Applied Microbiology and Biotechnology (2022) 106:2079–2089

https://doi.org/10.1007/s00253-022-11831-3

APPLIED GENETICS AND MOLECULAR BIOTECHNOLOGY

Validation of housekeeping genes as internal controls for gene

expression studies on biofilm formation in Bacillus velezensis
Lianmeng Liu1 · Zhiming Ji2 · Kehan Zhao1 · Yuan Zhao1 · Yilin Zhang1 · Shiwen Huang1,3

Received: 9 September 2021 / Revised: 19 January 2022 / Accepted: 10 February 2022 / Published online: 16 February 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2022

Abstract
Bacillus velezensis is an important bacterium widely applied in agriculture and industry, and biofilms play critical roles in
its environmental tolerance. The appropriate choice of reference genes is essential for key gene expression studies. Multiple
internal control genes were selected and validated from the 21 housekeeping genes of B. velezensis by expression stability
evaluation during biofilm formation and were used to study the expression of key genes involved in the process. The results
showed that pyk, gyrA, recA, and gyrB were stably expressed, and the expression of pyk was the most stable during biofilm
formation. A pair of two genes, pyk and gyrA, provided high-quality data when used as internal controls, and the combination
of three genes, pyk, gyrA, and recA, was even better. The expression levels of pyk, gyrA, and recA approximated those of five
key genes, abrB, epsD, kinC, sinR, and tasA, in biofilm formation, meeting the requirements of ideal internal control genes.
The expression patterns of 5 key genes were studied with 16S, pyk, the pair of 2 genes, pyk and gyrA, and the combination
of 3 genes, pyk, gyrA, and recA, as internal controls during the biofilm formation process. The results proved that pyk was
a suitable internal control, as were the pair of 2 genes, pyk and gyrA, and the combination of 3 genes, pyk, gyrA, and recA.
This study provided genes and gene combinations which were validated as suitable internal controls for gene expression
studies, especially those on the mechanism of biofilm formation in B. velezensis or even other Bacillus spp.

Key points
• Reference genes is necessary for gene expression study in biofilm formation of Bacillus velezensis
• Pyk and 2 gene combinations were selected and validated from 21 common used genes
• Expression of key genes in biofilm formation was normalized with the selected internal controls

Keywords Bacillus velezensis · Biofilm · Internal control · Pyk · RT-qPCR

Introduction et al. 2009; Khan et al. 2020), and commercial fermentation

species (Ruiz-García et al. 2005; Bafana et al. 2008) and is
Bacillus velezensis is an important bacterium with immense widely applied in agriculture and industry. To facilitate its
potential as a plant growth-promoting rhizobacterium (PGPR) development and application, molecular mechanism analysis
(Meng et al. 2016), biological control agent (BCA) (Nam of B. velezensis has become increasingly necessary and has
drawn increasing attentions from researchers (Liu et al. 2017;
* Lianmeng Liu Lee et al. 2021). There have already been 338 genomes for
liulianmeng@caas.cn this species published in the NCBI genome database (https://​
* Shiwen Huang www.​ncbi.​nlm.​nih.​gov/​genome). The stress tolerance of B.
huangshiwen@caas.cn velezensis to the environment and the underlying molecular
mechanisms are key points demanding intensive research
1
State Key Laboratory of Rice Biology, China National Rice because survival is generally critical for the application of B.
Research Institute, 311400 Hangzhou, China
velezensis, which is commonly done under harsh conditions
2
College of Biological and Food Engineering, Huaihua such as crop fields (Masmoudi et al. 2019; Liu et al. 2020).
University, Huaihua 418000, China
Biofilms are complex and stable structures composed of
3
College of Agriculture, Guangxi University, microbial communities and extracellular polymeric matrices,
530003 Nanning, China

13
Vol.:(0123456789)
2080 Applied Microbiology and Biotechnology (2022) 106:2079–2089

which are assemblies of metabolites, body lysates, and other In this study, housekeeping genes that were commonly
substances obtained from the environment, and act as adhe- used as internal controls in bacterial studies, especially those
sives for the biofilm itself and to carriers (Branda et al. 2005; of Bacillus spp., were retrieved from published literatures.
Flemming and Wingender 2010). For most bacteria, biofilms Multiple internal control genes were selected and validated
are shelters providing protection from adverse environments from the 21 housekeeping genes of B. velezensis by expres-
(Yin et al. 2019). Many reports have shown that Bacillus sion stability evaluation during biofilm formation and were
spp. can form biofilms with remarkable architectures in a used to study the expression of key genes related to this
process regulated by complex pathways (Branda et al. 2001; process.
Vlamakis et al. 2013).
To obtain direct insight into the mechanism of biofilm
formation, genes of Bacillus spp. expressed during this
process have drawn increasing attention from scientists Materials and methods
(Cairns et al. 2014; Kimura and Kobayashi 2020). Reverse
transcriptase PCR (RT-PCR) has been widely used to study Bacterial strain and growth conditions
gene expression in bacteria, including Bacillus spp. (Zhao
et al. 2016). Internal control genes, mostly selected among B. velezensis H168, deposited at the China General Micro-
housekeeping genes, are commonly used to quantify gene organisms Collection and Management Center (http://​www.​
expression by normalizing real-time PCR results. Expression cgmcc.​net) under the accession number CGMCC 1.19177,
stability is a prerequisite for internal control genes. Never- was isolated from the phyllosphere of pecan (Carya illinoen-
theless, many variables, such as material, growth stage, and sis) and showed great biological control potential against
treatment, could interfere with gene expression (Ritz et al. many rice diseases (Zhao 2020).
2009; Rocha et al. 2015; Feria-Romero et al. 2021). Conse-
quently, to avoid the risk of misinterpretation of expression DNA extraction and sequencing of housekeeping
data, the choice of internal control genes should depend on genes
certain conditions and be based on experimental validation
(Rocha et al. 2015). After culture on NA (nutrient agar, consisting of the fol-
In most published reports on biofilms as well as other lowing components per liter: peptone 5 g, yeast extract 1 g,
studies of Bacillus spp., the selection of reference genes was beef extract 3 g, sugar 15 g, and agar 20 g; the pH was
merely based on published literature rather than experimen- adjusted to 7.0 using NaOH solution) plates for 2–3 days, the
tal validation. Among the reference genes, the 16S rRNA plate surfaces were scraped with inoculation loops, and the
gene was the most commonly used (Rocha et al. 2015; Ma resulting bacteria were transferred to Matrix E tubes of the
et al. 2017; Huang et al. 2020). However, the expression of FastDNA Spin Kit for Soil (MP Biomedicals, LLC, USA),
the gene encoding 16S ribosomal RNA is more unstable than with 3 loops of the bacterial lawn per tube. Then, the DNA
that of genes belonging to some specific functional catego- was extracted following the protocol provided in the manual
ries, such as DNA replication (Rocha et al. 2015). Moreover, of the kit. A list of 21 housekeeping genes commonly used
the expression level of the 16S gene is often very high com- as internal controls in bacterial studies, especially those of
pared to that of other genes, commonly being a thousand- Bacillus spp., was made according to published literatures
fold higher than the genes to be analyzed, especially the key (Dheda et al. 2004; Janse et al. 2010; Metcalf et al. 2010;
genes expressed in the biofilm formation process. A large Galisa et al. 2012; Kirk et al. 2014; Lee et al. 2014; Rocha
expression difference increases the risk of error (Dheda et al. 2015; Zhao et al. 2016; Cozzolino et al. 2020). The
et al. 2004), and issues regularly occur when cDNA must corresponding sequences were obtained from the repre-
be diluted to create an ideal Ct value spectrum. Normally, sentative genome of B. velezensis (strain: ASM211716v1,
the use of only one reference gene is unreliable and poses a NZ_CP011937) in NCBI by searching for annotations with
risk of misinterpretation of expression data. Therefore, mul- the gene names in the list. Primers to amplify these genes
tiple reference genes are usually needed for normalization of were designed by Primer3 (Untergasser et al. 2012) based
RT-qPCR expression data (Vaudano et al. 2011; Sumby et al. on the corresponding sequences from the representative
2012). Considering the lack of experimentally validated genome. The primers were used to amplify these genes
internal control genes for biofilm studies of B. velezensis from the extracted DNA. The PCR products were purified,
as well as other Bacillus spp., it is necessary to select and cloned, and sequenced by Biosunya Biotechnology (Hang-
validate appropriate housekeeping genes as internal controls zhou, China). The assembled sequences were submitted to
for gene expression studies. GenBank.

13
Applied Microbiology and Biotechnology (2022) 106:2079–2089 2081

Samples and RNA extraction cycle was 95 °C for 10 s followed by 60 °C for 30 s, the
cycle number was set to 40. To verify the specificity of
After recovery from storage, the cells of H168 were the amplification, a thermal denaturing cycle followed to
streaked on an NA plate to form colonies. With a sterile derive the dissociation curves of the PCR products. Reac-
pipette tip that had been dipped into a colony, H168 was tions for each sample were performed in triplicate.
inoculated into a flask (100 ml) with 10 ml NB medium
(Nutrient Broth, NA without agar) and cultured at 37 °C
and 200 rpm rotation for 8–10 h. Then, the culture was Efficiency of qPCR
diluted to 50 NTU (2 × ­10 8 cfu ­m l −1 ) with minimal
medium (MSgg) (Branda et al. 2001) and used as inocu- cDNA was diluted 1-, 10-, 100-, 1000-, 10,000-, and
lum. The inoculum was mixed with fresh MSgg at a ratio 100,000-fold. The diluted cDNA was used as a template and
of 1:100 (v/v) following a published procedure (Ogran analyzed by qPCR in triplicate. The logarithm of relative
et al. 2019). The mixture was distributed to cell culture amounts of the targeted template and the mean of the cor-
plates (6 cells, Costar 3516, Corning Life Sciences Co. responding Ct values of each primer pair were tabulated and
Ltd, China), with 5 ml in each cell. The plates were stati- linearly regressed by the lm formula (Wilkinson and Rogers
cally cultured at 28 °C to form pellicle biofilms. Biofilms 1973) in R version 3.6.3. The qPCR efficiency (E) of each
were sampled at 36 h, 48 h, 72 h, 96 h, and 120 h after primer pair was calculated from the regression coefficients
inoculation. Because biofilms formed at 36 h were vulner- (R) in the results of the linear regression using the following
able, the biofilm samples were collected by centrifugation equation (Bustin et al. 2009; Gong et al. 2016).
for 1 min at 12,000 rpm ­m in −1. For biofilms formed at
other times, the MSgg medium was carefully pipetted from E(%) = (10−(1∕R) − 1) × 100
the edges. Then, the biofilms were rinsed three times with
fresh MSgg media, collected by sterile inoculation loops
and transferred to sterile centrifuge tubes. The bacterial Gene stability and minimum number of reference
cells were also collected by centrifugation for 1 min at genes
12,000 rpm ­min−1 from the inoculated suspension without
culture as samples of 0 h. RNA from sampled biofilms was The widely recognized reference gene normalization algo-
extracted with an RNeasy PowerBiofilm Kit (25,000–50, rithm geNorm (Vandesompele et al. 2002) was used to assess
Qiagen, Germany) following the provided manual. The the stability of the sequenced housekeeping genes. The Ct
residual DNA was removed by a TURBO DNA-free Kit values were transformed to quantities using the deltaCt
(AM1907, Ambion, Thermo Fisher Scientific, Lithuania). method with the highest relative quantities for each gene set
The incubation time was extended to 40 min, and DNase to 1 as described in the software manual and submitted to
was added in two steps to eliminate the residual DNA as the geNorm applet (Vandesompele et al. 2002). The average
much as possible. expression stability of the remaining reference genes was
indicated by the M value from the results, and genes with
lower M values had higher expression stability. Genes with
Reverse transcription and quantitative PCR M values lower than 1.5 were considered stably expressed.
Pairwise variations (V values) were used to determine the
First strand cDNA was synthesized by a RevertAid First ideal number of reference genes needed for optimal normali-
Strand cDNA Synthesis Kit (K1621, Thermo Fisher Sci- zation with 0.15 as a cutoff.
entific Inc.) with random hexamer oligonucleotide primers
following the manual. Real-time PCR was conducted in a
QuantStudio™ 1 System (ABI, Thermo Fisher Scientific Statistical analysis
Inc.). Analysis of 16S rDNA was performed using a pub-
lished primer pair (Feng et al. 2017), and the other prim- The comparative 2­ −ΔΔCt method was used to calculate the
ers were designed by Primer3 (Untergasser et al. 2012) relative fold changes in the expression of five previously
based on the sequences obtained in “DNA extraction and reported key genes of biofilm formation, including abrB,
sequencing of housekeeping genes.” The volume of each epsD, kinC, sinR, and tasA, in the samples collected at the
reaction mixture was 20 μl, containing 0.5 μl cDNA prod- different times during the process of biofilm formation, and
uct, 4 pmol each primer, and 10 μl 2 × PowerUp SYBR the expression was normalized with the recommended gene
Green Master Mix (A25742, ABI, Thermo Fisher Scien- and gene combinations of the most stable genes and the
tific Inc.). The experimental method was as follows: an common used internal control, 16S, for comparison.
initial denaturing step at 95 °C for 30 s; the amplification

13
2082 Applied Microbiology and Biotechnology (2022) 106:2079–2089

The data were compared with Student’s t test in R version sequences of 21 reference genes and are listed in Table 2,
3.6.3. GraphPad Prism 7.0 was used for variance analysis and the expected RT-qPCR amplification products ranged
and graph plotting. from 118 to 194 bp (Table 2). For each sample, only a
single band was obtained, the size of which was consist-
ent with the expectation for each of the 21 genes (Fig. 1A)
Results and only one peak was showed in melt curve of each of the
21 genes (Fig. 1B), which means that the amplification of
Housekeeping genes and RT‑qPCR amplification each gene was specific and precise.
products
Amplification efficiency
The candidate genes were selected from housekeeping
genes that were commonly used as internal controls for The amplification efficiency of 21 candidate reference genes
gene expression analysis in published bacterial stud- ranged from 95.3 to 104.6% (Table 2). The theoretical
ies, especially those of Bacillus spp. (Dheda et al. 2004; maximum efficiency of PCR is 100%, under which condi-
Janse et al. 2010; Metcalf et al. 2010; Galisa et al. 2012; tion the polymerase enzyme has already reaches maximum
Kirk et al. 2014; Lee et al. 2014; Rocha et al. 2015; Zhao capacity in an ideal PCR reaction. However, the amplifi-
et al. 2016; Cozzolino et al. 2020). Totally, 21 genes cation efficiency commonly exceeds 100% mainly due to
were selected and listed in Table 1. The selected house- the PCR inhibitor whose effects could be weakened after
keeping genes were amplified and sequenced with prim- dilution leading to calculated efficiency value higher than
ers designed from the reference genome of B. velezensis the actual value (Nybo 2011; Harshitha and Arunraj 2021).
(strain: ASM211716v1, NZ_CP011937) in NCBI and sub- Primers pyk, rpoB, and sdaAA showed higher amplification
mitted to GenBank under the accession numbers listed in efficiencies of more than 100%, whereas pta showed the
Table 1. Primers for RT-qPCR were designed from the

Table 1  Information on the candidate reference genes. Sequences and accession numbers of samples cloned and submitted in this study
Gene symbol Gene Functional category Accession number

16S 16S ribosomal RNA Component of the ribosome MW737443

23S 23S ribosomal RNA Component of the ribosome MW720958
ftsZ Cell division protein FtsZ Cell division (GO:0051301) MW736033
fusA Elongation factor G Cellular component disassembly (GO:0022411) MW736025
gapA Glyceraldehyde-3-phosphate dehydrogenase Glucose metabolic process(GO:0006006) MW736032
gmk Guanylate kinase Nucleobase-containing compound kinase MW736031
activity(GO:0019205)
gyrA DNA gyrase subunit A DNA conformation change(GO:0071103) MW736023
gyrB DNA topoisomerase (ATP-hydrolysing) subunit B DNA-dependent DNA replication (GO:0006261) MW736022
hemC Hydroxymethylbilane synthase Organonitrogen compound biosynthetic process MW736040
(GO:1901566)
PGDH Phosphoglycerate dehydrogenase Cellular amino acid metabolic process(GO:0006520) MW736030
polA DNA polymerase I DNA biosynthetic process (GO:0071897) MW736038
proS Proline-tRNA ligase Cellular amino acid metabolic process (GO:0006520) MW736037
pta Phosphate acetyltransferase Acetyl-CoA biosynthetic process (GO:0006085) MW736029
pyk Pyruvate kinase Oxidation–reduction process(GO:0055114) MW736036
pyrG CTP synthase (glutamine hydrolysing) Nucleobase metabolic process(GO:0009112) MW736026
recA Recombinase RecA DNA recombination (GO:0006310) MW736024
rho Transcription termination factor Rho DNA-templated transcription, termination MW736039
(GO:0006353)
rpoA DNA-directed RNA polymerase subunit alpha Transcription, DNA-templated (GO:0006351) MW736028
rpoB DNA-directed RNA polymerase subunit beta Transcription, DNA-templated (GO:0006351) MW736035
rpoD RNA polymerase sigma factor RpoD Transcription initiation from bacterial-type RNA MW736034
polymerase promoter (GO:0001123)
sdaAA L-serine ammonia-lyase, iron-sulfur-dependent, Lyase activity(GO:0016829) MW736027
subunit alpha

13
Applied Microbiology and Biotechnology (2022) 106:2079–2089 2083

Table 2  Information and amplification efficiency of primers designed for candidate reference genes. Primers for the 16S rRNA gene (rrn) were
obtained from a published report (Feng et al. 2017), and the other primers were designed in this study
Gene symbol Primers F Primers R Amplicon Amplification efficiency

16S GCG​TAG​AGA​TGT​GGA​GGA​AC CCA​GGC​GGA​GTG​CTT​AAT​G 190 98.3 ± 2.7

23S GGC​GAA​GTA​TAG​GGG​CTG​AC ACC​CGA​CAA​GGA​ATT​TCG​CT 141 98.7 ± 3.2
ftsZ AAA​CAC​ACC​GAT​GCT​TGA​AGC​ GCG​ATC​CCG​ATA​CCC​ATC​AAT​ 168 97.2 ± 0.9
fusA TGA​ACT​GAA​AGC​TGC​TAT​CCGT​ CGT​ACA​CCT​TTG​ATT​GCA​GCAA​ 159 99.7 ± 4.7
gapA CCG​GGC​AAA​AAT​GAA​GAC​GTT​ GTC​ATC​AGG​CCG​CTT​TCA​ATC​ 164 99.1 ± 3.4
gmk GAA​CGC​CCG​TTG​ATT​ATG​TCG​ GTC​GGT​TTC​AGT​CCC​TCT​TGT​ 182 97.6 ± 1.8
gyrA TTC​CTC​CGC​ATC​AGC​TTG​G CCG​GCC​CAA​AAT​CTG​ACC​T 134 98.7 ± 3.1
gyrB GAC​CAC​TCT​TGA​CGT​TAC​GGT​ GAC​ACG​GTT​TGA​AAG​CAG​GTC​ 193 96.4 ± 2.7
hemC GAT​TGA​TAT​GGC​GGT​TCA​CAGC​ ATG​CCC​CCT​TCT​TTT​ATG​TCGT​ 150 99.3 ± 3.0
PGDH ATT​ACC​GTC​CAC​ACG​CCT​TT GGC​GGT​TCC​ACT​TCA​AAC​AC 194 97.5 ± 2.2
polA AAT​TCA​CCA​AAA​CAG​CTC​GGC​ TGC​AGA​ATG​TAC​GCG​ACG​ATA​ 143 96.3 ± 3.3
proS CGC​AGC​CAA​TAC​GGA​AAT​GG CGA​AGC​GAG​CTC​TTC​AAT​GC 118 98.6 ± 1.1
pta GCT​GAT​TGC​GCG​ATC​AAC​AT TCA​GCC​ACT​TTG​TCC​GTC​TC 164 95.3 ± 2.9
pyk GGC​AGC​TAT​CGT​TAC​CCC​TAC​ ACC​GAA​TAC​AAG​GCC​GAG​TTT​ 130 104.6 ± 10.6
pyrG GAA​GGG​AAA​ATC​ATC​GCC​GC GGA​TCG​ATC​TCC​GCA​GAG​TG 143 98.3 ± 2.6
recA TTT​CCG​GCG​CCA​TCA​ATA​AATC​ CGT​TTT​ATT​CCC​CAT​GAC​GTCG​ 194 97.5 ± 2.2
rho GCC​GAA​AGA​AAA​CGA​ACG​CT AGT​CAG​CGC​AGG​AAA​ATG​GA 103 98.3 ± 2.9
rpoA GCA​CAA​AGA​GGA​CGT​GGG​TA CCT​GGC​CTA​CAC​GAG​TGT​TT 139 99.5 ± 3.5
rpoB TGG​GCG​ACA​CTG​ATG​ATA​TCG​ GAT​GTT​AAT​CAG​CTG​CTG​CGG​ 170 102.7 ± 4.8
rpoD CGA​CAC​CGG​AAG​AGA​TTG​CT GCA​TGA​TCC​GAC​GGA​GAT​GT 172 99.6 ± 3.9
sdaAA TAG​ACG​CCG​TTT​CTA​AAG​CTGT​ CTT​TAC​GGC​GAA​TAA​CGT​TCCC​ 119 100.5 ± 4.0

Fig. 1  Patterns of RT-qPCR products of candidate reference genes candidate genes in the order presented in Table 2 and amplified with
and the specificity of the primers. A Electrophoresis was performed the primers listed in Table 2. B Melt curves of the primers analyzed
on a 2% agarose gel with 5 µl of RT-qPCR products. M = 100-bp lad- at stage3 of qPCR by QuantStudio™ 1 System
der marker; sizes are listed on the left; 1–21 = RT-qPCR products of

13
2084 Applied Microbiology and Biotechnology (2022) 106:2079–2089

lowest amplification efficiency, 95.3%, but still more than software used for gene expression stability could not allow
95%. Normally, the accepted range for qPCR is generally blanks in the data. The gene expression stability of the other
80 to 110% (Nybo 2011). In practices, a variation of 5% is 18 genes was analyzed by geNorm with the Ct values of the
definitely acceptable, so no reference gene was excluded due genes in all 28 samples collected during the biofilm forma-
to low amplification efficiency (Rutledge and Stewart 2008; tion process of B. velezensis, and the results are displayed
Ruijter et al. 2009). in Fig. 2 and Fig. 3. GeNorm generated M values to repre-
sent the gene expression stability of the candidate reference
Gene expression stability genes, and lower M values indicated higher expression sta-
bility. According to Fig. 2, pyk and gyrA were the two most
Some genes could not be detected in all 28 samples because stably expressed genes in the biofilm formation process of
of low and unstable expression, such as hemC (4/28), polA B. velezensis because they have the lowest M values. RecA
(2/28), and rho (6/28). These 3 genes were excluded for the and gyrB were also stably expressed and ranked 3rd and 4th,
next step because a stable expression level and universality respectively.
are prerequisites for internal control genes. Moreover, the

Fig. 2  Candidate genes for

internal control ranked with
average expression stability
values (M). The figure was
produced by geNorm software
(Vandesompele et al. 2002), and
M values were also calculated
with the software

Fig. 3  Determination of the

optimal number of control genes
for normalization. The bars
illustrate values of “pairwise
variation (V).” V values were
calculated based on normaliza-
tion factor values ­(NFn and
­NFn+1) which started with the
two most stably expressed genes
(pyk and gyrA), then each of the
remaining reference gene were
separately added to the equation
in sequence of stability showed
in Fig. 2 resulting in a serials of
V values, from V2/3 to V17/18.
The lower V value indicated
more accurate and reliable
normalization

13
Applied Microbiology and Biotechnology (2022) 106:2079–2089 2085

Optimal number of reference genes Expression levels of reference genes

The optimal number of reference genes was determined by The expression levels of ideal reference genes should
a measure known as the “pairwise variation (V)” conducted approximate the genes to be measured. The Ct value differ-
by geNorm to estimate the least number of reference genes ences were used to present the expression level difference
for accurate and reliable normalization. V values were calcu- between the candidate reference genes and the 5 key genes,
lated based on normalization factor values ­(NFn and ­NFn+1) abrB, epsD, kinC, sinR, and tasA, in the pellicle biofilm
which started with the two most stably expressed genes (pyk formation process (Table 3). The 16S rRNA gene showed
and gyrA), then each of the remaining reference genes was the greatest Ct value difference, and the greatest difference,
separately added to the equation in sequence of stability 24.27, was achieved with epsD; even the least difference
showed in Fig. 2 resulting in a serials of V values, from was as high as 14.39 (with tasA), which means that the
V2/3 to V17/18 (Fig. 3). The lower V value indicated more expression level of the 16S rRNA gene was 2.1 × ­104- and
accurate and reliable normalisation. The result showed that 2.0 × ­107-fold higher than that of epsD and tasA, respec-
the V score did not exceed 0.15, the threshold recommended tively. The expression level of the 23S rRNA gene was
by the software, when the gene number was lower than 18 similar to that of the 16S rRNA gene and showed a Ct dif-
(Fig. 3). Value of V2/3 was 0.087 using two genes and lower ference spectrum of 13.25–22.54. The expression gaps with
than the threshold, which indicated that pair of the two most the two rRNA genes were not compared between the other
stably expressed genes was sufficient to provide high-quality reference genes and the key 5 genes, and the Ct value differ-
data. The three genes resulted in a lower V value of 0.062 ences were all below 10. The Ct value difference spectrum
(V3/4), which ranked third and was only higher than 0.061 was − 1.27–7.32 between pyk, gyrA, and recA, which were
for 10 genes (V10/11) and 0.059 for 12 genes (V12/13). A proven to be the most stably expressed reference genes dur-
lower V value corresponds to more accurate data normaliza- ing biofilm formation, and the 5 key genes. From the aspect
tion. However, more reference genes (10 genes or 12 genes) of the expression level, pyk, gyrA, and recA were also ideal
could increase more workload for PCR and cost more mate- reference genes.
rials. So, three reference genes would be ideal for optimal
normalization considering both data interpretation quality Expression normalization of key genes in biofilm
of gene expression and economy. Consequently, the use of formation
two genes, pyk and gyrA, would give high-quality data, and
three genes, namely pyk, gyrA, and recA, would be ideal. The expression of the 5 key genes was normalized to 16S,
pyk, the pair of 2 genes, pyk and gyrA, and the combination

Table 3  Expression differences Gene symbols Ct difference spectra with

represent Ct value differences
between the candidate reference abrB epsD kinC sinR tasA
genes and the 5 key genes
involved in biofilm formation. 16S 16.63–20.57 15.36–24.27 18.33–23.09 16.06–19.91 14.39–22.43
Ct value differences were 23S 15.60–20.54 14.38–21.86 15.76–22.54 13.25–17.61 13.57–21.84
calculated by Ct values of ftsZ 0.99–4.64 0.02–5.66 2.63–5.70 0.18–1.92 − 0.95–4.47
genes, measured separately,
minus the Ct values of the fusA 4.03–7.11 3.17–10.14 5.20–8.92 2.10–5.94 2.21–8.30
candidate reference genes gapA − 4.91–7.15 − 5.53–6.60 − 3.37–8.20 − 5.41–3.95 − 6.42–7.29
gmk − 2.00–4.25 − 3.43–4.41 − 0.31–3.34 − 3.23–0.27 − 4.24–3.99
gyrA 1.27–4.92 − 0.15–6.97 2.97–5.74 − 0.06–2.51 − 0.96–5.18
gyrB − 0.70–4.54 − 1.89–4.89 1.35–4.01 − 2.08–0.42 − 2.70–4.04
PGDH 1.07–6.23 1.62–6.92 3.54–5.34 − 1.33–3.26 0.22–5.60
proS − 0.58–2.96 − 1.46–5.77 0.95–4.16 − 1.55–2.06 − 2.43–3.93
pta − 4.13–3.24 − 3.07–4.15 − 1.35–2.45 − 6.01–0.05 − 3.80–2.61
pyk 1.99–5.33 0.61–7.32 3.69–6.46 0.75–2.86 − 0.20–5.17
pyrG − 1.46–1.14 − 1.89–4.73 − 0.71–3.08 − 3.21–0.79 − 2.70–2.89
recA 0.76–5.65 − 0.46–5.73 2.46–5.23 − 0.60–1.51 − 1.27–5.19
rpoA 3.54–7.72 2.97–9.63 5.82–8.09 2.45–5.90 2.00–7.79
rpoB 2.47–5.77 1.61–7.57 3.34–7.55 0.84–3.52 0.64–7.10
rpoD 0.89–5.44 0.31–5.45 2.11–6.93 − 0.40–2.05 − 0.50–5.94
sdaAA − 1.26–3.03 − 2.18–3.54 0.27–4.33 − 2.56– − 0.21 − 3.15–3.11

13
2086 Applied Microbiology and Biotechnology (2022) 106:2079–2089

Fig. 4  Expression patterns of the 5 key genes normalized with differ- bination of 3 genes, pyk, gyrA, and recA at 36 h, 48 h, 72 h, 96 h, and
ent internal control genes or gene combinations at the different times 120 h. Error bars represent means ± SDs. B Biofilms formed at differ-
during the process of biofilm formation. A The expression patterns ent sample times during the process of biofilm formation
normalized to 16S, pyk, pair of 2 genes, pyk and gyrA, and the com-

of the 3 genes, pyk, gyrA, and recA, at 36 h, 48 h, 72 h, 96 h, Discussion

and 120 h (Fig. 4). According to the results presented in
Fig. 4, the expression patterns of the key genes, except sinR, In this study, 21 housekeeping genes were sequenced and
were similar, and the expression levels showed a curved assessed to select and validate ideal internal controls for
shape and peaked at 72 h after inoculation; however, the gene expression studies of biofilm formation in B. velezen-
relative expression differences were more drastic between sis. Pyk, gyrA, recA, and gyrB were proven to be stably
samples collected at different times when normalized to 16S. expressed genes, and the expression of pyk was the most
The expression patterns of sinR were significantly differ- stable in the biofilm formation process. A pair of two genes,
ent between normalization to the 16S rRNA gene and the pyk and gyrA, provided high-quality data when used as inter-
other genes (combinations). For the expression patterns nor- nal controls, and the combination of three genes, pyk, gyrA,
malized to the 16S rRNA gene, the expression of sinR also and recA, proved even more ideal. The expression levels of
peaked at 72 h after inoculation and sharply decreased until pyk, gyrA, and recA approximated those of five known key
it was very low at 120 h after inoculation. However, in other genes of biofilm formation, meeting the requirements for
patterns, the expression of sinR remained relatively stable at ideal internal control genes. The expression patterns of 5 key
36 h and 48 h after inoculation at levels similar to or slightly genes, abrB, epsD, kinC, sinR, and tasA, were studied with
greater than that of the seed liquor (0 h) and decreased after 16S, pyk, the pair of 2 genes, pyk and gyrA, and the combi-
48 h. The latter pattern was more consistent with published nation of 3 genes, pyk, gyrA, and recA as internal controls
results (Fan et al. 2015; Cámara-Almirón et al. 2020; Futo during the biofilm formation process. The results proved
et al. 2021). that pyk was a suitable internal control, as were the pair of 2
genes, pyk and gyrA, and the combination of 3 genes, pyk,
gyrA, and recA. This study (1) identified and verified suit-
able genes and gene combinations for gene expression stud-
ies, especially those on the mechanism of biofilm formation

13
Applied Microbiology and Biotechnology (2022) 106:2079–2089 2087

of B. velezensis or even other Bacillus spp. and (2) analyzed be experimentally determined (Bustin et al. 2009). In this
the expression patterns of five key genes involved in the study, the results of geNorm analysis showed that a pair of
biofilm formation process. two genes, pyk and gyrA, was sufficient for normalization
Pyk, which encodes pyruvate kinase, catalyses the trans- (Fig. 3). To obtain more precise normalization, a 3-gene
fer of phosphate from phosphoenolpyruvate to adenosine combination, namely pyk, gyrA, and recA, would be needed
diphosphate (ADP), yielding ATP, which is the key and last considering that the value of V3/4 declined sharply. Addition-
step of glycolysis (Li et al. 2018); additionally, this gene is ally, V12/13 had the lowest V value. Obviously, normalization
ubiquitously and constitutively expressed in different cells with 12 internal genes is not economical; therefore, using
of almost all living beings (Israelsen and Vander Heiden two or three genes is suggested. In this study, there was no
2015), which makes it suitable as an internal control to study significant difference in the expression patterns of the 5
gene expression levels. Kinases are an important source of key genes during the biofilm formation process among the
internal control genes for prokaryotes. For example, gmk normalizations with pyk, the 2 paired genes pyk and gyrA,
(guanylate kinase) was used as an internal control gene for or the 3-gene combination of pyk, gyrA, and recA (Fig. 4).
expression studies in Zobellia galactanivorans (Thomas These results indicated that only pyk seemed to be sufficient
et al. 2011), Gluconacetobacter diazotrophicus (Galisa et al. for normalization among these 5 key genes of the biofilm
2012), Acidithiobacillus ferrooxidans (Nieto et al. 2009), formation process. Whether only pyk is sufficient for other
and Oenococcus oeni (Sumby et al. 2012). Adk (adenylate studies will be evaluated in our future work.
kinase) is also used in studies of many bacteria, such as B. velezensis belongs to the B. subtilis group, which
B. cereus (Reiter et al. 2011), Clostridium difficile (Metcalf contains several other closely related taxa, such as B. sub-
et al. 2010), and Clostridium botulinum (Kirk et al. 2014). tilis subsp. subtilis, B. licheniformis, B. vallismortis, and
However, pyk has not been used as frequently as gmk and B. sonorensis (Wang et al. 2007). These taxa share similar
adk for gene expression normalization. The only available gene sequences, so there is no variation in primer sequences
report on pyk included its selection as a suitable internal among them, and the primers designed and listed in this
control for Staphylococcus aureus, with the best expression study may be used in biofilm formation studies of these spe-
stability among those tested (Theis et al. 2007). Therefore, cies. However, this still needs to be proven in further studies
our study revealed that pyk could be a new internal control for risk of error. These primers may also be used to quantify
with great potential for use in gene expression analysis of gene expression of the related species by normalizing real-
bacteria. time PCR results in process other than biofilm formation.
Although housekeeping genes are believed to maintain
stable expression under all conditions, their expression Acknowledgements We are grateful to Prof. Lei Wang of the China
National Rice Research Institute for his kind support in the data
varies. According to the recommendations of the MIQE analysis.
standards, expression stability evaluation should be
performed for each newly selected gene used as an internal Author contribution LL and SH designed the experimentals. LL and
control (Bustin et al. 2009). However, the 16S rRNA ZJ performed most work of experiments. KZ, YZ, and YZ provided
gene is still used in most studies of gene expression in technical supports and assistances in material preparation and data
collection. The manuscript was initially written by LL, and proved
Bacillus spp. (Rocha et al. 2015; Ma et al. 2017; Huang and edited by SH. All authors read and approved the final manuscript.
et al. 2020). Rocha et al. surveyed and analyzed bacterial
reference genes for gene expression studies by RT-qPCR Funding The research in this paper was funded by the Natural Science
and found that 16S ribosomal RNA was not a good internal Foundation of Zhejiang Province (LY21C140001), Sinograin II: Tech-
control gene compared to genes belonging to some specific nological Innovation to support environmentally friendly food produc-
tion and food safety under a changing climate—Opportunities and chal-
functional categories (Rocha et al. 2015), even when used lenges for Norway-China cooperation, Technology Research Program
in all conditions. In this study, we selected and validated of Zhejiang Province (2022C02014), Key Research and Development
some functional genes, Pyk, gyrA, recA, and gyrB, to provide Program of Zhejiang Province (2019C02018), and Rice Pest Manage-
alternatives for the 16S rRNA gene. Moreover, compared ment Research Group of the Agricultural Science and Technology
Innovation Program of the Chinese Academy of Agricultural Science.
to the 16S rRNA gene, whose expression is tremendously
higher than that of other genes, the expression of these Data availability Available upon the request.
genes is relatively similar to that of functional genes, which
provides convincing evidence for cDNA dilution and makes Declarations
normalization more precise.
Normally, one gene is not enough for gene expression Ethics approval This article does not contain any studies performed
normalization, so more reference genes are usually needed with human participants or with animals by any of the authors.
(Vaudano et al. 2011; Sumby et al. 2012). Therefore, both
the optimal number and selection of reference genes must Conflict of interest The authors declare no conflict of interest.

13
2088 Applied Microbiology and Biotechnology (2022) 106:2079–2089

References Gong H, Sun L, Chen B, Han Y, Pang J, Wu W, Qi R, Zhang TM

(2016) Evaluation of candidate reference genes for RT-qPCR
studies in three metabolism related tissues of mice after caloric
Bafana A, Chakrabarti T, Devi SS (2008) Azoreductase and dye detoxi-
restriction. Sci Rep 6:38513. https://​doi.​org/​10.​1038/​srep3​8513
fication activities of Bacillus velezensis strain AB. Appl Microbiol
Harshitha R, Arunraj DR (2021) Real-time quantitative PCR: a tool
Biotechnol 77:1139. https://​doi.​org/​10.​1007/​s00253-​007-​1212-5
for absolute and relative quantification. Biochem Mol Biol Edu
Branda SS, González-Pastor JE, Ben-Yehuda S, Losick R, Kolter R
49:800–812. https://​doi.​org/​10.​1002/​bmb.​21552
(2001) Fruiting body formation by Bacillus subtilis. Proc Natl
Huang Z, Wu L, Li X, Ma L, Borriss R, Gao X (2020) Zn(II) sup-
Acad Sci U S A 98:11621–11626. https://​doi.​org/​10.​1073/​pnas.​
presses biofilm formation in Bacillus amyloliquefaciens by inac-
19138​4198
tivation of the Mn(II) uptake. Environ Microbiol 22:1547–1558.
Branda SS, Vik S, Friedman L, Kolter R (2005) Biofilms: the matrix
https://​doi.​org/​10.​1038/​srep3​8513
revisited. Trends Microbiol 13:20–26. https://​doi.​org/​10.​1016/j.​
Israelsen WJ, Vander Heiden MG (2015) Pyruvate kinase: function,
tim.​2004.​11.​006
regulation and role in cancer. Semin Cell Dev Biol 43:43–51.
Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M,
https://​doi.​org/​10.​1016/j.​semcdb.​2015.​08.​004
Mueller R, Nolan T, Pfaffl MW, Shipley GL, Vandesompele J,
Janse I, Hamidjaja RA, Bok JM, van Rotterdam BJ (2010) Reliable
Wittwer CT (2009) The MIQE guidelines: minimum information
detection of Bacillus anthracis, Francisella tularensis and Yers-
for publication of quantitative real-time PCR experiments. Clin
inia pestis by using multiplex qPCR including internal controls for
Chem 55:611–622. https://d​ oi.o​ rg/1​ 0.1​ 373/c​ linch​ em.2​ 008.1​ 12797
nucleic acid extraction and amplification. Bmc Microbiol 10:1–12.
Cairns LS, Hobley L, Stanley-Wall NR (2014) Biofilm formation
https://​doi.​org/​10.​1186/​1471-​2180-​10-​314
by Bacillus subtilis: new insights into regulatory strategies and
Khan MS, Gao J, Chen X, Zhang M, Zhang X (2020) The endophytic
assembly mechanisms. Mol Microbiol 93:587–598. https://​doi.​
bacteria Bacillus velezensis Lle-9, isolated from Lilium leucan-
org/​10.​1111/​mmi.​12697
thum harbors antifungal activity and plant growth-promoting
Cámara-Almirón J, Navarro Y, Díaz-Martínez L, Magno-Pérez-Bryan
effects. J Microbiol Biotechnol 30:668–680. https://​doi.​org/​10.​
MC, Molina-Santiago C, Pearson JR, de Vicente A, Pérez-García
4014/​jmb.​1910.​10021
A, Romero D (2020) Dual functionality of the amyloid protein
Kimura T, Kobayashi K (2020) Role of glutamate synthase in biofilm
TasA in Bacillus physiology and fitness on the phylloplane. Nat
formation by Bacillus subtilis. J Bacteriol 202:e00120-00120.
Commun 11:1859. https://​doi.​org/​10.​1038/​s41467-​020-​15758-z
https://​doi.​org/​10.​1128/​JB.​00120-​20
Cozzolino ME, Distel JS, García PA, Mascotti ML, Ayub MJ, Benazzi
Kirk DG, Palonen E, Korkeala H, Lindström M (2014) Evaluation of
LM, Di Masi SN, Silva PG (2020) Control of postharvest fungal
normalization reference genes for RT-qPCR analysis of spo0A
pathogens in pome fruits by lipopeptides from a Bacillus sp iso-
and four sporulation sigma factor genes in Clostridium botulinum
late SL-6. Sci Hortic-Amsterdam 261:108957. https://​doi.​org/​10.​
Group I strain ATCC 3502. Anaerobe 26:14–19. https://​doi.​org/​
1016/j.​scien​ta.​2019.​108957
10.​1016/j.​anaer​obe.​2013.​12.​003
Dheda K, Huggett JF, Bustin SA, Johnson MA, Rook G, Zumla A
Lee N, Kwon KY, Oh SK, Chang HJ, Chun HS, Choi SW (2014) A
(2004) Validation of housekeeping genes for normalizing RNA
multiplex PCR assay for simultaneous detection of Escherichia
expression in real-time PCR. BioTechniques 37:112–114, 116,
coli O157: H7, Bacillus cereus, Vibrio parahaemolyticus, Salmo-
118–119. https://​doi.​org/​10.​2144/​04371​RR03
nella spp., Listeria monocytogenes, and Staphylococcus aureus
Fan B, Li L, Chao Y, Förstner K, Vogel J, Borriss R, Wu XQ (2015)
in Korean ready-to-eat food. Foodborne Pathog Dis 11:574–580.
dRNA-Seq reveals genomewide TSSs and noncoding RNAs of
https://​doi.​org/​10.​1089/​fpd.​2013.​1638
plant beneficial Rhizobacterium Bacillus amyloliquefaciens
Lee H, Jung DH, Seo DH, Chung WH, Seo MJ (2021) Genome analy-
FZB42. PLoS ONE 10:e0142002. https://​doi.​org/​10.​1371/​journ​
sis of 1-deoxynojirimycin (1-DNJ)-producing Bacillus velezensis
al.​pone.​01420​02
K26 and distribution of Bacillus sp. harboring a 1-DNJ biosyn-
Feng Y, Liu S, Jiao Y, Gao H, Wang M, Du G, Chen J (2017) Enhanced
thetic gene cluster. Genomics 113:647–653. https://​doi.​org/​10.​
extracellular production of L-asparaginase from Bacillus subti-
1016/j.​ygeno.​2020.​09.​061
lis 168 by B. subtilis WB600 through a combined strategy. Appl
Li YH, Li XF, Liu JT, Wang H, Fan LL, Li J, Sun GP (2018) PKM2,
Microbiol Biotechnol 101:1509–1520. https://​doi.​org/​10.​1007/​
a potential target for regulating cancer. Gene 668:48–53. https://​
s00253-​016-​7816-x
doi.​org/​10.​1016/j.​gene.​2018.​05.​038
Feria-Romero IA, Bribiesca-Cruz I, Coyoy-Salgado A, Segura-Uribe
Liu G, Kong Y, Fan Y, Geng C, Peng D, Sun M (2017) Whole-genome
JJ, Bautista-Poblet G, Granados-Cervantes A, Guerra-Araiza C
sequencing of Bacillus velezensis LS69, a strain with a broad
(2021) Validation of housekeeping genes as an internal control
inhibitory spectrum against pathogenic bacteria. J Biotechnol
for gene expression studies in the brain of ovariectomized rats
249:20–24. https://​doi.​org/​10.​1016/j.​jbiot​ec.​2017.​03.​018
treated with tibolone. Gene 769:145255. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​
Liu L, Zhao Y, Lu J, Liang M, Sun L, Gao J, Wang L, Hou Y, Huang S
gene.​2020.​145255
(2020) Identification and evaluation of UV-Tolerant Streptomyces
Flemming HC, Wingender J (2010) The biofilm matrix. Nat Rev
araujoniae strain isolated from Tibet Plateau soil as biocontrol
Microbiol 8:623–633. https://​doi.​org/​10.​1038/​nrmic​ro2415
agent against rice blast. Rice Sci 27:359–362. https://​doi.​org/​10.​
Futo M, Opasic L, Koska S, Corak N, Siroki T, Ravikumar V, Thorsell
1016/j.​rsci.​2020.​03.​001
A, Lenuzzi M, Kifer D, Domazet-Loso M, Vlahovicek K, Mijako-
Ma W, Peng D, Walker SL, Cao B, Gao CH, Huang Q, Cai P (2017)
vic I, Domazet-Loso T (2021) Embryo-like features in developing
Bacillus subtilis biofilm development in the presence of soil clay
Bacillus subtilis biofilms. Mol Biol Evol 38:31–47. https://d​ oi.o​ rg/​
minerals and iron oxides. NPJ Biofilms Microbi 3:4. https://​doi.​
10.​1093/​molbev/​msaa2​17
org/​10.​1038/​s41522-​017-​0013-6
Galisa PS, da Silva HA, Macedo AV, Reis VM, Vidal MS, Baldani JI,
Masmoudi F, Abdelmalek N, Tounsi S, Dunlap CA, Trigui M (2019)
Simões-Araújo JL (2012) Identification and validation of refer-
Abiotic stress resistance, plant growth promotion and antifungal
ence genes to study the gene expression in Gluconacetobacter
potential of halotolerant bacteria from a Tunisian solar saltern.
diazotrophicus grown in different carbon sources using RT-qPCR.
Microbiol Res 229:126331. https://d​ oi.o​ rg/1​ 0.1​ 016/j.m
​ icres.2​ 019.​
J Microbiol Methods 91:1–7. https://​doi.​org/​10.​1016/j.​mimet.​
126331
2012.​07.​005

13
Applied Microbiology and Biotechnology (2022) 106:2079–2089 2089

Meng Q, Jiang H, Hao JJ (2016) Effects of Bacillus velezensis strain quantitative PCR, to study esterase gene expression in Oenococ-
BAC03 in promoting plant growth. Biol Control 98:18–26. https://​ cus oeni. Appl Microbiol Biotechnol 96:1039–1047. https://​doi.​
doi.​org/​10.​1016/j.​bioco​ntrol.​2016.​03.​010 org/​10.​1007/​s00253-​012-​4409-1
Metcalf D, Sharif S, Weese JS (2010) Evaluation of candidate reference Theis T, Skurray RA, Brown MH (2007) Identification of suitable
genes in Clostridium difficile for gene expression normalization. internal controls to study expression of a Staphylococcus aureus
Anaerobe 16:439–443. https://​doi.​org/​10.​1016/j.​anaer​obe.​2010.​ multidrug resistance system by quantitative real-time PCR. J
06.​007 Microbiol Methods 70:355–362. https://​doi.​org/​10.​1016/j.​mimet.​
Nam MH, Park MS, Hong GK, Yoo SJ (2009) Biological control of 2007.​05.​011
strawberry Fusarium wilt caused by Fusarium oxysporum f. sp. Thomas F, Barbeyron T, Michel G (2011) Evaluation of reference
fragariae using Bacillus velezensis BS87 and RK1 formulation. J genes for real-time quantitative PCR in the marine flavobacterium
Microbiol Biotechnol 19:520. https://​doi.​org/​10.​4014/​jmb.​0805.​ Zobellia galactanivorans. J Microbiol Methods 84:61–66. https://​
333 doi.​org/​10.​1016/j.​mimet.​2010.​10.​016
Nieto PA, Covarrubias PC, Jedlicki E, Holmes DS, Quatrini R (2009) Untergasser A, Cutcutache I, Koressaar T, Ye J, Faircloth BC, Remm
Selection and evaluation of reference genes for improved inter- M, Rozen SG (2012) Primer3–new capabilities and interfaces.
rogation of microbial transcriptomes: case study with the extre- Nucleic Acids Res 40:e115. https://​doi.​org/​10.​1093/​nar/​gks596
mophile Acidithiobacillus ferrooxidans. BMC Mol Biol 10:63. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe
https://​doi.​org/​10.​1186/​1471-​2199-​10-​63 A, Speleman F (2002) Accurate normalization of real-time quan-
Nybo K (2011) qPCR efficiency calculations. Biotechniques 51:401– titative RT-PCR data by geometric averaging of multiple internal
402. https://​doi.​org/​10.​2144/​00011​3776 control genes. Genome Biol 3: research0034.1–research0034.11.
Ogran A, Yardeni EH, Keren-Paz A, Bucher T, Jain R, Gilhar O, https://​doi.​org/​10.​1186/​gb-​2002-3-​7-​resea​rch00​34
Kolodkin-Gal I (2019) The plant host induces antibiotic produc- Vaudano E, Noti O, Costantini A, Garcia-Moruno E (2011) Identifica-
tion to select the most-beneficial colonizers. Appl Environ Micro- tion of reference genes suitable for normalization of RT-qPCR
biol 85:e00512-e519. https://​doi.​org/​10.​1128/​AEM.​00512-​19 expression data in Saccharomyces cerevisiae during alcoholic
Reiter L, Kolstø AB, Piehler AP (2011) Reference genes for quantita- fermentation. Biotechnol Lett 33:1593–1599. https://​doi.​org/​10.​
tive, reverse-transcription PCR in Bacillus cereus group strains 1007/​s10529-​011-​0603-y
throughout the bacterial life cycle. J Microbiol Methods 86:210– Vlamakis H, Chai Y, Beauregard P, Losick R, Kolter R (2013) Stick-
217. https://​doi.​org/​10.​1016/j.​mimet.​2011.​05.​006 ing together: building a biofilm the Bacillus subtilis way. Nat Rev
Ritz M, Garenaux A, Berge M, Federighi M (2009) Determination of Microbiol 11:157–168. https://​doi.​org/​10.​1038/​nrmic​ro2960
rpoA as the most suitable internal control to study stress response Wang LT, Lee FL, Tai CJ, Kasai H (2007) Comparison of gyrB gene
in C. jejuni by RT-qPCR and application to oxidative stress. J sequences, 16S rRNA gene sequences and DNA-DNA hybridi-
Microbiol Methods 76:196–200. https://​doi.​org/​10.​1016/j.​mimet.​ zation in the Bacillus subtilis group. Int J Syst Evol Microbiol
2008.​10.​014 57:1846–1850. https://​doi.​org/​10.​1099/​ijs.0.​64685-0
Rocha DJ, Santos CS, Pacheco LG (2015) Bacterial reference genes Wilkinson GN, Rogers CE (1973) Symbolic descriptions of factorial
for gene expression studies by RT-qPCR: survey and analysis. models for analysis of variance. Appl Stat 22:392–399. https://​
Anton Leeuw Int J G 108:685–693. https://​doi.​org/​10.​1007/​ doi.​org/​10.​2307/​23467​86
s10482-​015-​0524-1 Yin W, Wang Y, Liu L, He J (2019) Biofilms: The microbial “protective
Ruijter JM, Ramakers C, Hoogaars WM, Karlen Y, Bakker O, van den clothing” in extreme environments. Int J Mol Sci 20:3423. https://​
Hoff MJ, Moorman AF (2009) Amplification efficiency: linking doi.​org/​10.​3390/​ijms2​01434​23
baseline and bias in the analysis of quantitative PCR data. Nucleic Zhao J, Zhang C, Lu J, Lu Z (2016) Enhancement of fengycin pro-
Acids Res 37:e45. https://​doi.​org/​10.​1093/​nar/​gkp045 duction in Bacillus amyloliquefaciens by genome shuffling and
Ruiz-García C, Béjar V, Martínez-Checa F, Llamas I, Quesada E relative gene expression analysis using RT-PCR. Can J Microbiol
(2005) Bacillus velezensis sp. nov., a surfactant-producing bac- 62:431–436. https://​doi.​org/​10.​1139/​cjm-​2015-​0734
terium isolated from the river Vélez in Málaga, southern Spain. Zhao Y (2020) Studies on screening of biocontrol bacteria and con-
Int J Syst Evol Microbiol 55:191–195. https://​doi.​org/​10.​1099/​ trol technologies of rice bakanae disease. Dissertation, Guangxi
ijs.0.​63310-0 University. https://​doi.​org/​10.​27034/d.​cnki.​ggxiu.​2020.​000432
Rutledge RG, Stewart D (2008) Critical evaluation of methods used to
determine amplification efficiency refutes the exponential char- Publisher's note Springer Nature remains neutral with regard to
acter of real-time PCR. BMC Mol Biol 9:96. https://​doi.​org/​10.​ jurisdictional claims in published maps and institutional affiliations.
1186/​1471-​2199-9-​96
Sumby KM, Grbin PR, Jiranek V (2012) Validation of the use of
multiple internal control genes, and the application of real-time

13