig. 6. Pathway (PWY)-7208 as a pyrimidine nucleobases salvage, PWY-6317 as a D-galactose degradation, P124-PWY as a Bifidobactrerium shunt, P122-PWY as a heterolactic fermentation, PWY-5910 as a geranylgeranyldiphosphate biosynthesis I (via mevalonate), PWY-922 as a mevalonate pathway I, and LACTOSECAT-PWY as a lactose and galactose degradation I.

Fig. 5.

Comparison of functional profiles (pathways) of the fecal microbiomes of Marabou storks at the landfill area and at a slaughter house. The X axis shows the log₂ of the median abundance in a slaughter house flock over the median abundance in a landfill flock. The Y axis shows the –log₁₀ of the q value analyzed by the Wilcoxon rank sum test with a Benjamini-Hochberg false-discovery rate-correction. The red horizontal line indicates q=0.05. The size of the circle shows the median abundance of each pathway.

Fig. 6.

Comparison of functional profiles (pathway) of the fecal microbiomes of Marabou storks at the slaughter house and those at the landfill area. The figure shows the interquartile range (IQR) of the relative abundance of the pathways between the landfill flock and the slaughter house flock. Intergroup differences were analyzed by the Wilcoxon rank sum test with a Benjamini-Hochberg false-discovery rate-correction. Values in the figure show q-values. This box plot figure shows top 7 pathways, of which abundance was significantly different between two flocks.

Isolation of Lactobacillus spp. from the fecal samples of marabou storks

We obtained 25 sequences of formerly classified Lactobacillus spp. from meta 16S sequencing and have succeeded to isolate 5 species, L. agilis, L. aviaris, L. fermentum, L. mucosae, L. salivarius among them and additional 6 species; L. reuteri, L. vaginalis, L. oris, L. gasseri, L. johnsonii, and L. pontis, which were not suggested from the meta 16S sequencing. Novel nomenclature for those subjected to the Lactobacillus taxonomy change is used in Table 1. Vagococcus carniphilus was also isolated by the same media only from the slaughter house flocks. However, isolation of Bifidobacterium spp. was failed by EG and BL medium in this study.

Table 1. Detection of Lactobacillus spp. from feces of Marabou storks at two different locations.

	Landfill	Slaughter house
Lactobacillus agilis	+	-
Lactobacillus aviarius	+	+
Limosilactobacillus fermentum	+	+
Lactobacillus gasseri	+	-
Lactobacillus johnsonii	+	-
Limosilactobacillus mucosae	+	+
Limosilactobacillus oris	+	-
Limosilactobacillus pontis	+	-
Limosilactobacillus reuteri	+	+
Ligilactobacillus salivarius	+	-
Limosilactobacillus vaginalis	+	-
Vagococcus carniphilus*	-	+

+, success of isolation. -, not isolated from the feces. *Vagococcus carniphilus is involved in this table as an indicator of carnivorous feeding. Landfill and Slaughter house, see text and Fig. 1.

Concentration of organic acids and ammonia in the fecal samples of marabou storks

The concentrations of fecal organic acids and ammonia are shown in Table 2. The concentration of ammonia in the slaughter house samples was higher than that in the landfill samples (1,115 vs. 590 mg ammonia-N/kg feces, P=0.019). There was no statistical difference in the levels of acidic metabolites excreted in the two locations storks.

Table 2. Profile of fecal organic acid and ammonia of Marabou storks at two different locations.

Molar % of acid	Landfill (n=9)	Slaughter house (n=10)	Welch t-test
Succinate	3.57 ± 1.32	5.13 ± 2.18	NS
Lactate	1.01 ± 0.32	3.21 ± 1.8	NS
Formate	10.01 ± 2.13	12.42 ± 2.06	NS
Acetate	72.41 ± 3.09	65.07 ± 3.65	NS
Propionate	3.72 ± 0.79	3.55 ± 0.48	NS
iso-Butyrate	1.15 ± 0.13	2.33 ± 1.58	NS
n-Butyrate	7.29 ± 1.26	6.84 ± 1.58	NS
iso-Varelate	0.78 ± 0.19	0.63 ± 0.28	NS
Varelate	0.06 ± 0.04	0.81 ± 0.75	NS
Total organic acid (mmol/kg)	82.53 ± 17.91	79.02 ± 16.25	NS
Ammonia (mgN/kg)	590.35 ± 58.80	1,115.48 ± 217.51	P<0.05

Values are mean with SEM. Welch t-test was performed. NS, not significant (P>0.05). Landfill and Slaughter house, see text and Fig. 1.

DISCUSSION

The Marabou stork is a typical scavenging bird and formerly adapted to the Savannah environment, where they showed a carnivorous feeding style. However, Marabou stork recently acclimatized to the urban environment, where they modified their feeding habits toward an omnivorous type.

The gut microbes of scavenging birds were first studied by Roggenbuck et al. [22], who sought to understand the tolerance of the New World vultures (Coragyps atratus and Cathartes aura) to potentially harmful food such as decaying carcasses, which are deemed to have pathogens. These authors revealed that New World vultures fed mainly on road-killed mammals, Bovidae and Cervidae, in rural areas, and suggested that the acidity of the gut internal environment is a strong filter towards potentially harmful bacteria from carrion. In the study done by Meng et al. reported that 16S rRNA genes of pathogenic bacteria were considerably detected [14] in the Old World vultures (Gyps himalayensis, Gypaetus barbatus and Aegypius monachus), which fed on small rodents and human remains at the ritual site, typical in the Qinghai-Tibet Plateau. Unlike the New World vultures and the Old World vultures studied in those previous studies, the presently targeted Marabou storks adapt well to the city environment by changing their food habit [4].

The fecal microbiome structure was different between two locations as shown in PCoA (Fig. 2a). Shannon-Wiener alpha-diversity index was higher in the landfill flock compared to the slaughter house flocks (Fig. 2b). Simpler microbiome structure may depend on the diversity of food items and food chemical composition. We observed that the slaughter house flocks fed, the most of their day time on residual portions of pig carcasses found in the open abattoir water waste drainage, while the landfill flocks fed on city garbage which naturally consisted of a variety of human food residues. The residual portion of pig carcasses is composed mainly of protein and fat, and carbohydrate may be scarce. Therefore, fecal ammonia concentration was significantly higher in the slaughter house flocks (Table 2). On the contrary, the general composition of landfill garbage is relatively rich in organic waste [12] with a carbohydrate nutrient bias, which has also been noted with wastes in other town settings [3]. Although fecal organic acid profile might be different between two flocs due to the difference of the food in addition to the difference in microbiome composition, no significant difference was observed (Table 2). Fecal organic acid is the residual component after absorption of the acid from the intestinal mucosa, it may not correctly reflect the acid concentration in the gut. In vitro fermentation ability of the fecal bacteria is the one possible approach that gives more precise estimation fermentation ability of gut bacteria.

As shown in Fig. 3, Peptostreptococcus was the bacteria enriched in slaughter house flocks and Lactic acid bacteria such as Lactobacillus, Bifidobacterium, Streptococcus, Lactococcus and Weisella were enriched under landfill condition. This tendency is similar to the observation by Menke et al. [15] in which they showed the increase in Peptostreptococcus in strict carnivores, which may be sponsored by the higher fat intake [2]. In contrast to Peptostreptococcus, the same authors also mentioned that the prevalence of Lactobacillus spp. was enriched by omnivorous feeding. The feeding situation alters microbial composition in the gut, which may lead to the alteration in predicted metabolic pathways (Figs. 5 and 6). The enhanced metabolic pathways obviously depend on the enhanced prevalence of lactic acid bacteria under landfill condition, mostly related to the carbohydrate metabolism such as D-galactose degradation and lactose and galactose degradation. Decrease in several pathways for amino acid formation may indicate that the protein supply was limited as a feed nutrient in such a location.

In addition, the variety of Lactobacillal isolates were widely harbored in the landfill flock than in the slaughter house flock, which again indicates that the landfill omnivorous feeding condition supports the establishment of Lactobacillus spp. in the gut of the storks. Isolation of Vagococcus carniphilus from slaughter house flocks may explain their purer carnivore-feeding style, because Vagococci were isolated before from typical aquatic carnivore mammals [10, 13]. Lactic acid bacteria are recognized to have health promoting abilities [20]. Enhanced prevalence of Lactic acid bacteria may help the Marabou storks to live in the sanitary severe conditions, the landfill environment most probably by bacterial antagonism to restrict pathogenic invasion, which has been defined as a competitive exclusion of pathogens by probiotic Lactic acid bacteria in Poultry [1].

Regarding Archea, we used primer sets to amplify both Archaea and bacteria, but few archaeal sequences of Methanosphaera spp. were sporadically detected only from landfill individuals. Therefore, Marabou storks may harbor methanogenic archaea in their gut but archaea are not principal members of their gut microbiome and possibly a transient microbe. This is not surprising, because the methanogens generally require low rate of turnover and low redox potential. Such physicochemical conditions are found in the rumen of ruminant animals [32] or the cecum of the hind gut fermenters including particular species such as the ostrich [8], the galliformes like chicken [28] or rock ptarmigans [23].

Scavenging birds are recognized as a potential threat to public health, because they can carry and disseminate deadly pathogens such as Bacillus anthracis [9]. The status of scavenger birds, vultures (Gyps himalayensis, Gypaetus barbatus and Aegypius monachus) as a reservoir for pathogenic bacteria was suggested by 16S rRNA gene amplicon sequence diagnostics [14]. Due to their behavior in the city, Marabou storks can be recognized as potential threat to public health by playing a role as a reservoir and transmission vector of pathogens. In fact, several preliminary fecal analyses were conducted about the transmission of pathogens [19] and drug resistant bacteria [7, 26]. In this research, we did not detect particular serious pathogen sequences except for Erysipelothrix rhusiopathiae, the causative agent of swine erysipelosis, which is also regarded as a zoonotic pathogen under legal control in many countries (Supplementary Table 3). Sequences of several opportunistic nosocomials species were also detected; such as Erysipelatoclostridium ramosum, Hafnia paralvei, Leminorella richardii, and Terrisporobacter glycolicus. As we did not confirm the presence of these pathogens by performing isolation studies, we cannot fully conclude on the public health impact of Marabou storks in the community.

The present findings highlighted the role of feeding habits on the gut microbiome of chordates, with special emphasis on two flocks of Marabou storks. A slaughter house flock that practiced their original carnivorous feeding style and the landfill frock that have adapted more to the omnivorous feeding style. Due to their different feeding behavior, as the primary factor, the gut microbiota of the landfill flock showed high abundance and variation of Lactic acid bacteria, which plausibly protects their health under the difficult feeding environment and ensures the long-term survival of species.

CONFLICT OF INTEREST

The authors declare no conflict of interests.