Comparison of automatic methods MALDI-TOF, VITEK2 and manual methods for the identification of intestinal microbial communities on the example of samples from alpacas (Vicugna pacos)

Matrix-assisted laser desorption/ionisation–time-of-flight mass spectrometry (MALDI-TOF MS) technique is the result of a combination of mass spectrometry, matrix-assisted laser desorption/ ionisation and the technology of time of flight, a method in which charged ions are forced to fly from a source to a detector and are identified based on the time of their flight to the ion detector (13). This technique can be used for the detection of specific species of bacteria and yeast from human and animal samples (6). Thanks to its capability of detecting many species of bacteria or yeast at the same time, the technique is useful in diagnosing both pathogens and accompanying microbiota. The companion microbiota in human, animal or environmental materials for testing was treated by some researchers in the past as clinically insignificant contamination (15). Nowadays it is known that the digestive tract microbiota, as only one example, is a very important element in maintaining the health of humans and animals. Each unfavourable change in its composition might have an impact on the synthesis of B vitamins or even susceptibility to various infections and the course of the disease (20). The MALDI-TOF technique allows many microbial isolates to be identified in a short time because the throughput of testing can be high, which can significantly facilitate the characterisation of a given microbiota without the need to use advanced genetic techniques such as next-generation sequencing (NGS).

The MALDI-TOF development in mass spectrometry makes characterising the gastrointestinal tract microbiota a simpler proposition. Its utility can be exploited to elucidate the gut microbiota of the alpaca, which is a knowledge gap more noticeable in recent times. Alpacas are among the camelids of the New World, the natural habitat of which is mountainous regions of South America. Interest in this species has been increasing in various countries over the past three decades. Alpacas have gained popularity because of their unique fibre, mild temper, and the possibility of using them in psychotherapy. Also in Poland, alpacas are being bred on an larger scale (15). The meat of New World camelids may be also used for human consumption (28).

Although alpacas are becoming more and more popular, there is much that is not known about them. Knowledge of the pathogens present in alpacas is particularly important, from the public health hazard point of view as well as in making the right decision about treatment, particularly the use of antibiotics. The zoonotic diseases in alpacas are known to be tuberculosis, cryptosporidiosis, colibacteriosis (caused by E. coli VTEC O157), salmonellosis, campylobacteriosis, dermatophytosis, leptospirosis, listeriosis, streptococcosis (caused by Streptococcus equi subsp. zooepidemicus and Streptococcus suis), yersiniosis, brucellosis, sarcoptic mange and rabies (2, 12, 30).

Gut associated microbes may influence the alpaca health, but still little is known about the species’ microbiota (3, 5, 11). In research conducted by Guerrero-Olmos et al. (11) the most prevalent bacterial genus in the alpaca gut microbiome was Enterococcus and the dominant species was Enterococcus hirae (82%), followed by other members of the genus which were neither Enterococcus faecalis nor Enterococcus faecium (11). The latest reports concern the profile of faecal bacterial microbiota, characterised by sequencing of 16S rRNA amplicons or the isolation of specific pathogenic microorganisms (5, 29). However, there is a lack of research based on the isolation of the components of a living gut bacterial community in alpacas, which is highly variable impacted by the nutrition and health status of the animal (4, 28). The aims of this study were firstly to compare two automatic methods, MALDI-TOF MS and the VITEK2 fluorogenic system, and two manual biochemical methods, the analytical profile index (API) and RapID tests, for identification of isolated microorganisms from three fallen alpacas guts and faeces, and secondly to describe their gut microbiota.

In total, over 121 microorganisms belonging to 54 genera of bacteria and yeast were isolated in this study. Of these, 5 isolates could not be identified by MALDI-TOF MS or any other instrumental and manual method. Most of the isolates were Gram-positive, such as Bacillus spp., Enterococcus spp. (which were dominant) and Staphylococcus sciuri. Other Gram-positive bacteria isolated were Paenibacillus amylolyticus and Corynebacterium stationis. In addition, Gram-positive and Gram-variable Micrococcus cocci were isolated. There were also Gram-negative bacteria isolated: Enterobacter cloacae, Enterobacter hormaechei, Enterobacter ludwigii and Enterobacter gergoviae(currently Pluralibacter gergoviae); Citrobacter spp.; E. coli; Stenotrophomonas maltophilia; Serratia liquefaciens and Serratia odorifera were also found.

The bacterial species isolated from the individual sections of the intestine from which the swabs were taken differed. The following microorganisms were isolated from the small intestines of the alpacas: Candida albicans, Ochrobactrum intermedium, Enterobacter gergoviae, Serratia odorifera, Moraxella osloensis, Staphylococcus sciuri, Bacillus licheniformis, Bacilus altitudinis and Arcticiflavibacter luteus. The following microorganisms were isolated from the animals’ large intestines: Candida metapsilosis, Candida lipolytica (currently Yarrowia lipolytica) and Candida haemulonii; Enterobacter cloacae and Enterobacter hormaechei; E. coli; Serratia liquefaciens; Bacillus cecembensis, Bacillus pumilus, Bacillus cereus, Bacillus flexus, Bacillus licheniformis and Bacillus subtilis; Enterococcus hirae and Enterococcus casseliflavus; Staphylococcus sciuri; Brachybacterium conglomeratum; Corynebacterium stationis; Stenotrophomonas maltophilia; Sphingobacterium anhuiense; Thallasolituus oleivorans; Leeuwenhoekiella nanhaiensis; and Paenibacillus amylolyticus. The rectum and faeces yielded Citrobacter braakii and Citrobacter freundii; Serratia odorifera; Enterobacter ludwigii, Enterobacter cloacae, Enterococcus hirae and Enterococcus casseiflavus; Bacillus aeolinus and Bacillus licheniformis; Staphylococcus pseudintermedius; Micrococcus luteus; Leuconostoc citreum; and Cellulosimicrobium cellulans.

Enterobacter strains (61.5%) were the most numerous group in the Enterobacteriaceae family, followed by Citrobacter spp. (30.7%) and E. coli (7.7%). The largest group of Enterobacter were Enterobacter cloacae and single instances of Enterobacter gergoviae, Enterobacter hormaechei and Enterobacter ludwigii. Three of the four isolated Citrobacter spp. were confirmed as Citrobacter freundii by all identification methods, whereas one of these isolates was determined by VITEK2 GN and API 20E as Citrobacter braakii with 89% and 92% probabilities, respectively, and as Citrobacter freundii by MALDI-TOF MS. The spectrum which was not defined by MALDI-TOF MS was identified by the other three methods as Citrobacter freundii with >95% probabilities (Table 2). The biochemical identification results of E. coli isolates by the VITEK2, API and RapID systems were consistent with those from MALDI-TOF MS.

The Gram-negative Yersiniaceae strains in 60% of both discussed bacterial identification techniques indicated the same species, Serratia odorifera; however, in one case MALDI-TOF MS indicated Serratia liquefaciens and VITEK2 Serratia odorifera and in another case, MALDI-TOF MS did not indicate any bacteria or genus and VITEK2 found Serratia marcescens (Table 3). None of the Serratia species produced the red pigment prodigiosin. Colonies of Serratia odorifera and Serratia liquefaciens had a musty odour.

All Gram-negative Morganellaceae were isolated from the alpaca which was healthy until its death (AH). Of these isolates, 40% were Morganella morganii and 60% Providencia alcalifaciens. Morganella morganii, Providencia alcalifaciens and Pseudomonas aeruginosa were all correctly identified to the species level by the API, RapID, VITEK2 and MALDI-TOF MS techniques.

Among Gram-positive cocci in the MALDI-TOF method, 100% of the results with determination of the bacterial species were obtained, with the VITEK2 GP device 83.3% and the API Staph test, three Staphylococcus sp. strains and one Micrococcus sp.

Of all Gram-positive cocci, 58.3% were Enterococcus spp. of which five strains were Enterococcus hirae and two strains of Enterococcus casseliflavus, 25% were Staphylococcus spp., two of which belonged to Staphylococcus sciuri and one Staphylococcus pseudintermedius. Other strains were Brachybacterium conglomeratum and Micrococcus luteus. All details regarding the identification of Gram-positive cocci can be found in Tables 4 and 5.

Gram-positive cocci Brachybacterium conglomeratum were identified by MALDI-TOF, but neither by VITEK2 nor API Staph. Brachybacterium conglomeratum belongs to the Dermabacteraceae family and as originally classified as Micrococcus conglomeratus. Micrococcus luteus (family Micrococcaceae) was identified by MALDI-TOF and API Staph as Micrococcus sp. but not by VITEK2 GP. Leuconostoc citreum (family Lactobacillaceae) was identified by MALDI-TOF and VITEK2 as Leuconostoc mesenteroides but not by API Staph. VITEK2 identified Gram-positive bacteria (cocci and rods) but API Staph identified only Staphylococcus, Kocuria and Micrococcus.

Bacillus spp. has comparable results between MALDI-TOF and VITEK2 BCL with 78.8% identical results. 21.2% of the strains agreed on the genus but not the species (Table 6).

In one case, Bacillus “anthraci” (named by MALDI-TOF) was detected and VITEK2 BCL indicated as Bacillus cereus. These discrepancies required additional microbiological determinations. The tested Bacillus strain formed large, dull, opaque grey colonies with rough edges with circular zones of β-haemolysis surrounding the colonies on blood agar (Fig. 1a) and rough and dry with light pink colonies with halo egg yolk precipitation on MYP agar (Fig. 1b). Repeated biochemical results allowed its designation as Bacillus cereus.

Fig. 1a.

Fig. 1b.

For Clostridium perfringens identification both VITEK2 ANC card and MALDI-TOF MS provided identical results at species level for the four isolated strains. Thus both methods can be used for accurate routine anaerobe identification (Table 7).

Some of the isolated bacterial strains (Arcticiflavibacter luteus, Leeuwenhoekiella nanhaiensis, Sphingobacterium anhuiense, Thalassolituus oleivorans, Cellulosimicrobium cellulans, Brachybacterium conglomeratum) caused major diagnostic difficulties due to rare isolation and unspecific colony and cells morphology. These group of bacteria are presented in Table 8. In the MALDI-TOF study, 33.3% were reliable for species, 25% for reliable types and 41.6% for unreliable results. The results obtained in the MALDI-TOF for Stenotrophomonas sp. have been confirmed as Stenotrophomonas maltophilia by VITEK2 and API 20NE >96%, Moraxella osloensis as group level identification (Moraxella spp.) and RapID NF Plus as Moraxella osloensis. Ochrobactrum intermedium was classified as Ochrobactrum anthropi in biochemical tests (Table 8).

Leuconostoc citreum in the MALDI-TOF study was confirmed with VITEK2 GP as Leuconostoc citreum and it was similarly for Paenibacillus amylolyticus. Other bacteria Cellulosimicrobium cellulans, formerly known as Oerskovia xanthineolytica, grown as colonies less than 2 mm in size, glistening and yellow.

Corynebacterium stationis in the MALDI-TOF study was identified with VITEK2 CBC as Corynebacterium minutissimum (Table 9), but Corynebacterium xerosis was identified both by MALDI-TOF MS and VITEK2 CBC. Surprisingly, in Corynebacterium spp., Corynebacterium diphtheriae was identified by biochemical methods: manual API CORYNE and automated VITEK2 CBC card, but as Candida lipolytica by MALDI-TOF MS. Gram stain test results reveal presence of slightly curved rods visible in magnification 100 × instead of yeast identified as Candida lipolytica by MALDI-TOF MS.

A total of 8 yeast culture were isolated from two alpacas, mainly from 1A. Overall distribution of Candida yeast species was as follows: C. albicans, C. haemulonii, C. metapsilosis (C. ciferrii in VITEK2) and C. lipolytica. For Candida lipolytica MALDI-TOF score was 4.495, result unreliable (Table 10).

The smear prepare from putative Candida lipolytica shows typical Gram stain bacteria appearing as V-in Y-shaped arrangements or in clumps that resemble Chinese letters. VITEK2 CBC cards identified microorganism as Corynebacterium diphtheriae (Table 10).

For each alpaca, significant differences in microbial community compositions were observed. The small intestine, colon and rectum of alpaca 1A and 2A had the richest diversity of microorganisms than healthy alpaca (AH). In small intestine 1A alpaca 7 bacteria (Gram-negative rods, Gram-positive Enterococcus, and Corynebacterium) and 1 yeast Candida albicans, in 2A alpaca 6 bacteria (Gram-negative rods, Gram-positive Staphylococcus, and Bacillus) were found. In AH alpaca 5 bacteria (Gram-negative rods, Gram-positive Enterococcus, Bacillus and Clostridium) were found (Fig. 2). In colon in 1A alpaca 10 bacteria (Gram-negative rods, Gram-positive Staphylococcus, and Corynebacterium) and yeast Candida albicans, Candida. metapsilosis/ciferrii, in 2A alpaca 12 bacteria (Gram-negative rods, Gram-positive Enterococcus, and Bacillus) and Candida. haemulonii were found. In AH alpaca 3 bacteria (Gram-negative rods and Gram-positive Clostridium) were found (Fig. 3).

Fig. 2.

Fig. 3.

In rectum in 1A alpaca 6 bacteria (Gram-negative rods, Gram-positive Staphylococcus, Enterococcus), in 2A alpaca 11 bacteria (Gram-negative rods, Gram-positive Enterococcus, Staphylococcus and Bacillus) were found. In AH alpaca 4 bacteria (Gram-negative rods and Gram-positive Enterococcus) were found (Fig. 4).

Fig. 4.

Staphylococcus aureus, Staphylococcus CNS and Candida spp. were found in alpacas with clinical symptoms (alpaca 1A, 2A), whilst E. coli is primarily found as gut commensal in healthy animals (lack in 2A).

Bacterial identification was performed by MALDI-TOF MS, VITEK2, API 20E, 20NE, STAPH and RapID ONE and RapID NF Plus. Automatic and manual methods were compared (Tables 2–10).

All investigated automatic and manual methods turned out to be powerful tools for fast, accurate identification of microorganisms isolated from alpacas. However, several rare and poorly known bacteria were identified only by MALDI-TOF MS during the research: Ochrobactrum intermedium, Cellulosimicrobium cellulans, Brachybacterium conglomeratum, Thallasolituus oleivorans and Leeuwenhoekiella nanhaiensis.

Ochrobactrum intermedium is currently referred to as an emerging opportunistic pathogen of humans (16). It is Gram-negative, aerobic, short straight or slightly curved bacilli, belongs to the genus Ochrobactrum and the family Brucellaceae (1). Ochrobactrum intermedium is difficult to distinguish from other Ochrobactrum species by classical microbiological methods. More and more often identified with the use of MALDI-TOF (14, 22).

Cellulosimicrobium cellulans is Gram-positive bacilli belonging to the order Actinomycetales, widely distributed in the environment and have been isolated mainly from soil and grass. The bacterium was described as vancomycin resistant, opportunistic pathogen. In humans, specially, children up 5 years of age and adults with immunosuppression, such as HIV or over 70 years of age, the bacterium causes bacteremia, endophthalmitis, tenosynovitis and peritonitis. Many of Actinomycetales infections are characterised by a relatively long history with minimal clinical signs of infection at initial presentation, which often leads to a delay in diagnosis (26).

Brachybacterium conglomeratum is Gram-positive, nonmotile and nonsporing aerobic or facultatively anaerobic bacteria. Although Brachybacterium spp. was isolated from, among others, plants, salt-fermented seafood, and the surface of French Gruyère and Beaufort cheeses and even Brachybacterium. paraconglomeratum from human eye infection (10, 24, 25). However, there is no information available on the pathogenicity of Brachybacterium conglomeratum in humans and animals.

Thallasolituus oleivorans was isolated from sea water/sediment samples collected in the harbour of Milazzo, Sicily, Italy. A Gram-negative cell curved, vibrioid, occasionally screw-like morphology with characteristic monopolar, monotrichous flagella and was described that has been shown to play an important role in the biodegradation of crude oil (9). Until now, no isolation from animals has been described. In this study Thallasolituus oleivorans was isolated from the large intestine of alpacas.

Leeuwenhoekiella nanhaiensis pertaining to Nanhai from where the type strain was isolated (the Chinese name for the South China Sea). It is a Gram-negative, rod-shaped and yellow bacterium, motile by gliding and it was isolated from a water sample collected from the deep South China Sea (21). Until now, no isolation from animals has been described.

The MALDI-TOF identification is based on the spectral fingerprints which vary between microorganisms (compounds detected in the spectrum and specific molecular masses). The same species can give different mass spectra depending on growth conditions and extraction method used. Bacteria preparation for the study makes a huge impact on specificity and sensitivity. Most common systems of identifications based on biochemical reactions (VITEK2, API, RapID) has limitations due to focus on bacteria and yeasts presents in human medicine.

Moreover, if organisms are encountered frequently, MALDI-TOF can accurately identify most closely related species. However, MALDI-TOF is currently unable to differentiate E. coli from Shigella, Burkholderia mallei/pseudomallei and Burkholderia cepacia complex, Enterobacter cloacae complex, as well as the Mycobacterium tuberculosis complex (7, 8, 17, 18, 19).

Our research confirmed the results indicating that some MALDI-TOF devices are unable to distinguish Bacillus cereus from B acillus anthracis, however Manzulli et al. (23), showed slight differences in the pattern of the spectra of the B acillus cereus group, which in the future may contribute to the creation of an algorithm distinguishing these bacteria (21).

MALDI-TOF is useful tools for anaerobic bacteria identification (6). Conventional identification of anaerobes is mainly based on the detection of phenotypic characteristics, such as Gram staining, colony morphology, microscopic examination, differential growth on selective media and spore forming. Most of these conventional methods are arduous and time-consuming processes.

Another limitation may derive from lack of spectra in the database or when some members of a species complex are in the database, but others are not (8, 23). Our study found that there are still many bacteria not included in the databases currently available for VITEK2, API, RapID and MALDI-TOF systems. Manufacturers try to add newly detected spectra to database, however it is an ongoing process which needs further development (6). Developments of the databases to include an expanded number of new species and more robust mass spectra and biochemical profiles for new species will greatly improve the performance and utility of these systems for bacterial and yeast identification.

Therefore, using back up methods such as sequencing or VITEK2 can be highly effective for MALDI-TOF results confirmation. NGS analysis is most popular tool for microbiome identification and microorganism classification while the MALDI-TOF MS and biochemical systems are a popular commercial methods commonly used in clinical microbiology laboratories (6, 17). Gene sequence analysis has important limitation, detecting only genetic material, without information about liveness of identified microorganisms. Culture and biochemical identification are the “gold standard” for living bacterial and yeast. MALDI-TOF MS is showed to be a simple, rapid, accurate tool for identification of common and rare observed microorganisms (6). Identification of veterinary-specific microorganisms is limited and not common on all systems and restricts their usefulness in analyses of animals (23). MALDI-TOF is a quick method, which accurately identified most veterinary isolates of bacteria and yeast to the species level. Thus, MALDI-TOF constitutes a valuable diagnostic tool in the veterinary clinical laboratory (6).

The number of animals tested may seem to be a limitation in these study. However, obtaining autopsy material for microbiological tests from dozens of alpacas is impossible without deliberate slaughtering. Few animals go to the section in a fresh state, which allows the collection of intestinal material for microbiological tests. On the basis of these three cases, we showed the usefulness of the MALDI-TOF method in the study of the live intestinal microbiota of alpacas. This method can be used in the future to research the rapid identification of microbiota, which may be the basis for the diagnosis and treatment of diseases of the digestive system. This study demonstrates that both MALDI-TOF MS and popular systems based on biochemical reactions are powerful tools for fast, accurate identification of microorganisms isolated from alpacas. Moreover it indicated that the VITEK2 and API or RapID tests are accurate and inexpensive identification systems in comparison with MALDI -TOF. Especially in case of common bacteria such as E. coli, Staphylococcus, Enterococcus (23).

The microbiological methods for proper diagnosis and bacterial identification including underestimated opportunistic or rare pathogens, is still poorly studied in domestic animals. As these bacteria and fungi may represent a potential risk for human health further improvement of diagnostic methods and identification is required.