There are interesting aspects raised by this question, not the least at which end of a distant interaction does the adjustment take place. Should the dyslexic collaborator make adjustments or should the other party make accommodations? The issue of the influence any legislation is also interesting.
If the collaboration arrangement is a formalised one, does any anti-discrimination laws apply at the US end which would require the US party to make adjustments? Not knowing details of local laws I cannot say, but local advice might need to be obtained. Even if it is not a requirement to do anything there may be other imperatives such as the protection of university reputation that could be affected by not being able to enable effective collaborations with disabled students.
At the European end the location would indicate if their university should be providing resources to enable adjustments. If it was the UK (I know it is not) the the student's university would already have an obligation to support adjustments to enable the participation in the collaboration on equal terms.
Having discussed who might have to make adjustments, lets looks at possible means of adjustment. The most obvious mechanism is speech to text software, but even the best software has trouble with non-native (and some native) accents as well as the technical material often contained in Phd research discussions. Many regular uses of text-to-speech use a special "talking-to-the robot" form of speech to ensure accurate transcription.
This leaves two other approaches. The move to a reduced text content or human transcription.
Many dyslexic people find lengthy textual prose difficult. Often diagrams or speech are better. A tools that is often recommended are idea capture using brain-ware mapping. An example of such a Tool is Inspiration. The research ideas can be captured in diagram form and pasted into a discussion wiki. Using a diagram the dyslexic and non-dyslexic collaborators can communicate on equal terms.
If speech transcription seems to have a rôle then, perhaps, junior researchers, such as undergraduate volunteers could be used to transcribe. This could happen at either end depending on availability of labour or funding.
For a dyslexic student where funding is available there are software solutions that are very helpful in preparing text. Some tools assist in the conversion of non-linear expression into structured text (such as Inspiration as already mentioned), but there are also writing tools. I find Texthelp gold can be very useful, but others are available.
I final thought; It is often said we are two nations divided by a common language and as such there can be linguistic difficulties between the most erudite writers on different continents. Just think how much harder it can be when dyslexia is introduced into the mix. As with both issues dialect or dyslexia, a little tolerance goes a long way.