XENOFOOD—An Autoclaved Feed Supplement Containing Autoclavable Antimicrobial Peptides—Exerts Anticoccidial GI Activity, and Causes Bursa Enlargement, but Has No Detectable Harmful Effects in Broiler Cockerels despite In Vitro Detectable Cytotoxicity on LHM Cells

Entomopathogenic bacteria are obligate symbionts of entomopathogenic nematode (EPN) species. These bacteria biosynthesize and release non-ribosomal-templated hybrid peptides (NR-AMPs), with strong, and large-spectral antimicrobial potential, capable of inactivating pathogens belonging to different prokaryote, and eukaryote taxa. The cell-free conditioned culture media (CFCM) of Xenorhabdus budapestensis and X. szentirmaii efficiently inactivate poultry pathogens like Clostridium, Histomonas, and Eimeria. To learn whether a bio-preparation containing antimicrobial peptides of Xenorhabdus origin with accompanying (in vitro detectable) cytotoxic effects could be considered a safely applicable preventive feed supplement, we conducted a 42-day feeding experiment on freshly hatched broiler cockerels. XENOFOOD (containing autoclaved X. budapestensis, and X. szentirmaii cultures developed on chicken food) were consumed by the birds. The XENOFOOD exerted detectable gastrointestinal (GI) activity (reducing the numbers of the colony-forming Clostridium perfringens units in the lower jejunum. No animal was lost in the experiment. Neither the body weight, growth rate, feed-conversion ratio, nor organ-weight data differed between the control (C) and treated (T) groups, indicating that the XENOFOOD diet did not result in any detectable adverse effects. We suppose that the parameters indicating a moderate enlargement of bursas of Fabricius (average weight, size, and individual bursa/spleen weight-ratios) in the XENOFOOD-fed group must be an indirect indication that the bursa-controlled humoral immune system neutralized the cytotoxic ingredients of the XENOFOOD in the blood, not allowing to reach their critical cytotoxic concentration in the sensitive tissues.

oped on chicken food) were consumed by the birds. The XENOFOOD exerted detectable gastrointestinal (GI) activity (reducing the numbers of the colony-forming Clostridium perfringens units in the lower jejunum. No animal was lost in the experiment. Neither the body weight, growth rate, feed-conversion ratio, nor organ-weight data differed between the control (C) and treated (T) groups, indicating that the XENOFOOD diet did not result in any detectable adverse effects. That is in vitro toxic ingredients of the XENOFOOD seemed to be inactivated in vivo in chicken.
We suppose that the parameters indicating a moderate enlargement of bursas of Fabricius (average weight, size, and individual bursa/spleen weight-ratios) in the XENOFOODfed group must be an indirect indication that the bursa-controlled humoral immune system neutralized the cytotoxic ingredients of the XENOFOOD in the blood, not allowing to reach their critical cytotoxic concentration in the sensitive tissues.
This conclusion may have some significance concerning the application perspective. Whether the different NR-AMPs could ever be utilized as drugs not only against prokaryotic (bacteria) pathogens but eukaryotic (fungal pathogens, and parasitic protists) depends on the side effects. To get experimental experience with the option of applying EPB-produced antimicrobials to pathogens, and parasites of veterinary significance, we present here the results of an in vitro, and an accompanying in vivo study on chicken. In the in vitro study, we tested the cytotoxic potential of the cell-free conditioned culture media (CFCM) of three entomopathogenic bacterium species, -X. budapestensis, DSM16342 (EMA); X. szentirmaii DSM16338 (EMC); Photorhabdus luminescens ssp. akhurstii TT01-on chicken tissue culture cells, namely, on the Leghorn Male Hepatoma (LMH, see Materials and methods) Icells, (a permanent confluent hepato-carcinoma cell line). Each CFCM proved rather cytotoxic in this test. In the in vivo study, we fed freshly hatched male broiler chickens for 42 days with XENOFOOD [39] which contained autoclaved cultures of EMA, and EMC. These bacteria were grown on standard chicken (starter and grower) feed, and the whole culture was used as a "feed supplement". It had been known that these EPB species cannot grow at body temperature (above 33 • C).

Anti-Clostridial Activity of EMA and EMC In Vitro
Based on the previous results of repeated experiments, we choose X. budapestensis DSM16342 (EMA) and X. szentirmaii DSM 16338, (EMC) which had been identified in our laboratory, [34] and seem to be the most promising strains as antimicrobial producers we have ever come across so far. This statement also stands for their anti-Clostridial potential. For the exact description and history of these strains, see Ref. [35]. We retested the in vitro efficiency [36][37][38] in overlay bioassays and the pepsin resistance of the active ingredient that we previously published [39], (Figure 1a,b).
Both native and autoclaved CFCM samples of EMA and EMC bacterial strains were bioassayed both in liquid, and agar diffusion tests as described previously [36][37][38][39]. The anticoccidial activities of native and pepsin-treated CFCM samples were compared similarly as described previously [39]. All previous data on antimicrobial activity against each of the tested substances on different clinical isolates of Gram-positive Staphylococcus aureus [35][36][37][38][39], Ref. [42] was repeatedly confirmed, indicating the antimicrobial activity These data were especially important as confirming all previous findings published in [38]. The experiment also confirmed the thermostability and the pepsin durability of the CFCM of EMA, and EMC [40], since autoclaving of these CFCMs did not affect the antimicrobial activity against any of the tested organisms.

In Vitro Bioassays: Cytopathogenic Effects of Cell-Free Condition Media (CFCM) of 3 EPB Species on LMH Chicken Cell Monolayer
Two experiments were performed on this subject, (see Section 4). In both experiments, permanent, confluent monolayer culture samples (developed in different Falcon flasks) were used as target organisms, and different dilutions of different CFCMs were tested.
(As described in the Materials and Methods, confluent monolayers of LMH cells were parallelly developed in 18 Falcon flasks (for the first experiment), and 36 flasks (for the second experiment). The culture media were Medium RPMI 1640 supplemented with 10% FBS, penicillin G (200 IU/mL), and streptomycin (200 µg/mL). The incubation time was 72 h in both cases, in the controlled atmosphere of 5% CO 2 at 37 • C and around 85-90% humidity in 25 cm (2) flasks with filtered caps. After the layer formed, the liquid growth media was removed from all but 3 flasks (in the first experiment) and in 2 × 2 flasks (in the second experiment). Those few (3 × 1 and 1 × 2) flasks were referred to as "unchanged" control flasks. As for other flasks, 3 × 1 (in the first experiment) and 1 × 2 (in the second experiment) were refilled e fresh culture media (referred to as M199 + 15% FKS control flasks) while the rest of the flasks were refilled 100 v/v% of different CFCMs (EMA, EMC, TT01 yellow and TT01 red, in the first experiment, and different dilutions of EMA and EMC CFCMs in the second experiments. As for the replicates, in the first experiment we worked with triplicates, and in the second experiment with duplicates).

Experiment-1
The aim was to compare the cytopathogenic effects of 4 different undiluted CFCMs obtained from cultures of EMA, EMC, and TT01 yellow, and TT01 red colony color variants, respectively, (see Section 4). It was found that each CFCM exerted destructive effects on the cell layer. The initial lesions on the LMH monolayer were followed by more serious ones consisting of clumps attached to the monolayer, (Score 3).
The results (data given as empiric score values, (see Section 4) of the Experiment are presented in Table 1. We interpret these results as experimental proof of the presence of strong cytotoxic molecules in each of the 4 CFCMs. The four CFCMs differed in their cytotoxicity, the order of toxicity was: EMA > EMC > T0T1. Considering that the CFCMs filtrates contain not only AMPs but toxin protein molecules as well, we should know whether the experienced cytotoxicity was (a) a side effect(s) of the NR-AMP(s), or/and the toxin(s), and our results could not answer this question. Another question that remained open is whether the absence of any sign of in vivo toxicity was due to the denaturation of the toxic (protein-like) ingredients of the denatured CFCMs. We saw the antimicrobial activity of the CFCMs did not disappear after being autoclaved. There are the following experimental conclusions that can be drawn: (1) These experiments should be repeated with the peptide-rich fraction [38] and/or separated single NR-AMP molecules like (first of all) the fabclavines [43][44][45][46].

Experiment-2
The aim was to compare the cytopathogenic effects of serial dilutions of the CFCM filtrates of EMA, and EMC cultures, considering that both EMA and EMC CFCMs showed much stronger antimicrobial activity than the two TT01 CFCMs. The stock solutions, (and serial dilutions of them) used in this experiment were: EMA 60%, and EMC 80%, respectively. Confluent monolayer from LMH cells was developed again in Medium RPMI 1640 supplemented with 10% FBS, penicillin G (200 IU/mL), and streptomycin (200 µg/mL) for 72 h in a controlled atmosphere of 5% CO 2 at 37 • C and around 85-90% humidity in 25 cm 2 Falcon flasks with filtered caps (see also in Materials, and Methods). The same score system [47] (Amin, 2012, see Table S1) was used for evaluating the results. The results of Experiment 2 are presented in Figures 2 and 3, and the data and statistics are given in Supplementary Material (Tables S2 and S3).  [48] cell monolayer caused by serially diluted EMA CFCM samples. The scoring system is [47] shown in Table S1. (Abbreviations: EMA= Xenorhabdus budapestensis DSM16342(T) (Lengyel). CFCM= cell-free conditioned culture media). Rough data are given in Supplementary Table S3.) Above 1.5 v/v% of the EMA CFCM concentrations (that is below 1.5 v/v dilutions) the differences between the score values representing the degree of cytotoxicity differed significantly in a dose-dependent manner, at least at lower dilutions, (<8 v/v%), where the cytotoxic damages seem to be regenerating. The degree of significance was detectable at 0.001 < p < 0.005 level at each consecutive measurement between the following values: At  The experimental conclusion is that the cytotoxic effect of the CFCM of the EMA was not only severe but unambiguously dose-dependent throughout the experiments. The histological pictures of replicates seemed practically uniform.
The differences between the values represented by the 1st (orange) columns and the rest of them proved significant at (0.001 < p < 0.005 level) at each consecutive measurement. The replicate histological pictures seemed practically uniform, and below 12 v/v% the cytotoxic effect of the CFCM of the EMC was rather moderate, and not dose-dependent.
We interpret these results as experimental proof that the in vivo application either of EMA NR_AMP peptides or the EMC NR_AMP peptides did cause fatal or even detectable cytopathogenic in sensitive tissues (like liver cells) in a dose-dependent manner. The degree of destruction was serious and irreversible above 3.2 v/v% (1:25 dilution) EMA-CFCM and above 2.4 v/v% EMC. The experimental conclusion is that in each of both cases, there was a threshold concentration below which no cytotoxic destruction is observable. Consequently, if an antimicrobial compound with cytotoxic side effects (like an NR-AMP) had been orally administered in a higher dose, which is needed for its antimicrobial efficacy in the gut, but then reached its critical cytotoxic threshold concentration neither in the blood nor in any sensitives tissue (because one or another reason), then in vivo toxicity would not be detected.  [48] cell monolayer caused by serially diluted EMC CFCM samples. The scoring system is [47] shown in Table S1. (Abbreviations: EMC = Xenorhabdus budapestensis DSM16338(T) (Lengyel). CFCM = cell-free conditioned culture media). Data and statistics are given in Supplementary Table S2. EMC CFCM seems to be less cytotoxic than EMA CFCM. Above 12 v/v% of the EMC CFCM concentrations (that is below 12 v/v dilutions) the differences between the score values representing the degree of cytotoxicity did not differ significantly in a dose-dependent manner, at least at lower dilutions, (<12 v/v%), where the cytotoxic damages seem to be partially regenerating. The degree of significance was detectable at 0.001 < p < 0.005 level at each consecutive measurement between the following values:

Gastrointestinal Activity of XENOFOOD
To examine whether the gastrointestinal potential of the XENOFOOD was considerable we counted Clostridium CFUs in the lower ileum (see Figure 4) of XENOFOOD-fed and untreated animals.
Using ANOVA statistical analysis, we compared the average of the lower ileal Clostridium CFUs (see Figure 2) of the XENOFOOD-fed animals (48.1) with those of the controls (149.9). Although we found that despite the 3X difference, only a slightly significant could be demonstrated, probably due to the large standard deviation in the control values (Table 2). and the very low infestation rate in the animals chosen for the experiment during the 42-day long experiment. In other terms, only very few animals had been infected with Clostridium. concentration neither in the blood nor in any sensitives tissue (because one or another reason), then in vivo toxicity would not be detected.

Gastrointestinal Activity of XENOFOOD
To examine whether the gastrointestinal potential of the XENOFOOD was considerable we counted Clostridium CFUs in the lower ileum (see Figure 4) of XENOFOOD-fed and untreated animals.    We found that despite the 3X difference, only a slightly significant difference could be demonstrated, probably due to the large standard deviation in the control values (Table 3), and because of the very low infestation rate in the animals chosen for the experiment. Table 3. Absolute, and relative weights of immune organs (bursa of Fabricius, and spleens of XENOFOOD-fed (T) and Control (C) chicken. The data show that the average weights of the spleens were smaller, while the average measurements of the sizes of the bursas were larger in the XENOFOOD-fed group, and the differences were significant. The same can be stated about the Clostridium colonyforming units, (CFUs): there were fewer in the XENOFOOD-fed group (on average). The XENOFOOD exerted unambiguously detectable gastrointestinal antimicrobial activity. But the original infestation of the population from which the experimental animals were selected rate was very low. These results allow concluding that XENOFOOD feeding was harmless. We suppose that the bursa-dependent humoral immune system must have neutralized the cytotoxic peptide-like molecules (including the GI-active NR-APMPs) of the XENOFOOD in the blood, protecting the sensitive tissues and organs in such a way.

Growth Rate, Feed Consumption, Feed Conversion, and the Post-Mortal Data
During the 42-day-long experiment, no animal was lost. We did not detect any significant differences either in the feed intake, body weight gain, and fed conversion ratio (FCR) between the control and XENOFOOD-fed groups ( Table 4).
As for the organ weight, they followed the allometric rules, according to which the regular and systematic patterns of growth such as the mass or size of any organ or part of a body can be expressed about the total mass or size of the entire organism according to the allometric equation: Y = bxα, where Y = mass of the organ, x = mass of the organism, α = growth coefficient of the organ, and b = a constant (only bursa-and spleen data are given). This fact itself excludes any organ-development abnormality of the treated (XENOFOOD-fed) animals. But the larger bursa weight and the smaller spleen weights, as well as larger bursa/spleen ratios in the XENOFOOD-fed groups may indicate indirectly an intensified humoral immune activity, leading to the neutralization of the cytotoxic AMP-components in the blood (Table 4).
Summarizing the results of the in vivo feeding experiment, their interpretation is as follows: we wanted to learn two things: (1) whether the per os administered XENO-FOOD (1) has a preventive potential against Clostridium infection; (2) can be considered as a harmless or as a risky food supplement. The experimental conclusions are, that (1) XENOFOOD must have a preventive potential but the experiment should be repeated with experimentally infested birds, where infection inoculum is strong and standardized.
(2) The XENOFOOD proved a harmless food supplement, and despite the in vitro cytotoxicity of its CFCM ingredients, did not cause any detectable adverse effect in the animals of the XENOFOOD-fed (T) groups of the birds.

Discussion
When considering the potential benefits of studying and discovering AMPs of EPN/EPB symbioses we must remind of the natural role of these compounds. There are several biosynthetic (NR-AMPs) produced by Xenorhabdus species. They are end-products of nonribosomal peptide synthetases (NRPSs) which usually use terminal reductase domains for 2-electron reduction of the enzyme-bound thioester releasing the generated peptides as C-terminal aldehydes [49]. Each EPN/EPB symbiotic association has been, existing in a given environmental milieu, as a coevolutionary product selected by/in a respective polyxenic environment (insect cadaver in the given soil). The biological role of each NR-AMP molecule is a contribution to providing pathobiom) conditions for a given symbiotic complex [32]. We can give only a short list of those NR_AMPs discovered in our Xenorhabdus bacteria so far without having information about their modes of action. Structure elucidation of natural products including the absolute configuration is a complex task that involves different analytical methods like mass spectrometry, NMR spectroscopy, and chemical derivation, which are usually performed after the isolation of the compound of interest [50]. As for X. budapestensis DSM16342T(Nov) (EMA), in our previous experiments, we found an unusual arginine-rich hexapeptide bicornutin [37] which we mistakenly believed to be the active ingredient of the CFCM of EMA species, because it has always been present in the antimicrobial-active CFCM [37]. Our results drew the attention of Professor H. Bode and his associates. Continuing our search for strong AMPs in our EMA Fuchs and his associates (2012), proved that bicornutin exerted neither bacteriostatic nor bactericide activity, however [43], but this work initiated further search partly toward this strange molecule (studying the neutral loss fragmentation-pattern-based screening for arginine-rich natural products. They extended their studies on 16 different Xenorhabdus and Photorhabdus strains) and toward the substantive antibiotic component of the CFCM (the fabclavine). The bicornutin research led to the discovery of a novel group of related natural products comprising 25 different arginine-rich peptides. These molecules were identified due to their characteristic neutral loss fragmentation pattern, and the structures of eight of these compounds were elucidated. Two biosynthesis gene clusters encoding non-ribosomal peptide synthetases were also identified, emphasizing the possibility to identify a group of structurally and biosynthetically related natural products, [43]. The research was expanded by the same fellow scientists, leading to the discovery of the fabclavine [44] which seems to be the most significant AMP in the genus Xenorhabdus, and EMA and EMC proved the most efficient fabclavine-producers. Fabclavins, which were discovered in EMA and EMC seem to be the most important NR_AMP molecular family in the genus Xenorhabdus [45,46,51]. EMA and EMC seem to be the strongest fabclavine producers [35]. Surprisingly enough X. budapestensis which has been discovered in samples of Central European (Hungarian) soil has also been found in two different locations in China, and produce another type of NR-AMPs, two novel cyclic depsipeptides Xenematides F and G. but no one tested for fabclavine production this new isolate [52,53]. As for X. szentirmaii, it produces xenofuranones A and B [54], and phenazines [55], and szentiamide, [56,57].
The coexistence of so many antimicrobial compounds with large target spectra and cooperating potential comprises a potentially promising toolkit combatting MDR because antibiotic-resistant bacteria perform a high frequency of collateral sensitivity to antimicrobial peptides, [13,16,17]. X. budapestensis (EMA) and X. szentirmaii (EMC) are abundant sources of NR-AMP molecules of large target spectra and strong anti-pathogen potential, [35].
Each NR-AMP molecule in the Xenorhabdus CFCM is the end-product of a biosynthetic pathway. Each biosynthetic pathway is encoded by genes clustered in the respective operon (biosynthetic gene complex, BCG). In EMC 71 BCGs have been identified [58]. Their large spectral NR-AMP products of overlapping activities and cooperating potential are coexisting as ingredients of the CFCM. Theoretically, there are two alternative ways of research philosophy for the geneticist of agricultural or veterinary commitments.
One option is to focus on searching for individual molecules as new potential drugs, benefitting the global (co)regulation of the biosynthesis of different NR_AMPs, searching for or constructing EPB strains synthesizing only single NR-AMP molecules. As for this alternative, the recent discovery, according to which the different NRPS-encoding operons are under the global control of the post-transcriptional regulator, Hfq, provides an option for constructing Xenorhabdus strains producing single NR-AMP only [41]. It is a great challenge! Another strategy is to try to benefit from the co-existence of naturally cooperating NR_AMP molecules in the CFCMs by producing harmless but effective preventive biopreparations like food supplements either as food supplements for veterinary, or soil additives for plant protection goals against pathogens and parasites. This latter option may be discussed here in light of the experiment's results where chickens were fed by XENOFOOD, a food supplement, in which in vitro active ut heatsterilized NR-AMPs were present as ingredients. We know about these compounds that they were of peptides [28], in vitro anti-coccidial active [39], which did not lose their antimicrobial potential either when heat-sterilized [36], or when subjected to pepsin digestion [39].
But in the in vitro experiment we presented here we found that each EPB CCM we tested exerted strong cytopathogenic effects on LMH permanent confluent chicken cells. We do not know whether one single (most probably the fabclavine) [43][44][45][46] molecule or more than one molecule can be taken as responsible for the cyto-pathogenecity. There are several candidates (see Table 5).
It is not known whether the mechanism of the antimicrobial action of a given NR-AMP molecule is the same as those causing cytopathogenicity in the organisms.
The XENOFOOD food supplement proved in this experiment harmless and GI active against a bacterium pathogen, Clostridium perfringens.
However, in our in vivo feeding experiment the lower number of Clostridia does not automatically indicate a "protection" exerted by the XENOFOOD, since the natural infestation rate of the experimental animals was very low, making conclusions on the preventive potential of the XENOFOOD just hypothetical, probable, but not certain. Another plausible explanation however for the strong differences between the average CFU values between the control (C) and the XENOFOOD-fed (T) group cannot be given. The moderate but detectable anti-Clostridial effect without any detectable harmful side effects may be explained in more ways than one. An explanation may come from the previous experiments on Eimeria (unpublished) that the effective cytotoxic concentration is much higher than the effective anti-clostridial dose.
An alternative interpretation is that the actual local NP-AMR concentration (bioavailability) was much lower in the blood than in the GI system. Consequently, the local critical threshold cytotoxic concentration of the toxic compound(s) in other organs and tissues was not reached.
We suppose that the two alternatives cannot exclude each other but that the second option is more plausible because the immune system of the birds must be capable of neutralizing peptide-like compounds, either the toxins or the NR-AMPs or both. The higher average weights of the bursas in the T-groups suggest that humoral immunity is mainly responsible for that.
Perspectives are discussed in the Section 5 "Conclusions".
In detail, cells were inoculated into 25 cm (2) flasks with filtered caps (Sarstedt) containing an end volume of 7 mL culture and incubated in a controlled atmosphere of 5% CO2 at 37 • C and around 85-90% humidity. After 72 h of incubation, a confluent monolayer of LMH cells was obtained per flask.
These antibiotic-producing bacteria were freshly taken from frozen cultures and grown on the bacterial species grown on MacConkey agar plates before they were transferred to the liquid medium as described before, [35][36][37][38][39][40]. The bacterial species were grown on Mac-Conkey agar plates before they were transferred to the liquid medium, and unexpectedly two different types of colonies for TT01 were observed on the agar plates: red-brown colored colonies which adsorbed the neutral red from the MacConkey agar and yellow-colored colonies which did not so. Both types of colonies were tested for the effect of cell-free filtrates on LMH monolayers. In this particular experiment, the antibiotic-producing bacteria were cultured at 28 • C. The bacteria were incubated for 65 h at 30 • C in a shaker (225 rpm. Bacterial cultures were then centrifuged at 3300× g for 5 min and then the supernatants from the cultures were filtered through 0.22 µm cellulose acetate filters (Millipore).
Experimental design: From all but 3 of the Falcon flasks (with the 72-h old LMH-layers), the culture medium of the LMH monolayers was removed from the flasks and replaced by the 4 CFCMs. Each of the CFCMs was tested in triplicates. There were two sets of controls. There were also 3 flasks with fresh medium (fresh Medium 199 supplemented with 15% FBS) without CFCM, and another 3 with the original, ("unchanged", that is 72 h old) culture media. All these 4 × 3 experimental and 2 × 3 control flasks were incubated for another 72 h. The cultures were incubated in a controlled atmosphere of 5% CO 2 at 37 • C and around 85-90% humidity.
After 72 h of incubation, a confluent monolayer of LMH cells was obtained per flask. All but the so-called "Unchanged culture flasks" LMH monolayers, the culture medium of the LMH monolayers was removed from the flasks and replaced by something. The control flasks were refilled with fresh media (Medium RPMI 1640 (Invitrogen/GIBCO) supplemented with 10% heat-inactivated fetal bovine serum FBS (Invitrogen/GIBCO), penicillin G (200 IU/mL) and streptomycin (200 µg/mL). The experimental flasks were refilled with cell-free (centrifuged and filtered serially diluted stock solutions (EMA: stock solution: 60 v/v; EMC stock solution 80 v/v%). The dilutions were 1:2.5; 1:5; 1:10 1:.25; 1:.50; 1:75, and 1:100, respectively. The EPB cells had been also cultured in Medium RPMI 1640 supplemented with 10% FBS, penicillin G, without streptomycin, (Xenorhabdus are penicillin-resistant). Each of the different filtrate analyses and the controls was performed in duplicate. The cultures were incubated in a controlled atmosphere of 5% CO 2 at 37 • C and around 85-90% humidity.
Each monolayer was investigated visually by an inverted light microscope to detect the effect of the cell-free filtrates on LMH monolayers. According to the degree of monolayer destruction, the following scoring system was established and also published [47], (See Supplementary Material, Table S1).
XENOFOOD preparation: XENOFOOD contained 5% soy-meal, which had been suspended with an equal amount (w/w) of EMA and another 5% suspended in an equal amount (w/w) of EMC cells obtained from 5 days-old shaken (2000 rpm) liquid cultures by high-speed (Sorwall; for 30 min) centrifugation. The liquid cultures were in 2XLB; (DIFCO); supplemented with meat extract equivalent amount to the yeast extract. The 5 days had been proven optimal for antimicrobial substance production at 25 • C in these conditions). It had previously been discovered that both EMA and EMC grow and produce antimicrobial substances in autoclaved soy meal containing some water and yeast extract or autoclaved yeast, (in 0.5 w/w%). Therefore, the original chicken food served as a semi-solid culture media of Xenorhabdus cells. Both EMA and EMC culturing semi-solid chicken food, which was prepared daily, and have been incubated in sterile conditions for another five days; then the EMA and EMC culturing media were united; autoclaved (20 min, 121 • C), and then dried by heat overnight. The Xenorhabdus cells were killed in such a way, while the heat stabile [36] antimicrobial compounds remained active.
XENOFOOD in vivo feeding experiment: Experimental animals: One-day-old male broiler chickens (N = 2 × 34 = 68) were equally distributed into two groups: Control (C) and Treated (T) groups. The latter was fed with XENOFOOD. The C group was kept on a normal starter (1-10-d), grower (11-24-d), and finisher (25-42-d) Tables S4 and  S5. During the rearing period feed intake, growth rate, and feed conversion of animals were measured on a pen basis for the starter, grower, and finisher phases.
Dissection, post-mortal data: Not all but a sample (N = 2 × 10 = 20) of 42-day-old birds were dissected on the 42nd day and are presented here. We dissected a sample of 10 birds from the XENOFOOD-fed and a sample of 10 birds from the Control groups. That is, not all but samples (N = 2 × 10) of the 2 × 34) from 42-day-old birds were dissected to get post-mortem data about a few body organs as well as about the number of Clostridium germs in their ilea. The body weights of these selected animals did not differ from the average of their respective (C, or T) experimental groups. After dissection, the weights of the different organs were measured.
Statistical Analysis ANOVA procedure was carried out by using the respective propositions of the SAS 9.4, carried out partly by Sándor Józsa and partly by András Fodor. We considered our data as an unbalanced data set. The significant differences (α = 0.05) between treatment means were assessed using the Least Significant Difference (LSD). In detail, the data analysis was performed STATJ software, Version [9.4)  We used ANOVA and GLM Procedures following the requirements of the SAS 9.4 Software. The design of the experiment was a randomized complete block design with, the number of the respective treatments, concentrations, and replicates Data have been averaged to allow the analysis of variance (ANOVA). The significance of differences of the means (a. = 0.05) was determined by using t (Least Significant Difference LSD) tests or Duncan's Multiple Range Tests, depending on the respective experiment.

1.
The work presented here can be considered as a contribution to the field of application of biosynthetic non-ribosomal encoded antimicrobial peptides (NR-AMPs) controlling multi-drug resistance (MDR).

2.
Regarding the future potential application of natural compounds, several aspects have to be taken into consideration: (A) the antimicrobial potential; (B) the durability, thermotolerance, and bioavailability; (C) cytotoxicity, and another, unwanted side effects, especially if one group of target organisms are of eukaryotic (protist) pathogens.

3.
We have not found any publication concerning the comparisons of the in vitro and the in vivo effects of antimicrobial active NR-AMP molecules of EPB origin on the same eukaryote (host) organism to be protected from pathogens or parasites.

4.
Herein we demonstrated that both the heat-inactivated and the native cell-free conditioned culture media (CFCM) of Xenorhabdus budapestensis DSM16342 (EMA), and that of Xenorhabdus szentirmaii DSM16338 (EMC) which has broad-spectrum antimicrobial activity against several prokaryotic and eukaryotic microbial pathogens, also exerted a robust cytopathogenic effect on confluent (cultured) chicken cell layer.

5.
Whether cytotoxicity and antibiotic potential of the same NR-AMP molecule is an open question. 6.
The results of the in vivo experiment presented in Tables 2-4 may allow concluding that XENOFOOD feeding was harmless. It means that the NR-AMPs, present in the CFCM of X. budapestensis and X. szentirmaii-at least when applied as autoclaved food ingredients-did not cause any detectable side effects in vivo in broilers. 7.
The incorporation (by covalent chemical bounds, or complex formation) of non-decomposed NR-AMPs into proteins or other tissue components is biochemical nonsense. 8.
However, their deposition into the egg yolk cannot theoretically be ruled out, therefore, we may recommend XENOFOOD as protecting food supplement only for cockerels but not for pullets. 9.
The XENOFOOD exerted gastrointestinal antimicrobial activity in the GI has been reflected in the reduced average value of the Clostridium CFUs in the XENOFOODfed group. 10. We suppose that the bursa-dependent humoral immune system neutralized the cytotoxic side components (of peptide nature) of the XENOFOOD, (that is, cytotoxic NR-AMPs of CFCMs in the blood (but, naturally, not in the gut). 11. The presence of the active antimicrobial compounds in the GI is proved by the significantly reduced number of colony-forming germs in the lower jejunum of the XENOFOOD-fed (T-group of) animals compared to that of the controls, (C-group). Based on those facts, we may recommend XENOFOOD as a useful food supplement to protect broiler cockerels against coccidiosis. 12. However, before drawing that conclusion, we recommend another experiment on experimentally Clostridium-pre-infested chickens, (also on cockerels, not pullet) to reproducibly determine the in vivo anti-Clostridial effects of the XENOFOOD diet quantitatively. (As mentioned above, we do not recommend the use of pullets, because the deposition of the XENOFOOD ingredients is unpredictable).
We have been looking for a cooperating partner for carrying on this final experiment, and this is one of our strongest arguments for publishing these data in Pathogens.

Acknowledgments:
In Budapest, at the Department of Microbiology and Infectious Diseases, University of Veterinary Medicine, Budapest, Hungary, (where coauthor László Makrai works we would like to express our appreciation and thanks to László Fodor, who was the the Head of the Department at that time, for all his professional, financial and moral help what was essential to carry on the in vitro Clostridium, and provided the official protocol for in vivo feeding tests. We would like to express thanks and appreciation for their professional technical help to the lab technicians for taking care of the bacterium strains for long years, and their technical help; especially to Teréz Halasi, Éva Kolozsvári, and Alexandra Jutasi. In Vienna, at the Clinic for Poultry and Fish Medicine, Department for Farm Animals and Veterinary Public Health, University of Veterinary Medicine (Vetmeduni), Vienna), Austria; we would like to express thanks and appreciation not only for their professional technical but the intellectual help to the whole research team of the Hess laboratory, especially to Aziza Amin, and Irina Prokofieva for running the in vitro cytopathogenicity experiments, on the LMH cells. Also in Austria, we have to thank Klaus Teichmann and Eva Novák (Biomin, Tulln, Austria), for testing the same CFCM samples on Eimeria cells, repeating our cytotoxicity tests confirming our findings on chicken cells, and sharing their conclusions with us unselfishly, declaring that the effective anti-Eimeria was an order of magnitude lower than the cytotoxic dose. In Keszthely, at the Institute of Physiology and Nutrition, Georgikon Campus of Hungarian University of Agriculture and Life Sciences, (MATE), Hungary, where the in vivo feeding experiments were in fact going on, we would like to express thanks and appreciation not only for their professional technical but the intellectual help to the whole research team of the Dublecz team, especially for the technical help of Judit Varga, and Cseh, for her professional and administrative help Angéla Vörös-Mezei. We want to express our thanks and appreciation to Csaba Pintér (Institute of Plant Pathology, Georgikon Campus, the Hungarian University of Agriculture and Life Sciences, (MATE), Deák Ferenc utca 16, H-8360, Keszthely, Hungary; pintercsaba45@gmail.com) for the high-standard photography. Finally, the first author, A. Fodor would like to thank each of the coauthors as well as his wife Andrea Máthé-Fodor, for their professional, and intellectual, but for emotional-psychological support.