Exploring the Microbial Diversity and Composition of Three Cigar Product Categories

Abstract

Cigars and cigarillos are emerging as popular tobacco alternatives to cigarettes. However, these products may be equally harmful to human health than cigarettes and are associated with similar adverse health effects. We used 16S rRNA gene amplicon sequencing to extensively characterize the microbial diversity and investigate differences in microbial composition across 23 different products representing three different cigar product categories: filtered cigar, cigarillo, and large cigar. High throughput sequencing of the V4 hypervariable region of the 16 s rRNA gene revealed 2124 Operational Taxonomic Units (OTUs). Our findings showed that the three categories of cigars differed significantly in observed richness and Shannon diversity, with filtered cigars exhibiting lower diversity compared to large cigars and cigarillos. We also found a shared and unique microbiota among different product types. Firmicutes was the most abundant phylum in all product categories, followed by Actinobacteria. Among the 16 genera shared across all product types were Bacillus, Staphylococcus, Pseudomonas, and Pantoea. Nine genera were exclusively shared by large cigars and cigarillos and an additional thirteen genera were exclusive to filtered cigars. Analysis of individual cigar products showed consistent microbial composition across replicates for most large cigars and cigarillos while filtered cigars showed more inter-product variability. These findings provide important insights into the microbial diversity of the different cigar product types.

Introduction

Cigars in all product categories are emerging as popular tobacco products due to multiple factors that include lower cost as a result of the lower taxation rate when compared to cigarettes and the addition of flavors to cigar products [1, 2]. In recent years, cigars have become the most reported combustible tobacco product used by youth [3, 4]. A study conducted in the United States in 2020 of almost one million middle and high school students that self-reported having smoked cigars in the past 30 days, showed that the largest proportion of students reported using cigarillos (44.1%), followed by traditional cigars (33.1%), and filtered cigars (22.6%) [5, 6]. In this study, we focus on three different types of cigars, namely large (traditional) cigars, filtered cigars, and cigarillos. Large cigars come in a range of sizes and are made up of tobacco wrapped in leaf tobacco or homogenized tobacco leaf (HTL) material. Cigarillos are similar to large cigars but may be thinner and longer than a normal cigarette, while filtered cigars resemble cigarettes but are wrapped with HTL [7]. Cigar products are manufactured using specific varieties of tobacco which typically undergo a fermentation process that impacts the microbial population [8]. Additionally, the commercially available products used in this study use homogenized tobacco leaves for wrapper and binders, where applicable, as not all products contain both wrapper and binder. The products were expected to have differences as a result of the blend and ratio of the variety of cigar tobacco used to make the filler, the region of tobacco growth impacts the quality, and manufacturing processes. Contrary to popular belief, cigars are not less toxic than cigarettes but are associated with the same adverse health effects, including addiction to nicotine, oral lesions, oral cancer, lung cancer, cardiovascular disease, and chronic obstructive pulmonary disease [9–12]. Some studies have suggested that cigar products may be more harmful to human health than cigarettes [13–15]. Tobacco-specific nitrosamines (TSNAs) are a class of harmful compounds known to be carcinogenic and the formation of TSNAs has been shown to correlate with microbial metabolic activities [16, 17]. By studying microbial communities in cigars, organisms and pathways relating to the formation and possible reduction of TSNAs in tobacco products may be unveiled [18].

In this study, we use 16S rRNA gene amplicon to characterize the bacterial communities of cigar products in multiple product categories. 16S rRNA gene amplicon study is useful when examining multiple taxonomic domains simultaneously, especially when some of the microbes are not visually different and cannot be cultured in media [19]. Similar culture-independent methods have been used to investigate the microbial diversity of tobacco products, which revealed increased diversity compared to traditional culture-based approaches [18, 20]. We utilized the MiSeq platform to sequence the V4 region of samples representative of multiple cigar product categories which then allowed us to compare the similarities and differences between product types and analyze the associated microbial communities. The objective of the study is to extensively characterize microbial populations in cigar products that will fill a research gap and provide data to assess the potential health impact of cigars, providing important insights for the development of more effective tobacco control policies and public health interventions.

Materials and Methods

Sample Collection and DNA Extraction

A total of 23 different cigar products were obtained from online vendors and commercial distributors, including four research cigars (1C1, 1C2, 1C3, and 1C4) from the Center for Tobacco Reference Products (CTRP) at Kentucky Tobacco Research and Development Center (University of Kentucky, USA). The commercially available products were selected based on market share and sales within the United States to provide a comprehensive overview of the microbial compositions. The products were categorized into three groups: large standard cigars, filtered cigars, and cigarillos, consisting of five, six, and eight brands, respectively (Table 1). Cigars were randomly selected from the commercial packaging. Filtered cigars were selected from three packs within a carton, and 3 filtered cigars were randomly selected from various locations within each pack. Cigarillos were selected from three foil packs within an upright containing 15 foil packs, the industrial standard for packaging of cigarillos, and only one of the two cigarillos was used for analysis. The large cigars were randomly selected from a box of cigars containing up to 50 individually wrapped cigars. For all products tested, all tobacco material was homogenized prior to DNA extraction. DNA extraction was performed on freshly opened packages of all products using the ZymoBIOMICS™ DNA miniprep D4300 kit (Zymo Research, Irvine, CA, USA) and following the manufacturer’s recommended protocol. The extracted DNA was quantified using QUBIT 4.0. To minimize variability, we performed three technical replicates for each biological replicate, and three biological replicates were taken. The technical replicates were combined prior to sequencing to ensure that the data presented in the figures represent the averaged result of the technical replicates, thereby reducing technical variability. To ensure sterility, each product was opened under sterile conditions, homogenized in sterile saline solution (0.85%) using a bag mixer, filtered, and processed according to the kit instructions.

Table 1.

Tabulation of the different cigar products used with their name and category

Large cigars	Filtered cigars	Cigarillos
1C1	1C2	1C3
1C4	Phillies	Phillies Black
William Penn	Captain Black	Phillies Sweet
Garcia y Vega	King Edward	White Owl black
White Owl	Talon	White Owl silver
Swisher Sweets	Cheyenne	Black & Mild
Dutch Masters	Djarum	Pom Pom sweet
		Dutch Masters
		Swisher Sweets

Sequencing and Analysis of 16S rRNA Gene Sequences

We shipped the DNA samples from cigar products to The University of Michigan, Microbial Systems Molecular Biology Laboratory core sequencing facility (http://microbe.med.umich.edu/services/microbial-community-analysis) for PCR amplification and sequencing of the V4 region of the 16S rRNA gene on the Illumina MiSeq platform (dual-barcoded, paired-end reads, 2 × 250 flow cell) according to Kozich, et al. [21]. Sequence data from MiSeq sequencing was processed using MOTHUR software (v1.48.5) following the MiSeq SOP (https://www.mothur.org/wiki/MiSeq SOP, accessed January 2023) [22]. We followed the analysis methodology previously published by Law et al., 2020, using the SILVA reference alignment of the MOTHUR-formatted version of the RDP training set (SSU Silva 138 v.18) for classification [23, 24]. The cigar data set was subsampled and normalized to 3436 sequences per sample, which resulted in 2124 OTUs (operational taxonomic units) after classification at the 0.03 cutoff level. Raw sequence reads for all samples in this study were uploaded to the NCBI Bio Project database under accession number PRJNA1073920.

Statistical analysis was performed using the built-in functions in MOTHUR [22]. We compared bacterial community structure using the analysis of similarities function (ANOSIM) [25]. Observed Richness and Shannon’s diversity were calculated in MOTHUR and imported into the R program to compare alpha diversity measures and abundance using the Kruskal–Wallis rank sum test, with any significant results further tested by Dunn’s test [26]. The package ggplot2 was used to plot the data generated from MOTHUR [27]. The Venn diagram was made using VENNY 2.1 [28].

Results

Sequencing Dataset and Diversity

In this study, we extracted DNA from 69 different samples of cigar and cigarillo products, resulting in a total of 2,308,008 raw sequencing reads. After filtering out PCR and sequencing errors, 2,147,392 reads remained, leaving us with an error rate of 0.95%. We used Good’s coverage estimator to confirm that all samples had sufficient sampling depth, with values greater than 98.84%. We then calculated alpha diversity, which is based on the number of species within a sample, and measures for richness and Shannon diversity. We then compared them across cigar types using the non-parametric Kruskal–Wallis rank sum test. Our results indicate that the three categories of cigars (large, filtered, cigarillos) differ significantly in observed richness and Shannon diversity. Specifically, filtered cigars showed lower observed richness and Shannon diversity compared to large cigars and cigarillos, with large cigars having the highest measure of alpha diversity (Fig. 1). The Kruskal–Wallis rank sum test revealed a chi-squared value of 10.02 with a p-value of 0.01 for Shannon diversity. Using the Dunn test with Holm correction, we found a significant reduction in Shannon diversity in filtered cigars compared to both large cigars and cigarillos. Additionally, we observed that the microbial populations of filtered cigars were statistically different from those of cigarillos (p = 0.0388) and large cigars (p = 0.0026), while the microbial populations of cigarillos and large cigars were not statistically different.

Fig. 1.

Alpha diversity (a) Observed richness (b) Shannon diversity by product category (Lines connecting box plots indicate significant difference using the Dunn test. (** P < 0.01; *P < 0.05). Filtered cigars showed lower observed richness and Shannon diversity compared to large cigars and cigarillos, with large cigars having the highest measure of alpha diversity

Furthermore, we employed Principal Coordinate Analysis (PCoA) to assess the dissimilarity of various cigar samples based on Bray–Curtis distances plotted as the results in Fig. 2. Subsequently, we subjected the data to the ANOSIM test, which yielded an R statistic value of 0.193, and the significance level was set at 0.001. These findings indicate the presence of significant beta diversity, which compares the diversity across samples, among the three cigar categories. Beta diversity analysis is crucial for comparing the diversity between different environments or groups, even when raw materials, formulations, and processes differ. It highlights differences in the presence or absence and abundance of microbial species among cigar types. This approach allows us to understand the distinct microbial communities associated with each cigar product category. When comparing filtered cigars with cigarillos, the computed R-value was 0.29 at a significance level of 0.01, indicating a substantial dissimilarity between the two groups. Similarly, the comparison between filtered cigars and large cigars also exhibited a significant level of dissimilarity. However, in contrast, the comparison between large cigars and cigarillos did not reach a level of statistical significance (Table 2).

Fig. 2.

Principal component analysis (PCoA) of Bray–Curtis distances between cigar samples of different product categories. The ellipse is drawn at a 90% confidence level. PCoA analysis indicated significant differences in beta diversity measures between the three cigar categories

Table 2.

Analysis of Similarities (ANOSIM) test with distance and category of product. Pairwise comparison using Mothur for each product type. Experiment-wise error rate: 0.05, pairwise error rate (Bonferroni): 0.0167. Microbial populations of filtered cigars are statistically different from those of cigarillos and large cigars

Group Comparison	R-value	P-value
Cigarillos-Filter-Large	0.19	< 0.001*
Cigarillos-Filter	0.291	0.001*
Cigarillos-Large	0.044	0.087
Filter-Large	0.257	< 0.001*

Core and Shared Microbiomes Across Products

In our study, we clustered the overall sequences into 2124 Operational Taxonomic Units (OTUs) with 97% identity across all 69 cigar samples. The top five bacterial phyla identified across all cigar types were Firmicutes, Actinobacteria, Proteobacteria, Acidobacteria, and Bacteroidetes. The most abundant phyla in all product categories were Firmicutes, followed by Actinobacteria. Filtered cigars exhibited a slightly higher abundance of Firmicutes (91.12%) compared to large cigars (70.29%) and cigarillos (81.55%), while also showing a decrease in Actinobacteria (2.09%) compared to large cigars (16.42%) and cigarillos (8.05%) (Fig. 3a). When we compared the top 30 genera for the three cigar categories, 16 genera (33.3%) were shared among all product types, while nine genera were specifically shared only by large cigars and cigarillos. Apart from the core microbiome, there was no genus similarity between filtered cigars and large cigars (Fig. 3b). However, there were 13 exclusive genera only for filtered cigars. The list of shared and unique genera for each cigar type is provided in Table 3.

Fig. 3.

Core and shared microbiome in different categories in (a) Phylum level showing in each category in percentage (b) Genus level similarity and specific to each category in Venn diagram

Table 3.

List of genera shared and unique to specific cigar category types

16 common genera in all products	9 common genera in “Large Cigar” and “Cigarillos”	1 genus in “Filtered Cigar” and “Cigarillos”
Bacillus	Yaniella	Paenibacillaceae_1_unclassified
Acinetobacter	Lentibacillus
Aerococcus	Brachybacterium
Bacillaceae_2_unclassified	Gracilibacillus
Bacillales_unclassified	Atopostipes
Brevibacterium	Tetragenococcus
Corynebacterium	Geomicrobium
Enterobacteriaceae_unclassified	Aurantimonas
Oceanobacillus	Rhodobacteraceae_unclassified
Paenibacillus
Pantoea
Planococcaceae_unclassified
Pseudomonas
Sphingomonas
Staphylococcus
Terribacillus
13 genera exclusively in “Filtered Cigar” only	5 genera exclusively in “Large Cigar” only	4 genera exclusively in “Cigarillos” only
Tissierella	Ralstonia	Sedimentibacter
Clostridium_XlVa	Bacteria_unclassified	Weissella
Planococcaceae_incertae_sedis	Desemzia	Thermoactinomyces
Sporomusa	Stenotrophomonas	Actinomycetales_unclassified
Enterococcus	Staphylococcaceae_unclassified
Lachnospiraceae_unclassified
Veillonellaceae_unclassified
Firmicutes_unclassified
Garciella
Bacilli_unclassified
Clostridiales_unclassified
Clostridium_sensu_stricto
Dendrosporobacter

Microbial Populations Within each Category

Total Genus Population Within Filtered Cigars

We analyzed seven different filtered cigars including one research filtered cigar (1C2). The top twenty genera of each product type were assessed based on the most numerous OTUs. The filtered cigars 1C2 and Phillies had uniform microbiome composition for each replicate (Fig. 4). Other products including Captain Black, Djarum, Talon, and Cheyenne have similar microbiomes for two of the three replicates. Djarum, the only filtered cigar product tested that included cloves with tobacco, had a higher percentage of Bacillus and Staphylococcus populations compared to other filtered cigars, whereas 1C2 and Talon had higher Tissierella genus. Of the seven products in the filtered cigar category, only the King Edward samples were significantly different for each replicate tested, making it difficult to determine the microbial population of this product.

Fig. 4.

Stacked bar chart showing the percentage of the dominant top 20 genera within the microbial population of each filtered cigar sample

Total Genus Population Within Cigarillos

In the case of cigarillos, we tested nine products. Cigarillo samples had a more consistent and uniform microbiome composition between the replicates. The research cigarillo (1C3), Black and Mild, White Owl Black, White Owl Silver, and Swisher Sweets have uniform microbiomes for each replicate (Fig. 5). For Black and mild, Pseudomonas represented more than 40% of the microbial population, and Pantoea represented 12% of the population which was the highest compared to other cigarillo types. White Owl Black and White Owl Silver had the highest percentage of Staphylococcus compared to the other products tested. Swisher Sweets and PomPom had a higher percentage of Terribacillus compared to other products, and they had similar microbiomes for two of the three replicates. Bacillus was the most common genus present in most samples. Two out of three replicates of Phillies Sweet showed higher levels of Lentibacillus. Dutch Masters and Phillies Black samples were different for each replicate.

Fig. 5.

Stacked bar chart showing the dominant top 20 genera of each cigarillos

Total Genus Population Within Large Cigars

A total of seven products in the large cigar category were taken including two research cigar products (1C1 and 1C4). All products have a uniform microbiome for each replicate, except White Owl which had a consistent microbiome for two of the three replicates. All large cigar samples show a higher percentage (8.20–39.46%) of Staphylococcus genus except William Penn and Garcia y Vega (0.16–2.09%) (Fig. 6). Additionally, these products showed an increased percentage of Planococcaceae unclassified (17.62–36.70%) and Paenibacillus (0.72–4.3%) which was observed at low levels for the other products tested (0.01–0.08%, with one replicate on the 1C4 having 2.3%).

Fig. 6.

Stacked bar chart showing the dominant top 20 genera of each large cigar

Discussion

The use of 16S rRNA gene amplicon sequencing allowed for a comprehensive characterization of the microbial diversity of cigar and cigarillo products, revealing differences between product types and providing information on various microbial populations. This study explores the microbial diversity of large cigars, filtered cigars, and cigarillos, to provide insight into the composition of the microbial communities that are unique and shared in these products. Our study revealed a significant difference in microbial diversity among the three categories of cigars (large, filtered, cigarillos) based on alpha diversity measures for richness and Shannon diversity. Specifically, filtered cigars exhibited lower diversity compared to both large cigars and cigarillos. PCoA analysis with the ANOSIM test indicated significant differences in beta diversity measures between the three cigar categories, with the microbial populations of filtered cigars being statistically different from those of cigarillos and large cigars. Furthermore, when comparing the top 30 genera for the three cigar categories, the core microbiome analysis revealed that 16 genera were shared among all three categories of cigars, while nine genera were shared only by large cigars and cigarillos. These shared genera such as Oceanobacillus, Corynebacterium, Staphylococcus, Bacillus, and Pseudomonas are commonly found as the most predominant genera in tobacco products [29, 30]. The first three genera are shown to have a direct effect on the fermentation process [31–33]. Additionally, our study identified 13 exclusive genera only for filtered cigars, such as Tissierella, and Clostridium_XlVa. In contrast, there was no genus similarity between filtered cigars and large cigars (Table 3).

The presence of certain genera in all cigar types suggests the existence of a core microbiome in these products. However, there was no significant difference between the microbial populations of cigarillos and large cigars. The differences observed in the microbial composition of filtered cigars compared to large cigars and cigarillos could be due to the type of cigar tobacco (filler leaf, wrapper leaf), the region the tobacco was grown, the fermentation method during curing, or the manufacturing process. The most abundant phyla across all product categories were Firmicutes, followed by Actinobacteria, which agrees with similar studies to the one presented here [33–35]. Interestingly, filtered cigars had a slightly higher abundance of Firmicutes compared to large cigars and cigarillos, while also showing a decrease in Actinobacteria. A recent study showed that filtered cigar products have a unique bacterial signature and certain genera such as Enterobacteriaceae, Pantoea, Pseudomonas, and Staphylococcus were more abundant in the cigar tobacco used to produce the filtered cigar [1]. Microbial diversity studies conducted on cigarettes and various tobacco products, including smokeless forms, have uncovered a similar microbial population as found in the current study of cigar products [36]. Research on Burley and Flue-cured tobacco leaves has found Bacillus and Pseudomonas to be the predominant genera, while Proteobacteria is the dominant phylum, comprising over 90% of the operational taxonomic units (OTUs) in burley tobacco leaves [36, 37]. A recent study where sequencing was performed on the samples cultured from the mainstream smoke of unfiltered cigarettes showed Bacillus, Terribacillus, Paenibacillus, and Desulfotomaculum as their predominant genera [38]. Furthermore, bacterial community profiling in smokeless tobacco products revealed Firmicutes, Proteobacteria, Actinobacteria and Bacteroidetes as dominant phyla and Acinetobacter, Bacillus, Prevotella, Acetobacter, Lactobacillus were enriched bacterial genera. Additionally, microbial diversity and composition were determined to be variable across multiple smokeless tobacco product types and brands [18, 39–41].

We evaluated for inter-product variability within the microbiome composition of seven different filtered cigars, nine cigarillos, and seven large cigars. The filtered cigars 1C2 and Phillies had a uniform microbiome composition for each replicate, while other products showed some variability. Djarum had a higher percentage of Bacillus and Staphylococcus population, while 1C2 and Talon had a higher Tissierella genus. Cigarillo samples had a more consistent and uniform microbiome composition between the replicates, and Bacillus was the most common genus present in all samples. Compared to the other cigar categories, large cigars showed the highest percentage of Staphylococcus genus, except for William Penn and Gracia y Vega, which showed an increased percentage of Planococcaceae unclassified and Paenibacillus. There could be several factors causing the variation among the products, natural differences in the microbial populations between individual cigars as a result of the blend of tobacco used in manufacturing. The impact on the microbial population as a result of the manufacturer was not investigated for this study, as many manufacturers have production facilities in multiple locations, oftentimes in different countries. A study, conducted by Di Giacomo et al. (2007) on the microbial community in Italian Toscano cigar tobacco during fermentation, observed changes in the relative abundances of specific microbial groups over time [20]. They found that Bacillus species, including B. licheniformis and B. subtillis, could reduce NO₃ without producing N₂ gas, while Corynebacterium ammonia genes accumulated nitrite during subsequent stages of tobacco maturation. During the curing process of tobacco, changes in microbial communities were observed as nutrients were depleted, pH changed, moisture evaporated, and temperature increased. It also demonstrated a gradual decrease in tobacco pH, which was attributed to the metabolic byproducts of acid-producing species [20]. Various discrepancies in abundance and composition are not unexpected considering that brands are manufactured under different industrial conditions with proprietary tobacco blends. Another study showed that filtered cigar microbiome composition was dynamic and influenced by environmental factors such as temperature, humidity, and storage time [42].

Overall, our study provides new insights into the microbial diversity and composition of different categories of cigars, highlighting the impact of various cigar types and sizes on microbial populations. These findings hold significant implications for the regulation of tobacco products and the comprehension of potential health impacts arising from smoking diverse types of cigars. Prior research has demonstrated the abundance of human pathogens in tobacco products, particularly in cigarettes and smokeless tobacco, which could result in the development of chronic or infectious respiratory disease, as well as cancer [43–46]. Similarly, it is crucially important for further studies to elucidate the functional roles of these microorganisms in cigar products. In future studies, it may be possible to link pathogenicity to the microbial population as species with the genera Bacillus, Staphylococcus, and Pseudomonas have been identified as human pathogens. The use of MiSeq technology could be limited due to amplification bias, sequencing errors, and taxonomic resolutions. Utilizing advanced sequencing technologies and conducting experiments to identify metabolically active bacteria [34] could be one of the several approaches to dissecting the impact of microbes on human health. A recent study highlighted how viable bacteria could survive cigarette combustion and then be transferred to the upper respiratory system via mainstream smoke [38]. Additionally, future studies could focus on variation within microbiomes based on various production locations for specific manufacturers.

Author Contribution

SJ and RM contributed to the design and implementation of the research. SJ and KP contributed to the analysis of the sequences. SJ, KP, LM, and RM contributed to the writing and reviewing of the manuscript.

Funding

This work was supported by the Food and Drug Administration (grant UC2FD006890). The views expressed in written materials or publications and by speakers and moderators do not necessarily reflect the official policies of the Department of Health and Human Services nor does any mention of trade names, commercial practices, or organization imply endorsement by the United States Government.

Data Availability

Sequences can be found in the NCBI SRA database under the Bio project accession number PRJNA1073920.

Declarations

Competing Interests

The authors declare no competing interests.