Bacterial identification and characterisation

Genetic sequencing of the 16S rRNA encoding gene is a widely-accepted method for

determining bacterial identity, however the 16S rRNA encoding gene alone is not enough

to give a definitive identity at species level but is useful for identifying bacteria to at least

genus level, and can provide a lead to other genes to sequence for an accurate

identification. The bacteria isolated in this study had their 16S rRNA encoding gene

sequenced after amplification with PCR using universal primers.

Universal primers amplify the specific sequence they are designed to amplify no matter

the sample, and so one must be extremely careful that contamination doesn’t occur during

preparation of the PCR reaction, or there is a risk of amplifying the wrong product. To

reduce the possibility of contamination, isolated colonies of the bacteria to be identified

were selected and a negative control which contained no DNA was included in the PCR

step. If a band was observed on the resolving gel after PCR amplification, all of the

amplified products were disposed of and an investigation as to the contamination source

was carried out.

Sequencing PCR amplicons carries problems as the polymerase may lack proof reading

capability and can incorporate the incorrect nucleotide into the extending sequence, miss

a nucleotide or add additional nucleotides particularly in heavily repeated regions,

resulting in an erroneous sequence. Errors may also be introduced by the sequencing

method, and some reads may be unclear, relying on reading the chromatic graphic results

carefully and may still result in an incorrect base calling.

A more accurate and informative but costly method for bacterial identification is whole

genome sequencing using next generation technologies such as MiSeq. The Illumina

MiSeq platform was used to identify isolate 133. This technology is more accurate,

because more reads span a region of the genome, which when assembled overlap each

other, providing an extended sequence. The fact that every piece of DNA present gets

amplified can cause issues with the end result being impure, and the method of library

preparation can lead to information being lost if multiple samples are run.

Once the sequencing data is available it must be assembled. Assembly may lead to errors,

especially when there are multiple regions of repeat DNA. However, software designers

are continually improving the algorithms to increase the accuracy of assembly. The

assembled DNA sequences are split into contigs depending on how well the DNA could

be assembled. Ideally this results in a single contig representing the entire genome. A low

L50 score that represents where 50% of contigs assemble to 50% of the sequence length

and a high N50 score which represents the average length of all contigs that make up the

L50 are desirable.

The contigs resulting from the sequencing of isolate 133, were compared to the NCBI

database using BLAST, and this revealed that among the contigs of isolate 133 sequence

there was contamination from eukaryotic and viral sources. The eukaryotic sequences

were related to a range of families including Hominidae, Hylobatidae, Physeteridae, and

Plasmodiidae, while the viral sequence was that of PhiX a virus commonly used as a

control for next generation sequencing. This contamination raises questions about what

the accepted level of contamination is when carrying out next generation sequencing, and

although the sample sequenced in this case was bacterial and could be easily distinguished

from the contaminating sequences, samples of eukaryotic origin may encounter issues

with foreign sequences being included with their sequence.

The sequencing of isolate 133 using a paired end library preparation resulted in an

incomplete genome assembly comprised of 41 contigs of bacterial DNA. Resequencing

using a mate paired library preparation could be used for completion. From the

sequencing data housekeeping genes from isolate 133 which had been identified to genus

level as Raoultella could be aligned and compared to other known Raoultella sp. and

allowed for isolate 133 to be identified as R. terrigena. Although these housekeeping

genes were identical to the reference genome R. terrigena there were differences in non-

housekeeping genes. This and geographical separation of the isolate from the reference

led to a new strain name being assigned to isolate 133 of NZ133 where NZ stands for the

country of origin, New Zealand, and 133 being the isolates reference during the course of

the investigation. The full identification of isolate 133 is therefore Raoultella terrigena

strain NZ133.

Bacterial characterisation of R. terrigena NZ133 was carried out using commercial

biochemical test kits and the reference organisms, R. ornithinolytica and R. terrigena for

comparison. The biochemical test results revealed differences in substrate utilisation

between the tested bacteria, however there were inconsistencies between kits and tests,

with one kit testing positive for utilisation and the other testing negative, and repetition

of tests with the same kit having varying results. The differences seen between kits could

be due to differences in the principle used to test substrate utilisation. For example one

kit may test anaerobic fermentation of a substrate where another kit tests aerobic

utilisation. While both of these tests look at the utilisation of the same substrate, bacteria

capable of only one form of utilisation will test negative to one kit.

An issue surrounding interpretation of results from colourimetric biochemical test kits is

that unless an appropriate reader is available the results are open for interpretation and

will vary depending on the person reading the colour change. This is particularly true if

the difference in colour change between a positive and negative result is minor or

ambiguous, for example between aqua and turquoise.

Other issues surrounding the use of biochemical test kits are the deterioration of reagents,

and manual addition of reagents to kits. While some kits may avoid having to manually

add additional reagents by having the correct concentration and composition already

provided, these reagents deteriorate over time. The manual addition of additional reagents

may not be consistent and if not provide with the kit are subject to batch variation. This

can be overcome by using a carefully programmed fully automated system and frequent

preparation of reagents to prevent deterioration. However this would be expensive and

impractical when researching in the field environment.

Overall, miniaturised biochemical tests are useful for simple and quick biochemical

analysis, and can provide a guide as to other biochemical tests to be carried out. They can

also be used to distinguish species of a genus based on substrate utilisation. This must be

approached with some caution as bacteria within a species may test differently, as seen in

R. terrigena NZ133 and the reference R. terrigena.